Author: Ubuntu Git Importer
Author Date: 2019-07-17 09:53:03 UTC

DSC file for 7.2.6-1

Author: Ubuntu Git Importer
Author Date: 2019-07-17 09:52:32 UTC

DSC file for 7.2.6-1

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-applied version 7.2.6-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: fcb792d7d142fb6ea0f3d0629fe477ccf03ec836
Unapplied parent: 55045bbaa51c9a682c5371bc84f3a1480407c434

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-unapplied version 7.2.6-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c09773b977195fcc92488595ea75d7d5c999fd50

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-unapplied version 7.2.6-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c09773b977195fcc92488595ea75d7d5c999fd50

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-unapplied version 7.2.6-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c09773b977195fcc92488595ea75d7d5c999fd50

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-unapplied version 7.2.6-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c09773b977195fcc92488595ea75d7d5c999fd50

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-applied version 7.2.6-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: fcb792d7d142fb6ea0f3d0629fe477ccf03ec836
Unapplied parent: 55045bbaa51c9a682c5371bc84f3a1480407c434

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-applied version 7.2.6-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: fcb792d7d142fb6ea0f3d0629fe477ccf03ec836
Unapplied parent: 55045bbaa51c9a682c5371bc84f3a1480407c434

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-applied version 7.2.6-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: fcb792d7d142fb6ea0f3d0629fe477ccf03ec836
Unapplied parent: 55045bbaa51c9a682c5371bc84f3a1480407c434

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-unapplied version 7.2.6-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c09773b977195fcc92488595ea75d7d5c999fd50

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Troy Heber
Author Date: 2019-07-15 14:20:49 UTC

Import patches-applied version 7.2.6-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: fcb792d7d142fb6ea0f3d0629fe477ccf03ec836
Unapplied parent: 55045bbaa51c9a682c5371bc84f3a1480407c434

New changelog entries:
  * Two fixes for the Xen hypervisor; the first fixes a bug seen with Xen
    4.11.0 during initialization, which fails with the error message "crash:
    invalid kernel virtual address: <address> type: fill_pcpu_struct",
    followed by "WARNING: cannot fill pcpu_struct" and "crash: cannot read
    cpu_info". The second fix prevents a segmentation violation associated
    with a crash-7.1.1 commit that addressed the Xen 4.5.0 hypervisor symbol
    name change from "dom0" to "hardware_domain".
  * Fix for Linux 4.20 and later x86_64 kernels which are NOT configured with
    CONFIG_RANDOMIZE_BASE. Linux 4.20 introduced kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array". For
    kernels prior to Linux 4.20.0 which have backports of the kernel commit,
    the kernel's PAGE_OFFSET value must be manually specified via the command
    line option "--machdep page_offset=ffff888000000000" for kernels with
    4-level page tables, or "--machdep page_offset=ff11000000000000" for
    kernels with 5-level paging. (or alternatively the shorter version "-m
    page_offset=<address>" may be used). The command line option requirement
  * Fix for the "p" command if the expression contains more than one opening
    parenthesis character and a minus/dash sign. Without the patch, the
    minus/dash sign will get dropped from the command prior to it being passed
    on to gdb for evaluation, and the command will fail with the message "p:
    gdb request failed: <expression>", where the <expression> string will not
  * Fix for the internal parse_line() utility function to account for embedded
    sets of parentheses, which may be used for expressions that are passed to
    gdb by the "p" command. Without the patch, expressions containing
    embedded sets of parentheses are broken up into multiple argument tokens
    instead of just one. The previous commit has been reverted by this one.
  * First phase of support for ARM64 kernels that are configured with
    CONFIG_ARM64_USER_VA_BITS_52, which causes the PTRS_PER_PGD count to
    increase from 64 to 1024. Without the patch, "WARNING: cannot access
    vmalloc'd module memory" will be displayed during session initialization,
    and the translation of any mapped kernel virtual address that requires a
    page table walk will fail, leading to a myriad of other errors.
  * Support for configurable CONFIG_ARM64_PA_BITS values introduced in kernel
    commit 982aa7c5f0861bf56b2412ca341a13f44c238ba4, titled "arm64: add
    kconfig symbol to configure physical address size". Without the patch, it
    is impossible to determine the value of CONFIG_ARM64_PA_BITS is, and will
    require a new MAX_PHYSMEM_BITS vmcoreinfo entry to be exported. This
  * For live system analysis where there is no vmcoreinfo ELF note attached to
    /proc/kcore, or for dumpfile analysis where there is no vmcoreinfo ELF
    note attached to the dumpfile, this patch sets the internal
    pc->read_vmcoreinfo() function to a new plugin function that reads the
    data directly from the live kernel or dumpfile. Because the function is
    set much later during initialization than if the ELF note is attached to
    /proc/kcore or the dumpfile, it may not be available during very early
  * Fix for Linux 4.14.84 and later 4.14-based x86_64 kernels which are NOT
    configured with CONFIG_RANDOMIZE_BASE and have backported kernel commit
    d52888aa2753e3063a9d3a0c9f72f94aa9809c15, titled "x86/mm: Move LDT remap
    out of KASLR region on 5-level paging", which modified the 4-level and
    5-level paging PAGE_OFFSET values. Without this patch, the crash session
    fails during initialization with the error message "crash: read error:
    kernel virtual address: <address> type: tss_struct ist array".
  * Fix for determining the x86_64 "phys_base" value in dumpfiles created by
    the KVM "virsh dump" facility if the kernel is KASLR-enabled and does not
    have the phys_base value stored in vmcoreinfo data. Without the patch,
    the message "WARNING: cannot determine physical base address: defaulting
    to 0" is displayed, and the crash session fails to initialize.
  * 32-bit ARM kernels built with the Thumb-2 instruction set utilize the R7
    register instead of FP for unwinding stacks using the DWARF unwinder. On
    those kernels, without the patch, the "bt" command only shows the task
  * Fix for the "kmem -z" option on Linux 5.0 and later kernels that contain
    commit a921444382b49cc7fdeca3fba3e278bc09484a27, titled "mm: move zone
    watermark accesses behind an accessor". Without the patch, the command
    fails with the error message "kmem: invalid (optional) structure member
    offsets: zone_pages_min or zone_struct_pages_min".
  * Fix for the "kmem -i" option on Linux 5.0 and later kernels that contain
    commit ca79b0c211af63fa3276f0e3fd7dd9ada2439839 titled "mm: convert
    totalram_pages and totalhigh_pages variables to atomic". Without the
    patch, the command prints some incorrect values, and besides does not
    print high/low memory information on kernels which are configured with
  * Fix for the display of kernel module symbol types by the "sym" command in
    Linux 5.0 and later kernels if the module debuginfo data has not been
    loaded into the crash session. The st_info member of the Elf32_Sym or
    Elf64_Sym structures has changed so as to not contain ASCII symbol type
    characters, and as a result the "sym" command will show unprintable data
    as the symbol type. With the patch, only text types ("t" or "T") will be
  * First phase of support of the upcoming ARM64 kernel memory map changes to
    support 52-bit kernel virtual addressing, which allows the configuration
    of CONFIG_ARM64_VA_BITS to be 52, but where the actual number of VA bits
    may be downgraded during boot depending upon the hardware capability.
    This phase is only applicable for live system analysis.
  * Fix for the "dis <function>" option with kernel module text symbols on
    Linux 5.0 and later kernels. Without the patch, the disassembly may stop
    prematurely or extend into the next function because the st_size member of
    the Elf32_Sym or Elf64_Sym text symbol structures can no longer be used as
  * Commit dd12805ed1db7 in the linux-next kernel repository, titled "XArray:
    Remove radix tree compatibility", changes the definition of
    "radix_tree_root" back to be a struct. However, the content of the new
    structure differs from the original structure, so without the patch,
    current linux-next kernels fail during initialization with the error
    message "radix trees do not exist or have changed their format". Because
    the new "radix_tree_root" and "xarray" structures have nearly the same
    layout, the existing functionality for XArrays can be reused.
  * Fixes for the "trace.so" extension module: (1) The reader_page can be
    empty if it was never read, do not record it if it is empty. Better yet,
    do not record any page that is empty. The struct buffer_page "real_end"
    is not available in older kernels, so it needs to be tested if it exists
    before we can use it. (2) In newer kernels, the sp->type of kernel module
    symbols does not contain the symbol type character unless the module's
    debuginfo data has been loaded into the crash session. Writing a garbage
    type to the kallsyms file for trace-cmd to read causes it to crash, so
    just always write an 'm'. (3) Add the "trace dump -t <trace.dat>" option
  * Fix to find the kernel configuration data in Linux 5.1 kernels containing
    commit 13610aa908dcfce77135bb799c0a10d0172da6ba, titled "kernel/configs:
    use .incbin directive to embed config_data.gz". Without the patch, new
    kernels configured with CONFIG_IKCONFIG_PROC will display "WARNING: could
    not find MAGIC_START!" during session initialization, and also when
  * Fix for the PPC64 "bt" command running against kernels that are configured
    with CONFIG_THREAD_INFO_IN_TASK. Without the patch, the "bt" command
    fails with the message "bt: invalid/stale stack pointer for this task:
  * Fix for the "files -d <dentry>" option if the dentry.d_inode pointer is
    NULL. Without the patch, the command output does not display the
  * When the is_s390_dump() function is called to determine whether a file is
    an s390 dumpfile, it currently presumes that the fopen() call always
    works, and then tries to read it with using a NULL file pointer. Change
    it to verify that the fopen() was successful, and if not, print an error
    message as is done with the other dumpfile type verifier functions.
  * Implement support for ARM64 kernels that are configured with:
    CONFIG_ARM64_PA_BITS=52 CONFIG_ARM64_64K_PAGES CONFIG_PGTABLE_LEVELS=3 and
    that run on a host containing physical memory that utilizes any bit in the
    uppermost 4 bits of the 52-bit physical address range.
  * Extension of the "snap.so" extension module to pass a second
    architecture-specific value in the ELF header; its initial use is for
    support of the upcoming ARM64 52-bit kernel virtual address space by
  * Apply initial changes to support kernel address space layout randomization
    (KASLR) for s390X. This is the minimal patch-set required to process
    s390x dumps for the kernels configured with CONFIG_RANDOMIZE_BASE, and to
    accept the "--kaslr" command line option. Only dumpfiles whose headers
  * Fix for the "dev -[dD]" options on Linux 5.1-rc1 and later kernels that
    contain commit 570d0200123fb4f809aa2f6226e93a458d664d70, titled "driver
    core: move device->knode_class to device_private". Without the patch, the
    command options fail with the error message "dev: invalid structure member
  * Linux 4.18 kernels introduced a new CONFIG_PROC_VMCORE_DEVICE_DUMP
    configuration in commit 2724273e8fd00b512596a77ee063f49b25f36507, titled
    "vmcore: add API to collect hardware dump in second kernel", in which
    device drivers may collect a device specific snapshot of the
    hardware/firmware state of their underlying devices, and export the data
    as a kdump ELF note with type NT_VMCOREDD. This patch recognizes the new
    ELF note(s) in both ELF and compressed kdump vmcore dumpfiles. The "help
    -[nD]" option shows basic information about each note, and two new "dev"
    command options have been introduced. The "dev -V" option displays an
    indexed list of each note, showing the device name, the dumpfile offset,
    and the size of each note. The "dev -v index [file]" option either dumps
    the contents of a note to the display screen in a human-readable format,
    or copies the note data directly to a specified file.
  * If the kernel's "vmap_area_list" doubly-linked list is corrupt such that
    it does not link back to the global list_head, commands that require
    information regarding the range of virtually-mapped kernel addresses will
    display a generic list-handling error message such as "kmem: invalid list
    entry: 0", and the command will typically fail to fully complete.
    However, without the patch, there will also be "WARNING: malloc/free
    mismatch (29/30)" messages that get displayed after every subsequent
    command. This patch prevents the mismatch messages, and also adds an
    additional error message indicating "WARNING: invalid/corrupt
    vmap_area_list" to further clarify the generic list-handling error
  * Fix for the "dev" help page to remove the unused -r option letter.
  * If a duplicate list entry is encountered when using the "list -B" Brent
    algorithm, change the list loop length value from hexadecimal to decimal.
  * Update the README file to indicate the capability of building an x86_64
    crash binary with "make target=PPC64", which can be used to analyze
  * Fix for hybrid kernels that have backported support for the Xarray
    facility while allowing subsystems to continue to use radix trees.
    Without the patch, the crash session fails during initialization with the
    message "crash: xarray facility does not exist or has changed its format".

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-applied version 7.2.1-1ubuntu2 to applied/ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: da6b13a1a05bcef342b2bec3bc091e540c616953
Unapplied parent: 217c9bd02c3fc0216a4e4c3b51aa3e4da81d15cc

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-unapplied version 7.2.1-1ubuntu2 to ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: 55055ccbcdf24d08034f81681dcbdf5195679f23

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-unapplied version 7.2.1-1ubuntu2 to ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: 55055ccbcdf24d08034f81681dcbdf5195679f23

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-unapplied version 7.2.1-1ubuntu2 to ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: 55055ccbcdf24d08034f81681dcbdf5195679f23

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-applied version 7.2.1-1ubuntu2 to applied/ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: da6b13a1a05bcef342b2bec3bc091e540c616953
Unapplied parent: 217c9bd02c3fc0216a4e4c3b51aa3e4da81d15cc

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2019-02-15 08:58:17 UTC

Import patches-applied version 7.2.1-1ubuntu2 to applied/ubuntu/bionic-proposed

Imported using git-ubuntu import.

Changelog parent: da6b13a1a05bcef342b2bec3bc091e540c616953
Unapplied parent: 217c9bd02c3fc0216a4e4c3b51aa3e4da81d15cc

New changelog entries:
  * Add patch to fix ipcs command (LP: #1765660).
  * Add test for ipcs command.

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-unapplied version 7.2.5-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: b5b00bf2b5756d8e0d8f30adef4d84c199e4e816

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-unapplied version 7.2.5-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: b5b00bf2b5756d8e0d8f30adef4d84c199e4e816

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-applied version 7.2.5-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 022673fc6218fe75e9f59f7e1b1d358a5b77f737
Unapplied parent: a18ba791fd5f89546c64935b491242b11a426dc2

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-unapplied version 7.2.5-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: b5b00bf2b5756d8e0d8f30adef4d84c199e4e816

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-applied version 7.2.5-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 022673fc6218fe75e9f59f7e1b1d358a5b77f737
Unapplied parent: a18ba791fd5f89546c64935b491242b11a426dc2

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-applied version 7.2.5-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 022673fc6218fe75e9f59f7e1b1d358a5b77f737
Unapplied parent: a18ba791fd5f89546c64935b491242b11a426dc2

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-applied version 7.2.5-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 022673fc6218fe75e9f59f7e1b1d358a5b77f737
Unapplied parent: a18ba791fd5f89546c64935b491242b11a426dc2

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Troy Heber
Author Date: 2019-01-29 15:55:24 UTC

Import patches-unapplied version 7.2.5-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: b5b00bf2b5756d8e0d8f30adef4d84c199e4e816

New changelog entries:
  * New upstream (Closes: #919291)
  * Resurrection of the "dev -p" option for displaying PCI device data on
    Linux 2.6.26 and later kernels. The option was deprecated as of Linux
    2.6.26, and without the patch, the option would indicate "dev: -p option
    not supported or applicable on this architecture or kernel" when running
    against the newer kernel versions. PCI Bus information will also be
    displayed with this patch.
  * With Linux 4.19-rc1 commit 7d4340bb92a9df78e6e28152f3dd89d9bd82146b,
    titled "powerpc/mm: Increase MAX_PHYSMEM_BITS to 128TB with
    SPARSEMEM_VMEMMAP config", the PPC64 MAX_PHYSMEM_BITS value has been
    bumped up to 47. The appropriate update has been made in this patch.
  * Fix to allow piping command output to a shell script beginning with a
    shebang (#!) character sequence if the script pathname is specified with a
    preceding "./" or "/". Without the patch, the piped command fails with
    the message "crash: pipe operation failed".
  * Fix for the PPC64 "bt" command to recognize when a thread is running in
    OPAL firmware. Without the patch, the "bt" command indicates
    <task-address>: Invalid Stack Pointer <OPAL-firmware-address>"
  * As an addendum to the "dev -p" patch above, add the new structure member
    offsets for display by the "help -o" option.
  * Enhancement to the "kmem -n" option to dump memory block information if
    the kernel supports it. In addition, the memory section data block has a
    new "STATE" column added to it.
  * Addendum to the previous "kmem -n" patch to fix a FTBFS issue. Without
    the patch, certain architectures fail to compile with the error
    "memory.c:17315:16: error: ‘PAGE_SHIFT’ undeclared (first use in this
    function)"
  * Fix the calculation of the vmalloc memory region size to account for Linux
    4.17 commit a7412546d8cb5ad578805060b4006f2a021b5868, titled "x86/mm:
    Adjust vmalloc base and size at boot-time", which increases the region's
    size from 32TB to 1280TB when 5-level pagetables are enabled. Also
    presume that virtual addresses above the end of the vmalloc space up to
    the beginning of vmemmap space are translatable via 5-level page tables.
    Without the patch, mapped virtual addresses may fail translation in
    whatever command accesses them, with errors indicating "seek error: kernel
    virtual address: <mapped-address> type: <type-string>"
  * Address several Coverity Scan "RESOURCE_LEAK" issues in the following
    top-level source files: cmdline.c, kvmdump.c, lkcd_v8.c, xendump.c,
    symbols.c, unwind_x86_32_64.c, va_server.c and va_server_v1.c.
  * Modify the x86_64 "bt" behavior when a legitimate exception RIP value
    cannot be referenced symbolically, such as when the exception occurs while
    running in seccomp BPF filter code. Without the patch, the exception
    frame register dump is preceded by "[exception RIP: unknown or invalid
    address]", and then followed by "bt: WARNING: possibly bogus exception
    frame". With the patch applied, the translation of the exception RIP will
    show "[exception RIP: no symbolic reference]", and there will be no
    warning message.
  * Account for the /proc/kcore VMCOREINFO PT_NOTE in Linux 4.19 and later
    kernels having commit 23c85094fe1895caefdd19ef624ee687ec5f4507, titled
    "proc/kcore: add vmcoreinfo note to /proc/kcore". The PT_NOTE information
    is stored during session initialization for later display by "help
    -[n|D]"; a subsequent commit will make it available for use by the crash
    utility's internal pc->read_vmcoreinfo() function.
  * Second phase of support for the VMCOREINFO PT_NOTE added to the ELF header
    of /proc/kcore in Linux 4.19 and later kernels. This patch introduces
    support for live session /proc/kcore VMCOREINFO access by the crash
    utility's internal pc->read_vmcoreinfo() function. New usage include the
    initialization of the x86_64 phys_base value, and the arm64 phys_offset,
    page size, and VA bits count.
  * Fix for Linux 4.20-rc1 and later kernels that contain kernel commit
    5c83511bdb9832c86be20fb86b783356e2f58062, titled "x86/paravirt: Use a
    single ops structure". Without the patch, the kernel may be misidentified
    as an ARCH_XEN kernel, with the most noticeable result being the inability
    to read vmemmap'd page structures.
  * Implemented the functionality for a new MEMBER_TYPE_NAME() macro, which
    will return a pointer to the type name string of a structure member. It
    is being put in place for the support of Linux 4.20 radix tree to xarray
    replacements, where structure member types may be changed from
    radix_tree_root structures to xarray structures.
  * First phase of support for the XArray facility. The added support is
    similar to that of radix trees, but introduces completely separate
    functions, structures and #defines. None of the applicable radix tree
    users in the crash utility have been switched over, so this phase does not
    introduce any functional changes.
  * Second phase of support for the XArray facility, which handles the
    switch-over of PID handling from a radix tree to an XArray in Linux 4.20
    and later kernels. Without the patch, the crash session fails during
    session initialization with the message "crash: radix trees do not exist
    or have changed their format".
  * Third phase of support for the XArray facility, which consolidates the
    radix_tree_pair and xarray_pair structures into a unified list_pair
    structure that is used by both facilities, and fixes the "bpf" command.
    Without the patch, the command fails on Linux 4.20 or later kernels with
    the error message "bpf: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "files -p" option.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "files: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "irq" command.
    Without the patch, the command fails on Linux 4.20 and later kernels with
    the error message "irq: radix trees do not exist or have changed their
    format".
  * Added support for usage of the XArray facility by the "ipcs" command.
    Without the patch, the command may fail on Linux 4.20 and later kernels
    with the error message "irq: radix trees do not exist or have changed
    their format".
  * Added a new "tree -t xarray" option to display of the contents of an
    XArray in Linux 4.20 and later kernels. The implementation is similar to
    that of radix tree displays, but in addition, the "-p" option will also
    display the index value of each entry in a radix tree or XArray.
  * Fix for the "files -p <inode>" option on a file with a large number of
    pages. Without the patch, the command attempts to read radix tree node
    slot entries that are RADIX_TREE_EXCEPTIONAL_ENTRY types instead of page
    pointers, and as a result may fail with a dump of the internal buffer
    allocation stats followed by the message "files: cannot allocate any more
    memory!".
  * Fix for the "ps -s" option on ARM64 if the number of tasks exceeds 2000.
    Without the patch, the command ultimately fails with a dump of the
    internal buffer allocation stats, followed by the message "ps: cannot
    allocate any more memory!".
  * With Linux 4.20-rc1 commit 4ffe713b7587b14695c9bec26a000fc88ef54895,
    titled "powerpc/mm: Increase the max addressable memory to 2PB", the PPC64
    MAX_PHYSMEM_BITS value has been bumped up to 51 for
    CONFIG_SPARSEMEM_VMEMMAP and CONFIG_SPARSEMEM_EXTREME. The appropriate
    update has been made in this patch.
  * Implemented a new plugin function for the readline library's tab
    completion feature. Without the patch, the use of the default plugin from
    the embedded gdb module has been seen to cause segmentation violations or
    other fatal malloc/free/corruption assertions. The new plugin takes gdb
    out of the picture entirely, and also restricts the matching options to
    just symbol names, so as not to clutter the results with irrelevant
    filenames.
  * The RHEL8 kernel will contain a backport of the Linux 4.19 kernel commit
    7d4340bb92a9df78e6e28152f3dd89d9bd82146b, titled "powerpc/mm: Increase
    MAX_PHYSMEM_BITS to 128TB with SPARSEMEM_VMEMMAP config". As a result,
    the use of the THIS_KERNEL_VERSION() macro by the crash utility does not
    suffice for determining the MAX_PHYSMEM_BITS value for PPC64. The
    appropriate update has been made in this patch.
  * Fix for an initialization-time session failure when all three of the
    following conditions exist: (1) invoking the session with "crash -d2" or
    larger debug number (2) running against a Linux 3.3 or later kernel (3)
    using a post-7.2.4 crash utility that has the new "kmem -n" support above
    for the display of memory blocks Without the patch, the crash session
    fails with the error message "crash: invalid structure member offset:
    device_kobj".
  * Fix for an initialization-time segmentation violation when invoking
    crash-7.2.4 or later with "crash -d2" or larger debug number.
  * Add a write operation handler to the sample /dev/crash memory driver that
    enables writing to kernel memory via the "wr" command.
  * Prevent a SIGSEGV if a user attempts to input a command line that exceeds
    the maximum length of 1500 bytes. The patch displays an error message and
    ignores the command line.
  * Fix for the "dev -[dD]" options in kernels containing Linux 5.0-rc1 commit
    7ff4f8035695984c513598e2d49c8277d5d234ca, titled "block: remove dead queue
    members", in which the number of I/Os issued to a disk driver are no
    longer stored in the request_queue structure. Without the patch, the
    options indicate "dev: -d option not supported or applicable on this
    architecture or kernel". With the patch, the "DRV" column is not shown.
  * A crash-7.1.1 commit added support for Linux version 5.x. To prevent
    surprise failures due to unexpected kernel version bumps in the future,
    support has been added for version 6, keeping it one step ahead.
  * Fix for a gcc-9 compilation error that occurs if an inline asm statement
    clobbers the stack pointer. Without the patch, x86 and x86_64 builds will
    fail to compile gdb-7.6/gdb/common/linux-ptrace.c, generating an error
    that indicates "error: Stack Pointer register clobbered by '%rsp' in
    'asm'".

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-applied version 7.2.3+real-1~16.04.1 to applied/ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: 05f62c2f7652348f5b3698195c9991a500739441
Unapplied parent: 2682490cc129b3374ebe052d790365b8ee0f3d32

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-unapplied version 7.2.3+real-1~16.04.1 to ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: e7b8467cfc66ebb562ba84e5e4b76cd975e7a02b

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-unapplied version 7.2.3+real-1~16.04.1 to ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: e7b8467cfc66ebb562ba84e5e4b76cd975e7a02b

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-unapplied version 7.2.3+real-1~16.04.1 to ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: e7b8467cfc66ebb562ba84e5e4b76cd975e7a02b

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-applied version 7.2.3+real-1~16.04.1 to applied/ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: 05f62c2f7652348f5b3698195c9991a500739441
Unapplied parent: 2682490cc129b3374ebe052d790365b8ee0f3d32

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Thadeu Lima de Souza Cascardo
Author Date: 2018-06-26 17:32:30 UTC

Import patches-applied version 7.2.3+real-1~16.04.1 to applied/ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: 05f62c2f7652348f5b3698195c9991a500739441
Unapplied parent: 2682490cc129b3374ebe052d790365b8ee0f3d32

New changelog entries:
  * Backport to xenial. LP: #1746088
    - Build-Depends on debhelper 9.

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-applied version 7.2.3+real-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 85a666a561f0c04fa4adeb09d414476601c0a4a1
Unapplied parent: bc777cc0bcbb4f380e05c834ca3a826a129afd61

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-unapplied version 7.2.3+real-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c6e46b982274ec937ad55c06784a933dcaa48714

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-unapplied version 7.2.3+real-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c6e46b982274ec937ad55c06784a933dcaa48714

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-unapplied version 7.2.3+real-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: c6e46b982274ec937ad55c06784a933dcaa48714

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-applied version 7.2.3+real-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 85a666a561f0c04fa4adeb09d414476601c0a4a1
Unapplied parent: bc777cc0bcbb4f380e05c834ca3a826a129afd61

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Troy Heber
Author Date: 2018-06-22 21:39:16 UTC

Import patches-applied version 7.2.3+real-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 85a666a561f0c04fa4adeb09d414476601c0a4a1
Unapplied parent: bc777cc0bcbb4f380e05c834ca3a826a129afd61

New changelog entries:
  * Revert wget addition and re-include embedded gdb (Closes: #901863)

Author: Ubuntu Git Importer
Author Date: 2018-04-21 01:43:46 UTC

pristine-tar data for crash_7.2.1.orig.tar.gz

Author: Ubuntu Git Importer
Author Date: 2018-04-21 00:24:24 UTC

pristine-tar data for crash_7.2.1.orig.tar.gz

Author: Troy Heber
Author Date: 2018-02-16 17:47:33 UTC

Import patches-unapplied version 7.2.1-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: 33da4d481ee6cea64c45cf21dd86abc2743e009a

New changelog entries:
  * New upstream (closes: #890394)
  * Fix for the "runq" command on Linux 4.14 and later kernels that contain
    commit cd9e61ed1eebbcd5dfad59475d41ec58d9b64b6a, titled "rbtree: cache
    leftmost node internally". Without the patch, the command fails with the
    error message "runq: invalid structure member offset: cfs_rq_rb_leftmost".
  * Fix to prevent a useless message during session inialization. Without the
    patch, if the highest possible node bit in the node_states[N_ONLINE]
    multi-word bitmask is set, then a message such as "crash:
    next_online_node: 256 is too large!" will be displayed.
  * Additional fixes for the ARM64 "bt" command for Linux 4.14 kernels. The
    patch corrects the contents of in-kernel exception frame register dumps,
    and properly transitions the backtrace from the IRQ stack to the process
    stack.
  * Implemented a new "search -T" option, which is identical to the "search
    -t" option, except that the search is restricted to the kernel stacks of
    active tasks.
  * Removal of the ARM64 "bt -o" option for Linux 4.14 and later kernels,
    along with several cleanups/readability improvements.
  * Fix for support of KASLR enabled kernels captured by the SADUMP dumpfile
    facility. SADUMP dumpfile headers do not contain phys_base or VMCOREINFO
    notes, so without this patch, the crash session fails during
    initialization with the message "crash: seek error: kernel virtual
    address: <address> type: "page_offset_base". This patch calculates the
    phys_base value and the KASLR offset using the IDTR and CR3 registers from
    the dumpfile header.
  * Implemented a new "ps -y policy" option to filter the task display by
    scheduling policy. Applicable to both standalone ps invocation as well as
    via foreach.
  * Fix for the "kmem -[sS]" options on Linux 4.14 and later kernels that
    contain commit 2482ddec670fb83717d129012bc558777cb159f7, titled "mm: add
    SLUB free list pointer obfuscation". Without the patch, there will
    numerous error messages of the type "kmem: <cache name> slab: <address>
    invalid freepointer: <obfuscated address>" if the kernel is configured
    with CONFIG_SLAB_FREELIST_HARDENED.
  * Fix for the validation of the bits located in the least significant bits of
    mem_section.section_mem_map pointers. Without the patch, the validation
    functions always returned valid, due to a coding error found by clang.
    However, it was never really a problem because it is extremely unlikely
    that an existing mem_section would ever be invalid.
  * Fix for the x86_64 kernel virtual address to physical address translation
    mechanism. Without the patch, when verifying that the PAGE_PRESENT bit is
    set in the top-level page table, it would always test positively, and the
    translation would continue parsing the remainder of the page tables. This
    would virtually never be a problem in practice because if the top-level
    page table entry existed, its PAGE_PRESENT bit would be set.
  * Removed a check for a negative block_size value which is always a
    non-negative unsigned value in the SADUMP header parsing function.
  * Removed a check for an impossible negative value when calculating the
    beginning address when applying the context value specified by the "search
    -x <count>" option.
  * Implemented a new "timer -C <cpu-specifier>" option that restricts the
    timer or hrtimer output to the timer queue data associated with one or
    more cpus. For multiple cpus, the cpu-specifier uses the standard comma
    or dash separated list format.
  * Fix for a "ps -l" regression introduced by the new "ps -y" option
    introduced above. Without the patch, the -l option generates a
    segmentation violation if not accompanied by a -C cpu specifier option.
  * Fix for the "kmem -i" and "kmem -V" options in Linux 4.8 and later kernels
    containing commit 75ef7184053989118d3814c558a9af62e7376a58, titled "mm,
    vmstat: add infrastructure for per-node vmstats". Without the patch, the
    CACHED line of "kmem -i" shows 0, and the VM_STAT section of "kmem -V" is
    missing entirely.
  * Fix for Linux 4.11 and later kernels that contain kernel commit
    4b3ef9daa4fc0bba742a79faecb17fdaaead083b, titled "mm/swap: split swap
    cache into 64MB trunks". Without the patch, the CACHED line of "kmem -i"
    may show nonsensical data.
  * Implemented a new "dev -D" option that is the same as "dev -d", but
    filters out the display of disks that have no I/O in progress.
  * If a line number request for a module text address initially fails, force
    the embedded gdb module to complete its two-stage strategy used for
    reading debuginfo symbol tables from module object files, and then retry
    the line number extraction. This automatically does what the "mod -r" or
    "crash --readnow" options accomplish.
  * Update for support of Linux 4.12 and later PPC64 kernels where the hash
    page table geometry accommodates a larger virtual address range. Without
    the patch, the virtual-to-physical translation of user space virtual
    addresses by "vm -p", "vtop", and "rd -u" may generate an invalid
    translation or otherwise fail.
  * Implemented a new "runq -T" option that displays the time lag of each CPU
    relative to the most recent runqueue timestamp.
  * Fix to support Linux 4.15 and later kernels that contain kernel commit
    e8cfbc245e24887e3c30235f71e9e9405e0cfc39, titled "pid: remove pidhash".
    The kernel's traditional usage of a pid_hash[] array to store PIDs has
    been replaced by an IDR radix tree, requiring a new crash plug-in function
    to gather the system's task set. Without the patch, the crash session
    fails during initialization with the error message "crash: cannot resolve
    init_task_union".
  * Fix for the "net" command when the network device listing has an unusually
    large number of IP addresses. In that case, without the patch, the
    command may generate a segmentation violation.
  * Fix for Linux 4.15 and later kernels that are configured with
    CONFIG_SPARSEMEM_EXTREME, and that contain kernel commit
    83e3c48729d9ebb7af5a31a504f3fd6aff0348c4, titled "mm/sparsemem: Allocate
    mem_section at runtime for CONFIG_SPARSEMEM_EXTREME=y". Without the
    patch, kernels configured with SPARSEMEM_EXTREME have changed the data
    type of "mem_section" from an array to a pointer, leading to errors in
    commands such as "kmem -p", "kmem -n", "kmem -s", and any other command
    that translates a physical address to its page struct address.
  * With the latest PPC64 NMI IPI changes, crash_ipi_callback is found
    multiple times on the stack of active non-panic tasks. Ensure that the
    symbol reference relates to an actual backtrace stack frame.
  * Update the starting virtual address of vmalloc space for kernels
    configured with CONFIG_X86_5LEVEL.
  * Update the X86_64 VSYSCALL_END address to reflect that it only contains 1
    page.
  * Prevent the X86_64 FILL_PML() macro from updating the internal
    machdep->machspec->last_pml4_read address every time a vmalloc'd kernel
    virtual address is translated.
  * Fix for the "bt" command in x86_64 kernels that contain, or have backports
    of, kernel commit 4950d6d48a0c43cc61d0bbb76fb10e0214b79c66, titled
    "x86/dumpstack: Remove 64-byte gap at end of irq stack". Without the
    patch, backtraces fail to transition from the IRQ stack back to the
    process stack, showing an error message such as "bt: cannot transition
    exception stack to IRQ stack to current process stack".
  * Initial pass for support of kernel page table isolation. The x86_64 "bt"
    command may indicate "bt: cannot transition from exception stack to
    current process stack" if the crash callback NMI occurred while an active
    task was running on the new entry trampoline stack. This has only been
    tested on the RHEL7 backport of the upstream patch because as of this
    commit, crash does not run on 4.15-rc kernels. Further changes may be
    required for upstream kernels, and distributions that implement the kernel
    changes differently than upstream.
  * Fix for the "bt" command and the "ps -s" option for zombie tasks whose
    kernel stacks have been freed/detached. Without the patch, the "bt"
    command indicates "bt: invalid kernel virtual address: 0 type: stack
    contents" and "bt: read of stack at 0 failed"; it will be changed to
    display "(no stack)". The "ps -s" option would fail prematurely upon
    reaching such a task, indicating "ps: invalid kernel virtual address: 0
    type: stack contents" and "ps: read of stack at 0 failed".
  * Fix for running on live systems on 4.15-rc2 and later kernels that are
    configured with CONFIG_RANDOMIZE_BASE and contain kernel commit
    668533dc0764b30c9dd2baf3ca800156f688326b, titled "kallsyms: take advantage
    of the new '%px' format". Without the patch, a live crash session does
    not show the "WARNING: kernel relocated ..." message expected with KASLR,
    and then displays the message "crash: cannot set context for pid: <pid>"
    prior to generating a SIGSEGV.
  * Fix for 4.15-rc5 and later x86_64 kernels that contain kernel commit
    c482feefe1aeb150156248ba0fd3e029bc886605, titled "x86/entry/64: Make
    cpu_entry_area.tss read-only". Without the patch, the addresses and sizes
    of the x86_64 exception stacks cannot be determined; therefore if a
    backtrace starts on one of the exception stacks, then the "bt" command
    will fail.
  * Additional fix for support of KASLR enabled kernels captured by the SADUMP
    dumpfile facility, where this patch fixes a problem when Page Table
    Isolation(PTI) is enabled. When PTI is enabled, bit 12 of CR3 register is
    used to split user space and kernel space. Also bit 11:0 is used for
    Process Context IDentifiers(PCID). To open an SADUMP dumpfile, the value
    of CR3 is used to calculate KASLR offset and phys_base; this patch masks
    the CR3 register value correctly for a PTI enabled kernel.
  * Second phase of future support for x86_64 5-level page tables. This patch
    is a cleanup/collaboration of the original logic used by the various vtop
    functions, where several new common functions have been added for
    extracting page table entries from PGD, P4D, PUD, PMD and PTE pages. The
    usage of the former PML4 and UPML pages have been replaced with the use of
    the common PGD page, and use the PUD page in 4-level page table
    translation. Support for 5-level page tables has been incorporated into
    the the existing x86_64_kvtop() and x86_64_uvtop_level4() functions.
    Backwards compatibility for older legacy kernels has been maintained. The
    third phase of support will automatically detect whether the kernel
    proper, and whether an individual user task, is utilizing 5-level page
    tables. This patch enables support for kernel-only 5-level page tables by
    entering the command line option "--machdep vm=5level".
  * Xen commit 615588563e99a23aaf37037c3fee0c413b051f4d (Xen 4.0.0.) extended
    the direct mapping to 5 TB. This area was previously reserved for future
    use, so it is OK to simply change the upper bound unconditionally.
  * Add a new "foreach gleader" qualifier option, restricting the output to
    user-space tasks that are thread group leaders.
  * Since Xen commit 666aca08175b ("sched: use the auto-generated list of
    schedulers") crash cannot open Xen vmcores because the "schedulers" symbol
    no longer exists. Xen 4.7 implemented schedulers as its own section in
    "xen/arch/x86/xen.lds.S", delimited by the two symbols
    "__start_schedulers_array" and "__end_schedulers_array". Without the
    patch, the crash session fails during initialization with the error
    message "crash: cannot resolve schedulers"
  * Fix the sample crash.ko memory driver to prevent an s390X kernel
    addressing exception. Legitimate pages of RAM that successfully pass the
    page_is_ram() and pfn_valid() verifier functions may not be provided by
    the s390x hypervisor, and the memcpy() from the non-existent memory to the
    bounce buffer panics the kernel. The patch replaces the the memcpy() call
    with probe_kernel_read().
  * Fix for the ARM64 "bt" command running against Linux 4.14 and later
    kernels. Without the patch, the backtraces of the active tasks in a
    kdump-generated dumpfile are truncated. Without the patch, the panic task
    will just show the "crash_kexec" frame and the kernel-entry user-space
    exception frame; the non-panic tasks will show their backtraces starting
    from the stackframe addresses captured in the per-cpu NT_PRSTATUS notes,
    and will not display the exception frame generated by the NMI callback,
    nor any stackframes on the IRQ stack.
  * Fix for the ARM64 "bt" command in kernels that contain commit
    30d88c0e3ace625a92eead9ca0ad94093a8f59fe, titled "arm64: entry: Apply BP
    hardening for suspicious interrupts from EL0". Without the patch, there
    may be invalid kernel kernel exception frames displayed on an active
    task's kernel stack, often below a stackframe of the
    "do_el0_ia_bp_hardening" function; the address translation of the PC and
    LR values in the the bogus exception frame will display "[unknown or
    invalid address]".

Author: Troy Heber
Author Date: 2018-02-16 17:47:33 UTC

Import patches-applied version 7.2.1-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 1e18628f4fc04b834801bff9fed812159ce1087a
Unapplied parent: 55055ccbcdf24d08034f81681dcbdf5195679f23

New changelog entries:
  * New upstream (closes: #890394)
  * Fix for the "runq" command on Linux 4.14 and later kernels that contain
    commit cd9e61ed1eebbcd5dfad59475d41ec58d9b64b6a, titled "rbtree: cache
    leftmost node internally". Without the patch, the command fails with the
    error message "runq: invalid structure member offset: cfs_rq_rb_leftmost".
  * Fix to prevent a useless message during session inialization. Without the
    patch, if the highest possible node bit in the node_states[N_ONLINE]
    multi-word bitmask is set, then a message such as "crash:
    next_online_node: 256 is too large!" will be displayed.
  * Additional fixes for the ARM64 "bt" command for Linux 4.14 kernels. The
    patch corrects the contents of in-kernel exception frame register dumps,
    and properly transitions the backtrace from the IRQ stack to the process
    stack.
  * Implemented a new "search -T" option, which is identical to the "search
    -t" option, except that the search is restricted to the kernel stacks of
    active tasks.
  * Removal of the ARM64 "bt -o" option for Linux 4.14 and later kernels,
    along with several cleanups/readability improvements.
  * Fix for support of KASLR enabled kernels captured by the SADUMP dumpfile
    facility. SADUMP dumpfile headers do not contain phys_base or VMCOREINFO
    notes, so without this patch, the crash session fails during
    initialization with the message "crash: seek error: kernel virtual
    address: <address> type: "page_offset_base". This patch calculates the
    phys_base value and the KASLR offset using the IDTR and CR3 registers from
    the dumpfile header.
  * Implemented a new "ps -y policy" option to filter the task display by
    scheduling policy. Applicable to both standalone ps invocation as well as
    via foreach.
  * Fix for the "kmem -[sS]" options on Linux 4.14 and later kernels that
    contain commit 2482ddec670fb83717d129012bc558777cb159f7, titled "mm: add
    SLUB free list pointer obfuscation". Without the patch, there will
    numerous error messages of the type "kmem: <cache name> slab: <address>
    invalid freepointer: <obfuscated address>" if the kernel is configured
    with CONFIG_SLAB_FREELIST_HARDENED.
  * Fix for the validation of the bits located in the least significant bits of
    mem_section.section_mem_map pointers. Without the patch, the validation
    functions always returned valid, due to a coding error found by clang.
    However, it was never really a problem because it is extremely unlikely
    that an existing mem_section would ever be invalid.
  * Fix for the x86_64 kernel virtual address to physical address translation
    mechanism. Without the patch, when verifying that the PAGE_PRESENT bit is
    set in the top-level page table, it would always test positively, and the
    translation would continue parsing the remainder of the page tables. This
    would virtually never be a problem in practice because if the top-level
    page table entry existed, its PAGE_PRESENT bit would be set.
  * Removed a check for a negative block_size value which is always a
    non-negative unsigned value in the SADUMP header parsing function.
  * Removed a check for an impossible negative value when calculating the
    beginning address when applying the context value specified by the "search
    -x <count>" option.
  * Implemented a new "timer -C <cpu-specifier>" option that restricts the
    timer or hrtimer output to the timer queue data associated with one or
    more cpus. For multiple cpus, the cpu-specifier uses the standard comma
    or dash separated list format.
  * Fix for a "ps -l" regression introduced by the new "ps -y" option
    introduced above. Without the patch, the -l option generates a
    segmentation violation if not accompanied by a -C cpu specifier option.
  * Fix for the "kmem -i" and "kmem -V" options in Linux 4.8 and later kernels
    containing commit 75ef7184053989118d3814c558a9af62e7376a58, titled "mm,
    vmstat: add infrastructure for per-node vmstats". Without the patch, the
    CACHED line of "kmem -i" shows 0, and the VM_STAT section of "kmem -V" is
    missing entirely.
  * Fix for Linux 4.11 and later kernels that contain kernel commit
    4b3ef9daa4fc0bba742a79faecb17fdaaead083b, titled "mm/swap: split swap
    cache into 64MB trunks". Without the patch, the CACHED line of "kmem -i"
    may show nonsensical data.
  * Implemented a new "dev -D" option that is the same as "dev -d", but
    filters out the display of disks that have no I/O in progress.
  * If a line number request for a module text address initially fails, force
    the embedded gdb module to complete its two-stage strategy used for
    reading debuginfo symbol tables from module object files, and then retry
    the line number extraction. This automatically does what the "mod -r" or
    "crash --readnow" options accomplish.
  * Update for support of Linux 4.12 and later PPC64 kernels where the hash
    page table geometry accommodates a larger virtual address range. Without
    the patch, the virtual-to-physical translation of user space virtual
    addresses by "vm -p", "vtop", and "rd -u" may generate an invalid
    translation or otherwise fail.
  * Implemented a new "runq -T" option that displays the time lag of each CPU
    relative to the most recent runqueue timestamp.
  * Fix to support Linux 4.15 and later kernels that contain kernel commit
    e8cfbc245e24887e3c30235f71e9e9405e0cfc39, titled "pid: remove pidhash".
    The kernel's traditional usage of a pid_hash[] array to store PIDs has
    been replaced by an IDR radix tree, requiring a new crash plug-in function
    to gather the system's task set. Without the patch, the crash session
    fails during initialization with the error message "crash: cannot resolve
    init_task_union".
  * Fix for the "net" command when the network device listing has an unusually
    large number of IP addresses. In that case, without the patch, the
    command may generate a segmentation violation.
  * Fix for Linux 4.15 and later kernels that are configured with
    CONFIG_SPARSEMEM_EXTREME, and that contain kernel commit
    83e3c48729d9ebb7af5a31a504f3fd6aff0348c4, titled "mm/sparsemem: Allocate
    mem_section at runtime for CONFIG_SPARSEMEM_EXTREME=y". Without the
    patch, kernels configured with SPARSEMEM_EXTREME have changed the data
    type of "mem_section" from an array to a pointer, leading to errors in
    commands such as "kmem -p", "kmem -n", "kmem -s", and any other command
    that translates a physical address to its page struct address.
  * With the latest PPC64 NMI IPI changes, crash_ipi_callback is found
    multiple times on the stack of active non-panic tasks. Ensure that the
    symbol reference relates to an actual backtrace stack frame.
  * Update the starting virtual address of vmalloc space for kernels
    configured with CONFIG_X86_5LEVEL.
  * Update the X86_64 VSYSCALL_END address to reflect that it only contains 1
    page.
  * Prevent the X86_64 FILL_PML() macro from updating the internal
    machdep->machspec->last_pml4_read address every time a vmalloc'd kernel
    virtual address is translated.
  * Fix for the "bt" command in x86_64 kernels that contain, or have backports
    of, kernel commit 4950d6d48a0c43cc61d0bbb76fb10e0214b79c66, titled
    "x86/dumpstack: Remove 64-byte gap at end of irq stack". Without the
    patch, backtraces fail to transition from the IRQ stack back to the
    process stack, showing an error message such as "bt: cannot transition
    exception stack to IRQ stack to current process stack".
  * Initial pass for support of kernel page table isolation. The x86_64 "bt"
    command may indicate "bt: cannot transition from exception stack to
    current process stack" if the crash callback NMI occurred while an active
    task was running on the new entry trampoline stack. This has only been
    tested on the RHEL7 backport of the upstream patch because as of this
    commit, crash does not run on 4.15-rc kernels. Further changes may be
    required for upstream kernels, and distributions that implement the kernel
    changes differently than upstream.
  * Fix for the "bt" command and the "ps -s" option for zombie tasks whose
    kernel stacks have been freed/detached. Without the patch, the "bt"
    command indicates "bt: invalid kernel virtual address: 0 type: stack
    contents" and "bt: read of stack at 0 failed"; it will be changed to
    display "(no stack)". The "ps -s" option would fail prematurely upon
    reaching such a task, indicating "ps: invalid kernel virtual address: 0
    type: stack contents" and "ps: read of stack at 0 failed".
  * Fix for running on live systems on 4.15-rc2 and later kernels that are
    configured with CONFIG_RANDOMIZE_BASE and contain kernel commit
    668533dc0764b30c9dd2baf3ca800156f688326b, titled "kallsyms: take advantage
    of the new '%px' format". Without the patch, a live crash session does
    not show the "WARNING: kernel relocated ..." message expected with KASLR,
    and then displays the message "crash: cannot set context for pid: <pid>"
    prior to generating a SIGSEGV.
  * Fix for 4.15-rc5 and later x86_64 kernels that contain kernel commit
    c482feefe1aeb150156248ba0fd3e029bc886605, titled "x86/entry/64: Make
    cpu_entry_area.tss read-only". Without the patch, the addresses and sizes
    of the x86_64 exception stacks cannot be determined; therefore if a
    backtrace starts on one of the exception stacks, then the "bt" command
    will fail.
  * Additional fix for support of KASLR enabled kernels captured by the SADUMP
    dumpfile facility, where this patch fixes a problem when Page Table
    Isolation(PTI) is enabled. When PTI is enabled, bit 12 of CR3 register is
    used to split user space and kernel space. Also bit 11:0 is used for
    Process Context IDentifiers(PCID). To open an SADUMP dumpfile, the value
    of CR3 is used to calculate KASLR offset and phys_base; this patch masks
    the CR3 register value correctly for a PTI enabled kernel.
  * Second phase of future support for x86_64 5-level page tables. This patch
    is a cleanup/collaboration of the original logic used by the various vtop
    functions, where several new common functions have been added for
    extracting page table entries from PGD, P4D, PUD, PMD and PTE pages. The
    usage of the former PML4 and UPML pages have been replaced with the use of
    the common PGD page, and use the PUD page in 4-level page table
    translation. Support for 5-level page tables has been incorporated into
    the the existing x86_64_kvtop() and x86_64_uvtop_level4() functions.
    Backwards compatibility for older legacy kernels has been maintained. The
    third phase of support will automatically detect whether the kernel
    proper, and whether an individual user task, is utilizing 5-level page
    tables. This patch enables support for kernel-only 5-level page tables by
    entering the command line option "--machdep vm=5level".
  * Xen commit 615588563e99a23aaf37037c3fee0c413b051f4d (Xen 4.0.0.) extended
    the direct mapping to 5 TB. This area was previously reserved for future
    use, so it is OK to simply change the upper bound unconditionally.
  * Add a new "foreach gleader" qualifier option, restricting the output to
    user-space tasks that are thread group leaders.
  * Since Xen commit 666aca08175b ("sched: use the auto-generated list of
    schedulers") crash cannot open Xen vmcores because the "schedulers" symbol
    no longer exists. Xen 4.7 implemented schedulers as its own section in
    "xen/arch/x86/xen.lds.S", delimited by the two symbols
    "__start_schedulers_array" and "__end_schedulers_array". Without the
    patch, the crash session fails during initialization with the error
    message "crash: cannot resolve schedulers"
  * Fix the sample crash.ko memory driver to prevent an s390X kernel
    addressing exception. Legitimate pages of RAM that successfully pass the
    page_is_ram() and pfn_valid() verifier functions may not be provided by
    the s390x hypervisor, and the memcpy() from the non-existent memory to the
    bounce buffer panics the kernel. The patch replaces the the memcpy() call
    with probe_kernel_read().
  * Fix for the ARM64 "bt" command running against Linux 4.14 and later
    kernels. Without the patch, the backtraces of the active tasks in a
    kdump-generated dumpfile are truncated. Without the patch, the panic task
    will just show the "crash_kexec" frame and the kernel-entry user-space
    exception frame; the non-panic tasks will show their backtraces starting
    from the stackframe addresses captured in the per-cpu NT_PRSTATUS notes,
    and will not display the exception frame generated by the NMI callback,
    nor any stackframes on the IRQ stack.
  * Fix for the ARM64 "bt" command in kernels that contain commit
    30d88c0e3ace625a92eead9ca0ad94093a8f59fe, titled "arm64: entry: Apply BP
    hardening for suspicious interrupts from EL0". Without the patch, there
    may be invalid kernel kernel exception frames displayed on an active
    task's kernel stack, often below a stackframe of the
    "do_el0_ia_bp_hardening" function; the address translation of the PC and
    LR values in the the bogus exception frame will display "[unknown or
    invalid address]".

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-applied version 7.1.9-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: f302b5d376cad2f307b7bf770a871b20e24f9b09
Unapplied parent: a0f49e170b6893435ca7c6f1a4e7e752d5352b83

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-applied version 7.1.9-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: f302b5d376cad2f307b7bf770a871b20e24f9b09
Unapplied parent: a0f49e170b6893435ca7c6f1a4e7e752d5352b83

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-unapplied version 7.1.9-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: 078e879eaa0cb4459792a7e41fe36059a050e536

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-unapplied version 7.1.9-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: 078e879eaa0cb4459792a7e41fe36059a050e536

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-unapplied version 7.1.9-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: 078e879eaa0cb4459792a7e41fe36059a050e536

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Troy Heber
Author Date: 2017-07-26 18:37:08 UTC

Import patches-applied version 7.1.9-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: f302b5d376cad2f307b7bf770a871b20e24f9b09
Unapplied parent: a0f49e170b6893435ca7c6f1a4e7e752d5352b83

New changelog entries:
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Build crash on all Linux
    architectures (Closes: #763856, #757450)
  * Patch from Balint Reczey <rbalint@ubuntu.com>: Continuous integration
    tests can fail due to missing packages for the running kernel and missing
    *-updates packages (Closes: #869367)
  * Fixes to address three gcc-7.0.1 compiler warnings that are generated when
    building with "make warn". The warning types are "[-Wnonnull]" in
    filesys.c, and "[-Wformat-overflow=]" in kernel.c and cmdline.c.
  * Fix for the PPC64 "mach -o" option to update the OPAL console buffer size
    from 256K to 1MB, based upon the latest skiboot firmware source.
  * Fix for the "mod -[sS]" option to prevent the erroneous reassignment of
    one or more symbol values of a kernel module. Without the patch, when
    loading a kernel module, a message may indicate "mod: <module>: last
    symbol: <symbol> is not _MODULE_END_<module>?" may be displayed, and one
    or more symbols may be reassigned an incorrect symbol value. If none of
    the erroneous symbol value reassignments are beyond the end of the
    module's address space, then there will be no message.
  * Linux 4.10 commit 401721ecd1dcb0a428aa5d6832ee05ffbdbffbbe finally exports
    the x86_64 "phys_base" value in the VMCOREINFO note, so utilize it
    whenever it exists.
  * Implemented a new "log -a" option that dumps the audit logs remaining in
    kernel audit buffers that have not been copied out to the user-space audit
    daemon.
  * Fix for the "kmem <address>" option and the "search" command in x86_64
    kernels that contain, or have backports of, kernel commit
    7c1da8d0d046174a4188b5729d7579abf3d29427, titled "crypto: sha - SHA1
    transform x86_64 AVX2", which introduced an "_end" text symbol. Without
    the patch, if a base kernel symbol address that is larger than the "_end"
    text symbol is passed to "kmem <address>", its symbol/filename information
    will not be displayed. Also, when the "search" command scans the
    __START_KERNEL_map region that contains kernel text and static data, the
    search will be truncated to stop at the "_end" text symbol address.
  * Enhancement for the determination of the ARM64 "kimage_voffset" value in
    Linux 4.6 and later kernels if an ELF format dumpfile does not contain its
    value in a VMCOREINFO note, or when running against live systems using
    /dev/mem, /proc/kcore, or an older version of /dev/crash.
  * Optimization of the "kmem -f <address>" and "kmem <r;address>" options to
    significantly reduce the amount of time to complete the buddy allocator
    free-list scan for the target address. On very large memory systems, the
    patch may reduce the time spent by several orders of magnitude.
  * Fix for a compilation error if glibc-2.25 or later has been installed on
    the host build machine. Without the patch, the build fails with the error
    message "amd64-linux-nat.c:496:1: error: conflicting types for
    'ps_get_thread_area'".
  * Fix for the "list -[hH]" options if a list_head.next pointer is
    encountered that contains an invalid NULL pointer. Without the patch, the
    "list -[hH]" options would complete/continue as if the NULL were a
    legitimate end-of-list indicator, and no error would be reported.
  * Provide basic Huge Page usage as part of "kmem -i" output, showing the
    total amount of memory allocated for huge pages, and the amount of the
    total that is free.
  * Fix for the determination of the x86_64 "phys_base" value when it is not
    passed in the VMCOREINFO data of ELF vmcores. Without the patch, it is
    possible that the base address of the vmalloc region is unknown and
    initialized to an incorrect default address during the very early stages
    of initialization, which causes the parsing of the PT_LOAD segments for
    the START_KERNEL_map region to fail.
  * Fix for the "dis" command to detect duplicate symbols in the case of a
    "symbol+offset" argument where the duplicates are contiguous in the symbol
    list. In addition, reject "symbol+offset" arguments if the resultant
    address goes beyond the end of the function.
  * Fix for the "set scope" option if the kernel was configured with
    CONFIG_RANDOMIZE_BASE. Without the patch, the command fails with the
    message "set: gdb cannot find text block for address: <symbol>". This
    also affects extension modules that call gdb_set_crash_scope() when
    running with KASLR kernels.
  * Fix for the extensions/trace.c extension module to account for Linux 4.7
    kernel commit 9b94a8fba501f38368aef6ac1b30e7335252a220, which changed the
    ring_buffer_per_cpu.nr_pages member from an int to a long. Without the
    patch, the trace.so extension module fails to load on big-endian machines,
    indicating "extend: Num of pages is less than 0".
  * Fix for the extensions/trace.c extension module when running on the ppc64
    architecture. Without the patch, the trace.so extension module fails to
    load, indicating "extend: invalid text address: ring_buffer_read". On the
    ppc64 architecture, the text symbol is ".ring_buffer_read".
  * Fix for the ARM64 "bt" command. Without the patch, the backtrace of a
    non-panicking active task generates a segmentation violation when
    analyzing Android 4.4-based dumpfiles.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-unapplied version 7.1.8-1ubuntu1 to ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 3ee74a11843470c3639aef93b6132dceab3e3415

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-applied version 7.1.8-1ubuntu1 to applied/ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 4366aa0505592b2feb88e58d51bb9914f29888cb
Unapplied parent: c6062b27744af58a5c2817a0abda92feecc003ed

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-applied version 7.1.8-1ubuntu1 to applied/ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 4366aa0505592b2feb88e58d51bb9914f29888cb
Unapplied parent: c6062b27744af58a5c2817a0abda92feecc003ed

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-applied version 7.1.8-1ubuntu1 to applied/ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 4366aa0505592b2feb88e58d51bb9914f29888cb
Unapplied parent: c6062b27744af58a5c2817a0abda92feecc003ed

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-unapplied version 7.1.8-1ubuntu1 to ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 3ee74a11843470c3639aef93b6132dceab3e3415

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Stefan Bader
Author Date: 2017-03-17 13:03:05 UTC

Import patches-unapplied version 7.1.8-1ubuntu1 to ubuntu/zesty-proposed

Imported using git-ubuntu import.

Changelog parent: 3ee74a11843470c3639aef93b6132dceab3e3415

New changelog entries:
  * Merge from Debian unstable (LP: #1672361). Remaining changes:
    - Build for armhf, ppc64el.
    - Minor fixes for live autopkgtest.

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-unapplied version 7.0.3-3ubuntu4.5 to ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: 99b70f051d24a3aadaa9922cf4dd9eab57cb947d

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-applied version 7.0.3-3ubuntu4.5 to applied/ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: b68a1d914b6db1ddc375292e179fce65f08670fe
Unapplied parent: a16ddb03f0495ec8c20e271baaa0be5696344d40

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-applied version 7.0.3-3ubuntu4.5 to applied/ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: b68a1d914b6db1ddc375292e179fce65f08670fe
Unapplied parent: a16ddb03f0495ec8c20e271baaa0be5696344d40

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-applied version 7.0.3-3ubuntu4.5 to applied/ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: b68a1d914b6db1ddc375292e179fce65f08670fe
Unapplied parent: a16ddb03f0495ec8c20e271baaa0be5696344d40

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-unapplied version 7.0.3-3ubuntu4.5 to ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: 99b70f051d24a3aadaa9922cf4dd9eab57cb947d

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Robert Hooker
Author Date: 2016-12-01 17:03:17 UTC

Import patches-unapplied version 7.0.3-3ubuntu4.5 to ubuntu/trusty-proposed

Imported using git-ubuntu import.

Changelog parent: 99b70f051d24a3aadaa9922cf4dd9eab57cb947d

New changelog entries:
  * Fix for: Unable to analyse vmcore/dump via crash due to bad kernel
    debug info build (LP: #1607894).

Author: Troy Heber
Author Date: 2016-12-02 18:57:29 UTC

Import patches-unapplied version 7.1.7-1 to debian/sid

Imported using git-ubuntu import.

Changelog parent: b40a457c318a8ef0fa67cb5ca12404133a670aa8

New changelog entries:
  * New upstream 7.1.7 (Closes: #843731):
  * Remove -pie compile option (Closes: #844831)
  * Set the default 32-bit MIPS HZ value to 100 if the in-kernel config data
    is unavailable, and have the "mach" command display the value.
  * Enable SPARSEMEM support on 32-bit MIPS by setting SECTION_SIZE_BITS and
    MAX_PHYSMEM_BITS.
  * Fix for Linux 4.9-rc1 commits 15f4eae70d365bba26854c90b6002aaabb18c8aa and
    c65eacbe290b8141554c71b2c94489e73ade8c8d, which have introduced a new
    CONFIG_THREAD_INFO_IN_TASK configuration. This configuration moves each
    task's thread_info structure from the base of its kernel stack into its
    task_struct. Without the patch, the crash session fails during
    initialization with the error "crash: invalid structure member offset:
    thread_info_cpu".
  * Fixes for the gathering of the active task registers from 32-bit MIPS
    dumpfiles: (1) If ELF notes are not available, read them from the kernel's
    crash_notes. (2) If an online CPUs did not save its ELF notes, then
    adjust the mapping of each ELF note to its CPU accordingly.
  * Add support for "help -r" on 32-bit MIPS to display the registers for each
    CPU from a dumpfile.
  * Fix for Linux 4.9-rc1 commit 0100301bfdf56a2a370c7157b5ab0fbf9313e1cd,
    which rewrote the x86_64 switch_to() code by embedding the call to
    __switch_to() inside a new __switch_to_asm() assembly code ENTRY()
    function. Without the patch, the message "crash: cannot determine thread
    return address" gets displayed during initialization, and the "bt" command
    shows frame #0 starting at "schedule" instead of "__schedule".
  * When each x86_64 per-cpu cpu_tss.x86_tss.ist[] array member (or in older
    kernels, each per-cpu init_tss.x86_hw_tss.ist[] array member), is compared
    with its associated per-cpu orig_ist.ist[] array member, ensure that both
    exception stack pointers have been initialized (non-NULL) before printing
    a WARNING message if they don't match.
  * Fix for a possible segmentation violation when analyzing Linux 4.7 x86_64
    kernels that are configured with CONFIG_RANDOMIZE_BASE. Depending upon
    the randomized starting address of the kernel text and static data, a
    segmentation violation may occur during session initialization, just after
    the patching of the gdb minimal_symbol values message.
  * Restore the x86_64 "dis" command's symbolic translation of jump or call
    target addresses if the kernel was configured with CONFIG_RANDOMIZE_BASE.
  * Fix for the 32-bit MIPS "bt" command to prevent an empty display (task
    header only) for an active task if the epc register in its exception frame
    contains 00000000.
  * Fix for support of Linux 4.7 and later x86_64 ELF kdump vmcores from
    kernels configured with CONFIG_RANDOMIZE_BASE. Without the patch, the
    crash session may fail during initialization with the message "crash:
    vmlinux and vmcore do not match!".
  * Fix for the x86_64 "mach" command display of the vmemmap base address in
    Linux 4.9 and later kernels configured with CONFIG_RANDOMIZE_BASE.
    Without the patch, the command shows a value of ffffea0000000000 next to
    "KERNEL VMEMMAP BASE".
  * Since the Linux 3.10 release, the kernel has offered the ability to create
    multiple independent ftrace buffers. At present, however, the "trace.c"
    extension module is only able to extract the primary buffer. This patch
    refactors the trace.c extension module so that the global instance is
    passed around as a parameter rather than accessing it directly, and then
    locates all of the available instances and extracts the data from each of
    them.
  * Fix for the s390x "bt" command for active tasks. Since the commit above
    in this crash-7.1.7 release that added support for the new
    CONFIG_THREAD_INFO_IN_TASK configuration, the backtrace of active tasks
    can be incomplete.
  * In collaboration with an update to the /dev/crash kernel driver, fix for
    Linux 4.6 commit a7f8de168ace487fa7b88cb154e413cf40e87fc6, which allows
    the ARM64 kernel image to be loaded anywhere in physical memory. Without
    the patch, attempting to run live on an ARM64 Linux 4.6 and later kernel
    may display the warning message "WARNING: cannot read linux_banner
    string", and then fails with the message "crash: vmlinux and /dev/crash do
    not match!". Version 1.3 of the crash driver is required, which
    introduces a new ioctl command that retrieves the ARM64-only
    "kimage_voffset" value that is required for virtual-to-physical address
    translation.
  * Update of the sample memory_driver/crash.c /dev/crash kernel driver to
    version 1.3, which adds support for Linux 4.6 and later ARM64 kernels,
    kernels configured with CONFIG_HARDENED_USERCOPY, and S390X kernels use
    xlate_dev_mem_ptr() and unxlate_dev_mem_ptr() instead of kmap() and
    kunmap().
  * New upstream 7.1.5 (Closes: #843731):
  * Introduction of support for "live" ramdump files, such as those that are
    specified by the QEMU mem-path argument of a memory-backend-file object.
    This allows the running of a live crash session against a QEMU guest from
    the host machine. In this example, the /tmp/MEM file on a QEMU host
    represents the guest's physical memory:
     $ qemu-kvm ...other-options... \
     -object memory-backend-file,id=MEM,size=128m,mem-path=/tmp/MEM,share=on \
     -numa node,memdev=MEM -m 128
    and a live session run can be run against the guest kernel like so:
     $ crash <path-to-guest-vmlinux> live:/tmp/MEM@0
    By prepending the ramdump image name with "live:", the crash session will
    act as if it were running a normal live session.
  * Fix for the support of ELF vmcores created by the KVM "virsh dump
    --memory-only" facility if the guest kernel was not configured with
    CONFIG_KEXEC, or CONFIG_KEXEC_CORE in Linux 4.3 and later kernels.
    Without the patch, the crash session fails during initialization with the
    message "crash: cannot resolve kexec_crash_image".
  * Added support for x86_64 ramdump files. Without the patch, the crash
    session fails immediately with the message "ramdump: unsupported machine
    type: X86_64".
  * Fix for a "[-Werror=misleading-indentation]" compiler warning that is
    generated by gdb-7.6/bfd/elf64-s390.c when building S390X in a Fedora
    Rawhide environment with gcc-6.0.0
  * Recognize and parse the new QEMU_VM_CONFIGURATION and QEMU_VM_FOOTER
    sections used for live migration of KVM guests, which are seen in the
    "kvmdump" format generated if "virsh dump" is used without the
    "--memory-only" option.
  * Fix for Linux commit edf14cdbf9a0e5ab52698ca66d07a76ade0d5c46, which has
    appended a NULL entry as the final member of the pageflag_names[] array.
    Without the patch, a message that indicates "crash: failed to read
    pageflag_names entry" is displayed during session initialization in Linux
    4.6 kernels.
  * Fix for Linux commit 0139aa7b7fa12ceef095d99dc36606a5b10ab83a, which
    renamed the page._count member to page._refcount. Without the patch,
    certain "kmem" commands fail with the "kmem: invalid structure member
    offset: page_count".
  * Fix for an ARM64 crash-7.1.5 "bt" regression for a task that has called
    panic(). Without the patch, the backtrace may fail with a message such as
    "bt: WARNING: corrupt prstatus? pstate=0x20000000, but no user frame
    found" followed by "bt: WARNING: cannot determine starting stack frame for
    task <address>". The pstate register warning will still be displayed (as
    it is essentially a kdump bug), but the backtrace will proceed normally.
  * Fix for the ARM64 "bt" command in Linux 4.5 and later kernels which use
    per-cpu IRQ stacks. Without the patch, if an active non-crashing task was
    running in user space when it received the shutdown IPI from the crashing
    task, the "-- <IRQ stack> ---" transition marker from the IRQ stack to the
    process stack is not displayed, and a message indicating "bt: WARNING:
    arm64_unwind_frame: on IRQ stack: oriq_sp: <address> fp: 0 (?)" gets
    displayed.
  * Fix for the ARM64 "bt" command in Linux 4.5 and later kernels which are
    not configured with CONFIG_FUNCTION_GRAPH_TRACER. Without the patch,
    backtraces that originate from a per-cpu IRQ stack will dump an invalid
    exception frame before transitioning to the process stack.
  * Introduction of ARM64 support for 4K pages with 4-level page tables and 48
    VA bits.
  * Implemented support for the redesigned ARM64 kernel virtual memory layout
    and associated KASLR support that was introduced in Linux 4.6. The kernel
    text and static data has been moved from unity-mapped memory into the
    vmalloc region, and its start address can be randomized if
    CONFIG_RANDOMIZE_BASE is configured. Related support is being put into
    the kernel's kdump code, the kexec-tools package, and makedumpfile(8);
    with that in place, the analysis of Linux 4.6 ARM64 dumpfiles with or
    without KASLR enabled should work normally by entering "crash vmlinux
    vmcore". On live systems, Linux 4.6 ARM64 kernels will only work
    automatically if CONFIG_RANDOMIZE_BASE is not configured. Unfortunately,
    if CONFIG_RANDOMIZE_BASE is configured on a live system, two --machdep
    command line arguments are required, at least for the time being. The
    arguments are:
     --machdep phys_offset=<base physical address>
     --machdep kimage_voffset=<kernel kimage_voffset value>
    Without the patch, any attempt to analyze a Linux 4.6 ARM64 kernel fails
    during initialization with a stream of "read error" messages followed by
    "crash: vmlinux and vmcore do not match!".
  * Linux 3.15 and later kernels configured with CONFIG_RANDOMIZE_BASE could
    be identified because of the "randomize_modules" kernel symbol, and if it
    existed, the "--kaslr=<offset>" and/or "--kaslr=auto" options were
    unnecessary. Since the "randomize_modules" symbol was removed in Linux
    4.1, this patch has replaced the KASLR identifier with the
    "module_load_offset" symbol, which was also introduced in Linux 3.15, but
    still remains.
  * Improvement of the ARM64 "bt -f" display such that in most cases, each
    stack frame level delimiter will be set to the stack address location
    containing the old FP and old LR pair.
  * Fix for the introduction of ARM64 support for 64K pages with 3-level page
    tables in crash-7.1.5, which fails to translate user space virtual
    addresses. Without the patch, "vtop <user-space address>" fails to
    translate all user-space addresses, and any command that needs to either
    translate or read user-space memory, such as "vm -p", "ps -a", and "rd -u"
    will fail.
  * Enhancement of the error message generated by the "tree -t radix" option
    when a duplicate entry is encountered. Without the patch, the error
    message shows the address of the radix_tree_node that contains the
    duplicate entry, for example, "tree: duplicate tree entry:
    <radix_tree_node>". It has been changed to also display the
    radix_tree_node.slots[] array index and the duplicate entry value, for
    example, "tree: duplicate tree entry: radix_tree_node: <radix_tree_node>
    slots[<index>]: <entry>".
  * Introduction of a new "bt -v" option that checks the kernel stack of all
    tasks for evidence of stack overflows. It does so by verifying the
    thread_info.task address, ensuring the thread_info.cpu value is a valid
    cpu number, and checking the end of the stack for the STACK_END_MAGIC
    value.
  * Fix to recognize a kernel thread that has user space virtual memory
    attached to it. While kernel threads typically do not have an mm_struct
    referencing a user-space virtual address space, they can either
    temporarily reference one for a user-space copy operation, or in the case
    of KVM "vhost" kernel threads, keep a reference to the user space of the
    "quem-kvm" task that created them. Without the patch, they will be
    mistaken for user tasks; the "bt" command will display an invalid
    kernel-entry exception frame that indicates "[exception RIP: unknown or
    invalid address]", the "ps" command will not enclose the command name with
    brackets, and the "ps -[uk]" and "foreach [user|kernel]" options will show
    the kernel thread as a user task.
  * Fix for the "bt -[eE]" options on ARM64 to recognize kernel exception
    frames in VHE enabled systems, in which the kernel runs in EL2.
  * Fix for the extensions/trace.c extension module to account for the Linux
    4.7 kernel commit dcb0b5575d24 that changed the bit index for the
    TRACE_EVENT_FL_TRACEPOINT flag. Without the patch, the "extend" command
    fails to load the trace.so module, with the error message "extend:
    /path/to/crash/extensions/trace.so: no commands registered: shared object
    unloaded". The patch reads the flag's enum value dynamically instead of
    using a hard-coded value.
  * Incorporated Takahiro Akashi's alternative backtrace method as a "bt"
    option, which can be accessed using "bt -o", and where "bt -O" will toggle
    the original and optional methods as the default. The original backtrace
    method has adopted two changes/features from the optional method:
     (1) ORIG_X0 and SYSCALLNO registers are not displayed in kernel
         exception frames.
     (2) stackframe entry text locations are modified to be the PC
         address of the branch instruction instead of the subsequent
         "return" PC address contained in the stackframe link register.
   Accordingly, these are the essential differences between the original and
   optional methods:
     (1) optional: the backtrace will start with the IPI exception frame
         located on the process stack.
     (2) original: the starting point of backtraces for the active,
         non-crashing, tasks, will continue to have crash_save_cpu()
         on the IRQ stack as the starting point.
     (3) optional: the exception entry stackframe adjusted to be located
         farther down in the IRQ stack.
     (4) optional: bt -f does not display IRQ stack memory above the
         adjusted exception entry stackframe.
     (5) optional: may display "(Next exception frame might be wrong)".
  * Fix for the failure of the "sym <symbol>" option in the extremely unlikely
    case where the symbol's name string is composed entirely of hexadecimal
    characters. For example, without the patch, "sym e820" fails with the
    error message "sym: invalid address: e820".
  * Fix for the failure of the "dis <symbol>" option in the extremely unlikely
    case where the symbol's name string is composed entirely of hexadecimal
    characters. For example, without the patch, "dis f" fails with the error
    message "dis: WARNING: f: no associated kernel symbol found" followed by
    "0xf: Cannot access memory at address 0xf".
  * Fix for the X86_64 "bt -R <symbol>" option if the only reference to the
    kernel text symbol in a backtrace is contained within the "[exception RIP:
    <symbol+offset>]" line of an exception frame dump. Without the patch, the
    reference will only be picked up if the exception RIP's hexadecimal
    address value is used.
  * Fix for the ARM64 "bt -R <symbol>" option if the only reference to the
    kernel text symbol in a backtrace is contained within the "[PC: <address>
    [<symbol+offset>]" line of an exception frame dump. Without the patch,
    the reference will only be picked up if the PC's hexadecimal address value
    is used.
  * Fix for the gathering of module symbol name strings during session
    initialization. In the unlikely case where the ordering of module symbol
    name strings does not match the order of the kernel_symbol structures, a
    faulty module symbol list entry may be created that contains a bogus name
    string.
  * Fix the PERCENTAGE of total output of the "kmem -i" SWAP USED line when
    the system has no swap pages at all. Without the patch, both the PAGES
    and TOTAL columns show values of zero, but it confusingly shows "100% of
    TOTAL SWAP", which upon first glance may seem to indicate potential memory
    pressure.
  * Enhancement to determine structure member data if the member is contained
    within an anonymous structure or union. Without the patch, it is
    necessary to parse the output of a discrete gdb "printf" command to
    determine the offset of such a structure member.
  * Speed up session initialization by attempting MEMBER_OFFSET_INIT() before
    falling back to ANON_MEMBER_OFFSET_INIT() in several known cases of
    structure members that are contained within anonymous structures.
  * Implemented new "list -S" and "tree -S" options that are similar to each
    command's -s option, but instead of parsing gdb output, member values are
    read directly from memory, so the command is much faster for 1-, 2-, 4-,
    and 8-byte members.
  * Fix to recognize and support x86_64 Linux 4.8-rc1 and later kernels that
    are configured with CONFIG_RANDOMIZE_MEMORY, which randomizes the base
    addresses of the kernel's unity-map address (PAGE_OFFSET), and the vmalloc
    region. Without the patch, the crash utility fails with a segmentation
    violation during session initialization on a live system, or will generate
    a number of WARNING messages followed by the fatal error message "crash:
    vmlinux and <dumpfile name> do not match!" with dumpfiles.
  * Fix for Linux 4.1 commit d0a0de21f82bbc1737ea3c831f018d0c2bc6b9c2, which
    renamed the x86_64 "init_tss" per-cpu variable to "cpu_tss". Without the
    patch, the addresses of the 4 per-cpu exception stacks cannot be
    determined, which causes backtraces that originate on any of the per-cpu
    DOUBLEFAULT, NMI, DEBUG, or MCE stacks to be truncated.
  * With the introduction of radix MMU in Power ISA 3.0, there are changes in
    kernel page table management accommodating it. This patch series makes
    appropriate changes here to work for such kernels. Also, this series
    fixes a few bugs along the way:
     ppc64: fix vtop page translation for 4K pages
     ppc64: Use kernel terminology for each level in 4-level page table
     ppc64/book3s: address changes in kernel v4.5
     ppc64/book3s: address change in page flags for PowerISA v3.0
     ppc64: use physical addresses and unfold pud for 64K page size
     ppc64/book3s: support big endian Linux page tables
    The patches are needed for Linux v4.5 and later kernels on all ppc64
    hardware.
  * Fix for Linux 4.8-rc1 commit 500462a9de657f86edaa102f8ab6bff7f7e43fc2, in
    which Thomas Gleixner redesigned the kernel timer mechanism to switch to a
    non-cascading wheel. Without the patch, the "timer" command fails with
    the message "timer: zero-size memory allocation! (called from <address>)"
  * Support for PPC64/BOOK3S virtual address translation for radix MMU. As
    both radix and hash MMU are supported in a single kernel on Power ISA 3.0
    based server processors, identify the current MMU type and set page table
    index values accordingly. Also, in Linux 4.7 and later kernels,
    PPC64/BOOK3S uses the same masked bit values in page table entries for 4K
    and 64K page sizes.
  * Change the RESIZEBUF() macro so that it will accept buffer pointers that
    are not declared as "char *" types. Change two prior direct callers of
    resizebuf() to use RESIZEBUF(), and fix two prior users of RESIZEBUF() to
    correctly calculate the need to resize their buffers.
  * Fix for the "trace.so" extension module to properly recognize Linux 3.15
    and later kernels. In crash-7.1.6, the MEMBER_OFFSET() macro has been
    improved so that it is able to recognize members of embedded anonymous
    structures. However, the module's manner of recognizing Linux 3.15 and
    later kernels depended upon MEMBER_OFFSET() failing to handle anonymous
    members, and therefore the improvement prevented the module from
    successfully loading.
  * If a "struct" command address argument is expressed using the per-cpu
    "symbol:cpuspec" format, and the symbol is a pointer type, i.e., not the
    address of the structure, display a WARNING message.
  * Exclude ARM64 kernel module linker mapping symbols like "$d" and "$x" as
    is done with 32-bit ARM. Without the patch, a crash session may fail
    during the "gathering module symbol data" stage with a message similar to
    "crash: store_module_symbols_v2: total: 15 mcnt: 16".
  * Enhancement to the ARM64 "dis" command when the kernel has enabled KASLR.
    When KASLR is enabled on ARM64, a function call between a module and the
    base kernel code will be done via a veneer (PLT) if the displacement is
    more than +/-128MB. As a result, disassembled code will show a branch to
    the in-module veneer location instead of the in-kernel target location.
    To avoid confusion, the output of the "dis" command will translate the
    veneer location to the target location preceded by "plt:", for example,
    "<plt:printk>".
  * Improvement of the "dev -d" option to display I/O statics for disks whose
    device driver uses the blk-mq interface. Currently "dev -d" always
    displays 0 in all fields for the blk-mq disk because blk-mq does not
    increment/decrement request_list.count[2] on I/O creation and I/O
    completion. The following values are used in blk-mq in such situations:
     - I/O creation: blk_mq_ctx.rq_dispatched[2]
     - I/O completion: blk_mq_ctx.rq_completed[2]
    So, we can get the counter of in-progress I/Os as follows: in progress
    I/Os == rq_dispatched - rq_completed This patch displays the result of
    above calculation for the disk. It determines whether the device driver
    uses blk-mq if the request_queue.mq_ops is not NULL. The "DRV" field is
    displayed as "N/A(MQ)" if the value for in-flight in the device driver
    does not exist for blk-mq.

Author: Troy Heber
Author Date: 2016-12-02 18:57:29 UTC

Import patches-applied version 7.1.7-1 to applied/debian/sid

Imported using git-ubuntu import.

Changelog parent: 65bb0b813289fbfab63c7fd14d70382d12625653
Unapplied parent: 8f75153918a4558e502f277a14ef7f3ceb54d5ff

New changelog entries:
  * New upstream 7.1.7 (Closes: #843731):
  * Remove -pie compile option (Closes: #844831)
  * Set the default 32-bit MIPS HZ value to 100 if the in-kernel config data
    is unavailable, and have the "mach" command display the value.
  * Enable SPARSEMEM support on 32-bit MIPS by setting SECTION_SIZE_BITS and
    MAX_PHYSMEM_BITS.
  * Fix for Linux 4.9-rc1 commits 15f4eae70d365bba26854c90b6002aaabb18c8aa and
    c65eacbe290b8141554c71b2c94489e73ade8c8d, which have introduced a new
    CONFIG_THREAD_INFO_IN_TASK configuration. This configuration moves each
    task's thread_info structure from the base of its kernel stack into its
    task_struct. Without the patch, the crash session fails during
    initialization with the error "crash: invalid structure member offset:
    thread_info_cpu".
  * Fixes for the gathering of the active task registers from 32-bit MIPS
    dumpfiles: (1) If ELF notes are not available, read them from the kernel's
    crash_notes. (2) If an online CPUs did not save its ELF notes, then
    adjust the mapping of each ELF note to its CPU accordingly.
  * Add support for "help -r" on 32-bit MIPS to display the registers for each
    CPU from a dumpfile.
  * Fix for Linux 4.9-rc1 commit 0100301bfdf56a2a370c7157b5ab0fbf9313e1cd,
    which rewrote the x86_64 switch_to() code by embedding the call to
    __switch_to() inside a new __switch_to_asm() assembly code ENTRY()
    function. Without the patch, the message "crash: cannot determine thread
    return address" gets displayed during initialization, and the "bt" command
    shows frame #0 starting at "schedule" instead of "__schedule".
  * When each x86_64 per-cpu cpu_tss.x86_tss.ist[] array member (or in older
    kernels, each per-cpu init_tss.x86_hw_tss.ist[] array member), is compared
    with its associated per-cpu orig_ist.ist[] array member, ensure that both
    exception stack pointers have been initialized (non-NULL) before printing
    a WARNING message if they don't match.
  * Fix for a possible segmentation violation when analyzing Linux 4.7 x86_64
    kernels that are configured with CONFIG_RANDOMIZE_BASE. Depending upon
    the randomized starting address of the kernel text and static data, a
    segmentation violation may occur during session initialization, just after
    the patching of the gdb minimal_symbol values message.
  * Restore the x86_64 "dis" command's symbolic translation of jump or call
    target addresses if the kernel was configured with CONFIG_RANDOMIZE_BASE.
  * Fix for the 32-bit MIPS "bt" command to prevent an empty display (task
    header only) for an active task if the epc register in its exception frame
    contains 00000000.
  * Fix for support of Linux 4.7 and later x86_64 ELF kdump vmcores from
    kernels configured with CONFIG_RANDOMIZE_BASE. Without the patch, the
    crash session may fail during initialization with the message "crash:
    vmlinux and vmcore do not match!".
  * Fix for the x86_64 "mach" command display of the vmemmap base address in
    Linux 4.9 and later kernels configured with CONFIG_RANDOMIZE_BASE.
    Without the patch, the command shows a value of ffffea0000000000 next to
    "KERNEL VMEMMAP BASE".
  * Since the Linux 3.10 release, the kernel has offered the ability to create
    multiple independent ftrace buffers. At present, however, the "trace.c"
    extension module is only able to extract the primary buffer. This patch
    refactors the trace.c extension module so that the global instance is
    passed around as a parameter rather than accessing it directly, and then
    locates all of the available instances and extracts the data from each of
    them.
  * Fix for the s390x "bt" command for active tasks. Since the commit above
    in this crash-7.1.7 release that added support for the new
    CONFIG_THREAD_INFO_IN_TASK configuration, the backtrace of active tasks
    can be incomplete.
  * In collaboration with an update to the /dev/crash kernel driver, fix for
    Linux 4.6 commit a7f8de168ace487fa7b88cb154e413cf40e87fc6, which allows
    the ARM64 kernel image to be loaded anywhere in physical memory. Without
    the patch, attempting to run live on an ARM64 Linux 4.6 and later kernel
    may display the warning message "WARNING: cannot read linux_banner
    string", and then fails with the message "crash: vmlinux and /dev/crash do
    not match!". Version 1.3 of the crash driver is required, which
    introduces a new ioctl command that retrieves the ARM64-only
    "kimage_voffset" value that is required for virtual-to-physical address
    translation.
  * Update of the sample memory_driver/crash.c /dev/crash kernel driver to
    version 1.3, which adds support for Linux 4.6 and later ARM64 kernels,
    kernels configured with CONFIG_HARDENED_USERCOPY, and S390X kernels use
    xlate_dev_mem_ptr() and unxlate_dev_mem_ptr() instead of kmap() and
    kunmap().
  * New upstream 7.1.5 (Closes: #843731):
  * Introduction of support for "live" ramdump files, such as those that are
    specified by the QEMU mem-path argument of a memory-backend-file object.
    This allows the running of a live crash session against a QEMU guest from
    the host machine. In this example, the /tmp/MEM file on a QEMU host
    represents the guest's physical memory:
     $ qemu-kvm ...other-options... \
     -object memory-backend-file,id=MEM,size=128m,mem-path=/tmp/MEM,share=on \
     -numa node,memdev=MEM -m 128
    and a live session run can be run against the guest kernel like so:
     $ crash <path-to-guest-vmlinux> live:/tmp/MEM@0
    By prepending the ramdump image name with "live:", the crash session will
    act as if it were running a normal live session.
  * Fix for the support of ELF vmcores created by the KVM "virsh dump
    --memory-only" facility if the guest kernel was not configured with
    CONFIG_KEXEC, or CONFIG_KEXEC_CORE in Linux 4.3 and later kernels.
    Without the patch, the crash session fails during initialization with the
    message "crash: cannot resolve kexec_crash_image".
  * Added support for x86_64 ramdump files. Without the patch, the crash
    session fails immediately with the message "ramdump: unsupported machine
    type: X86_64".
  * Fix for a "[-Werror=misleading-indentation]" compiler warning that is
    generated by gdb-7.6/bfd/elf64-s390.c when building S390X in a Fedora
    Rawhide environment with gcc-6.0.0
  * Recognize and parse the new QEMU_VM_CONFIGURATION and QEMU_VM_FOOTER
    sections used for live migration of KVM guests, which are seen in the
    "kvmdump" format generated if "virsh dump" is used without the
    "--memory-only" option.
  * Fix for Linux commit edf14cdbf9a0e5ab52698ca66d07a76ade0d5c46, which has
    appended a NULL entry as the final member of the pageflag_names[] array.
    Without the patch, a message that indicates "crash: failed to read
    pageflag_names entry" is displayed during session initialization in Linux
    4.6 kernels.
  * Fix for Linux commit 0139aa7b7fa12ceef095d99dc36606a5b10ab83a, which
    renamed the page._count member to page._refcount. Without the patch,
    certain "kmem" commands fail with the "kmem: invalid structure member
    offset: page_count".
  * Fix for an ARM64 crash-7.1.5 "bt" regression for a task that has called
    panic(). Without the patch, the backtrace may fail with a message such as
    "bt: WARNING: corrupt prstatus? pstate=0x20000000, but no user frame
    found" followed by "bt: WARNING: cannot determine starting stack frame for
    task <address>". The pstate register warning will still be displayed (as
    it is essentially a kdump bug), but the backtrace will proceed normally.
  * Fix for the ARM64 "bt" command in Linux 4.5 and later kernels which use
    per-cpu IRQ stacks. Without the patch, if an active non-crashing task was
    running in user space when it received the shutdown IPI from the crashing
    task, the "-- <IRQ stack> ---" transition marker from the IRQ stack to the
    process stack is not displayed, and a message indicating "bt: WARNING:
    arm64_unwind_frame: on IRQ stack: oriq_sp: <address> fp: 0 (?)" gets
    displayed.
  * Fix for the ARM64 "bt" command in Linux 4.5 and later kernels which are
    not configured with CONFIG_FUNCTION_GRAPH_TRACER. Without the patch,
    backtraces that originate from a per-cpu IRQ stack will dump an invalid
    exception frame before transitioning to the process stack.
  * Introduction of ARM64 support for 4K pages with 4-level page tables and 48
    VA bits.
  * Implemented support for the redesigned ARM64 kernel virtual memory layout
    and associated KASLR support that was introduced in Linux 4.6. The kernel
    text and static data has been moved from unity-mapped memory into the
    vmalloc region, and its start address can be randomized if
    CONFIG_RANDOMIZE_BASE is configured. Related support is being put into
    the kernel's kdump code, the kexec-tools package, and makedumpfile(8);
    with that in place, the analysis of Linux 4.6 ARM64 dumpfiles with or
    without KASLR enabled should work normally by entering "crash vmlinux
    vmcore". On live systems, Linux 4.6 ARM64 kernels will only work
    automatically if CONFIG_RANDOMIZE_BASE is not configured. Unfortunately,
    if CONFIG_RANDOMIZE_BASE is configured on a live system, two --machdep
    command line arguments are required, at least for the time being. The
    arguments are:
     --machdep phys_offset=<base physical address>
     --machdep kimage_voffset=<kernel kimage_voffset value>
    Without the patch, any attempt to analyze a Linux 4.6 ARM64 kernel fails
    during initialization with a stream of "read error" messages followed by
    "crash: vmlinux and vmcore do not match!".
  * Linux 3.15 and later kernels configured with CONFIG_RANDOMIZE_BASE could
    be identified because of the "randomize_modules" kernel symbol, and if it
    existed, the "--kaslr=<offset>" and/or "--kaslr=auto" options were
    unnecessary. Since the "randomize_modules" symbol was removed in Linux
    4.1, this patch has replaced the KASLR identifier with the
    "module_load_offset" symbol, which was also introduced in Linux 3.15, but
    still remains.
  * Improvement of the ARM64 "bt -f" display such that in most cases, each
    stack frame level delimiter will be set to the stack address location
    containing the old FP and old LR pair.
  * Fix for the introduction of ARM64 support for 64K pages with 3-level page
    tables in crash-7.1.5, which fails to translate user space virtual
    addresses. Without the patch, "vtop <user-space address>" fails to
    translate all user-space addresses, and any command that needs to either
    translate or read user-space memory, such as "vm -p", "ps -a", and "rd -u"
    will fail.
  * Enhancement of the error message generated by the "tree -t radix" option
    when a duplicate entry is encountered. Without the patch, the error
    message shows the address of the radix_tree_node that contains the
    duplicate entry, for example, "tree: duplicate tree entry:
    <radix_tree_node>". It has been changed to also display the
    radix_tree_node.slots[] array index and the duplicate entry value, for
    example, "tree: duplicate tree entry: radix_tree_node: <radix_tree_node>
    slots[<index>]: <entry>".
  * Introduction of a new "bt -v" option that checks the kernel stack of all
    tasks for evidence of stack overflows. It does so by verifying the
    thread_info.task address, ensuring the thread_info.cpu value is a valid
    cpu number, and checking the end of the stack for the STACK_END_MAGIC
    value.
  * Fix to recognize a kernel thread that has user space virtual memory
    attached to it. While kernel threads typically do not have an mm_struct
    referencing a user-space virtual address space, they can either
    temporarily reference one for a user-space copy operation, or in the case
    of KVM "vhost" kernel threads, keep a reference to the user space of the
    "quem-kvm" task that created them. Without the patch, they will be
    mistaken for user tasks; the "bt" command will display an invalid
    kernel-entry exception frame that indicates "[exception RIP: unknown or
    invalid address]", the "ps" command will not enclose the command name with
    brackets, and the "ps -[uk]" and "foreach [user|kernel]" options will show
    the kernel thread as a user task.
  * Fix for the "bt -[eE]" options on ARM64 to recognize kernel exception
    frames in VHE enabled systems, in which the kernel runs in EL2.
  * Fix for the extensions/trace.c extension module to account for the Linux
    4.7 kernel commit dcb0b5575d24 that changed the bit index for the
    TRACE_EVENT_FL_TRACEPOINT flag. Without the patch, the "extend" command
    fails to load the trace.so module, with the error message "extend:
    /path/to/crash/extensions/trace.so: no commands registered: shared object
    unloaded". The patch reads the flag's enum value dynamically instead of
    using a hard-coded value.
  * Incorporated Takahiro Akashi's alternative backtrace method as a "bt"
    option, which can be accessed using "bt -o", and where "bt -O" will toggle
    the original and optional methods as the default. The original backtrace
    method has adopted two changes/features from the optional method:
     (1) ORIG_X0 and SYSCALLNO registers are not displayed in kernel
         exception frames.
     (2) stackframe entry text locations are modified to be the PC
         address of the branch instruction instead of the subsequent
         "return" PC address contained in the stackframe link register.
   Accordingly, these are the essential differences between the original and
   optional methods:
     (1) optional: the backtrace will start with the IPI exception frame
         located on the process stack.
     (2) original: the starting point of backtraces for the active,
         non-crashing, tasks, will continue to have crash_save_cpu()
         on the IRQ stack as the starting point.
     (3) optional: the exception entry stackframe adjusted to be located
         farther down in the IRQ stack.
     (4) optional: bt -f does not display IRQ stack memory above the
         adjusted exception entry stackframe.
     (5) optional: may display "(Next exception frame might be wrong)".
  * Fix for the failure of the "sym <symbol>" option in the extremely unlikely
    case where the symbol's name string is composed entirely of hexadecimal
    characters. For example, without the patch, "sym e820" fails with the
    error message "sym: invalid address: e820".
  * Fix for the failure of the "dis <symbol>" option in the extremely unlikely
    case where the symbol's name string is composed entirely of hexadecimal
    characters. For example, without the patch, "dis f" fails with the error
    message "dis: WARNING: f: no associated kernel symbol found" followed by
    "0xf: Cannot access memory at address 0xf".
  * Fix for the X86_64 "bt -R <symbol>" option if the only reference to the
    kernel text symbol in a backtrace is contained within the "[exception RIP:
    <symbol+offset>]" line of an exception frame dump. Without the patch, the
    reference will only be picked up if the exception RIP's hexadecimal
    address value is used.
  * Fix for the ARM64 "bt -R <symbol>" option if the only reference to the
    kernel text symbol in a backtrace is contained within the "[PC: <address>
    [<symbol+offset>]" line of an exception frame dump. Without the patch,
    the reference will only be picked up if the PC's hexadecimal address value
    is used.
  * Fix for the gathering of module symbol name strings during session
    initialization. In the unlikely case where the ordering of module symbol
    name strings does not match the order of the kernel_symbol structures, a
    faulty module symbol list entry may be created that contains a bogus name
    string.
  * Fix the PERCENTAGE of total output of the "kmem -i" SWAP USED line when
    the system has no swap pages at all. Without the patch, both the PAGES
    and TOTAL columns show values of zero, but it confusingly shows "100% of
    TOTAL SWAP", which upon first glance may seem to indicate potential memory
    pressure.
  * Enhancement to determine structure member data if the member is contained
    within an anonymous structure or union. Without the patch, it is
    necessary to parse the output of a discrete gdb "printf" command to
    determine the offset of such a structure member.
  * Speed up session initialization by attempting MEMBER_OFFSET_INIT() before
    falling back to ANON_MEMBER_OFFSET_INIT() in several known cases of
    structure members that are contained within anonymous structures.
  * Implemented new "list -S" and "tree -S" options that are similar to each
    command's -s option, but instead of parsing gdb output, member values are
    read directly from memory, so the command is much faster for 1-, 2-, 4-,
    and 8-byte members.
  * Fix to recognize and support x86_64 Linux 4.8-rc1 and later kernels that
    are configured with CONFIG_RANDOMIZE_MEMORY, which randomizes the base
    addresses of the kernel's unity-map address (PAGE_OFFSET), and the vmalloc
    region. Without the patch, the crash utility fails with a segmentation
    violation during session initialization on a live system, or will generate
    a number of WARNING messages followed by the fatal error message "crash:
    vmlinux and <dumpfile name> do not match!" with dumpfiles.
  * Fix for Linux 4.1 commit d0a0de21f82bbc1737ea3c831f018d0c2bc6b9c2, which
    renamed the x86_64 "init_tss" per-cpu variable to "cpu_tss". Without the
    patch, the addresses of the 4 per-cpu exception stacks cannot be
    determined, which causes backtraces that originate on any of the per-cpu
    DOUBLEFAULT, NMI, DEBUG, or MCE stacks to be truncated.
  * With the introduction of radix MMU in Power ISA 3.0, there are changes in
    kernel page table management accommodating it. This patch series makes
    appropriate changes here to work for such kernels. Also, this series
    fixes a few bugs along the way:
     ppc64: fix vtop page translation for 4K pages
     ppc64: Use kernel terminology for each level in 4-level page table
     ppc64/book3s: address changes in kernel v4.5
     ppc64/book3s: address change in page flags for PowerISA v3.0
     ppc64: use physical addresses and unfold pud for 64K page size
     ppc64/book3s: support big endian Linux page tables
    The patches are needed for Linux v4.5 and later kernels on all ppc64
    hardware.
  * Fix for Linux 4.8-rc1 commit 500462a9de657f86edaa102f8ab6bff7f7e43fc2, in
    which Thomas Gleixner redesigned the kernel timer mechanism to switch to a
    non-cascading wheel. Without the patch, the "timer" command fails with
    the message "timer: zero-size memory allocation! (called from <address>)"
  * Support for PPC64/BOOK3S virtual address translation for radix MMU. As
    both radix and hash MMU are supported in a single kernel on Power ISA 3.0
    based server processors, identify the current MMU type and set page table
    index values accordingly. Also, in Linux 4.7 and later kernels,
    PPC64/BOOK3S uses the same masked bit values in page table entries for 4K
    and 64K page sizes.
  * Change the RESIZEBUF() macro so that it will accept buffer pointers that
    are not declared as "char *" types. Change two prior direct callers of
    resizebuf() to use RESIZEBUF(), and fix two prior users of RESIZEBUF() to
    correctly calculate the need to resize their buffers.
  * Fix for the "trace.so" extension module to properly recognize Linux 3.15
    and later kernels. In crash-7.1.6, the MEMBER_OFFSET() macro has been
    improved so that it is able to recognize members of embedded anonymous
    structures. However, the module's manner of recognizing Linux 3.15 and
    later kernels depended upon MEMBER_OFFSET() failing to handle anonymous
    members, and therefore the improvement prevented the module from
    successfully loading.
  * If a "struct" command address argument is expressed using the per-cpu
    "symbol:cpuspec" format, and the symbol is a pointer type, i.e., not the
    address of the structure, display a WARNING message.
  * Exclude ARM64 kernel module linker mapping symbols like "$d" and "$x" as
    is done with 32-bit ARM. Without the patch, a crash session may fail
    during the "gathering module symbol data" stage with a message similar to
    "crash: store_module_symbols_v2: total: 15 mcnt: 16".
  * Enhancement to the ARM64 "dis" command when the kernel has enabled KASLR.
    When KASLR is enabled on ARM64, a function call between a module and the
    base kernel code will be done via a veneer (PLT) if the displacement is
    more than +/-128MB. As a result, disassembled code will show a branch to
    the in-module veneer location instead of the in-kernel target location.
    To avoid confusion, the output of the "dis" command will translate the
    veneer location to the target location preceded by "plt:", for example,
    "<plt:printk>".
  * Improvement of the "dev -d" option to display I/O statics for disks whose
    device driver uses the blk-mq interface. Currently "dev -d" always
    displays 0 in all fields for the blk-mq disk because blk-mq does not
    increment/decrement request_list.count[2] on I/O creation and I/O
    completion. The following values are used in blk-mq in such situations:
     - I/O creation: blk_mq_ctx.rq_dispatched[2]
     - I/O completion: blk_mq_ctx.rq_completed[2]
    So, we can get the counter of in-progress I/Os as follows: in progress
    I/Os == rq_dispatched - rq_completed This patch displays the result of
    above calculation for the disk. It determines whether the device driver
    uses blk-mq if the request_queue.mq_ops is not NULL. The "DRV" field is
    displayed as "N/A(MQ)" if the value for in-flight in the device driver
    does not exist for blk-mq.

Author: Chris J Arges
Author Date: 2016-09-27 19:58:43 UTC

Import patches-unapplied version 7.1.5-1ubuntu2 to ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: cbaab12aa1563ab637d3d923e52b26e001913e5c

New changelog entries:
  * Crash cannot parse Linux 4.8 vmcore files on ppc64el. (LP: #1628173)

Author: Chris J Arges
Author Date: 2016-09-27 19:58:43 UTC

Import patches-applied version 7.1.5-1ubuntu2 to applied/ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: 474536a777202803f1a6c4cc85e23d0b64a71103
Unapplied parent: 28f67005038abbe23a11506d6f02aa4f8fca2f22

New changelog entries:
  * Crash cannot parse Linux 4.8 vmcore files on ppc64el. (LP: #1628173)

Author: Chris J Arges
Author Date: 2016-09-27 19:58:43 UTC

Import patches-applied version 7.1.5-1ubuntu2 to applied/ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: 474536a777202803f1a6c4cc85e23d0b64a71103
Unapplied parent: 28f67005038abbe23a11506d6f02aa4f8fca2f22

New changelog entries:
  * Crash cannot parse Linux 4.8 vmcore files on ppc64el. (LP: #1628173)

Author: Chris J Arges
Author Date: 2016-09-27 19:58:43 UTC

Import patches-unapplied version 7.1.5-1ubuntu2 to ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: cbaab12aa1563ab637d3d923e52b26e001913e5c

New changelog entries:
  * Crash cannot parse Linux 4.8 vmcore files on ppc64el. (LP: #1628173)

Author: Chris J Arges
Author Date: 2016-05-03 11:23:16 UTC

Import patches-applied version 7.1.5-1ubuntu1 to applied/ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: fe177d039b4b9ee72a70cb289ab2d4ec1fcf82f0
Unapplied parent: cbaab12aa1563ab637d3d923e52b26e001913e5c

New changelog entries:
  * Merge from Debian unstable. Remaining changes:
    - Build for armhf, arm64, ppc64el.
    - Minor fixes for live autopkgtest.
    - debian/tests/live: Redirect gpg stderr to stdout, as that's the only
      known and expected stderr source.
    - Simplify ddeb archive requirements.
  * Dropped patches:
    - 0001-Fix-for-the-changes-made-to-the-kernel-module-struct.patch
    - 0002-Fix-for-the-changes-made-to-the-kernel-module-struct.patch

Author: Chris J Arges
Author Date: 2016-05-03 11:23:16 UTC

Import patches-unapplied version 7.1.5-1ubuntu1 to ubuntu/yakkety-proposed

Imported using git-ubuntu import.

Changelog parent: daed1b8c29ff47dc5875b8394037238a029f3e42

New changelog entries:
  * Merge from Debian unstable. Remaining changes:
    - Build for armhf, arm64, ppc64el.
    - Minor fixes for live autopkgtest.
    - debian/tests/live: Redirect gpg stderr to stdout, as that's the only
      known and expected stderr source.
    - Simplify ddeb archive requirements.
  * Dropped patches:
    - 0001-Fix-for-the-changes-made-to-the-kernel-module-struct.patch
    - 0002-Fix-for-the-changes-made-to-the-kernel-module-struct.patch

Author: Chris J Arges
Author Date: 2016-04-01 13:46:07 UTC

Import patches-unapplied version 7.1.4-1ubuntu4 to ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: 486ca617068be761cbe5f2bcc003e25e75fb7c45

New changelog entries:
  * d/p/0002-Fix-for-the-changes-made-to-the-kernel-module-struct.patch:
    Cherry-pick additional crash commit needed to fix issues related to
    kernel commit 8244062e. (LP: #1564487)

Author: Chris J Arges
Author Date: 2016-04-01 13:46:07 UTC

Import patches-applied version 7.1.4-1ubuntu4 to applied/ubuntu/xenial-proposed

Imported using git-ubuntu import.

Changelog parent: 802cf374974bca1d5a681677704bb4ab4ba44dbe
Unapplied parent: cb09e67791468748810eea52d9f476876fceb7d2

New changelog entries:
  * d/p/0002-Fix-for-the-changes-made-to-the-kernel-module-struct.patch:
    Cherry-pick additional crash commit needed to fix issues related to
    kernel commit 8244062e. (LP: #1564487)

Author: Dave Chiluk
Author Date: 2015-12-15 22:31:54 UTC

Import patches-unapplied version 7.1.1-1ubuntu4.1 to ubuntu/wily-proposed

Imported using git-ubuntu import.

Changelog parent: 5dc7cd5125bab6917b3338f164b3d692b50bd7ba

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-unapplied version 7.0.8-1ubuntu3.2 to ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 73faa3e5322ed8bbe449828c76a7ac13cbbde6eb

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-unapplied version 7.0.8-1ubuntu3.2 to ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 73faa3e5322ed8bbe449828c76a7ac13cbbde6eb

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-applied version 7.0.8-1ubuntu3.2 to applied/ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 96e28d55d8b9926c7a70e4c00ddd51d3d6be3620
Unapplied parent: 5eb8848f4325a08bb68a7937d7234ad1a6a61b53

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-applied version 7.0.8-1ubuntu3.2 to applied/ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 96e28d55d8b9926c7a70e4c00ddd51d3d6be3620
Unapplied parent: 5eb8848f4325a08bb68a7937d7234ad1a6a61b53

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-applied version 7.0.8-1ubuntu3.2 to applied/ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 96e28d55d8b9926c7a70e4c00ddd51d3d6be3620
Unapplied parent: 5eb8848f4325a08bb68a7937d7234ad1a6a61b53

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:18:33 UTC

Import patches-unapplied version 7.0.8-1ubuntu3.2 to ubuntu/vivid-proposed

Imported using git-ubuntu import.

Changelog parent: 73faa3e5322ed8bbe449828c76a7ac13cbbde6eb

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)

Author: Dave Chiluk
Author Date: 2015-12-15 22:31:54 UTC

Import patches-applied version 7.1.1-1ubuntu4.1 to applied/ubuntu/wily-proposed

Imported using git-ubuntu import.

Changelog parent: 87ca04f45eee4704219587de8e27a447e734893e
Unapplied parent: cf43e9568dea2763463728cfa23624ce9adf1f3a

New changelog entries:
  * Fix for crash: page excluded: kernel virtual address: ffff881052fa8000
    type: "fill_task_struct" (LP: #1523606)