Various failures of kernel_security suite on Xenial kernel on s390x arch
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
QA Regression Testing |
Fix Released
|
Undecided
|
Unassigned | ||
linux (Ubuntu) |
Fix Released
|
Undecided
|
Tim Gardner | ||
Xenial |
Fix Released
|
Undecided
|
Tim Gardner |
Bug Description
Partial log:
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR enabled ... ok
01/05 13:07:05 ERROR| utils:0153| [stderr] test_020_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of stack ... ok
01/05 13:07:05 ERROR| utils:0153| [stderr] test_021_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of libs ... FAIL
01/05 13:07:05 ERROR| utils:0153| [stderr] test_021_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of mmap ... FAIL
01/05 13:07:05 ERROR| utils:0153| [stderr] test_022_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of text ... FAIL
01/05 13:07:05 ERROR| utils:0153| [stderr] test_022_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of vdso ... FAIL
01/05 13:07:05 ERROR| utils:0153| [stderr] test_022_
01/05 13:07:05 ERROR| utils:0153| [stderr] ASLR of brk ... FAIL
Full log is attached (client.DEBUG.txt).
Related branches
Changed in linux (Ubuntu): | |
status: | Incomplete → Confirmed |
Changed in linux (Ubuntu): | |
assignee: | nobody → Tim Gardner (timg-tpi) |
status: | Confirmed → Fix Committed |
tags: | added: s390x |
Hi Brad,
I've addressed most of the issues with commit 2464 to the qa-regression- testing tree. However, there are a few remaining failures that are legit issues with the kernel on 390x. They are:
FAIL: test_074_ config_ security_ default_ mmap_min_ addr (__main_ _.KernelSecurit yTest) DEFAULT_ MMAP_MIN_ ADDR
CONFIG_
CONFIG_ DEFAULT_ MMAP_MIN_ ADDR is set to 4096 on s390x, not 65536 like on other arches.
FAIL: test_075_ config_ stack_protector (__main_ _.KernelSecurit yTest) CC_STACKPROTECT OR set stack_guard_ kernel (__main_ _.KernelSecurit yTest)
CONFIG_
FAIL: test_082_
Kernel stack guard
CONFIG_ CC_STACKPROTECT OR is not set for s390x, unlike for our other arches. As far as I can tell, gcc and the kernel support the option on s390x (but if not, then I'll adjust our tests).
I'm also seeing the module tainting test test_140_ kernel_ modules_ not_tainted failing on s390x. Checking the contents of /proc/modules, it looks like all the modules on the s390x kernel (4.4.0- 2.16-generic) aren't being signed (i.e. have an (E) listed for them). I'm not sure why this is, if it indicates an issue with the kernel build process on s390x.
Thanks.