~canonical-hwe-team/hwe-next/+git/sof:1.1-stable

Last commit made on 2018-04-05
Get this branch:
git clone -b 1.1-stable https://git.launchpad.net/~canonical-hwe-team/hwe-next/+git/sof

Branch merges

Branch information

Name:
1.1-stable
Repository:
lp:~canonical-hwe-team/hwe-next/+git/sof

Recent commits

017c40b... by Liam Girdwood <email address hidden>

dist: fix make dist for SRC

Signed-off-by: Liam Girdwood <email address hidden>

082c86f... by Pierre-Louis Bossart <email address hidden>

mailbox: move multi-line macros to static inlines

Detected with Coverity:

In do_notify: The indentation of this code suggests it is nested when
it is not. (CWE-483)

multi_stmt_macro: The macro on this line expands into multiple
statements, only the first of which is nested within the preceding
parent while the rest are not.

 if (msg->rx_size && msg->rx_size < SOF_IPC_MSG_MAX_SIZE)
  mailbox_dspbox_read(msg->rx_data, 0, msg->rx_size);

Move mailbox macros to inline functions to remove the issue, keep
indentation the same and add typecasts as needed

Signed-off-by: Pierre-Louis Bossart <email address hidden>

669e7c9... by Pierre-Louis Bossart <email address hidden>

sof: solve circular dependencies, move rmemcpy/arch_memcpy definitions

The preprocessor doesn't seem to care about circular dependencies,
the compiler does. So before we replace multi-line macros with
static inlines, move rmemcpy/arch_memcpy to a reef/string.h and
arch/string.h header file.

This is mostly useful for the mailbox functions, which depend on rmemcpy
but are also called from arch/reef.h dump/panic utilities

Signed-off-by: Pierre-Louis Bossart <email address hidden>

f7471ab... by Pierre-Louis Bossart <email address hidden>

sof: scripts: force PATH to compile xtensa without environment dependencies

The ROOT is already assumed to be as $pwd/../xtensa-root/$ROOT,
set the PATH as well

Signed-off-by: Pierre-Louis Bossart <email address hidden>

c858615... by Keyon Jie <email address hidden>

Host: fix error return handlings for some dma invokings.

Add hanle for error on dma_set_config(), dma_start() and dma_copy().

Signed-off-by: Pierre-Louis Bossart <email address hidden>
Signed-off-by: Keyon Jie <email address hidden>

970bd97... by Yan Wang <email address hidden>

Use tracev instead of trace for dai irq event.

This event is in pipeline and will produce many trace data in a short
time. It should use tracev to disable it by default.

Signed-off-by: Yan Wang <email address hidden>
Signed-off-by: Ranjani Sridharan <email address hidden>

d3d5b31... by Liam Girdwood <email address hidden>

Revert "dai: copy: Move buffer pointer updates into dai_copy()"

This reverts commit 7f9110c4c0ca26a67e0e4baa91e5f0bcc31a1c3e.

This does not work when both buffer data consumer and producer are DMACs.

Signed-off-by: Liam Girdwood <email address hidden>

7f9110c... by Wu Zhigang <email address hidden>

dai: copy: Move buffer pointer updates into dai_copy()

Move the DAI's buffer pointer update location from dai_dma_cb()
to dai_copy(). This fixes an occasional race where xrun recovery can
become out of sync with other pipeline components.

Signed-off-by: Wu Zhigang <email address hidden>
Tested-by: Keyon Jie <email address hidden>

9052f91... by Keyon Jie <email address hidden>

dai: handle xrun with more clear logic

1. xrun happens.
2. trigger pipeline xrun and recover.
3. set dd->xrun in XRUN cmd from pipeline recover.
4. stop dma and reset buffer to 0s.
5. dai prepare and clear dd->xrun in prepare calling from recover.
6. start dma again in START cmd from pipeline recover.

This fix APL pause/release Xrun issue in my test.

Signed-off-by: Keyon Jie <email address hidden>

08f8570... by Liam Girdwood <email address hidden>

cnl: core: Add support for xcc

xcc has more strict rules than gcc:
- Add missing sections to linker scripts
- Fix compile warnings
- Extend memory enabling timeout to infinite (too small delay for xcc)

Signed-off-by: Tomasz Lauda <email address hidden>

# Conflicts:
# src/platform/apollolake/platform.c
# src/platform/cannonlake/platform.c