ptl: userspace: temporarily switch to stack pool#10632
ptl: userspace: temporarily switch to stack pool#10632lyakh wants to merge 1 commit intothesofproject:mainfrom
Conversation
This fixes sporadic CI endurance test failures. Once Zephyr merges "kernel: dynamic stack to cached area if coherence" we should be able to switch back to dynamic stack allocation. Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
lyakh
requested review from
dbaluta,
kv2019i,
lbetlej,
lgirdwood,
mmaka1 and
plbossart
as code owners
There was a problem hiding this comment.
Pull request overview
This PR adjusts the Panther Lake ACE30 userspace configuration to avoid sporadic CI endurance failures by switching dynamic thread stack provisioning away from runtime allocation and toward a preallocated pool.
Changes:
- Disable dynamic thread stack allocation preference and enable pool preference for dynamic threads.
- Introduce a fixed-size dynamic thread stack pool configuration.
- Reduce configured dynamic thread stack size and
CONFIG_SOF_STACK_SIZEfrom 8192 to 4096.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
lgirdwood
left a comment
lgirdwood
left a comment
There was a problem hiding this comment.
LGTM, but pls review/resolve copliot comments
|
I'm not convinced about reducing |
@softwarecki I don't think we need 8k stack so far, we can check it with something like the debug stream. And the "Internal Intel CI" failure is still the infamous "random HDA DMA" test. But jenkins failure https://sof-ci.01.org/sofpr/PR10632/build20140/devicetest/index.html?model=PTLH_RVP_NOCODEC&testcase=test-speaker is indeed interesting... looking |
This fixes sporadic CI endurance test failures. Once Zephyr merges "kernel: dynamic stack to cached area if coherence" we should be able to switch back to dynamic stack allocation.