Skip to content

Conversation

@lyakh
Copy link
Collaborator

@lyakh lyakh commented Nov 24, 2021

When resuming after a suspend, sometimes a spurious EDF interrupt arrives, causing panic in assert() in edf_scheduler_run(). Clear the interrupt before enabling it to avoid that.

When resuming after a suspend, sometimes a spurious EDF interrupt
arrives, causing panic in assert() in edf_scheduler_run(). Clear the
interrupt before enabling it to avoid that.

Signed-off-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
return edf_sch->irq;

interrupt_mask(edf_sch->irq, cpu_get_id());
interrupt_unmask(edf_sch->irq, cpu_get_id());
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@bkokoszx @mwasko fyi - as discussed on Tuesday.
@lyakh is masking and unmasking flow optimal, should we not be clearing the IRQ with interrupt_clear() ? We also need a comment here explaining why we are doing this.

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@lgirdwood of course it isn't optimal and yes, interrupt_clear() would be the right function to call, but currently it only clears DSP "native" interrupts, and this is a level-2 interrupt. Let's test if this helps, if it does, we can try to extend interrupt_clear() to also work on "level interrupts."

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@lyakh I did the stress test on this PR. The bug #4538 still can be reproduced. On the meantime, I did the stress test on the PR #4874. The bug #4538 can be fixed and no other regression is found. So I will ask Dell team to test PR4874.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the test. @lyakh @lgirdwood I would suggest to make the #4874 as hotfix for 1.9 and keep main unchanged as root causing the #4538 could be very time consuming.

@gkbldcig
Copy link
Collaborator

gkbldcig commented Dec 4, 2021

Can one of the admins verify this patch?

@lgirdwood
Copy link
Member

@lyakh @keyonjie do we still need this for main ?

@keyonjie
Copy link
Contributor

@lyakh @keyonjie do we still need this for main ?

@libinyang told me that this didn't help anything, so I would suggest keeping it intact.

@lyakh lyakh closed this Jun 9, 2022
@lyakh lyakh deleted the edf branch April 11, 2024 13:10
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants