Repository navigation
drivers: Intel: hda-dma: only check DGCS_BUSY with delayed stop - #4938
Conversation
|
@lgirdwood this is the temporary fix I suggested, we should add a rebased version of #4931 when the changes from @lyakh are also merged, so that we don't add a poll with interrupts disabled for too long :-) |
@plbossart do you mean a revived version of #4724? |
When an older kernel is used, we use the immediate stop. On some platforms, we see an -EBUSY error that prevents the trigger stop from working. This patch adds a check to be 'bug-compatible' with previous releases and older kernels. An additional fix would be to insert a poll timeout to make sure the BUSY bit is cleared. BugLink: thesofproject#4919 Fixes: 7e54f45 ('drivers: Intel: hda-dma: ensure DMA channel is idle after stop') Suggested-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com> Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
93668fe to
ea2fd40
Compare
|
@lgirdwood I think this needs to go in 1.9.1, otherwise the new firmware will break with older kernels. |
Ack, I think you are right. @ranj063 agree ? @keyonjie did we get this in the branch ? We may need a v1.9.2. @mengdonglin |
|
SOFCI TEST |
Yes @lgirdwood this is needed for 1.9.1 |
|
@plbossart running CI again as there is a failure popping up the the CML PM, I'm seeing this intermittently over the last week and it's mainly on the CML RVP. Not related to this PR (since the CI will be using the updated kernel). |
@plbossart @ranj063 is this really mandatory for 1.9.1? We already have daily "linux/for-next + main" recipe coverage and no issue has been observed, see test plan result 7900. Anything I missed here? |
|
@ranj063 just told me that our test-case in daily_lite doesn't have this coverage. |
|
The BYT kmod failure looks unrelated, merging. |
When an older kernel is used, we use the immediate stop. On some
platforms, we see an -EBUSY error that prevents the trigger stop from
working.
This patch adds a check to be 'bug-compatible' with previous releases
and older kernels. An additional fix would be to insert a poll timeout
to make sure the BUSY bit is cleared.
BugLink: #4919
Fixes: 7e54f45 ('drivers: Intel: hda-dma: ensure DMA channel is idle after stop')
Suggested-by: Ranjani Sridharan ranjani.sridharan@linux.intel.com
Signed-off-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com