pipeline: disable xrun recovery #1439
Conversation
slawblauciak
requested review from
lbetlej,
lgirdwood,
mmaka1,
plbossart,
RanderWang,
xiulipan and
a team
as code owners
|
@keyonjie this flag will have to be handled accordingly on the kernel side now |
LGTM, thanks. |
|
SOFCI TEST |
slawblauciak
force-pushed
the
no_xrun_recovery
branch
from
5a9e3e3
to
c0a6500
Compare
slawblauciak
force-pushed
the
no_xrun_recovery
branch
from
c0a6500
to
2e18fdb
Compare
added a flag to FW ready message to inform whether the internal xrun recovery is enabled or not Signed-off-by: Slawomir Blauciak <[email protected]>
slawblauciak
force-pushed
the
no_xrun_recovery
branch
from
2e18fdb
to
c01b683
Compare
|
@slawblauciak is there a kernel PR for ABI addition ? |
|
@slawblauciak MINOR 8 atm, if you can get kernel PR created soon (as it can merge with other PRs on the dashboard), otherwise it will be MINOR 9 if kernel PR cant be done until late next week. Let me know and I'll update as required. |
|
@lgirdwood what would the kernel do with this anyways? If it's not set, then the firmware recovers on its own? Why isn't the ALSA prepare/restart not good enough ? Didn't we have this conversation already? |
|
@lgirdwood @plbossart Please let us know what you decided. Almost week passed and we don't know if it's good to merge for you. Thanks. |
|
@plbossart user here is not ALSA, but an RTOS or stand alone mode. |
|
@lgirdwood no, there's no change ready yet. I believe it might require a functional change in the kernel, not just an ABI change. |
Ah, get back to this, yes, no ABI change needed in Linux driver side, we will need to enable CONFIG_SND_SOC_SOF_DEBUG_XRUN_STOP to handle Xrun IPC come from DSP. |
|
On 5/23/19 8:16 PM, Keyon wrote:
@lgirdwood <https://github.com/lgirdwood> no, there's no change
ready yet. I believe it might require a functional change in the
kernel, not just an ABI change.
@keyonjie <https://github.com/keyonjie> mind shedding some light on
the situation on your side?
Ah, get back to this, yes, no ABI change needed in Linux driver side, we
will need to enable CONFIG_SND_SOC_SOF_DEBUG_XRUN_STOP to handle Xrun
IPC come from DSP.
This PR LGTM.
That doesn't look quite right to me. If the kernel behavior needs to
depend on how the firmware is compiled, then the kernel needs to
discover this capability and not use a Kconfig at all.
|
Yep, @slawblauciak @keyonjie we need a similar PR for the kernel that has the new flag and @keyonjie that will stop ALSA attempting to recover XRUNs (this would be core ALSA). Alternatively, we could postpone this PR until we have IPC2.0 and a user ? Do we have a user for this PR today ? |
|
@lgirdwood frankly, this is such an old discussed issue that I'm no longer sure whether it's really necessary to have it merged ASAP. There was another problem that's been gated by this, but it's no longer relevant since it's been solved. |
|
@slawblauciak ok, please close if you want too and nobody objects. |
|
Closing for now, will revisit later. |
added a flag to FW ready message to inform
whether the internal xrun recovery is enabled or not
Signed-off-by: Slawomir Blauciak [email protected]