dai: change to use new version of dai_config_get #6939
Conversation
juimonen
requested review from
libinyang,
lgirdwood,
plbossart,
mmaka1,
lbetlej and
dbaluta
as code owners
|
depends on zephyr side: zephyrproject-rtos/zephyr#53708 |
juimonen
force-pushed
the
get_config_fix
branch
from
9f24adf to
fde89fc
Compare
src/lib/dai.c
Outdated
| cfg = dai_config_get(zephyr_dev[i], dir); | ||
| if (cfg && cfg->type == type && cfg->dai_index == index) | ||
| if (dai_config_get(zephyr_dev[i], &cfg, dir)) | ||
| return NULL; |
There was a problem hiding this comment.
@juimonen is this correct? Should you continue here instead?
There was a problem hiding this comment.
I mean dai_config_get should return 0 on success, otherwise negative. You kind of always get some config back, it is just testing the arguments you give are not bogus... so in that sense this can't fail, just added the check for completeness.
There was a problem hiding this comment.
even if this is not possible in practice, the right thing to do here is continue instead of return NULL right?
juimonen
force-pushed
the
get_config_fix
branch
from
fde89fc to
5f512d1
Compare
|
@juimonen some conflicts. |
|
@juimonen Would be good to add some tag/note that this cannot be merged until the Zephyr dependency is merged (and a link to it). |
Nice way to do that is to add list with dependencies in pull request description. I also think that good idea is to point the zephyr Pull Request in the west manifest. This way the code will compile and pass validation. |
|
@juimonen any ETA for the Zephyr PR ? |
juimonen
force-pushed
the
get_config_fix
branch
from
5f512d1 to
53b0ed5
Compare
|
zephyr side merged. Updated/rebased for sof main and added west update for zephyr main/head. |
|
hmm, seems we are failing into a watchdog warning (treated as error), this commit is before set_config change, so can't really roll back. @softwarecki fyi. |
src/audio/dai-zephyr.c
Outdated
|
|
||
| comp_dbg(dev, "dai_hw_params()"); | ||
|
|
||
| if (!cfg) | ||
| if (dai_config_get(dd->dai->dev, &cfg, dir)) |
There was a problem hiding this comment.
is dai_config_get() now returning an error code? Could we propagate it to the caller instead of -EINVAL?
juimonen
force-pushed
the
get_config_fix
branch
2 times, most recently
from
0873b09 to
7f59b61
Compare
|
So Intel System/merge/build passes (#7132 is NOT hit which is surprising), but the SOF CI fails due to the DMA API changes https://sof-ci.01.org/sofpr/PR6939/build4024/devicetest/index.html |
Start using new version of dai_config_get where config struct is given as pointer argument. Update west.yaml to point to correct zephyr version for this change. Signed-off-by: Jaska Uimonen <[email protected]>
juimonen
force-pushed
the
get_config_fix
branch
from
7f59b61 to
6d75857
Compare
|
One failure in CI https://sof-ci.01.org/sofpr/PR6939/build4057/devicetest/index.html , but no errors in the FW logs. |
|
SOFCI TEST |
Exactly that. This PR should have been combined with a manifest update. Instead it broke the daily tests ( https://github.com/thesofproject/sof/actions/runs/4228168969/jobs/7343348093) and caused very confusing PR results (e.g. https://github.com/thesofproject/sof/actions/runs/4233634492/jobs/7354897229) |
|
Daily compilation is green again https://github.com/thesofproject/sof/actions/workflows/daily-tests.yml |
|
@marc-hb wrote:
? This pr WAS combined with a manifest update, in a single git commit, just because of this. |
|
Then why did daily tests fail? (for just one day) https://github.com/thesofproject/sof/actions/runs/4228168969/jobs/7343348093 |
|
My bad, daily tests failed only when NOT using the manifest but the latest Zephyr. Sorry for the noise. @aborisovich I cannot see the warning in the Windows build, |
Start using new version of dai_config_get where config struct is given as pointer argument from outside.
Signed-off-by: Jaska Uimonen [email protected]