| Message ID | 1483022987-28843-1-git-send-email-nikita.yoush@cogentembedded.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 12/29/2016 03:49 PM, Nikita Yushchenko wrote: > This reverts commit 1a653aa44725668590b36bbe2d7fe4736a69f055 ("ASoC: > core: replace aux_comp_list to component_dev_list"). > > That commit tries to remove card->aux_comp_list, using flagged entries > in card->component_dev_list instead. > > However, components are added to card->component_dev_list in > soc_probe_component(), which is called for aux devices by > soc_probe_aux_devices(). Before that commit, it traversed > card->aux_comp_list and found aux devices added by soc_bind_aux_dev(). > After that commit, it traverses card->component_dev_list and finds no > aux devices, because soc_probe_component() was not yet called for them. > > Thus all aux devices are lost and any setup that needs them no longer > works. I think we can't simply revert that commit now, after commit 9178feb ASoC: add Component level suspend/resume component_dev_list is also used for the component level suspend/resume, if auxiliary devices would have been on a separate list they would be missed in the suspend/resume sequences. I run into same issue and posted a patch: "ASoC: Fix binding and probing of auxiliary components". -- Thanks, Sylwester
29.12.2016 19:18, Sylwester Nawrocki пишет: > On 12/29/2016 03:49 PM, Nikita Yushchenko wrote: >> This reverts commit 1a653aa44725668590b36bbe2d7fe4736a69f055 ("ASoC: >> core: replace aux_comp_list to component_dev_list"). >> >> That commit tries to remove card->aux_comp_list, using flagged entries >> in card->component_dev_list instead. >> >> However, components are added to card->component_dev_list in >> soc_probe_component(), which is called for aux devices by >> soc_probe_aux_devices(). Before that commit, it traversed >> card->aux_comp_list and found aux devices added by soc_bind_aux_dev(). >> After that commit, it traverses card->component_dev_list and finds no >> aux devices, because soc_probe_component() was not yet called for them. >> >> Thus all aux devices are lost and any setup that needs them no longer >> works. > > I think we can't simply revert that commit now, after commit > > 9178feb ASoC: add Component level suspend/resume > component_dev_list is also used for the component level suspend/resume, > if auxiliary devices would have been on a separate list they would be > missed in the suspend/resume sequences. I believe aux devices are on both lists.
>> This reverts commit 1a653aa44725668590b36bbe2d7fe4736a69f055 ("ASoC: >> core: replace aux_comp_list to component_dev_list"). >> >> That commit tries to remove card->aux_comp_list, using flagged entries >> in card->component_dev_list instead. >> >> However, components are added to card->component_dev_list in >> soc_probe_component(), which is called for aux devices by >> soc_probe_aux_devices(). Before that commit, it traversed >> card->aux_comp_list and found aux devices added by soc_bind_aux_dev(). >> After that commit, it traverses card->component_dev_list and finds no >> aux devices, because soc_probe_component() was not yet called for them. >> >> Thus all aux devices are lost and any setup that needs them no longer >> works. > > I think we can't simply revert that commit now, after commit > > 9178feb ASoC: add Component level suspend/resume > component_dev_list is also used for the component level suspend/resume, > if auxiliary devices would have been on a separate list they would be > missed in the suspend/resume sequences. I believe aux devices are on both lists.
On 12/29/2016 05:20 PM, Nikita Yushchenko wrote:
> I believe aux devices are on both lists.
Indeed, you are right, they are on both lists.
So there should be no issues after reverting.
On Thu, Dec 29, 2016 at 05:18:21PM +0100, Sylwester Nawrocki wrote: > I run into same issue and posted a patch: > "ASoC: Fix binding and probing of auxiliary components". Which I applied so I guess things are fine now without the revert? I'm still working through backlog from the holidays so didn't check properly yet.
On Sat, Jan 7, 2017 at 2:18 AM, Mark Brown <broonie@kernel.org> wrote: > On Thu, Dec 29, 2016 at 05:18:21PM +0100, Sylwester Nawrocki wrote: > >> I run into same issue and posted a patch: > >> "ASoC: Fix binding and probing of auxiliary components". > > Which I applied so I guess things are fine now without the revert? I'm > still working through backlog from the holidays so didn't check > properly yet. This fixes the Allwinner H3 codec analog controls aux_device for me. Thanks. ChenYu > > _______________________________________________ > Alsa-devel mailing list > Alsa-devel@alsa-project.org > http://mailman.alsa-project.org/mailman/listinfo/alsa-devel >
>>> I run into same issue and posted a patch: >> >>> "ASoC: Fix binding and probing of auxiliary components". >> >> Which I applied so I guess things are fine now without the revert? I'm >> still working through backlog from the holidays so didn't check >> properly yet. > > This fixes the Allwinner H3 codec analog controls aux_device for me. > Thanks. But this is not yet in 4.10-rc3 ?
On Mon, Jan 09, 2017 at 10:41:37AM +0300, Nikita Yushchenko wrote:
> But this is not yet in 4.10-rc3 ?
Yes. It's not yet been sent to Linus, it'll go whenever the next batch
of fixes goes.
diff --git a/include/sound/soc.h b/include/sound/soc.h index 2b502f6cc6d0..8172a512632a 100644 --- a/include/sound/soc.h +++ b/include/sound/soc.h @@ -809,10 +809,10 @@ struct snd_soc_component { unsigned int ignore_pmdown_time:1; /* pmdown_time is ignored at stop */ unsigned int registered_as_component:1; - unsigned int auxiliary:1; /* for auxiliary component of the card */ unsigned int suspended:1; /* is in suspend PM state */ struct list_head list; + struct list_head list_aux; /* for auxiliary component of the card */ struct list_head card_list; struct snd_soc_dai_driver *dai_drv; @@ -1152,6 +1152,7 @@ struct snd_soc_card { */ struct snd_soc_aux_dev *aux_dev; int num_aux_devs; + struct list_head aux_comp_list; const struct snd_kcontrol_new *controls; int num_controls; @@ -1547,6 +1548,7 @@ static inline void snd_soc_initialize_card_lists(struct snd_soc_card *card) INIT_LIST_HEAD(&card->widgets); INIT_LIST_HEAD(&card->paths); INIT_LIST_HEAD(&card->dapm_list); + INIT_LIST_HEAD(&card->aux_comp_list); INIT_LIST_HEAD(&card->component_dev_list); } diff --git a/sound/soc/soc-core.c b/sound/soc/soc-core.c index f1901bb1466e..15657a65e4ca 100644 --- a/sound/soc/soc-core.c +++ b/sound/soc/soc-core.c @@ -1747,8 +1747,7 @@ static int soc_bind_aux_dev(struct snd_soc_card *card, int num) } component->init = aux_dev->init; - component->auxiliary = 1; - + list_add(&component->list_aux, &card->aux_comp_list); return 0; err_defer: @@ -1764,10 +1763,7 @@ static int soc_probe_aux_devices(struct snd_soc_card *card) for (order = SND_SOC_COMP_ORDER_FIRST; order <= SND_SOC_COMP_ORDER_LAST; order++) { - list_for_each_entry(comp, &card->component_dev_list, card_list) { - if (!comp->auxiliary) - continue; - + list_for_each_entry(comp, &card->aux_comp_list, list_aux) { if (comp->driver->probe_order == order) { ret = soc_probe_component(card, comp); if (ret < 0) { @@ -1791,14 +1787,11 @@ static void soc_remove_aux_devices(struct snd_soc_card *card) for (order = SND_SOC_COMP_ORDER_FIRST; order <= SND_SOC_COMP_ORDER_LAST; order++) { list_for_each_entry_safe(comp, _comp, - &card->component_dev_list, card_list) { - - if (!comp->auxiliary) - continue; - + &card->aux_comp_list, list_aux) { if (comp->driver->remove_order == order) { soc_remove_component(comp); - comp->auxiliary = 0; + /* remove it from the card's aux_comp_list */ + list_del(&comp->list_aux); } } }
This reverts commit 1a653aa44725668590b36bbe2d7fe4736a69f055 ("ASoC: core: replace aux_comp_list to component_dev_list"). That commit tries to remove card->aux_comp_list, using flagged entries in card->component_dev_list instead. However, components are added to card->component_dev_list in soc_probe_component(), which is called for aux devices by soc_probe_aux_devices(). Before that commit, it traversed card->aux_comp_list and found aux devices added by soc_bind_aux_dev(). After that commit, it traverses card->component_dev_list and finds no aux devices, because soc_probe_component() was not yet called for them. Thus all aux devices are lost and any setup that needs them no longer works. Signed-off-by: Nikita Yushchenko <nikita.yoush@cogentembedded.com> --- Fixed lost line in commit message ;) include/sound/soc.h | 4 +++- sound/soc/soc-core.c | 17 +++++------------ 2 files changed, 8 insertions(+), 13 deletions(-)