Re: [PATCH for-6.1.y] Revert "ASoC: codecs: lpass: register mclk after runtime pm"

[Srinivas Kandagatla <srinivas.kandagatla@linaro.org>]

Hi Amit,

sorry for late reply.

On 16/03/2023 11:29, Amit Pundir wrote:
> On Thu, 16 Mar 2023 at 13:55, Greg KH <gregkh@linuxfoundation.org> wrote:
>>
>> On Thu, Mar 16, 2023 at 12:13:40AM +0530, Amit Pundir wrote:
>>> On Wed, 15 Mar 2023 at 23:54, Greg KH <gregkh@linuxfoundation.org> wrote:
>>>>
>>>> On Wed, Mar 15, 2023 at 11:49:00PM +0530, Amit Pundir wrote:
>>>>> This reverts commit 7b642273438cf500d36cffde145b9739fa525c1d which is
>>>>> commit 1dc3459009c33e335f0d62b84dd39a6bbd7fd5d2 upstream.
>>>>>
>>>>> This patch broke RB5 (Qualcomm SM8250) devboard. The device
>>>>> reboots into USB crash dump mode after following error:
>>>>>
>>>>>      qcom_q6v5_pas 17300000.remoteproc: fatal error received: \
>>>>>      ABT_dal.c:278:ABTimeout: AHB Bus hang is detected, \
>>>>>      Number of bus hang detected := 2 , addr0 = 0x3370000 , addr1 = 0x0!!!
>>>>>
>>>>> Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
>>>>> ---
>>>>>   sound/soc/codecs/lpass-rx-macro.c  |  8 ++++----
>>>>>   sound/soc/codecs/lpass-tx-macro.c  |  8 ++++----
>>>>>   sound/soc/codecs/lpass-va-macro.c  | 20 ++++++++++----------
>>>>>   sound/soc/codecs/lpass-wsa-macro.c |  9 +++++----
>>>>>   4 files changed, 23 insertions(+), 22 deletions(-)
>>>>
>>>> Is this also reverted in Linus's tree?  If not, why not?
>>>
>>> I couldn't reproduce this crash on Linus's tree. It was first reported
>>> on android14-6.1 and then I could reproduce it on v6.1.19 as well,
>>> hence this revert.
>>>
>>> A quick search points out that this patch is a part of a 8 patch
>>> series https://lore.kernel.org/lkml/20230209122806.18923-1-srinivas.kandagatla@linaro.org/
>>> while only 5 of them landed on v6.1.y. May be we need the remaining
>>> fixes on v6.1.y as well? I can give the remaining patches a quick shot
>>> tomorrow if that helps.
>>
>> Yes please, we would much rather take whatever is in Linus's tree than a
>> special revert as that will keep the trees in sync better.  If you can
>> provide the missing git ids, I can just queue them up if you have tested
>> them.
>>
> 
> Cherry-picking the rest of the relevant fixes from the
> https://lore.kernel.org/lkml/20230209122806.18923-1-srinivas.kandagatla@linaro.org/
> series didn't help.
> 
> Srini, does this patch series has a dependency on other upstream fixes
> as well? With the above patch series on v6.1.y, I see the following
> crash on RB5:
> 
This looks like uncovered an issue with 
drivers/clk/qcom/lpass-gfm-sm8250.c driver which seems to not do runtime 
pm correctly in some case.
This driver seems to get lucky with clocks.

I already sent a fix for this 
https://lore.kernel.org/lkml/c5273d67493cbb008f13d7538837828a.sboyd@kernel.org/T/

That should fix the reported issue.

--srini
> qcom-q6afe aprsvc:apr-service:4:4: cmd = 0x100f6 returned error = 0x1
> q6asm-dai 17300000.remoteproc:glink-edge:apr:apr-service@7:dais:
> Adding to iommu group 25
> qcom-q6afe aprsvc:apr-service:4:4: Unknown cmd 0x100f6
> qcom,apr 17300000.remoteproc:glink-edge.apr_audio_svc.-1.-1: Adding
> APR/GPR dev: aprsvc:apr-service:4:8
> qcom-q6afe aprsvc:apr-service:4:4: cmd = 0x100f6 returned error = 0x1
> qcom-q6afe aprsvc:apr-service:4:4: Unknown cmd 0x100f6
> qcom-q6afe aprsvc:apr-service:4:4: cmd = 0x100f6 returned error = 0x1
> qcom-q6afe aprsvc:apr-service:4:4: Unknown cmd 0x100f6
> wsa881x-codec sdw:0:0217:2110:00:4: nonexclusive access to GPIO for powerdown
> qcom-soundwire 3250000.soundwire-controller: Qualcomm Soundwire
> controller v1.5.1 Registered
> qcom_q6v5_pas 17300000.remoteproc: fatal error received:
> ABT_dal.c:278:ABTimeout: AHB Bus hang is detected, Number of bus hang
> detected := 2 , addr0 = 0x3370000 , addr1 = 0x0!!!
> remoteproc remoteproc1: crash detected in 17300000.remoteproc: type fatal error
> remoteproc remoteproc1: handling crash #1 in 17300000.remoteproc
> remoteproc remoteproc1: recovering 17300000.remoteproc
> platform 17300000.remoteproc:glink-edge:fastrpc:compute-cb@5: Removing
> from iommu group 23
> platform 17300000.remoteproc:glink-edge:fastrpc:compute-cb@4: Removing
> from iommu group 22
> platform 17300000.remoteproc:glink-edge:fastrpc:compute-cb@3: Removing
> from iommu group 21
> qcom-q6afe aprsvc:apr-service:4:4: AFE set params failed -110
> clk_unregister: unregistering prepared clock: LPASS_HW_DCODEC
> clk_unregister: unregistering prepared clock: LPASS_HW_MACRO
> platform 17300000.remoteproc:glink-edge:apr:apr-service@7:dais:
> Removing from iommu group 25
> remoteproc remoteproc1: stopped remote processor 17300000.remoteproc
> remoteproc remoteproc1: remote processor 17300000.remoteproc is now up
> qcom,fastrpc-cb 17300000.remoteproc:glink-edge:fastrpc:compute-cb@3:
> Adding to iommu group 21
> qcom,fastrpc-cb 17300000.remoteproc:glink-edge:fastrpc:compute-cb@4:
> Adding to iommu group 22
> qcom,fastrpc-cb 17300000.remoteproc:glink-edge:fastrpc:compute-cb@5:
> Adding to iommu group 23
> qcom,apr 17300000.remoteproc:glink-edge.apr_audio_svc.-1.-1: Adding
> APR/GPR dev: aprsvc:apr-service:4:3
> qcom,apr 17300000.remoteproc:glink-edge.apr_audio_svc.-1.-1: Adding
> APR/GPR dev: aprsvc:apr-service:4:4
> SError Interrupt on CPU6, code 0x00000000be000411 -- SError
> <RB5 reboots into crash mode here..>
> 
> 
> Regards,
> Amit Pundir
> 
>> thanks,
>>
>> greg k-h