diff mbox series

[V3,05/15] clk: imx: imx8mp-audiomix: remove sdma root clock

Message ID 20240607133347.3291040-6-peng.fan@oss.nxp.com (mailing list archive)
State New, archived
Headers show
Series clk: imx: misc update/fix | expand

Commit Message

Peng Fan (OSS) June 7, 2024, 1:33 p.m. UTC 
From: Peng Fan <peng.fan@nxp.com>

There is an issue:
SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2] (bit-26) to 1

The workaround is:
As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled,
we just need to keep it on as reset state, don't touch it
in kernel, then every thing is same as before, if we register
the clock in clk-audiomix, then kernel will try to disable
it in idle.

Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block control")
Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
---
 drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
 1 file changed, 1 deletion(-)

Comments

Adam Ford July 14, 2024, 5:02 p.m. UTC | #1 
On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS) <peng.fan@oss.nxp.com> wrote:
>
> From: Peng Fan <peng.fan@nxp.com>
>
> There is an issue:
> SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2] (bit-26) to 1
>
> The workaround is:
> As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled,
> we just need to keep it on as reset state, don't touch it
> in kernel, then every thing is same as before, if we register
> the clock in clk-audiomix, then kernel will try to disable
> it in idle.
>
> Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block control")
> Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> Signed-off-by: Peng Fan <peng.fan@nxp.com>
> ---

With this patch,I found it broke the imx8mp-beacon board when running
audio through a codec connected to sai3.  Reverting this patch made
the crash go away.  Is there a way to mark the clock as critical so it
doesn't get idled?

adam
>  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
>  1 file changed, 1 deletion(-)
>
> diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c b/drivers/clk/imx/clk-imx8mp-audiomix.c
> index b381d6f784c8..88d8ba975b5a 100644
> --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> @@ -172,7 +172,6 @@ static struct clk_imx8mp_audiomix_sel sels[] = {
>         CLK_GATE("ocrama", OCRAMA_IPG),
>         CLK_GATE("aud2htx", AUD2HTX_IPG),
>         CLK_GATE("earc_phy", EARC_PHY),
> -       CLK_GATE("sdma2", SDMA2_ROOT),
>         CLK_GATE("sdma3", SDMA3_ROOT),
>         CLK_GATE("spba2", SPBA2_ROOT),
>         CLK_GATE("dsp", DSP_ROOT),
> --
> 2.37.1
>
>
Peng Fan July 15, 2024, 1:11 a.m. UTC | #2 
> Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> sdma root clock
> 
> On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS) <peng.fan@oss.nxp.com>
> wrote:
> >
> > From: Peng Fan <peng.fan@nxp.com>
> >
> > There is an issue:
> > SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2] (bit-
> 26) to 1
> >
> > The workaround is:
> > As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled, we just
> need
> > to keep it on as reset state, don't touch it in kernel, then every
> > thing is same as before, if we register the clock in clk-audiomix,
> > then kernel will try to disable it in idle.
> >
> > Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block
> control")
> > Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> > Signed-off-by: Peng Fan <peng.fan@nxp.com>
> > ---
> 
> With this patch,I found it broke the imx8mp-beacon board when
> running audio through a codec connected to sai3.  Reverting this patch
> made the crash go away.  Is there a way to mark the clock as critical so
> it doesn't get idled?

Mark it as critical means the blk ctrl needs to be powered on always.

You driver touched the SDMA2_ROOT clock?

Regards,
Peng.

> 
> adam
> >  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
> >  1 file changed, 1 deletion(-)
> >
> > diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > index b381d6f784c8..88d8ba975b5a 100644
> > --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > @@ -172,7 +172,6 @@ static struct clk_imx8mp_audiomix_sel sels[]
> = {
> >         CLK_GATE("ocrama", OCRAMA_IPG),
> >         CLK_GATE("aud2htx", AUD2HTX_IPG),
> >         CLK_GATE("earc_phy", EARC_PHY),
> > -       CLK_GATE("sdma2", SDMA2_ROOT),
> >         CLK_GATE("sdma3", SDMA3_ROOT),
> >         CLK_GATE("spba2", SPBA2_ROOT),
> >         CLK_GATE("dsp", DSP_ROOT),
> > --
> > 2.37.1
> >
> >
Adam Ford July 16, 2024, 12:02 a.m. UTC | #3 
On Sun, Jul 14, 2024 at 8:11 PM Peng Fan <peng.fan@nxp.com> wrote:
>
> > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> > sdma root clock
> >
> > On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS) <peng.fan@oss.nxp.com>
> > wrote:
> > >
> > > From: Peng Fan <peng.fan@nxp.com>
> > >
> > > There is an issue:
> > > SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2] (bit-
> > 26) to 1
> > >
> > > The workaround is:
> > > As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled, we just
> > need
> > > to keep it on as reset state, don't touch it in kernel, then every
> > > thing is same as before, if we register the clock in clk-audiomix,
> > > then kernel will try to disable it in idle.
> > >
> > > Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block
> > control")
> > > Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> > > Signed-off-by: Peng Fan <peng.fan@nxp.com>
> > > ---
> >
> > With this patch,I found it broke the imx8mp-beacon board when
> > running audio through a codec connected to sai3.  Reverting this patch
> > made the crash go away.  Is there a way to mark the clock as critical so
> > it doesn't get idled?
>
> Mark it as critical means the blk ctrl needs to be powered on always.
>
> You driver touched the SDMA2_ROOT clock?

I have a WM8962 CODEC connected through sai3.  Without this patch, the
sound plays just fine.  With this patch, I get the following:

[   73.245199] imx-sdma 30e10000.dma-controller: Timeout waiting for CH0 ready
[   73.252197] ------------[ cut here ]------------
[   73.256816] WARNING: CPU: 1 PID: 1080 at kernel/dma/mapping.c:586
dma_free_attrs+0x94/0xc0
[   73.256832] Modules linked in: overlay af_alg dw_hdmi_gp_audio
dw_hdmi_cec tpm_tis_spi snd_soc_hdmi_codec caam_jr tpm_tis_core
caamhash_desc caamalg_desc crypto_engine authenc libdes hantro_vpu
v4l2_jpeg v4l2_vp9 v4l2_h264 v4l2_mem2mem videobuf2_dma_contig
videobuf2_memops videobuf2_v4l2 videodev imx8mp_hdmi_tx btnxpuart nvme
videobuf2_common dw_hdmi mc nvme_core etnaviv imx_lcdif bluetooth
fsl_imx8_ddr_perf drm_display_helper dwmac_imx gpu_sched
phy_fsl_samsung_hdmi imx8mp_hdmi_pvi drm_dma_helper samsung_dsim
snd_soc_fsl_sai stmmac_platform snd_soc_fsl_xcvr snd_soc_fsl_micfil
snd_soc_fsl_aud2htx stmmac adv7511 ecdh_generic imx_pcm_dma
snd_soc_fsl_utils ecc pcs_xpcs cec tcpci flexcan tcpm hd3ss3220
snd_soc_wm8962 spi_imx can_dev typec at24 caam rtc_pcf85363 rtc_snvs
error snvs_pwrkey imx8mm_thermal crct10dif_ce snd_soc_imx_hdmi
coresight_tmc snd_soc_imx_card snd_soc_simple_card coresight_funnel
snd_soc_simple_card_utils display_connector snd_soc_dmic coresight
drm_kms_helper imx_cpufreq_dt imx_sdma mwifiex_sdio
[   73.256997]  mwifiex cfg80211 rfkill fuse drm backlight ipv6
[   73.257015] CPU: 1 UID: 0 PID: 1080 Comm: speaker-test Not tainted
6.10.0-rc7-next-20240709-ga90ee5931efe-dirty #9
[   73.257022] Hardware name: Beacon EmbeddedWorks i.MX8MPlus
Development kit (DT)
[   73.257025] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[   73.257031] pc : dma_free_attrs+0x94/0xc0
[   73.257037] lr : dma_free_attrs+0x50/0xc0
[   73.257041] sp : ffff800084f6b980
[   73.257043] x29: ffff800084f6b980 x28: 0000000000000010 x27: 00000000000003c2
[   73.257051] x26: 0000000000000005 x25: 0000000000000010 x24: 0000000000000000
[   73.257057] x23: 00000000c5504000 x22: 0000000000000000 x21: ffff800083183000
[   73.257064] x20: 00000000000000c0 x19: ffff0000c0e54410 x18: 0000000000000006
[   73.257071] x17: 0000000000000000 x16: 0000000000000000 x15: ffff800084f6b330
[   73.257078] x14: 0000000000000000 x13: ffff8000826845d8 x12: 0000000000000639
[   73.257085] x11: 0000000000000213 x10: ffff8000826dc5d8 x9 : ffff8000826845d8
[   73.257092] x8 : 00000000ffffefff x7 : ffff8000826dc5d8 x6 : 0000000000000040
[   73.257098] x5 : 0000000000000000 x4 : 0000000000000000 x3 : 00000000c5504000
[   73.257105] x2 : ffff800083183000 x1 : 0000000000000000 x0 : 00000000000000c0
[   73.257112] Call trace:
[   73.257115]  dma_free_attrs+0x94/0xc0
[   73.257121]  sdma_free_bd+0x60/0x6c [imx_sdma]
[   73.257130]  sdma_transfer_init+0x1e8/0x270 [imx_sdma]
[   73.257137]  sdma_prep_dma_cyclic+0x74/0x200 [imx_sdma]
[   73.257143]  snd_dmaengine_pcm_trigger+0xd8/0x18c
[   73.257152]  dmaengine_pcm_trigger+0x18/0x24
[   73.257159]  snd_soc_pcm_component_trigger+0x170/0x21c
[   73.257168]  soc_pcm_trigger+0xdc/0x1c8
[   73.257175]  snd_pcm_do_start+0x44/0x70
[   73.257183]  snd_pcm_action_single+0x48/0xa4
[   73.257189]  snd_pcm_action+0x80/0x9c
[   73.257195]  snd_pcm_start+0x24/0x30
[   73.257203]  __snd_pcm_lib_xfer+0x6a4/0x7d8
[   73.257208]  snd_pcm_common_ioctl+0x1140/0x1780
[   73.257215]  snd_pcm_ioctl+0x34/0x4c
[   73.257222]  __arm64_sys_ioctl+0xac/0xf0
[   73.257231]  invoke_syscall+0x48/0x114
[   73.257239]  el0_svc_common.constprop.0+0x40/0xe0
[   73.257246]  do_el0_svc+0x1c/0x28
[   73.257252]  el0_svc+0x34/0xd8
[   73.257260]  el0t_64_sync_handler+0x120/0x12c
[   73.257267]  el0t_64_sync+0x190/0x194
[   73.257273] ---[ end trace 0000000000000000 ]---
[   73.565659] fsl-sai 30c30000.sai: ASoC: error at
soc_component_trigger on 30c30000.sai: -12

Write error: -5,Input/output erro[   76.767120] imx-sdma
30e10000.dma-controller: Timeout waiting for CH0 ready

No sound is heard.

I haven't looked far into the driver to see what was touched, but I
was able to reproduce this by enabling and disabling the patch several
times, so it's repeatable. imx-sdma 30e10000.dma-controller seems to
correspond to sdma2.

adam
>
> Regards,
> Peng.
>
> >
> > adam
> > >  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
> > >  1 file changed, 1 deletion(-)
> > >
> > > diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > index b381d6f784c8..88d8ba975b5a 100644
> > > --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > @@ -172,7 +172,6 @@ static struct clk_imx8mp_audiomix_sel sels[]
> > = {
> > >         CLK_GATE("ocrama", OCRAMA_IPG),
> > >         CLK_GATE("aud2htx", AUD2HTX_IPG),
> > >         CLK_GATE("earc_phy", EARC_PHY),
> > > -       CLK_GATE("sdma2", SDMA2_ROOT),
> > >         CLK_GATE("sdma3", SDMA3_ROOT),
> > >         CLK_GATE("spba2", SPBA2_ROOT),
> > >         CLK_GATE("dsp", DSP_ROOT),
> > > --
> > > 2.37.1
> > >
> > >
Peng Fan July 16, 2024, 1:24 a.m. UTC | #4 
> Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> sdma root clock
> 
> On Sun, Jul 14, 2024 at 8:11 PM Peng Fan <peng.fan@nxp.com> wrote:
> >
> > > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> sdma
> > > root clock
> > >
> > > On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS)
> <peng.fan@oss.nxp.com>
> > > wrote:
> > > >
> > > > From: Peng Fan <peng.fan@nxp.com>
> > > >
> > > > There is an issue:
> > > > SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2]
> (bit-
> > > 26) to 1
> > > >
> > > > The workaround is:
> > > > As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled, we
> just
> > > need
> > > > to keep it on as reset state, don't touch it in kernel, then every
> > > > thing is same as before, if we register the clock in clk-audiomix,
> > > > then kernel will try to disable it in idle.
> > > >
> > > > Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block
> > > control")
> > > > Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> > > > Signed-off-by: Peng Fan <peng.fan@nxp.com>
> > > > ---
> > >
> > > With this patch,I found it broke the imx8mp-beacon board when
> > > running audio through a codec connected to sai3.  Reverting this
> > > patch made the crash go away.  Is there a way to mark the clock as
> > > critical so it doesn't get idled?
> >
> > Mark it as critical means the blk ctrl needs to be powered on always.
> >
> > You driver touched the SDMA2_ROOT clock?
> 
> I have a WM8962 CODEC connected through sai3.  Without this patch,
> the sound plays just fine.  With this patch, I get the following:

Try use this for SDMA2.
                                clocks = <&clk IMX8MP_CLK_AUDIO_AHB_ROOT>,                          
                                         <&clk IMX8MP_CLK_AUDIO_AHB_ROOT>;

I will give a look and see how to address the issue.

Thanks for the report.

Thanks,
Peng.

> 
> [   73.245199] imx-sdma 30e10000.dma-controller: Timeout waiting
> for CH0 ready
> [   73.252197] ------------[ cut here ]------------
> [   73.256816] WARNING: CPU: 1 PID: 1080 at
> kernel/dma/mapping.c:586
> dma_free_attrs+0x94/0xc0
> [   73.256832] Modules linked in: overlay af_alg dw_hdmi_gp_audio
> dw_hdmi_cec tpm_tis_spi snd_soc_hdmi_codec caam_jr tpm_tis_core
> caamhash_desc caamalg_desc crypto_engine authenc libdes
> hantro_vpu v4l2_jpeg v4l2_vp9 v4l2_h264 v4l2_mem2mem
> videobuf2_dma_contig videobuf2_memops videobuf2_v4l2 videodev
> imx8mp_hdmi_tx btnxpuart nvme videobuf2_common dw_hdmi mc
> nvme_core etnaviv imx_lcdif bluetooth fsl_imx8_ddr_perf
> drm_display_helper dwmac_imx gpu_sched phy_fsl_samsung_hdmi
> imx8mp_hdmi_pvi drm_dma_helper samsung_dsim snd_soc_fsl_sai
> stmmac_platform snd_soc_fsl_xcvr snd_soc_fsl_micfil
> snd_soc_fsl_aud2htx stmmac adv7511 ecdh_generic imx_pcm_dma
> snd_soc_fsl_utils ecc pcs_xpcs cec tcpci flexcan tcpm hd3ss3220
> snd_soc_wm8962 spi_imx can_dev typec at24 caam rtc_pcf85363
> rtc_snvs error snvs_pwrkey imx8mm_thermal crct10dif_ce
> snd_soc_imx_hdmi coresight_tmc snd_soc_imx_card
> snd_soc_simple_card coresight_funnel snd_soc_simple_card_utils
> display_connector snd_soc_dmic coresight drm_kms_helper
> imx_cpufreq_dt imx_sdma mwifiex_sdio
> [   73.256997]  mwifiex cfg80211 rfkill fuse drm backlight ipv6
> [   73.257015] CPU: 1 UID: 0 PID: 1080 Comm: speaker-test Not
> tainted
> 6.10.0-rc7-next-20240709-ga90ee5931efe-dirty #9
> [   73.257022] Hardware name: Beacon EmbeddedWorks i.MX8MPlus
> Development kit (DT)
> [   73.257025] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO -DIT -SSBS
> BTYPE=--)
> [   73.257031] pc : dma_free_attrs+0x94/0xc0
> [   73.257037] lr : dma_free_attrs+0x50/0xc0
> [   73.257041] sp : ffff800084f6b980
> [   73.257043] x29: ffff800084f6b980 x28: 0000000000000010 x27:
> 00000000000003c2
> [   73.257051] x26: 0000000000000005 x25: 0000000000000010 x24:
> 0000000000000000
> [   73.257057] x23: 00000000c5504000 x22: 0000000000000000 x21:
> ffff800083183000
> [   73.257064] x20: 00000000000000c0 x19: ffff0000c0e54410 x18:
> 0000000000000006
> [   73.257071] x17: 0000000000000000 x16: 0000000000000000 x15:
> ffff800084f6b330
> [   73.257078] x14: 0000000000000000 x13: ffff8000826845d8 x12:
> 0000000000000639
> [   73.257085] x11: 0000000000000213 x10: ffff8000826dc5d8 x9 :
> ffff8000826845d8
> [   73.257092] x8 : 00000000ffffefff x7 : ffff8000826dc5d8 x6 :
> 0000000000000040
> [   73.257098] x5 : 0000000000000000 x4 : 0000000000000000 x3 :
> 00000000c5504000
> [   73.257105] x2 : ffff800083183000 x1 : 0000000000000000 x0 :
> 00000000000000c0
> [   73.257112] Call trace:
> [   73.257115]  dma_free_attrs+0x94/0xc0
> [   73.257121]  sdma_free_bd+0x60/0x6c [imx_sdma]
> [   73.257130]  sdma_transfer_init+0x1e8/0x270 [imx_sdma]
> [   73.257137]  sdma_prep_dma_cyclic+0x74/0x200 [imx_sdma]
> [   73.257143]  snd_dmaengine_pcm_trigger+0xd8/0x18c
> [   73.257152]  dmaengine_pcm_trigger+0x18/0x24
> [   73.257159]  snd_soc_pcm_component_trigger+0x170/0x21c
> [   73.257168]  soc_pcm_trigger+0xdc/0x1c8
> [   73.257175]  snd_pcm_do_start+0x44/0x70
> [   73.257183]  snd_pcm_action_single+0x48/0xa4
> [   73.257189]  snd_pcm_action+0x80/0x9c
> [   73.257195]  snd_pcm_start+0x24/0x30
> [   73.257203]  __snd_pcm_lib_xfer+0x6a4/0x7d8
> [   73.257208]  snd_pcm_common_ioctl+0x1140/0x1780
> [   73.257215]  snd_pcm_ioctl+0x34/0x4c
> [   73.257222]  __arm64_sys_ioctl+0xac/0xf0
> [   73.257231]  invoke_syscall+0x48/0x114
> [   73.257239]  el0_svc_common.constprop.0+0x40/0xe0
> [   73.257246]  do_el0_svc+0x1c/0x28
> [   73.257252]  el0_svc+0x34/0xd8
> [   73.257260]  el0t_64_sync_handler+0x120/0x12c
> [   73.257267]  el0t_64_sync+0x190/0x194
> [   73.257273] ---[ end trace 0000000000000000 ]---
> [   73.565659] fsl-sai 30c30000.sai: ASoC: error at
> soc_component_trigger on 30c30000.sai: -12
> 
> Write error: -5,Input/output erro[   76.767120] imx-sdma
> 30e10000.dma-controller: Timeout waiting for CH0 ready
> 
> No sound is heard.
> 
> I haven't looked far into the driver to see what was touched, but I was
> able to reproduce this by enabling and disabling the patch several
> times, so it's repeatable. imx-sdma 30e10000.dma-controller seems to
> correspond to sdma2.
> 
> adam
> >
> > Regards,
> > Peng.
> >
> > >
> > > adam
> > > >  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
> > > >  1 file changed, 1 deletion(-)
> > > >
> > > > diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > index b381d6f784c8..88d8ba975b5a 100644
> > > > --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > @@ -172,7 +172,6 @@ static struct clk_imx8mp_audiomix_sel
> sels[]
> > > = {
> > > >         CLK_GATE("ocrama", OCRAMA_IPG),
> > > >         CLK_GATE("aud2htx", AUD2HTX_IPG),
> > > >         CLK_GATE("earc_phy", EARC_PHY),
> > > > -       CLK_GATE("sdma2", SDMA2_ROOT),
> > > >         CLK_GATE("sdma3", SDMA3_ROOT),
> > > >         CLK_GATE("spba2", SPBA2_ROOT),
> > > >         CLK_GATE("dsp", DSP_ROOT),
> > > > --
> > > > 2.37.1
> > > >
> > > >
Adam Ford July 16, 2024, 1:50 a.m. UTC | #5 
On Mon, Jul 15, 2024 at 8:25 PM Peng Fan <peng.fan@nxp.com> wrote:
>
> > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> > sdma root clock
> >
> > On Sun, Jul 14, 2024 at 8:11 PM Peng Fan <peng.fan@nxp.com> wrote:
> > >
> > > > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> > sdma
> > > > root clock
> > > >
> > > > On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS)
> > <peng.fan@oss.nxp.com>
> > > > wrote:
> > > > >
> > > > > From: Peng Fan <peng.fan@nxp.com>
> > > > >
> > > > > There is an issue:
> > > > > SDMA3 can't work without setting AUDIOMIX_CLKEN0[SDMA2]
> > (bit-
> > > > 26) to 1
> > > > >
> > > > > The workaround is:
> > > > > As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled, we
> > just
> > > > need
> > > > > to keep it on as reset state, don't touch it in kernel, then every
> > > > > thing is same as before, if we register the clock in clk-audiomix,
> > > > > then kernel will try to disable it in idle.
> > > > >
> > > > > Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block
> > > > control")
> > > > > Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> > > > > Signed-off-by: Peng Fan <peng.fan@nxp.com>
> > > > > ---
> > > >
> > > > With this patch,I found it broke the imx8mp-beacon board when
> > > > running audio through a codec connected to sai3.  Reverting this
> > > > patch made the crash go away.  Is there a way to mark the clock as
> > > > critical so it doesn't get idled?
> > >
> > > Mark it as critical means the blk ctrl needs to be powered on always.
> > >
> > > You driver touched the SDMA2_ROOT clock?
> >
> > I have a WM8962 CODEC connected through sai3.  Without this patch,
> > the sound plays just fine.  With this patch, I get the following:
>
> Try use this for SDMA2.
>                                 clocks = <&clk IMX8MP_CLK_AUDIO_AHB_ROOT>,
>                                          <&clk IMX8MP_CLK_AUDIO_AHB_ROOT>;
>

Unfortunately, that didn't work either.  Changing the SDMA2 clocks to
the above yields the following:

[   20.714987] fsl-micfil-dai 30ca0000.audio-controller: failed to pcm register
[   20.728904] fsl-aud2htx 30cb0000.aud2htx: failed to pcm register
[   20.755070] fsl,imx8mp-audio-xcvr 30cc0000.xcvr: failed to pcm register
[   20.766373] platform 30c50000.sai: deferred probe pending: fsl-sai:
PCM DMA init failed
[   20.774497] platform 30ca0000.audio-controller: deferred probe
pending: (reason unknown)
[   20.786387] platform 30cb0000.aud2htx: deferred probe pending:
(reason unknown)
[   20.793761] platform sound-dmic: deferred probe pending:
asoc-simple-card: parse error
[   20.806374] platform sound-wm8962: deferred probe pending:
asoc-simple-card: parse error
[   20.814526] platform sound-hdmi: deferred probe pending: imx-hdmi:
snd_soc_register_card failed
[   20.826370] platform 30c30000.sai: deferred probe pending: fsl-sai:
PCM DMA init failed
[   20.834429] platform sound-xcvr: deferred probe pending: imx-card:
XCVR PCM: error getting cpu dai info
[   20.846402] platform 30cc0000.xcvr: deferred probe pending: (reason unknown)
[   20.858378] platform sound-adv7535: deferred probe pending:
asoc-simple-card: parse error


adam

> I will give a look and see how to address the issue.
>
> Thanks for the report.
>
> Thanks,
> Peng.
>
> >
> > [   73.245199] imx-sdma 30e10000.dma-controller: Timeout waiting
> > for CH0 ready
> > [   73.252197] ------------[ cut here ]------------
> > [   73.256816] WARNING: CPU: 1 PID: 1080 at
> > kernel/dma/mapping.c:586
> > dma_free_attrs+0x94/0xc0
> > [   73.256832] Modules linked in: overlay af_alg dw_hdmi_gp_audio
> > dw_hdmi_cec tpm_tis_spi snd_soc_hdmi_codec caam_jr tpm_tis_core
> > caamhash_desc caamalg_desc crypto_engine authenc libdes
> > hantro_vpu v4l2_jpeg v4l2_vp9 v4l2_h264 v4l2_mem2mem
> > videobuf2_dma_contig videobuf2_memops videobuf2_v4l2 videodev
> > imx8mp_hdmi_tx btnxpuart nvme videobuf2_common dw_hdmi mc
> > nvme_core etnaviv imx_lcdif bluetooth fsl_imx8_ddr_perf
> > drm_display_helper dwmac_imx gpu_sched phy_fsl_samsung_hdmi
> > imx8mp_hdmi_pvi drm_dma_helper samsung_dsim snd_soc_fsl_sai
> > stmmac_platform snd_soc_fsl_xcvr snd_soc_fsl_micfil
> > snd_soc_fsl_aud2htx stmmac adv7511 ecdh_generic imx_pcm_dma
> > snd_soc_fsl_utils ecc pcs_xpcs cec tcpci flexcan tcpm hd3ss3220
> > snd_soc_wm8962 spi_imx can_dev typec at24 caam rtc_pcf85363
> > rtc_snvs error snvs_pwrkey imx8mm_thermal crct10dif_ce
> > snd_soc_imx_hdmi coresight_tmc snd_soc_imx_card
> > snd_soc_simple_card coresight_funnel snd_soc_simple_card_utils
> > display_connector snd_soc_dmic coresight drm_kms_helper
> > imx_cpufreq_dt imx_sdma mwifiex_sdio
> > [   73.256997]  mwifiex cfg80211 rfkill fuse drm backlight ipv6
> > [   73.257015] CPU: 1 UID: 0 PID: 1080 Comm: speaker-test Not
> > tainted
> > 6.10.0-rc7-next-20240709-ga90ee5931efe-dirty #9
> > [   73.257022] Hardware name: Beacon EmbeddedWorks i.MX8MPlus
> > Development kit (DT)
> > [   73.257025] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO -DIT -SSBS
> > BTYPE=--)
> > [   73.257031] pc : dma_free_attrs+0x94/0xc0
> > [   73.257037] lr : dma_free_attrs+0x50/0xc0
> > [   73.257041] sp : ffff800084f6b980
> > [   73.257043] x29: ffff800084f6b980 x28: 0000000000000010 x27:
> > 00000000000003c2
> > [   73.257051] x26: 0000000000000005 x25: 0000000000000010 x24:
> > 0000000000000000
> > [   73.257057] x23: 00000000c5504000 x22: 0000000000000000 x21:
> > ffff800083183000
> > [   73.257064] x20: 00000000000000c0 x19: ffff0000c0e54410 x18:
> > 0000000000000006
> > [   73.257071] x17: 0000000000000000 x16: 0000000000000000 x15:
> > ffff800084f6b330
> > [   73.257078] x14: 0000000000000000 x13: ffff8000826845d8 x12:
> > 0000000000000639
> > [   73.257085] x11: 0000000000000213 x10: ffff8000826dc5d8 x9 :
> > ffff8000826845d8
> > [   73.257092] x8 : 00000000ffffefff x7 : ffff8000826dc5d8 x6 :
> > 0000000000000040
> > [   73.257098] x5 : 0000000000000000 x4 : 0000000000000000 x3 :
> > 00000000c5504000
> > [   73.257105] x2 : ffff800083183000 x1 : 0000000000000000 x0 :
> > 00000000000000c0
> > [   73.257112] Call trace:
> > [   73.257115]  dma_free_attrs+0x94/0xc0
> > [   73.257121]  sdma_free_bd+0x60/0x6c [imx_sdma]
> > [   73.257130]  sdma_transfer_init+0x1e8/0x270 [imx_sdma]
> > [   73.257137]  sdma_prep_dma_cyclic+0x74/0x200 [imx_sdma]
> > [   73.257143]  snd_dmaengine_pcm_trigger+0xd8/0x18c
> > [   73.257152]  dmaengine_pcm_trigger+0x18/0x24
> > [   73.257159]  snd_soc_pcm_component_trigger+0x170/0x21c
> > [   73.257168]  soc_pcm_trigger+0xdc/0x1c8
> > [   73.257175]  snd_pcm_do_start+0x44/0x70
> > [   73.257183]  snd_pcm_action_single+0x48/0xa4
> > [   73.257189]  snd_pcm_action+0x80/0x9c
> > [   73.257195]  snd_pcm_start+0x24/0x30
> > [   73.257203]  __snd_pcm_lib_xfer+0x6a4/0x7d8
> > [   73.257208]  snd_pcm_common_ioctl+0x1140/0x1780
> > [   73.257215]  snd_pcm_ioctl+0x34/0x4c
> > [   73.257222]  __arm64_sys_ioctl+0xac/0xf0
> > [   73.257231]  invoke_syscall+0x48/0x114
> > [   73.257239]  el0_svc_common.constprop.0+0x40/0xe0
> > [   73.257246]  do_el0_svc+0x1c/0x28
> > [   73.257252]  el0_svc+0x34/0xd8
> > [   73.257260]  el0t_64_sync_handler+0x120/0x12c
> > [   73.257267]  el0t_64_sync+0x190/0x194
> > [   73.257273] ---[ end trace 0000000000000000 ]---
> > [   73.565659] fsl-sai 30c30000.sai: ASoC: error at
> > soc_component_trigger on 30c30000.sai: -12
> >
> > Write error: -5,Input/output erro[   76.767120] imx-sdma
> > 30e10000.dma-controller: Timeout waiting for CH0 ready
> >
> > No sound is heard.
> >
> > I haven't looked far into the driver to see what was touched, but I was
> > able to reproduce this by enabling and disabling the patch several
> > times, so it's repeatable. imx-sdma 30e10000.dma-controller seems to
> > correspond to sdma2.
> >
> > adam
> > >
> > > Regards,
> > > Peng.
> > >
> > > >
> > > > adam
> > > > >  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
> > > > >  1 file changed, 1 deletion(-)
> > > > >
> > > > > diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > index b381d6f784c8..88d8ba975b5a 100644
> > > > > --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > @@ -172,7 +172,6 @@ static struct clk_imx8mp_audiomix_sel
> > sels[]
> > > > = {
> > > > >         CLK_GATE("ocrama", OCRAMA_IPG),
> > > > >         CLK_GATE("aud2htx", AUD2HTX_IPG),
> > > > >         CLK_GATE("earc_phy", EARC_PHY),
> > > > > -       CLK_GATE("sdma2", SDMA2_ROOT),
> > > > >         CLK_GATE("sdma3", SDMA3_ROOT),
> > > > >         CLK_GATE("spba2", SPBA2_ROOT),
> > > > >         CLK_GATE("dsp", DSP_ROOT),
> > > > > --
> > > > > 2.37.1
> > > > >
> > > > >
Peng Fan July 17, 2024, 12:21 p.m. UTC | #6 
> Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> sdma root clock
> 
> On Mon, Jul 15, 2024 at 8:25 PM Peng Fan <peng.fan@nxp.com>
> wrote:
> >
> > > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix: remove
> sdma
> > > root clock
> > >
> > > On Sun, Jul 14, 2024 at 8:11 PM Peng Fan <peng.fan@nxp.com>
> wrote:
> > > >
> > > > > Subject: Re: [PATCH V3 05/15] clk: imx: imx8mp-audiomix:
> remove
> > > sdma
> > > > > root clock
> > > > >
> > > > > On Fri, Jun 7, 2024 at 8:28 AM Peng Fan (OSS)
> > > <peng.fan@oss.nxp.com>
> > > > > wrote:
> > > > > >
> > > > > > From: Peng Fan <peng.fan@nxp.com>
> > > > > >
> > > > > > There is an issue:
> > > > > > SDMA3 can't work without setting
> AUDIOMIX_CLKEN0[SDMA2]
> > > (bit-
> > > > > 26) to 1
> > > > > >
> > > > > > The workaround is:
> > > > > > As the reset state of AUDIOMIX_CLKEN0[SDMA2] is enabled,
> we
> > > just
> > > > > need
> > > > > > to keep it on as reset state, don't touch it in kernel, then
> > > > > > every thing is same as before, if we register the clock in
> > > > > > clk-audiomix, then kernel will try to disable it in idle.
> > > > > >
> > > > > > Fixes: 6cd95f7b151c ("clk: imx: imx8mp: Add audiomix block
> > > > > control")
> > > > > > Reviewed-by: Shengjiu Wang <shengjiu.wang@nxp.com>
> > > > > > Signed-off-by: Peng Fan <peng.fan@nxp.com>
> > > > > > ---
> > > > >
> > > > > With this patch,I found it broke the imx8mp-beacon board
> when
> > > > > running audio through a codec connected to sai3.  Reverting
> this
> > > > > patch made the crash go away.  Is there a way to mark the clock
> > > > > as critical so it doesn't get idled?
> > > >
> > > > Mark it as critical means the blk ctrl needs to be powered on
> always.
> > > >
> > > > You driver touched the SDMA2_ROOT clock?
> > >
> > > I have a WM8962 CODEC connected through sai3.  Without this
> patch,
> > > the sound plays just fine.  With this patch, I get the following:
> >
> > Try use this for SDMA2.
> >                                 clocks = <&clk IMX8MP_CLK_AUDIO_AHB_ROOT>,
> >                                          <&clk
> > IMX8MP_CLK_AUDIO_AHB_ROOT>;
> >
> 
> Unfortunately, that didn't work either.  Changing the SDMA2 clocks to
> the above yields the following:

Shengjiu,

Do you have any ideas?

Thanks,
Peng.

> 
> [   20.714987] fsl-micfil-dai 30ca0000.audio-controller: failed to pcm
> register
> [   20.728904] fsl-aud2htx 30cb0000.aud2htx: failed to pcm register
> [   20.755070] fsl,imx8mp-audio-xcvr 30cc0000.xcvr: failed to pcm
> register
> [   20.766373] platform 30c50000.sai: deferred probe pending: fsl-sai:
> PCM DMA init failed
> [   20.774497] platform 30ca0000.audio-controller: deferred probe
> pending: (reason unknown)
> [   20.786387] platform 30cb0000.aud2htx: deferred probe pending:
> (reason unknown)
> [   20.793761] platform sound-dmic: deferred probe pending:
> asoc-simple-card: parse error
> [   20.806374] platform sound-wm8962: deferred probe pending:
> asoc-simple-card: parse error
> [   20.814526] platform sound-hdmi: deferred probe pending: imx-
> hdmi:
> snd_soc_register_card failed
> [   20.826370] platform 30c30000.sai: deferred probe pending: fsl-sai:
> PCM DMA init failed
> [   20.834429] platform sound-xcvr: deferred probe pending: imx-card:
> XCVR PCM: error getting cpu dai info
> [   20.846402] platform 30cc0000.xcvr: deferred probe pending:
> (reason unknown)
> [   20.858378] platform sound-adv7535: deferred probe pending:
> asoc-simple-card: parse error
> 
> 
> adam
> 
> > I will give a look and see how to address the issue.
> >
> > Thanks for the report.
> >
> > Thanks,
> > Peng.
> >
> > >
> > > [   73.245199] imx-sdma 30e10000.dma-controller: Timeout
> waiting
> > > for CH0 ready
> > > [   73.252197] ------------[ cut here ]------------
> > > [   73.256816] WARNING: CPU: 1 PID: 1080 at
> > > kernel/dma/mapping.c:586
> > > dma_free_attrs+0x94/0xc0
> > > [   73.256832] Modules linked in: overlay af_alg
> dw_hdmi_gp_audio
> > > dw_hdmi_cec tpm_tis_spi snd_soc_hdmi_codec caam_jr
> tpm_tis_core
> > > caamhash_desc caamalg_desc crypto_engine authenc libdes
> hantro_vpu
> > > v4l2_jpeg v4l2_vp9 v4l2_h264 v4l2_mem2mem
> videobuf2_dma_contig
> > > videobuf2_memops videobuf2_v4l2 videodev imx8mp_hdmi_tx
> btnxpuart
> > > nvme videobuf2_common dw_hdmi mc nvme_core etnaviv
> imx_lcdif
> > > bluetooth fsl_imx8_ddr_perf drm_display_helper dwmac_imx
> gpu_sched
> > > phy_fsl_samsung_hdmi imx8mp_hdmi_pvi drm_dma_helper
> samsung_dsim
> > > snd_soc_fsl_sai stmmac_platform snd_soc_fsl_xcvr
> snd_soc_fsl_micfil
> > > snd_soc_fsl_aud2htx stmmac adv7511 ecdh_generic imx_pcm_dma
> > > snd_soc_fsl_utils ecc pcs_xpcs cec tcpci flexcan tcpm hd3ss3220
> > > snd_soc_wm8962 spi_imx can_dev typec at24 caam rtc_pcf85363
> rtc_snvs
> > > error snvs_pwrkey imx8mm_thermal crct10dif_ce
> snd_soc_imx_hdmi
> > > coresight_tmc snd_soc_imx_card snd_soc_simple_card
> coresight_funnel
> > > snd_soc_simple_card_utils display_connector snd_soc_dmic
> coresight
> > > drm_kms_helper imx_cpufreq_dt imx_sdma mwifiex_sdio
> > > [   73.256997]  mwifiex cfg80211 rfkill fuse drm backlight ipv6
> > > [   73.257015] CPU: 1 UID: 0 PID: 1080 Comm: speaker-test Not
> > > tainted
> > > 6.10.0-rc7-next-20240709-ga90ee5931efe-dirty #9
> > > [   73.257022] Hardware name: Beacon EmbeddedWorks
> i.MX8MPlus
> > > Development kit (DT)
> > > [   73.257025] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO -DIT -
> SSBS
> > > BTYPE=--)
> > > [   73.257031] pc : dma_free_attrs+0x94/0xc0
> > > [   73.257037] lr : dma_free_attrs+0x50/0xc0
> > > [   73.257041] sp : ffff800084f6b980
> > > [   73.257043] x29: ffff800084f6b980 x28: 0000000000000010 x27:
> > > 00000000000003c2
> > > [   73.257051] x26: 0000000000000005 x25: 0000000000000010
> x24:
> > > 0000000000000000
> > > [   73.257057] x23: 00000000c5504000 x22: 0000000000000000
> x21:
> > > ffff800083183000
> > > [   73.257064] x20: 00000000000000c0 x19: ffff0000c0e54410 x18:
> > > 0000000000000006
> > > [   73.257071] x17: 0000000000000000 x16: 0000000000000000
> x15:
> > > ffff800084f6b330
> > > [   73.257078] x14: 0000000000000000 x13: ffff8000826845d8
> x12:
> > > 0000000000000639
> > > [   73.257085] x11: 0000000000000213 x10: ffff8000826dc5d8 x9 :
> > > ffff8000826845d8
> > > [   73.257092] x8 : 00000000ffffefff x7 : ffff8000826dc5d8 x6 :
> > > 0000000000000040
> > > [   73.257098] x5 : 0000000000000000 x4 : 0000000000000000
> x3 :
> > > 00000000c5504000
> > > [   73.257105] x2 : ffff800083183000 x1 : 0000000000000000 x0 :
> > > 00000000000000c0
> > > [   73.257112] Call trace:
> > > [   73.257115]  dma_free_attrs+0x94/0xc0
> > > [   73.257121]  sdma_free_bd+0x60/0x6c [imx_sdma]
> > > [   73.257130]  sdma_transfer_init+0x1e8/0x270 [imx_sdma]
> > > [   73.257137]  sdma_prep_dma_cyclic+0x74/0x200 [imx_sdma]
> > > [   73.257143]  snd_dmaengine_pcm_trigger+0xd8/0x18c
> > > [   73.257152]  dmaengine_pcm_trigger+0x18/0x24
> > > [   73.257159]  snd_soc_pcm_component_trigger+0x170/0x21c
> > > [   73.257168]  soc_pcm_trigger+0xdc/0x1c8
> > > [   73.257175]  snd_pcm_do_start+0x44/0x70
> > > [   73.257183]  snd_pcm_action_single+0x48/0xa4
> > > [   73.257189]  snd_pcm_action+0x80/0x9c
> > > [   73.257195]  snd_pcm_start+0x24/0x30
> > > [   73.257203]  __snd_pcm_lib_xfer+0x6a4/0x7d8
> > > [   73.257208]  snd_pcm_common_ioctl+0x1140/0x1780
> > > [   73.257215]  snd_pcm_ioctl+0x34/0x4c
> > > [   73.257222]  __arm64_sys_ioctl+0xac/0xf0
> > > [   73.257231]  invoke_syscall+0x48/0x114
> > > [   73.257239]  el0_svc_common.constprop.0+0x40/0xe0
> > > [   73.257246]  do_el0_svc+0x1c/0x28
> > > [   73.257252]  el0_svc+0x34/0xd8
> > > [   73.257260]  el0t_64_sync_handler+0x120/0x12c
> > > [   73.257267]  el0t_64_sync+0x190/0x194
> > > [   73.257273] ---[ end trace 0000000000000000 ]---
> > > [   73.565659] fsl-sai 30c30000.sai: ASoC: error at
> > > soc_component_trigger on 30c30000.sai: -12
> > >
> > > Write error: -5,Input/output erro[   76.767120] imx-sdma
> > > 30e10000.dma-controller: Timeout waiting for CH0 ready
> > >
> > > No sound is heard.
> > >
> > > I haven't looked far into the driver to see what was touched, but I
> > > was able to reproduce this by enabling and disabling the patch
> > > several times, so it's repeatable. imx-sdma 30e10000.dma-
> controller
> > > seems to correspond to sdma2.
> > >
> > > adam
> > > >
> > > > Regards,
> > > > Peng.
> > > >
> > > > >
> > > > > adam
> > > > > >  drivers/clk/imx/clk-imx8mp-audiomix.c | 1 -
> > > > > >  1 file changed, 1 deletion(-)
> > > > > >
> > > > > > diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > > b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > > index b381d6f784c8..88d8ba975b5a 100644
> > > > > > --- a/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > > +++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
> > > > > > @@ -172,7 +172,6 @@ static struct
> clk_imx8mp_audiomix_sel
> > > sels[]
> > > > > = {
> > > > > >         CLK_GATE("ocrama", OCRAMA_IPG),
> > > > > >         CLK_GATE("aud2htx", AUD2HTX_IPG),
> > > > > >         CLK_GATE("earc_phy", EARC_PHY),
> > > > > > -       CLK_GATE("sdma2", SDMA2_ROOT),
> > > > > >         CLK_GATE("sdma3", SDMA3_ROOT),
> > > > > >         CLK_GATE("spba2", SPBA2_ROOT),
> > > > > >         CLK_GATE("dsp", DSP_ROOT),
> > > > > > --
> > > > > > 2.37.1
> > > > > >
> > > > > >
diff mbox series

Patch

diff --git a/drivers/clk/imx/clk-imx8mp-audiomix.c b/drivers/clk/imx/clk-imx8mp-audiomix.c
index b381d6f784c8..88d8ba975b5a 100644
--- a/drivers/clk/imx/clk-imx8mp-audiomix.c
+++ b/drivers/clk/imx/clk-imx8mp-audiomix.c
@@ -172,7 +172,6 @@  static struct clk_imx8mp_audiomix_sel sels[] = {
 	CLK_GATE("ocrama", OCRAMA_IPG),
 	CLK_GATE("aud2htx", AUD2HTX_IPG),
 	CLK_GATE("earc_phy", EARC_PHY),
-	CLK_GATE("sdma2", SDMA2_ROOT),
 	CLK_GATE("sdma3", SDMA3_ROOT),
 	CLK_GATE("spba2", SPBA2_ROOT),
 	CLK_GATE("dsp", DSP_ROOT),