Re: [PATCH v2 2/2] ASoC: SOF: mediatek: add adsp debug dump

From: Trevor Wu (吳文良)
Date: Wed May 31 2023 - 22:29:16 EST


On Tue, 2023-05-30 at 15:12 +0100, Mark Brown wrote:
> On Tue, May 23, 2023 at 10:59:33AM +0800, Trevor Wu wrote:
> > Add mt8188 and mt8186 .dbg_dump callback to print some information
> > when
> > DSP panic occurs.
>
> This breaks an x86_64 allmodconfig build:
>
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c: In
> function ‘mt8186_adsp_dump’:
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:483:56:
> error: ‘DSP_PDEBUGDATA’ undeclared (first use in this function); did
> you mean ‘DSP_PDEBUGPC’?
> 483 | dbg_data = snd_sof_dsp_read(sdev, DSP_REG_BAR,
> DSP_PDEBUGDATA);
> | ^~~~~~
> ~~~~~~~~
> | DSP_PD
> EBUGPC
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:483:56:
> note: each undeclared identifier is reported only once for each
> function it appears in
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:484:56:
> error: ‘DSP_PDEBUGINST’ undeclared (first use in this function); did
> you mean ‘DSP_PDEBUGPC’?
> 484 | dbg_inst = snd_sof_dsp_read(sdev, DSP_REG_BAR,
> DSP_PDEBUGINST);
> | ^~~~~~
> ~~~~~~~~
> | DSP_PD
> EBUGPC
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:485:59:
> error: ‘DSP_PDEBUGLS0STAT’ undeclared (first use in this function)
> 485 | dbg_ls0stat = snd_sof_dsp_read(sdev, DSP_REG_BAR,
> DSP_PDEBUGLS0STAT);
> | ^~~
> ~~~~~~~~~~~~~~
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:486:58:
> error: ‘DSP_PDEBUGSTATUS’ undeclared (first use in this function);
> did you mean ‘DSP_PDEBUGPC’?
> 486 | dbg_status = snd_sof_dsp_read(sdev, DSP_REG_BAR,
> DSP_PDEBUGSTATUS);
> | ^~~~
> ~~~~~~~~~~~~
> | DSP_
> PDEBUGPC
> /build/stage/linux/sound/soc/sof/mediatek/mt8186/mt8186.c:487:57:
> error: ‘DSP_PFAULTINFO’ undeclared (first use in this function)
> 487 | faultinfo = snd_sof_dsp_read(sdev, DSP_REG_BAR,
> DSP_PFAULTINFO);
> | ^~~~~
> ~~~~~~~~~


Sorry, I didn't notice that I missed the header which possibly caused
by the wrong format-patch command. I checked the same patch in v1, and
the header was included in the series. I will send v3 to resolve the
problem.

Thanks,
Trevor