Re: [PATCH 5.10 384/390] Revert "drm/amdgpu: move nbio sdma_doorbell_range() into sdma code for vega"

[Greg Kroah-Hartman]

On Tue, Oct 25, 2022 at 11:02:33AM +0200, Salvatore Bonaccorso wrote:
> Hi Greg,
> 
> On Mon, Oct 24, 2022 at 01:33:01PM +0200, Greg Kroah-Hartman wrote:
> > From: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx>
> > 
> > This reverts commit 9f55f36f749a7608eeef57d7d72991a9bd557341 which is
> > commit e3163bc8ffdfdb405e10530b140135b2ee487f89 upstream.
> > 
> > This commit causes repeated WARN_ONs from
> > 
> > drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amd
> > gpu_dm.c:7391 amdgpu_dm_atomic_commit_tail+0x23b9/0x2430 [amdgpu]
> > 
> > dmesg fills up with the following messages and drm initialization takes
> > a very long time.
> > 
> > Cc: <stable@xxxxxxxxxxxxxxx>    # 5.10
> > Signed-off-by: Shuah Khan <skhan@xxxxxxxxxxxxxxxxxxx>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
> > ---
> >  drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c |    5 -----
> >  drivers/gpu/drm/amd/amdgpu/soc15.c     |   25 +++++++++++++++++++++++++
> >  2 files changed, 25 insertions(+), 5 deletions(-)
> > 
> > --- a/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v4_0.c
> > @@ -1475,11 +1475,6 @@ static int sdma_v4_0_start(struct amdgpu
> >  		WREG32_SDMA(i, mmSDMA0_CNTL, temp);
> >  
> >  		if (!amdgpu_sriov_vf(adev)) {
> > -			ring = &adev->sdma.instance[i].ring;
> > -			adev->nbio.funcs->sdma_doorbell_range(adev, i,
> > -				ring->use_doorbell, ring->doorbell_index,
> > -				adev->doorbell_index.sdma_doorbell_range);
> > -
> >  			/* unhalt engine */
> >  			temp = RREG32_SDMA(i, mmSDMA0_F32_CNTL);
> >  			temp = REG_SET_FIELD(temp, SDMA0_F32_CNTL, HALT, 0);
> > --- a/drivers/gpu/drm/amd/amdgpu/soc15.c
> > +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c
> > @@ -1332,6 +1332,25 @@ static int soc15_common_sw_fini(void *ha
> >  	return 0;
> >  }
> >  
> > +static void soc15_doorbell_range_init(struct amdgpu_device *adev)
> > +{
> > +	int i;
> > +	struct amdgpu_ring *ring;
> > +
> > +	/* sdma/ih doorbell range are programed by hypervisor */
> > +	if (!amdgpu_sriov_vf(adev)) {
> > +		for (i = 0; i < adev->sdma.num_instances; i++) {
> > +			ring = &adev->sdma.instance[i].ring;
> > +			adev->nbio.funcs->sdma_doorbell_range(adev, i,
> > +				ring->use_doorbell, ring->doorbell_index,
> > +				adev->doorbell_index.sdma_doorbell_range);
> > +		}
> > +
> > +		adev->nbio.funcs->ih_doorbell_range(adev, adev->irq.ih.use_doorbell,
> > +						adev->irq.ih.doorbell_index);
> > +	}
> > +}
> > +
> >  static int soc15_common_hw_init(void *handle)
> >  {
> >  	struct amdgpu_device *adev = (struct amdgpu_device *)handle;
> > @@ -1351,6 +1370,12 @@ static int soc15_common_hw_init(void *ha
> >  
> >  	/* enable the doorbell aperture */
> >  	soc15_enable_doorbell_aperture(adev, true);
> > +	/* HW doorbell routing policy: doorbell writing not
> > +	 * in SDMA/IH/MM/ACV range will be routed to CP. So
> > +	 * we need to init SDMA/IH/MM/ACV doorbell range prior
> > +	 * to CP ip block init and ring test.
> > +	 */
> > +	soc15_doorbell_range_init(adev);
> >  
> >  	return 0;
> >  }
> 
> Can you please as well revert 7b0db849ea030a70b8fb9c9afec67c81f955482e
> on top?
> 
> See https://lore.kernel.org/stable/BL1PR12MB5144F3CC640A18DF0C36E414F72E9@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx/
> 
> Both of these reverts need to be applied to fix regressions which were
> reported in https://gitlab.freedesktop.org/drm/amd/-/issues/2216 and
> downstream in Debian (https://bugs.debian.org/1022025).
> 
> If it is now not anymore possible for 5.10.150 can you pick the revert
> for 5.10.151?

Now queued up.

greg k-h