| Message ID | 1411061204-721-1-git-send-email-alexander.deucher@amd.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Am 18.09.2014 um 19:26 schrieb Alex Deucher: > Otherwise we may lose the DMA golden settings which can > lead to hangs, etc. Don't we still want the soft reset at some point, e.g. when the engine hangs? I would rather move that to another function and call it before loading the golden values. Christian. > > bug: > https://www.libreoffice.org/bugzilla/show_bug.cgi?id=83500 > > Signed-off-by: Alex Deucher <alexander.deucher@amd.com> > Cc: stable@vger.kernel.org > --- > drivers/gpu/drm/radeon/ni_dma.c | 6 ------ > 1 file changed, 6 deletions(-) > > diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c > index 8a3e622..f26f0a9 100644 > --- a/drivers/gpu/drm/radeon/ni_dma.c > +++ b/drivers/gpu/drm/radeon/ni_dma.c > @@ -191,12 +191,6 @@ int cayman_dma_resume(struct radeon_device *rdev) > u32 reg_offset, wb_offset; > int i, r; > > - /* Reset dma */ > - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); > - RREG32(SRBM_SOFT_RESET); > - udelay(50); > - WREG32(SRBM_SOFT_RESET, 0); > - > for (i = 0; i < 2; i++) { > if (i == 0) { > ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
On Thu, Sep 18, 2014 at 1:51 PM, Christian König <deathsimple@vodafone.de> wrote: > Am 18.09.2014 um 19:26 schrieb Alex Deucher: >> >> Otherwise we may lose the DMA golden settings which can >> lead to hangs, etc. > > > Don't we still want the soft reset at some point, e.g. when the engine > hangs? > > I would rather move that to another function and call it before loading the > golden values. We have separate soft reset code for dma already. This is just a one-off soft reset in the dma init code. The golden registers are reloaded after a GPU reset so we should be fine. Alex > > Christian. > > >> >> bug: >> https://www.libreoffice.org/bugzilla/show_bug.cgi?id=83500 >> >> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> >> Cc: stable@vger.kernel.org >> --- >> drivers/gpu/drm/radeon/ni_dma.c | 6 ------ >> 1 file changed, 6 deletions(-) >> >> diff --git a/drivers/gpu/drm/radeon/ni_dma.c >> b/drivers/gpu/drm/radeon/ni_dma.c >> index 8a3e622..f26f0a9 100644 >> --- a/drivers/gpu/drm/radeon/ni_dma.c >> +++ b/drivers/gpu/drm/radeon/ni_dma.c >> @@ -191,12 +191,6 @@ int cayman_dma_resume(struct radeon_device *rdev) >> u32 reg_offset, wb_offset; >> int i, r; >> - /* Reset dma */ >> - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); >> - RREG32(SRBM_SOFT_RESET); >> - udelay(50); >> - WREG32(SRBM_SOFT_RESET, 0); >> - >> for (i = 0; i < 2; i++) { >> if (i == 0) { >> ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; > >
Am 18.09.2014 um 20:01 schrieb Alex Deucher: > On Thu, Sep 18, 2014 at 1:51 PM, Christian König > <deathsimple@vodafone.de> wrote: >> Am 18.09.2014 um 19:26 schrieb Alex Deucher: >>> Otherwise we may lose the DMA golden settings which can >>> lead to hangs, etc. >> >> Don't we still want the soft reset at some point, e.g. when the engine >> hangs? >> >> I would rather move that to another function and call it before loading the >> golden values. > We have separate soft reset code for dma already. This is just a > one-off soft reset in the dma init code. The golden registers are > reloaded after a GPU reset so we should be fine. Of course! In this case the patches are Reviewed-by: Christian König <christian.koenig@amd.com> > > Alex > >> Christian. >> >> >>> bug: >>> https://www.libreoffice.org/bugzilla/show_bug.cgi?id=83500 >>> >>> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> >>> Cc: stable@vger.kernel.org >>> --- >>> drivers/gpu/drm/radeon/ni_dma.c | 6 ------ >>> 1 file changed, 6 deletions(-) >>> >>> diff --git a/drivers/gpu/drm/radeon/ni_dma.c >>> b/drivers/gpu/drm/radeon/ni_dma.c >>> index 8a3e622..f26f0a9 100644 >>> --- a/drivers/gpu/drm/radeon/ni_dma.c >>> +++ b/drivers/gpu/drm/radeon/ni_dma.c >>> @@ -191,12 +191,6 @@ int cayman_dma_resume(struct radeon_device *rdev) >>> u32 reg_offset, wb_offset; >>> int i, r; >>> - /* Reset dma */ >>> - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); >>> - RREG32(SRBM_SOFT_RESET); >>> - udelay(50); >>> - WREG32(SRBM_SOFT_RESET, 0); >>> - >>> for (i = 0; i < 2; i++) { >>> if (i == 0) { >>> ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX]; >>
diff --git a/drivers/gpu/drm/radeon/ni_dma.c b/drivers/gpu/drm/radeon/ni_dma.c index 8a3e622..f26f0a9 100644 --- a/drivers/gpu/drm/radeon/ni_dma.c +++ b/drivers/gpu/drm/radeon/ni_dma.c @@ -191,12 +191,6 @@ int cayman_dma_resume(struct radeon_device *rdev) u32 reg_offset, wb_offset; int i, r; - /* Reset dma */ - WREG32(SRBM_SOFT_RESET, SOFT_RESET_DMA | SOFT_RESET_DMA1); - RREG32(SRBM_SOFT_RESET); - udelay(50); - WREG32(SRBM_SOFT_RESET, 0); - for (i = 0; i < 2; i++) { if (i == 0) { ring = &rdev->ring[R600_RING_TYPE_DMA_INDEX];
Otherwise we may lose the DMA golden settings which can lead to hangs, etc. bug: https://www.libreoffice.org/bugzilla/show_bug.cgi?id=83500 Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org --- drivers/gpu/drm/radeon/ni_dma.c | 6 ------ 1 file changed, 6 deletions(-)