On 2019-06-18 1:37, Christoph Hellwig wrote:
> On Mon, Jun 17, 2019 at 09:45:09PM -0300, Jason Gunthorpe wrote:
>> Am I looking at the wrong thing? Looks like it calls it through a work
>> queue should should be OK..
> Yes, it calls it through a work queue. I guess that is fine because
> it needs
On 2019-06-18 17:34, Kim, Jonathan wrote:
> v2: fixed missing break in switch statement
>
> add pmu counters
>
> Change-Id: I1aca271fd12cabce0ccfc076f771cde2d4cadd54
> Signed-off-by: Jonathan Kim
> ---
> drivers/gpu/drm/amd/amdgpu/Makefile| 2 +-
>
On Tue, Jun 18, 2019 at 1:41 PM Hawking Zhang wrote:
>
> From: Tao Zhou
>
> AMDGPU_VRAM_TYPE_GDDR6 is a new vram type for navi10
>
> Change-Id: I6789230f8f7f5bdcb0aec82cc764d10d72c4cba8
Can you remove the Change-Id from these patches?
Otherwise,
Reviewed-by: Bas Nieuwenhuizen
for patches
v2: fixed missing break in switch statement
add pmu counters
Change-Id: I1aca271fd12cabce0ccfc076f771cde2d4cadd54
Signed-off-by: Jonathan Kim
---
drivers/gpu/drm/amd/amdgpu/Makefile| 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 3 +
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
Sorry, I caught another problem with error handling. See below.
On 2019-06-18 15:24, Kim, Jonathan wrote:
> v3: fixed cleanup by adding fini to free up adev df config counters
>
> v2: simplified by removing xgmi references in function names and moving to
> generic df function names. fixed issue
v3: fixed cleanup by adding fini to free up adev df config counters
v2: simplified by removing xgmi references in function names and moving to
generic df function names. fixed issue by removing hardcoded cake tx data
events. streamlined error handling by having df_v3_6_pmc_get_ctrl return
error
Am 18.06.19 um 20:26 schrieb StDenis, Tom:
Minor indentation and simplifications for the DF 3.6 driver. No
functional changes.
Signed-off-by: Tom St Denis
Acked-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/df_v3_6.c | 50 ++--
1 file changed, 18
Minor indentation and simplifications for the DF 3.6 driver. No
functional changes.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/df_v3_6.c | 50 ++--
1 file changed, 18 insertions(+), 32 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/df_v3_6.c
i think should be drop this patch from git tree,
this patch only for bringing up stage.
thanks.
Best Regards,
Kevin
From: Christian K?nig
Sent: Tuesday, June 18, 2019 3:58:28 PM
To: Alex Deucher; amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander; Wang,
This way we can even pipeline imported BO evictions.
v2: Limit this to only cases when the parent object uses a separate
reservation object as well. This fixes another OOM problem.
Signed-off-by: Christian König
---
drivers/gpu/drm/ttm/ttm_bo_util.c | 20 +++-
1 file
Avoid that we ping/pong the buffers when we stop to pin DMA-buf
exports by using the allowed domains for exported buffers.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git
Instead of relying on the DRM functions just implement our own import
functions. This prepares support for taking care of unpinned DMA-buf.
v2: enable for all exporters, not just amdgpu, fix invalidation
handling, lock reservation object while setting callback
v3: change to new dma_buf attach
The caching of SGT's is actually quite harmful and should probably removed
altogether when all drivers are audited.
Start by providing a separate DMA-buf export implementation in amdgpu.
This is also a prerequisite of unpinned DMA-buf handling.
v2: fix unintended recursion, remove debugging
Pipeline removal of the BOs backing store when no placement is given
during validation.
Signed-off-by: Christian König
---
drivers/gpu/drm/ttm/ttm_bo.c | 12
1 file changed, 12 insertions(+)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index
On the exporter side we add optional explicit pinning callbacks. If those
callbacks are implemented the framework no longer caches sg tables and the
map/unmap callbacks are always called with the lock of the reservation object
held.
On the importer side we add an optional invalidate callback.
From: Leo Liu
With different register offsets from VCN1.0
Signed-off-by: Leo Liu
Reviewed-by: James Zhu
---
tests/amdgpu/vcn_tests.c | 50 +++-
1 file changed, 37 insertions(+), 13 deletions(-)
diff --git a/tests/amdgpu/vcn_tests.c
pa_sc_tile_steering_override is a new member introduced for gfx10
Change-Id: I1482a5ef22cc4564eea63e09b1c40e9be3900e1f
Signed-off-by: Hawking Zhang
Reviewed-by: Marek Olšák
---
include/drm/amdgpu_drm.h | 2 ++
1 file changed, 2 insertions(+)
diff --git a/include/drm/amdgpu_drm.h
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Hawking Zhang
---
include/drm/amdgpu_drm.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/drm/amdgpu_drm.h b/include/drm/amdgpu_drm.h
index 3d0318e..b28fee4 100644
--- a/include/drm/amdgpu_drm.h
+++ b/include/drm/amdgpu_drm.h
From: Tao Zhou
AMDGPU_VRAM_TYPE_GDDR6 is a new vram type for navi10
Change-Id: I6789230f8f7f5bdcb0aec82cc764d10d72c4cba8
Signed-off-by: Tao Zhou
Reviewed-by: Hawking Zhang
---
include/drm/amdgpu_drm.h | 1 +
1 file changed, 1 insertion(+)
diff --git a/include/drm/amdgpu_drm.h
Am 18.06.19 um 12:31 schrieb Xiao, Jack:
job->preamble_status was accidently dropped from the combination value
of status varaiable which is the parameter of some ring level function.
The patch is to add it back.
Signed-off-by: Jack Xiao
For now Acked-by: Christian König , cause we
need to
I sent out a separate patch to add job->preamble_status back.
Regards,
Jack
-Original Message-
From: Koenig, Christian
Sent: Tuesday, June 18, 2019 5:03 PM
To: Xiao, Jack ; Alex Deucher ;
amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander ; Zhang, Hawking
Subject: Re: [PATCH
job->preamble_status was accidently dropped from the combination value
of status varaiable which is the parameter of some ring level function.
The patch is to add it back.
Signed-off-by: Jack Xiao
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 1 +
1 file changed, 1 insertion(+)
diff --git
On Thu, Jun 13, 2019 at 10:13:54PM -0700, Kees Cook wrote:
> On Thu, Jun 13, 2019 at 04:26:32PM +0100, Catalin Marinas wrote:
> > On Thu, Jun 13, 2019 at 12:02:35PM +0100, Dave P Martin wrote:
> > > On Wed, Jun 12, 2019 at 01:43:20PM +0200, Andrey Konovalov wrote:
> > > > +static int zero;
> > > >
Yeah thought so, cause that would break immediately and won't work at all.
It must be something like a merge or rebase artifact.
What's the best way to fix that up? In other words how should it look like?
Christian.
Am 18.06.19 um 10:32 schrieb Xiao, Jack:
> In original patch, there is no
In original patch, there is no deletion on job->preamble_status.
Regards,
Jack
-Original Message-
From: Christian König
Sent: Tuesday, June 18, 2019 4:04 PM
To: Alex Deucher ; amd-gfx@lists.freedesktop.org
Cc: Deucher, Alexander ; Xiao, Jack
; Zhang, Hawking
Subject: Re: [PATCH
Am 17.06.19 um 21:30 schrieb Alex Deucher:
From: Hawking Zhang
Signed-off-by: Hawking Zhang
Signed-off-by: Alex Deucher
Since the fields where removed this should most likely be squashed into
the original commit adding the code, or otherwise everything in between
won't compile any more.
Can probably be squashed into the original adding of the unit test.
Christian.
Am 17.06.19 um 21:29 schrieb Alex Deucher:
From: Jack Xiao
It forgot to put job fence, caused fence memory leak.
Signed-off-by: Jack Xiao
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
Am 17.06.19 um 21:10 schrieb Alex Deucher:
From: Jack Xiao
Add ib preemption status in amdgpu_job, so that ring level function
can detect preemption and program for resuming it.
Acked-by: Alex Deucher
Reviewed-by: Hawking Zhang
Signed-off-by: Jack Xiao
Signed-off-by: Alex Deucher
---
Am 17.06.19 um 21:30 schrieb Alex Deucher:
From: Kevin Wang
This reverts commit fd9c75d217d5b4ed72672722b6621e2635363dfe.
Signed-off-by: Kevin Wang
Reviewed-by: Huang Rui
Signed-off-by: Alex Deucher
Would it be possible to squash that into the original commit which
disabled things?
On Sat, Jun 15, 2019 at 07:12:11AM -0700, Christoph Hellwig wrote:
> > + spin_lock(>page_table_lock);
> > + if (mm->hmm) {
> > + if (kref_get_unless_zero(>hmm->kref)) {
> > + spin_unlock(>page_table_lock);
> > + return mm->hmm;
> > + }
>
On Sat, Jun 15, 2019 at 07:18:26AM -0700, Christoph Hellwig wrote:
> On Thu, Jun 13, 2019 at 09:44:48PM -0300, Jason Gunthorpe wrote:
> > range->hmm = hmm;
> > kref_get(>kref);
> > - list_add_rcu(>list, >ranges);
> > + list_add(>list, >ranges);
> >
> > /*
> > * If there are
On Sat, Jun 15, 2019 at 07:21:06AM -0700, Christoph Hellwig wrote:
> On Thu, Jun 13, 2019 at 09:44:49PM -0300, Jason Gunthorpe wrote:
> > From: Jason Gunthorpe
> >
> > hmm_release() is called exactly once per hmm. ops->release() cannot
> > accidentally trigger any action that would recurse back
On Mon, Jun 17, 2019 at 09:45:09PM -0300, Jason Gunthorpe wrote:
> Am I looking at the wrong thing? Looks like it calls it through a work
> queue should should be OK..
Yes, it calls it through a work queue. I guess that is fine because
it needs to take the lock again.
> Though very strange that
num_gb_pipes was set to a wrong value using r420_pipe_config
This have lead to HyperZ glitches on fast Z clearing.
See: https://bugs.freedesktop.org/show_bug.cgi?id=110897
Signed-off-by: Richard Thier
---
drivers/gpu/drm/radeon/r300.c | 3 ++-
drivers/gpu/drm/radeon/rs400.c | 21
34 matches
Mail list logo