Hi Christian,
Please apply this patch if you think this is the correct solution. This
is currently blocking the amd-kfd-staging merge from amd-staging-4.11.
Kent will try another merge tomorrow morning our time.
I'm quite sure this also breaks userptrs in the graphics driver.
Thanks,
Felix
Otherwise bo->shadow_list (which is aliased by bo->mn_list) will not
appear empty in amdgpu_ttm_bo_destroy and cause an oops when freeing
former userptr BOs.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c | 2 +-
1 file changed, 1 insertion(+),
On Wed, Aug 02, 2017 at 10:07:01AM +1000, Dave Airlie wrote:
> Running amd-staging-4.11 at 0ae168080db88fa3a2ec84741a7293a25e4ecb42
> on
> VEGA10 0x1002:0x687F pciid
>
> with firmware from
> https://people.freedesktop.org/~agd5f/radeon_ucode/vega10/
>
> PSP firmware loading fails.
>
> Dave.
>
On Tue, Aug 1, 2017 at 1:46 PM, Laurent Pinchart <
laurent.pinch...@ideasonboard.com> wrote:
> Hi Joe,
>
> On Tuesday 01 Aug 2017 10:24:25 Joe Kniss wrote:
> > On Tue, Aug 1, 2017 at 5:09 AM, Laurent Pinchart wrote:
> > > On Monday 31 Jul 2017 11:29:13 Joe Kniss wrote:
> > >> New getfb2
Running amd-staging-4.11 at 0ae168080db88fa3a2ec84741a7293a25e4ecb42
on
VEGA10 0x1002:0x687F pciid
with firmware from
https://people.freedesktop.org/~agd5f/radeon_ucode/vega10/
PSP firmware loading fails.
Dave.
[1.690928] [drm] initializing kernel modesetting (VEGA10
0x1002:0x687F
Hi Joe,
On Tuesday 01 Aug 2017 10:24:25 Joe Kniss wrote:
> On Tue, Aug 1, 2017 at 5:09 AM, Laurent Pinchart wrote:
> > On Monday 31 Jul 2017 11:29:13 Joe Kniss wrote:
> >> New getfb2 functionality uses drm_mode_fb_cmd2 struct to be symmetric
> >> with addfb2.
> >
> > What's the use case for this
Am 01.08.2017 um 19:00 schrieb Tom St Denis:
On 01/08/17 11:41 AM, Christian König wrote:
We can turn of the trace subsystem during compile time already.
Additional to that the trace points use self modifying code to make
sure that they don't have any overhead as long as they are turned
off.
Thanks for the quick review!
On Tue, Aug 1, 2017 at 5:09 AM, Laurent Pinchart wrote:
> Hi Joe,
>
> Thank you for the patch.
>
> On Monday 31 Jul 2017 11:29:13 Joe Kniss wrote:
> > New getfb2 functionality uses drm_mode_fb_cmd2 struct to be symmetric
> > with
On 01/08/17 11:41 AM, Christian König wrote:
We can turn of the trace subsystem during compile time already.
Additional to that the trace points use self modifying code to make sure
that they don't have any overhead as long as they are turned off. But I
don't think this works with the "if
In v2 of the patch I used the _enabled() function around the blocks so
the for loop is only reached if the trace is enabled.
Tom
On 01/08/17 11:56 AM, Xie, AlexBin wrote:
I don't know if compiler is smart enough to optimize the following for
statement out...
+ for (i = 0; i <
We can turn of the trace subsystem during compile time already.
Additional to that the trace points use self modifying code to make sure
that they don't have any overhead as long as they are turned off. But I
don't think this works with the "if (trace_*_enabled()" as well.
Anyway, not a
Because this is used by a debug tool, can we use a macro to
conditionally compile this feature?
On 2017-08-01 10:54 AM, Christian König wrote:
Am 01.08.2017 um 16:26 schrieb Tom St Denis:
On 01/08/17 10:10 AM, Christian König wrote:
You need to cover multiple code path here:
1. The one you
This helps map DMA addresses back to physical addresses.
Signed-off-by: Tom St Denis
(v2): Added tracepoints for USERPTR, SG mappings, and
SWIOTBL mappings. Reformatted trace call perform
PCI decoding internal to the trace.
---
Am 01.08.2017 um 16:26 schrieb Tom St Denis:
On 01/08/17 10:10 AM, Christian König wrote:
You need to cover multiple code path here:
1. The one you currently implemented which uses ttm_dma_populate()
and pci_map_page().
2. The one using ttm_dma_populate().
I'll have to look into this one
On 01/08/17 10:10 AM, Christian König wrote:
You need to cover multiple code path here:
1. The one you currently implemented which uses ttm_dma_populate() and
pci_map_page().
2. The one using ttm_dma_populate().
I'll have to look into this one though it's my understanding that code
path is
On Tue, Aug 1, 2017 at 5:19 AM, Michel Dänzer wrote:
> From: Michel Dänzer
>
> Preparation for following changes, no functional change intended yet.
>
> Reviewed-by: Alex Deucher # v1
> Signed-off-by: Michel Dänzer
On Tue, Aug 01, 2017 at 06:00:42AM -0400, Tom St Denis wrote:
> We correctly grabbed the bits but did not shift them
> back to their correct place.
>
> Signed-off-by: Tom St Denis
Tested-by: Huang Rui
> ---
> src/lib/ring_decode.c | 2 +-
> 1 file
On 01/08/17 07:55 AM, Christian König wrote:
Am 01.08.2017 um 13:51 schrieb Tom St Denis:
This helps map DMA addresses back to physical addresses.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h | 28
Hi Joe,
Thank you for the patch.
On Monday 31 Jul 2017 11:29:13 Joe Kniss wrote:
> New getfb2 functionality uses drm_mode_fb_cmd2 struct to be symmetric
> with addfb2.
What's the use case for this ? We haven't needed such an ioctl for so long
that it seemed to me that userspace doesn't really
Am 01.08.2017 um 13:51 schrieb Tom St Denis:
This helps map DMA addresses back to physical addresses.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h | 28
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 8
2
This helps map DMA addresses back to physical addresses.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h | 28
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 8
2 files changed, 36 insertions(+)
diff --git
We correctly grabbed the bits but did not shift them
back to their correct place.
Signed-off-by: Tom St Denis
---
src/lib/ring_decode.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/lib/ring_decode.c b/src/lib/ring_decode.c
index
From: Michel Dänzer
Preparation for following changes, no functional change intended yet.
Reviewed-by: Alex Deucher # v1
Signed-off-by: Michel Dänzer
---
v2: Add drmmode_screen_damage_destroy to prevent use-after-free
From: Michel Dänzer
As long as the CRTC we're synchronizing to doesn't.
Reviewed-by: Alex Deucher
Signed-off-by: Michel Dänzer
---
v2: Remove redundant checks from can_exchange which still prevented DRI2
page
24 matches
Mail list logo