From: Michel Dänzer
Xorg doesn't use the following functionality of GPU screens, so don't
bother initializing it:
* DRI page flipping
* DRI3 / Present / SYNC fences
* XVideo / XvMC
* Root window with background None
(Ported from radeon commit
Anyone review the 4 patches and another libdrm patch ?
-邮件原件-
发件人: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] 代表 Monk Liu
发送时间: Wednesday, March 08, 2017 3:54 PM
收件人: amd-gfx@lists.freedesktop.org
抄送: Liu, Monk
主题: [PATCH 4/4] drm/amdgpu:enable MCBP for
On Wed, Mar 8, 2017 at 4:03 AM, Michel Dänzer wrote:
> From: Michel Dänzer
>
> This is the only place we call drmmode_set_desired_modes for GPU screens
> during server startup. Without this change, the display outputs of
> secondary GPUs may stay on
On Wed, Mar 8, 2017 at 3:58 AM, Michel Dänzer wrote:
> From: Michel Dänzer
>
> Signed-off-by: Michel Dänzer
Reviewed-by: Alex Deucher
> ---
> configure.ac| 9 +++--
>
On 28 February 2017 at 11:46, zhoucm1 wrote:
> Hi Dave,
>
> The attached is our semaphore implementation, amdgpu_cs.c is drm file, the
> others are kernel file.
> Any suggestion?
Thanks,
I've built a tree with all these in it, and started looking into the interface.
I do
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 9 +
1 file changed, 9 insertions(+)
diff --git
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 29 +
From: Leo Liu
And initialize them
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 3 ++-
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 30 +-
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h | 1 +
1 file changed, 1 insertion(+)
diff --git
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/include/amd_shared.h | 1 +
1 file changed, 1 insertion(+)
diff --git
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/include/amd_shared.h | 1 +
1 file changed, 1 insertion(+)
diff --git
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_ring.h | 1 +
2 files
Newer asics use different registers so abstract it.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 3 +++
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
drivers/gpu/drm/amd/amdgpu/cik.c | 6 ++
From: Huang Rui
This will be used for newer asics.
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ucode.c | 7 +--
This set of patches updates various infrastructure in the driver
to support new asics.
Alex Deucher (3):
drm/amdgpu: switch ih handling to two levels
drm/amdgpu/ih: store the full context id
drm/amdgpu: add asic callback to get memsize register
Alex Xie (2):
drm/amdgpu: set GART PTE asic
Newer asics have a two levels of irq ids now:
client id - the IP
src id - the interrupt src within the IP
Signed-off-by: Alex Deucher
Signed-off-by: Ken Wang
Reviewed-by: Ken Wang
Signed-off-by: Alex Deucher
From: Junwei Zhang
Signed-off-by: Flora Cui
Reviewed-by: Jammy Zhou
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 6 ++
From: Ken Wang
Newer asics need 64 bit doorbells.
Signed-off-by: Ken Wang
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 4
The contextID field (formerly known as src_data) of the IH
vector stores client specific information about an interrupt.
It was expanded from 32 bits to 128 on newer asics. Expand the
src_id field to handle this.
Reviewed-by: Harry Wentland
Reviewed-by: Christian König
From: Ken Wang
Newer asics need 64 bit writeback slots.
Signed-off-by: Ken Wang
Reviewe-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 2 ++
Reviewed-by: Edward O'Callaghan
On 03/09/2017 07:03 AM, Tom St Denis wrote:
> When counting size of IBs count in dwords...
>
> Signed-off-by: Tom St Denis
> ---
> src/lib/ring_decode.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
On 03/09/2017 10:18 AM, Alex Deucher wrote:
> We added new gem ioctl flags and the new fences ioctl, but forgot
> to bump the version.
Reviewed-by: Edward O'Callaghan
Form which commit Alex? perhaps CC stable?
>
> Signed-off-by: Alex Deucher
Reject it if there are any invalid flags or domains.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 21 +
1 file changed, 21 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
We added new gem ioctl flags and the new fences ioctl, but forgot
to bump the version.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
Acked-by: Edward O'Callaghan
On 03/09/2017 03:31 AM, Tom St Denis wrote:
> This allows you to specify (say)
>
> -R gfx[16:32] to read words 16 through 32 which is
> consistent with how the data is presented.
>
> This patch also correctly enables the PM4 decoder
>
From: Andrey Grodzovsky
Change-Id: I50db672b5b1f2eed7933863a2e901466dcb636d0
Signed-off-by: Andrey Grodzovsky
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
Reviewed-by: Jordan Lazare
From: Zeyu Fan
Change-Id: I47456746d26307356b62379b1cf3c221e5935eb5
Signed-off-by: Zeyu Fan
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/dc/core/dc_link.c | 9 -
From: Andrey Grodzovsky
Headless mode set needs to be synchronized against outstanding nonblocking
commits. This achieved by building atomic state and commiting it.
Change-Id: Ie5e778afb33dd5c303a169216a7bb8c2d857037e
Signed-off-by: Andrey Grodzovsky
From: Charlene Liu
Change-Id: Icb5d285ed455e7081f88cd09df3845c08f7ceb20
Signed-off-by: Charlene Liu
Acked-by: Harry Wentland
Reviewed-by: Jordan Lazare
---
drivers/gpu/drm/amd/display/dc/core/dc.c
From: Charlene Liu
Change-Id: I334d4dfa6f5db2366457b1db9062f8ae6ec04e19
Signed-off-by: Charlene Liu
Acked-by: Harry Wentland
Reviewed-by: Wenjing Liu
Reviewed-by: Tony Cheng
---
From: Dmytro Laktyushkin
Change-Id: I1d7db6e097c5a0d307e4d08d70f52542afba0f93
Signed-off-by: Dmytro Laktyushkin
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
From: Charlene Liu
Change-Id: Ib84c5ff3887c4b096c8716279a7741b46350e433
Signed-off-by: Charlene Liu
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/dc/core/dc.c |
Don't tell DRM otherwise.
Change-Id: I6c01701b3be35cb5badf236459ee5d9a1ab8c309
Signed-off-by: Harry Wentland
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_types.c | 5
From: Andrey Grodzovsky
Modify amdgpu_dm_atomic_comit to implement
atomic_comit_tail hook.
Unify Buffer objects allocation and dealocation
for surface updates and page flips.
Simplify wait for fences and target_vbank logic
for non blockiing commit.
Remove hacky update
From: Yongqiang Sun
Issue:
In case of two 4K@60 + one non-4k monitor, when unplug non-4k monitor,
the remain two 4k monitor don't work properly.
Reason:
In that case, two 4k use two pipes and no split, when unplug happens,
those two monitor will use 4 pipes and split, but
From: Dmytro Laktyushkin
Change-Id: I603f1f04c9bbac3236dfd56b3587b6943e25acd1
Signed-off-by: Dmytro Laktyushkin
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
From: Yongqiang Sun
Change-Id: I30658257832be3865f3f3af0b471fd490fab4426
Signed-off-by: Yongqiang Sun
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
From: Tony Cheng
always take update lock instead of using HW built in update lock trigger with
write to primary_addr_lo.
we will be a little more inefficient with the extra registers write to lock,
but this simplify code and make it always correct.
Will revisit locking
From: Leon Elazar
1.current_context memory wasn't released at
dc_post_update_surfaces_to_stream during context swap.
Change-Id: Iff814cf675856708daadfa540ef2b57d633139df
Signed-off-by: Leon Elazar
Acked-by: Harry Wentland
From: Dmytro Laktyushkin
Viewport would be incorrectly adjusted when surface was used
for multiple displays
Change-Id: Iffd7e03973edc72f63d21aa83b0482f6013d2ecd
Signed-off-by: Dmytro Laktyushkin
Acked-by: Harry Wentland
From: Amy Zhang
Change-Id: I3fbfab5a617bed56bb5013e820b0e37c507e0bfe
Signed-off-by: Amy Zhang
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/dc/dce/dce_clocks.c | 1 +
From: Charlene Liu
Change-Id: I1f8de5415aac149d85b73a28fb630c50696a9250
Signed-off-by: Charlene Liu
Acked-by: Harry Wentland
Reviewed-by: Jordan Lazare
---
From: Leon Elazar
Exposing DC Api dc_check_update_surfaces_for_stream
validation will return the answer which type of update is required,
so upper layers can is it safe to call the update API fro high IRQ yes/no.
Change-Id: I094592c5df4227ed2fea2ceb5de5b2604173fa20
Change-Id: Idfd0d0d6e537eddbc75378ba394b0f36bd89dd50
Signed-off-by: Harry Wentland
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_types.h | 5 -
1 file changed, 5
From: Andrey Grodzovsky
This function makes a lot of assumptions to try and deduce
tha an atomic_commit came from FLIP ioctl. Those assumptions
are not explicit contract with DRM API and might lead to wrong
result as the code changes. Instead use
No need to assign them dynamically. This is much more readable.
Change-Id: I2d8a356e8d916800c4553c4e9b19ce42f7f1c391
Signed-off-by: Harry Wentland
Acked-by: Harry Wentland
Reviewed-by: Tony Cheng
---
From: Charlene Liu
Change-Id: Iccf6c436dc0807f4d5e0ce96ae1e64ed31c20a38
Signed-off-by: Charlene Liu
Acked-by: Harry Wentland
Reviewed-by: Charlene Liu
---
drivers/gpu/drm/amd/display/dc/core/dc.c | 19
* Use atomic helpers for commit, suspend/resume, and gamma
* Whole bunch of fixes
* Updated scaler calculations
Amy Zhang (1):
drm/amd/display: Simplify some DMCU waits
Andrey Grodzovsky (4):
drm/amd/display: Refactor atomic commit implementation. (v2)
drm/amd/display: Refactor headless to
I'm working on a follow-up that has Christian's extra requests to remove
the job/fence pointers. But that is coming as a separate commit.
Regards,
Andres
___
amd-gfx mailing list
amd-gfx@lists.freedesktop.org
Is this patch ready to land or is there more work required?
Thanks,
Alex
On Sun, Feb 26, 2017 at 5:54 AM, Christian König
wrote:
> Am 25.02.2017 um 18:28 schrieb Andres Rodriguez:
>
>
>
> On Feb 25, 2017 4:40 AM, "Christian König" wrote:
>
>
On Sat, Feb 25, 2017 at 4:33 AM, Christian König
wrote:
> Am 24.02.2017 um 19:20 schrieb Andres Rodriguez:
>>
>> Most of the traces have uniform format except for two of them. Having
>> all the traces match makes it simple to run awk on the ftrace output.
>>
>>
On Wed, Mar 8, 2017 at 12:01 AM, Lukas Wunner wrote:
> On Tue, Mar 07, 2017 at 03:30:30PM -0500, Alex Deucher wrote:
>> On Fri, Feb 24, 2017 at 2:19 PM, Lukas Wunner wrote:
>> > An external Thunderbolt GPU can neither drive the laptop's panel nor be
>> > powered
When counting size of IBs count in dwords...
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 401d6caaa088..a991f5ffa237 100644
---
Reviewed-by: Christian König for this one and
#19.
Christian.
Am 08.03.2017 um 15:12 schrieb Daniel Vetter:
Again no apparent explanation for the split except hysterical raisins.
Merging them also makes it a bit more obviuos what's going on wrt the
runtime pm
Again no apparent explanation for the split except hysterical raisins.
Merging them also makes it a bit more obviuos what's going on wrt the
runtime pm refdancing.
Cc: Alex Deucher
Cc: Christian König
Cc: amd-gfx@lists.freedesktop.org
On Wed, Mar 08, 2017 at 11:46:33AM +0100, Peter Wu wrote:
> On Wed, Mar 08, 2017 at 06:01:54AM +0100, Lukas Wunner wrote:
> > On Tue, Mar 07, 2017 at 03:30:30PM -0500, Alex Deucher wrote:
> > > On Fri, Feb 24, 2017 at 2:19 PM, Lukas Wunner wrote:
> > > > An external Thunderbolt
On Wed, Mar 08, 2017 at 06:01:54AM +0100, Lukas Wunner wrote:
> On Tue, Mar 07, 2017 at 03:30:30PM -0500, Alex Deucher wrote:
> > On Fri, Feb 24, 2017 at 2:19 PM, Lukas Wunner wrote:
> > > An external Thunderbolt GPU can neither drive the laptop's panel nor be
> > > powered off
Am 08.03.2017 um 11:22 schrieb Liu, Monk:
Rebase done: will send out another patch for review
Additional to that I would provide the implementation (e.g. forwarding it to
the IB command) for the PREEMPT flag in the same patch.
No, we don't have such imple, because this flag will be tagged by
PREEMPT flag is used for MCBP
Change-Id: I2ddf0f3912c219663e80680f76d671c7856e
Signed-off-by: Monk Liu
---
include/drm/amdgpu_drm.h | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/include/drm/amdgpu_drm.h b/include/drm/amdgpu_drm.h
index
Rebase done: will send out another patch for review
>Additional to that I would provide the implementation (e.g. forwarding it to
>the IB command) for the PREEMPT flag in the same patch.
No, we don't have such imple, because this flag will be tagged by UMD, instead
of by libdrm
BR Monk
Please rebase that on the amd-staging-4.9 branch, the GDS flag never
made it upstream.
Additional to that I would provide the implementation (e.g. forwarding
it to the IB command) for the PREEMPT flag in the same patch.
Christian.
Am 08.03.2017 um 09:13 schrieb Monk Liu:
we don't use
From: Michel Dänzer
This is the only place we call drmmode_set_desired_modes for GPU screens
during server startup. Without this change, the display outputs of
secondary GPUs may stay on even while Xorg isn't using them.
Signed-off-by: Michel Dänzer
From: Michel Dänzer
Xorg doesn't use the following functionality of GPU screens, so don't
bother initializing it:
* DRI page flipping
* DRI3 / Present / SYNC fences
* XVideo / XvMC
* Root window with background None
Signed-off-by: Michel Dänzer
From: Michel Dänzer
Signed-off-by: Michel Dänzer
---
configure.ac| 9 +++--
src/amdgpu_glamor.c | 7 +++
2 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/configure.ac b/configure.ac
index 00249a599..591c2b9f6 100644
we don't use FLAG_GDS actually, and we need PREEMPT flag
Change-Id: I87f704e3fa97d7a7ffa956f0c94b77c333de380d
Signed-off-by: Monk Liu
---
include/drm/amdgpu_drm.h | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/include/drm/amdgpu_drm.h
67 matches
Mail list logo