From: Michel Dänzer
Preparatory, no functional change intended yet.
Also inline amdgpu_pixmap_create_fb into amdgpu_pixmap_get_fb, since
there's only one call-site.
(Ported from radeon commit 20f6b56fdb74d88086e8e094013fedbb14e50a24)
Signed-off-by: Michel Dänzer
---
src/amdgpu_pixmap.h | 46
From: Michel Dänzer
Preparatory, no functional change intended yet.
(Ported from radeon commit 4bc992c31059eb50e22df4ebf5b92d08411f41ef)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 82 ++-
src/drmmode_display.h | 3 ++
2 files chan
From: Michel Dänzer
And destroy all other FBs. This is so that other DRM masters can only
get access to this all-black FB, not to any other FB we created, while
we're switched away and not DRM master.
Fixes: b09fde0d81e0 ("Use reference counting for tracking KMS
framebuffer
From: Michel Dänzer
Not used anymore.
(Cherry picked from radeon commit e4a3df19d588a4310fcb889ef34e205d0e92e4d7)
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 10 --
src/drmmode_display.h | 1 -
2 files changed, 11 deletions(-)
diff --git a/src/drmmode_display.c b/src/dr
From: Michel Dänzer
Preparatory, no functional change intended yet.
(Ported from radeon commit 3f6210ca2c8ef60d59efc8139151d3b9838bb875)
Signed-off-by: Michel Dänzer
---
src/amdgpu_bo_helper.c | 20
src/amdgpu_bo_helper.h | 2 ++
src/drmmode_display.c | 14 +
From: Michel Dänzer
Preparatory, no functional change intended yet.
(Ported from radeon commit 7f0cd68d1b0c132e32ae736371bce3e12ed33c7a)
Signed-off-by: Michel Dänzer
---
src/amdgpu_drv.h | 14 ++
1 file changed, 10 insertions(+), 4 deletions(-)
diff --git a/src/amdgpu_drv.h b/src
From: Michel Dänzer
The timer fires 1 second after LeaveVT. This gives the next DRM master
enough time to set up scanout of its own buffers.
Fixes prolonged intermittent black screen when switching from Xorg to
e.g. the GDM Wayland mode login VT.
Fixes: 06a465484101 ("Make all active CRTCs scan
This feature is for UMD to run benchmark in a
power state that is as steady as possible. kmd
need to fix the power state as stable as possible.
now, kmd support four level:
profile_standard,peak,min_sclk,min_mclk
move common related code to amd_powerplay.c
Change-Id: Ie06c122199b7246f5b1951c354cf
Change-Id: Ibd74590c3fe9dbdeac924b697d18448bddbefcdb
---
drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 11 +--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
index a125e30..10
Hi Michel,
On 28 August 2017 at 10:23, Michel Dänzer wrote:
> From: Michel Dänzer
>
> And destroy all other FBs. This is so that other DRM masters can only
> get access to this all-black FB, not to any other FB we created, while
> we're switched away and not DRM master.
>
Isn't the issue applica
Allows writing data to vram via debugfs.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 41 -
1 file changed, 40 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.
Am 29.08.2017 um 14:43 schrieb Tom St Denis:
Allows writing data to vram via debugfs.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 41 -
1 file changed, 40 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgp
On 29/08/17 08:46 AM, Christian König wrote:
Am 29.08.2017 um 14:43 schrieb Tom St Denis:
Allows writing data to vram via debugfs.
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 41
-
1 file changed, 40 insertions(+), 1 deletion(-
Allows writing data to vram via debugfs.
Signed-off-by: Tom St Denis
(v2): Call get_user before holding spinlock.
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 42 -
1 file changed, 41 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_
I asked you how you keep the access of base.moved is safely in last
thread,
Sorry, I've missed that question somehow.
I check it just now, it depends on the shared resv lock.
Actually that's not 100% correct. The moved member is protected by the
BOs resv lock.
That can be the shared one (in
Am 29.08.2017 um 14:49 schrieb Tom St Denis:
Allows writing data to vram via debugfs.
Signed-off-by: Tom St Denis
Reviewed-by: Christian König
(v2): Call get_user before holding spinlock.
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 42 -
1 file change
Am 29.08.2017 um 15:21 schrieb Himanshu Jha:
Kfree on NULL pointer is a no-op and therefore checking is redundant.
Signed-off-by: Himanshu Jha
Reviewed-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c | 6 ++
drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr
Kfree on NULL pointer is a no-op and therefore checking is redundant.
Signed-off-by: Himanshu Jha
---
drivers/gpu/drm/amd/amdgpu/amdgpu_connectors.c | 6 ++
drivers/gpu/drm/amd/powerplay/smumgr/smu7_smumgr.c | 6 ++
2 files changed, 4 insertions(+), 8 deletions(-)
diff --git a/drive
kfree on NULL pointer is a no-op and therefore checking is redundant.
Signed-off-by: Himanshu Jha
---
drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 6 +-
.../gpu/drm/amd/powerplay/hwmgr/processpptables.c | 96 --
drivers/gpu/drm/amd/powerplay/hwmgr/rv_hwmgr.c | 5
On 2017-08-29 09:12 AM, Himanshu Jha wrote:
> kfree on NULL pointer is a no-op and therefore checking is redundant.
>
> Signed-off-by: Himanshu Jha
Reviewed-by: Harry Wentland
Harry
> ---
> drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 6 +-
> .../gpu/drm/amd/powerplay/hwmgr/processp
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Tuesday, August 29, 2017 4:31 AM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH xf86-video-amdgpu 6/6] Remove drmmode_scanout_free
>
> From: Michel Dänzer
>
> N
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Tuesday, August 29, 2017 5:18 AM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH xf86-video-ati] Use a timer for unreferencing the all-black FB
>
> From: Michel
Also add --vram-read and -vr which will eventually replace --vram and -v.
Signed-off-by: Tom St Denis
---
doc/umr.1 | 6 -
src/app/main.c | 43 +---
src/lib/read_vram.c | 72 ++---
src/umr.h
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Rex Zhu
> Sent: Tuesday, August 29, 2017 5:14 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Zhu, Rex
> Subject: [PATCH 1/2] drm/amd/powerplay: add UMD P-state in powerplay.
>
> This feature is
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Rex Zhu
> Sent: Tuesday, August 29, 2017 5:14 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Zhu, Rex
> Subject: [PATCH 2/2] drm/amd/powerplay: set uvd/vce/nb/mclk level as
> pstate requested
>
Ok, found something that works. Xonotic in lowest resolution, lowest
effects quality (e.g. totally CPU bound):
Without per process BOs:
Xonotic 0.8:
pts/xonotic-1.4.0 [Resolution: 800 x 600 - Effects Quality: Low]
Test 1 of 1
Estimated Trial Run Count:3
Estimated Time To Com
Reviewed-by: Rex Zhu mailto:rex@amd.com>>
Best Regards
Rex
From: Wentland, Harry
Sent: Tuesday, August 29, 2017 9:34:01 PM
To: Himanshu Jha; airl...@linux.ie
Cc: linux-ker...@vger.kernel.org; dri-de...@lists.freedesktop.org;
amd-gfx@lists.freedesktop.org;
Need to select dpm0 to avoid clock fluctuations.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 47 +-
1 file changed, 1 insertion(+), 46 deletions(-)
diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c
b/drivers/gpu/drm/amd/pow
Am 29.08.2017 um 15:12 schrieb Himanshu Jha:
kfree on NULL pointer is a no-op and therefore checking is redundant.
Signed-off-by: Himanshu Jha
Reviewed-by: Christian König
---
drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c | 6 +-
.../gpu/drm/amd/powerplay/hwmgr/processpptables.c
From: Christian König
Don't allow them to be GEM imported into another process.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c | 6 ++
1 file changed, 6 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.
From: Christian König
Per VM BOs are handled like VM PDs and PTs. They are always valid and don't
need to be specified in the BO lists.
v2: validate PDs/PTs first
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 79 +++
From: Christian König
Signed-off-by: Christian König
---
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
b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 21116fc..5035305c 100644
--- a/drivers/
From: Christian König
Add the IOCTL interface so that applications can allocate per VM BOs.
Still WIP since not all corner cases are tested yet, but this reduces average
CS overhead for 10K BOs from 21ms down to 48us.
v2: add some extra checks, remove the WIP tag
Signed-off-by: Christian König
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Christian König
> Sent: Tuesday, August 29, 2017 1:08 PM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH 3/4] drm/amdgpu: add IOCTL interface for per VM BOs v2
>
> From: Christian König
calling memcpy immediately after memset with the same region of memory
makes memset redundant.
Signed-off-by: Himanshu Jha
---
drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c
b/drive
Hi Dave,
A few fixes for drm-next for 4.14. Nothing too major. I'll check in
again in a week or so.
The following changes since commit 7c0059dd832cc686bf0febefdcf8295cdd93007f:
Merge branch 'linux-4.14' of git://github.com/skeggsb/linux into drm-next
(2017-08-23 05:32:26 +1000)
are availa
On Tue, Aug 29, 2017 at 9:28 AM, Christian König
wrote:
> Am 29.08.2017 um 15:21 schrieb Himanshu Jha:
>>
>> Kfree on NULL pointer is a no-op and therefore checking is redundant.
>>
>> Signed-off-by: Himanshu Jha
>
>
> Reviewed-by: Christian König
>
Applied. thanks!
Alex
>
>> ---
>> drive
On Tue, Aug 29, 2017 at 9:34 AM, Harry Wentland wrote:
> On 2017-08-29 09:12 AM, Himanshu Jha wrote:
>> kfree on NULL pointer is a no-op and therefore checking is redundant.
>>
>> Signed-off-by: Himanshu Jha
>
> Reviewed-by: Harry Wentland
Applied. thanks!
Alex
>
> Harry
>
>> ---
>> drivers
Rebased on the public drm-next-4.15-wip. Patch 8 from the WIP patch
series did not apply at all, because upstream KFD doesn't support
GPUVM yet.
The "lib: Closed hash table ..." change is updated and the same as
what I sent to LKML yesterday. Changes are mainly in the way the self
test is hooked u
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_module.c | 6 ---
drivers/gpu/drm/amd/amdkfd/kfd_pasid.c | 90 ++---
2 files changed, 38 insertions(+), 58 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_module.c
b/drivers/gpu/drm/amd/amdkf
Allows assigning a PASID to a VM for identifying VMs involved in page
faults. The global PASID manager is also exported in the KFD
interface so that AMDGPU and KFD can share the PASID space.
PASIDs of different sizes can be requested. On APUs, the PASID size
is deterined by the capabilities of the
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/radeon/radeon_kfd.c | 31 +++
1 file changed, 31 insertions(+)
diff --git a/drivers/gpu/drm/radeon/radeon_kfd.c
b/drivers/gpu/drm/radeon/radeon_kfd.c
index f6578c9..a2ac8ac 100644
--- a/drivers/gpu/drm/radeon/radeon_k
PASID management is moving into KGD. Limiting the PASID range to the
number of doorbell pages is no longer practical.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c | 7 -
drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c | 50 +--
drivers/gp
Make sure vm->root.bo is not left reserved if amdgpu_bo_kmap fails.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
index
IH tracks pending retry faults in a hash table for fast lookup in
interrupt context. Each VM has a short FIFO of pending VM faults for
processing in a bottom half.
The IH prescreening stage adds retry faults and filters out repeated
retry interrupts to minimize the impact of interrupt storms.
It'
This adds a statically sized closed hash table implementation with
low memory and CPU overhead. The API is inspired by kfifo.
Storing, retrieving and deleting data does not involve any dynamic
memory management, which makes it ideal for use in interrupt context.
Static memory usage per entry compr
To filter out high-frequency interrupts that can be safely ignored.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c | 6 ++
drivers/gpu/drm/amd/amdgpu/cik_ih.c | 14 ++
drivers/gpu/drm/amd/amdgpu/cz_i
I've send a related kernel crash log to amd-devel some days ago without
any answer, yet...
Was:
[amd-staging-drm-next] kernel crash with amdgpu on RX580 in
'drm_object_property_get_value'
I get this in _all_ current 'amd-staging-drm-next' versions. ;-(
[16301.515079] [ cut here ]
It might be interesting to try glmark2.
Marek
On Tue, Aug 29, 2017 at 3:59 PM, Christian König
wrote:
> Ok, found something that works. Xonotic in lowest resolution, lowest effects
> quality (e.g. totally CPU bound):
>
> Without per process BOs:
>
> Xonotic 0.8:
> pts/xonotic-1.4.0 [Resoluti
On Tue, Aug 29, 2017 at 4:54 PM, Jérôme Glisse wrote:
>
> Note this is barely tested. I intend to do more testing of next few days
> but i do not have access to all hardware that make use of the mmu_notifier
> API.
Thanks for doing this.
> First 2 patches convert existing call of mmu_notifier_in
(Sorry for so many list cross-posting and big cc)
Please help testing !
The invalidate_page callback suffered from 2 pitfalls. First it used to
happen after page table lock was release and thus a new page might have
been setup for the virtual address before the call to invalidate_page().
This is
Call to mmu_notifier_invalidate_page() are replaced by call to
mmu_notifier_invalidate_range() and thus call are bracketed by
call to mmu_notifier_invalidate_range_start()/end()
Remove now useless invalidate_page callback.
Signed-off-by: Jérôme Glisse
Cc: amd-gfx@lists.freedesktop.org
Cc: Felix
On Tue, Aug 29, 2017 at 05:11:24PM -0700, Linus Torvalds wrote:
> On Tue, Aug 29, 2017 at 4:54 PM, Jérôme Glisse wrote:
> >
> > Note this is barely tested. I intend to do more testing of next few days
> > but i do not have access to all hardware that make use of the mmu_notifier
> > API.
>
> Than
On Tue, Aug 29, 2017 at 6:39 PM, Dieter Nützel wrote:
> I've send a related kernel crash log to amd-devel some days ago without any
> answer, yet...
>
> Was:
> [amd-staging-drm-next] kernel crash with amdgpu on RX580 in
> 'drm_object_property_get_value'
>
> I get this in _all_ current 'amd-staging
this can get performance improvement for some cases
Change-Id: Ibb58bb3099f7e8c4b5da90da73a03544cdb2bcb7
Signed-off-by: Roger He
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 98 +++---
1 file changed, 79 insertions(+), 19 deletions(-)
diff --git a/drivers/gpu/drm/amd
Am 29.08.2017 um 19:20 schrieb Deucher, Alexander:
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
Of Christian König
Sent: Tuesday, August 29, 2017 1:08 PM
To: amd-gfx@lists.freedesktop.org
Subject: [PATCH 3/4] drm/amdgpu: add IOCTL interface for
Hi Roger,
I think when you can select fragment automatically, you shouldn't
involve the vm_manager.fragment_size to calculation, then we can use the
properest fragment for every segment and get the best performance.
So vm_manager.fragment_size should be always be -1, if it becomes valid
fra
Am 30.08.2017 um 01:54 schrieb Jérôme Glisse:
Call to mmu_notifier_invalidate_page() are replaced by call to
mmu_notifier_invalidate_range() and thus call are bracketed by
call to mmu_notifier_invalidate_range_start()/end()
Remove now useless invalidate_page callback.
Signed-off-by: Jérôme Glis
Hi David & Roger,
I think when you can select fragment automatically, you shouldn't
involve the vm_manager.fragment_size to calculation, then we can use
the properest fragment for every segment and get the best performance.
No, that won't work. I've already tried this and it decreases
perform
On 30/08/17 03:09 PM, Christian König wrote:
> Am 29.08.2017 um 19:20 schrieb Deucher, Alexander:
>>> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
>>> Of Christian König
>>>
>>> @@ -89,6 +89,8 @@ extern "C" {
>>> #define AMDGPU_GEM_CREATE_SHADOW(1 << 4)
>>> /*
On 30/08/17 03:42 PM, Michel Dänzer wrote:
> On 30/08/17 03:09 PM, Christian König wrote:
>> Am 29.08.2017 um 19:20 schrieb Deucher, Alexander:
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
Of Christian König
@@ -89,6 +89,8 @@ extern "C" {
#define
61 matches
Mail list logo
