Could you stop the clean for syncing? and wait we turn back
performance first?
Performance is not that important here, what we need is correct handling
and to simplify the code first.
If the closed source UMD is concerned about the performance hit with per
VM BOs it should simply revert to not
I really very surprise that you prefer to expand sync_obj to get
syncfile fd!!!
Why not directly generate syncfile fd and use it? Which one is simple is
so obvious.
Regards,
David Zhou
On 2017年09月13日 14:57, Christian König wrote:
Hi guys,
Mareks IOCTL proposal looks really good to me.
Plea
syncfile is the older implementation, sync_obj is the replacement for it.
I don't think we should implement syncfile in the CS any more, sync_obj
is already done and can be converted to/from syncfiles.
Mareks IOCTL should cover the case when we need to create a syncfile or
syncobj from a fenc
On 2017年09月13日 15:09, Christian König wrote:
syncfile is the older implementation, sync_obj is the replacement for it.
Are you sure sync_obj is a replacement for syncfile? Anyone can clarify it?
sync_obj is mainly for semaphore usage, we can see sync_obj docs from
Dave in drm_syncobj.c:
"/
Am 13.09.2017 um 09:39 schrieb zhoucm1:
On 2017年09月13日 15:09, Christian König wrote:
syncfile is the older implementation, sync_obj is the replacement for
it.
Are you sure sync_obj is a replacement for syncfile? Anyone can
clarify it?
sync_obj is mainly for semaphore usage, we can see syn
Am 13.09.2017 um 04:23 schrieb zhoucm1:
On 2017年09月12日 23:59, Deucher, Alexander wrote:
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
Of Christian König
Sent: Tuesday, September 12, 2017 5:09 AM
To: amd-gfx@lists.freedesktop.org
Subject: [PAT
From: Christian König
Just set the CPU access required flag when we pin it.
Signed-off-by: Christian König
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 23 ---
1 file changed, 8 insertions(+), 15 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
b/d
From: Christian König
We need to free the reservation object before we take the BO
from the delayed delete list.
Signed-off-by: Christian König
---
drivers/gpu/drm/ttm/ttm_bo.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
ind
amdgpu_bo_create() doesn't necessarily allocate anything, it just
creates the BO structure.
The backing memory for GTT and CPU domain is only allocated on first
use, only VRAM is allocated directly.
So just call amdgpu_bo_create() with AMDGPU_GEM_DOMAIN_CPU and then the
pin with AMDGPU_GEM_D
but we use VRAM domain in the first bo_create stage, so this BO is allocated
directly ...
I think our current approach is incorrect ... because the space from
bo_create(VRAM_DOMAIN) is missed somehow
From: Christian König
Sent: Wednesday, September 13, 2017 5
SRIOV need to reserve a memory at an offset that set by GIM/hypervisor side,
but I'm not sure how to do it perfectly, currently we call bo_create to
allocate a VRAM BO, and call pin_restrict with "offset" as parameter for "min"
and "offset + size" as "max",
I feel strange of above approach fr
but we use VRAM domain in the first bo_create stage, so this BO is
allocated directly ...
Yeah, but just stop doing so. amdgpu_bo_create can just be used to
allocate the BO structure without any backing storage.
So what you do is calling amdgpu_bo_create() with AMDGPU_GEM_DOMAIN_CPU
(or jus
verified work,
Reviewed-by: Monk Liu
From: amd-gfx on behalf of Christian
König
Sent: Wednesday, September 13, 2017 4:47:34 PM
To: amd-gfx@lists.freedesktop.org; dri-de...@lists.freedesktop.org
Subject: [PATCH] drm/ttm: fix memory leak while individualizing B
yeah, that's right !
Thanks
BR Monk
From: Christian König
Sent: Wednesday, September 13, 2017 5:30 PM
To: Liu, Monk; Deucher, Alexander; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH 4/5] drm/amdgpu: cleanup amdgpu_bo_pin_restricted
but we use VRAM domai
On 2017年09月13日 16:07, Christian König wrote:
Am 13.09.2017 um 09:39 schrieb zhoucm1:
On 2017年09月13日 15:09, Christian König wrote:
syncfile is the older implementation, sync_obj is the replacement
for it.
Are you sure sync_obj is a replacement for syncfile? Anyone can
clarify it?
sync_o
On Wed, Sep 13, 2017 at 5:03 AM, zhoucm1 wrote:
> Hi Marek,
>
> You're doing same things with me, see my "introduce syncfile as fence
> reuturn" patch set, which makes things more simple, we just need to directly
> return syncfile fd to UMD when CS, then the fence UMD get will be always
> syncfile
Could you describe how difficult to directly use CS syncfile fd in Mesa
compared with concerting CS seq to syncfile fd via several syncobj ioctls?
Regards,
David Zhou
发自坚果 Pro
Marek Ol?醟 于 2017年9月13日 下午6:11写道：
On Wed, Sep 13, 2017 at 5:03 AM, zhoucm1 wrote:
> Hi Marek,
>
> You're doing same
On Wed, Sep 13, 2017 at 1:32 PM, Zhou, David(ChunMing)
wrote:
> Could you describe how difficult to directly use CS syncfile fd in Mesa
> compared with concerting CS seq to syncfile fd via several syncobj ioctls?
It just simplifies things. Mesa primarily uses seq_no-based fences and
will continue
Yes, to be comptibility, I kept both seq_no and syncfile fd in the patch set,
you can take a look, which really is simpler and effective way.
syncfile indeed be a good way to pass fence for user space, which already is
proved in Android and is upstreamed.
Regards,
David Zhou
发自坚果 Pro
Marek O
syncfile indeed be a good way to pass fence for user space, which already is
proved in Android and is upstreamed.
Not really. syncfile needs a file descriptor for each handle it
generates, that's quite a show stopper if you want to use it in general.
Android syncfile are meant to be used for i
Change-Id: Ifd9693470534d96bfa99a6de2fa79c84d6b97662
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/hwmgr/hwmgr.c| 52 +
drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.c | 84 +-
drivers/gpu/drm/amd/powerplay/hwmgr/ppatomctrl.h | 8 +-
drivers/gpu/drm
Change-Id: I47a456f403b99ddce0b82ac28e038d8a2ce8a7d1
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_powerplay.c
Change-Id: I307465ec2fe8fe02e19c76d979be0a1af30fed0c
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h
b/drivers/gpu/drm/amd/powerplay/inc/hardware
For android using mesa instance, egl draw will dequeue an android buffer, after
egl draw, the buffer will back to android bufffer queue, but need append a
syncfile fd. If getting syncfile fd for every egl draw always needs several
syncobj ioctls, the io overhead isn't small. But if we directly r
On Wed, Sep 13, 2017 at 3:46 PM, Zhou, David(ChunMing)
wrote:
> For android using mesa instance, egl draw will dequeue an android buffer,
> after egl draw, the buffer will back to android bufffer queue, but need
> append a syncfile fd. If getting syncfile fd for every egl draw always needs
> sever
Am 13.09.2017 um 16:06 schrieb Marek Olšák:
On Wed, Sep 13, 2017 at 3:46 PM, Zhou, David(ChunMing)
wrote:
For android using mesa instance, egl draw will dequeue an android buffer,
after egl draw, the buffer will back to android bufffer queue, but need
append a syncfile fd. If getting syncfile f
v2: Clarify help text for pre-vega config
v3: Refer to older ASIC in more generic DCE 8,10,11 fashion
Signed-off-by: Harry Wentland
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 ++
drivers/gpu/drm/amd/display/Kconfig| 8
2 files changed, 10 insertions(+)
diff --git a/driv
On Wed, Sep 13, 2017 at 10:13 AM, Harry Wentland wrote:
> v2: Clarify help text for pre-vega config
> v3: Refer to older ASIC in more generic DCE 8,10,11 fashion
>
> Signed-off-by: Harry Wentland
Reviewed-by: Alex Deucher
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 ++
> drivers/gp
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Rex Zhu
> Sent: Wednesday, September 13, 2017 9:09 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Zhu, Rex
> Subject: [PATCH 2/4] drm/amd/powerplay: add ci support in hwmgr
>
> Change-Id: Ifd96
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Rex Zhu
> Sent: Wednesday, September 13, 2017 9:09 AM
> To: amd-gfx@lists.freedesktop.org
> Cc: Zhu, Rex
> Subject: [PATCH 4/4] drm/amd/powerplay: fix pcie max lane define error
>
> Change-Id
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c | 12 ++--
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 0d4c2c5a7948..2180ed3eae3b 100644
--- a/drivers/gpu/drm
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Tom St Denis
> Sent: Wednesday, September 13, 2017 12:38 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: StDenis, Tom
> Subject: [PATCH] drm/amd/amdgpu: Change vram debugfs to NO_KIQ for
> VM env
From: Akshu Agrawal
Amdgpu asic types will be required for other drivers too.
Hence, its better to keep it in a separate include file.
Reviewed-by: Alex Deucher
Signed-off-by: Akshu Agrawal
Signed-off-by: Alex Deucher
---
v1: New patch to share asic_type definitions between GPU and audio dri
From: Vijendar Mukunda
Added DMA driver changes for Stoney platform.
Below are the key differences between Stoney and CZ
In Stoney, Memory Gating is disabled.SRAM Banks won't
be turned off.No Of SRAM Banks reduced to 6.
DAGB Garlic Interface used and 16 bit resolution is supported.
SRAM bank 1 &
From: Vijendar Mukunda
asic_type information is passed to ACP DMA Driver as platform data.
Reviewed-by: Alex Deucher
Signed-off-by: Vijendar Mukunda
Signed-off-by: Alex Deucher
---
v2: Removed asic_type local variable and directly passing asic_type
instance to ACP DMA driver as platform data
From: Vijendar Mukunda
Stoney uses 16kb SRAM memory for playback and 16Kb
for capture.Modified Max buffer size to have the
correct mapping between System Memory and SRAM.
Added snd_pcm_hardware structures for playback
and capture for Stoney.
Reviewed-by: Alex Deucher
Signed-off-by: Vijendar Mu
From: Akshu Agrawal
The driver is used for AMD board using rt5650 codec.
Reviewed-by: Alex Deucher
Signed-off-by: Akshu Agrawal
Signed-off-by: Alex Deucher
Signed-off-by: Dylan Reid
---
v2: Change ACPI ID to AMDI1002
v3: add R-b, resend
v4: Fix audio jack issue (Dylan Reid)
Fix Kconfig
From: Vijendar Mukunda
For Stoney platform, Memory gating is disabled.i.e SRAM Banks
won't be turned off. By Default, SRAM Bank state set to ON.
Added condition checks to skip SRAM Bank state set logic for
Stoney platform.
Acked-by: Mark Brown
Reviewed-by: Alex Deucher
Signed-off-by: Vijendar
This patch set updates the AMD GPU and Audio CoProcessor (ACP)
audio drivers and the designware i2s driver for Stoney (ST).
ST is an APU similar to Carrizo (CZ) which already has ACP audio
support. The i2s controller and ACP audio DMA engine are part of
the GPU and both need updating so I would li
On 11/09/17 03:46 AM, Christian König wrote:
Am 09.09.2017 um 12:49 schrieb Nicolai Hähnle:
From: Nicolai Hähnle
This is already hooked up to the "amdgpu_gpr" debugfs file used by
the umr userspace debugging tool.
Signed-off-by: Nicolai Hähnle
Acked-by: Christian König
Reviewed-by: Tom
On Wed, Sep 13, 2017 at 01:57:49PM -0400, Alex Deucher wrote:
> I just need acks from audio on the remaining patches that touch audio
> (1, 5, 6), and then I can send a pull request.
No, send me a pull request for the dependencies so we can do a cross
merge. I want to make sure that when this ge
Signed-off-by: Allen Pais
---
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
index 4395a4f..0487936 100644
--- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_
Signed-off-by: Tom St Denis
---
doc/umr.1 | 4 ++--
src/app/main.c | 26 ++
2 files changed, 24 insertions(+), 6 deletions(-)
diff --git a/doc/umr.1 b/doc/umr.1
index b990ff2c412f..e865e131fa51 100644
--- a/doc/umr.1
+++ b/doc/umr.1
@@ -94,13 +94,13 @@ The VMID can
Am 13.09.2017 um 18:47 schrieb Deucher, Alexander:
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
Of Tom St Denis
Sent: Wednesday, September 13, 2017 12:38 PM
To: amd-gfx@lists.freedesktop.org
Cc: StDenis, Tom
Subject: [PATCH] drm/amd/amdgpu: Cha
On Wed, Sep 13, 2017 at 2:12 PM, Mark Brown wrote:
> On Wed, Sep 13, 2017 at 01:57:49PM -0400, Alex Deucher wrote:
>
>> I just need acks from audio on the remaining patches that touch audio
>> (1, 5, 6), and then I can send a pull request.
>
> No, send me a pull request for the dependencies so we
On Wed, Sep 13, 2017 at 4:01 AM, Allen Pais wrote:
> Signed-off-by: Allen Pais
Applied. thanks!
Alex
> ---
> drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c
> b/drivers/gpu/drm/amd/a
On 13/09/17 02:47 PM, Christian König wrote:
Am 13.09.2017 um 18:47 schrieb Deucher, Alexander:
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
Of Tom St Denis
Sent: Wednesday, September 13, 2017 12:38 PM
To: amd-gfx@lists.freedesktop.org
Cc: StD
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> Make sure vm->root.bo is not left reserved if amdgpu_bo_kmap fails.
>
> Signed-off-by: Felix Kuehling
Reviewed-by: Alex Deucher
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> d
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> 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 Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> Signed-off-by: Felix Kuehling
Acked-by: Alex Deucher
> ---
> 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/
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> 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
Maybe you want this before patches 2 and 3?
Acked-by: Alex Deucher
> ---
> drivers
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> Signed-off-by: Felix Kuehling
Acked-by: Alex Deucher
> ---
> drivers/gpu/drm/amd/amdkfd/kfd_module.c | 6 ---
> drivers/gpu/drm/amd/amdkfd/kfd_pasid.c | 90
> ++---
> 2 files changed, 38 insertions(+), 58
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> To filter out high-frequency interrupts that can be safely ignored.
>
> Signed-off-by: Felix Kuehling
Reviewed-by: Alex Deucher
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu.h | 2 ++
> drivers/gpu/drm/amd/amdgpu/amdgpu_ih.c | 6
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> 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
On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling wrote:
> 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
Reviewed-By : Shaoyun.liu
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of Tom
St Denis
Sent: Wednesday, September 13, 2017 2:27 PM
To: amd-gfx@lists.freedesktop.org
Cc: StDenis, Tom
Subject: [PATCH umr] Add --vm-write/--vm-read to replace the
On 2017-09-13 03:40 PM, Alex Deucher wrote:
> On Tue, Sep 12, 2017 at 7:05 PM, Felix Kuehling
> wrote:
>> 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
> Maybe you want this before patc
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Felix Kuehling
> Sent: Wednesday, September 13, 2017 5:12 PM
> To: Alex Deucher
> Cc: amd-gfx list
> Subject: Re: [PATCH 4/8] drm/amdkfd: Separate doorbell allocation from
> PASID
>
> On 2017
v2: drop hdp invalidate/flush.
Signed-off-by: Samuel Li
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 3 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c | 77 ++-
3 files changed, 81 insertions(+), 2 deletions(-)
dif
On 14 September 2017 at 00:16, Christian König wrote:
> Am 13.09.2017 um 16:06 schrieb Marek Olšák:
>>
>> On Wed, Sep 13, 2017 at 3:46 PM, Zhou, David(ChunMing)
>> wrote:
>>>
>>> For android using mesa instance, egl draw will dequeue an android buffer,
>>> after egl draw, the buffer will back to
On 2017年09月14日 09:52, Dave Airlie wrote:
On 14 September 2017 at 00:16, Christian König wrote:
Am 13.09.2017 um 16:06 schrieb Marek Olšák:
On Wed, Sep 13, 2017 at 3:46 PM, Zhou, David(ChunMing)
wrote:
For android using mesa instance, egl draw will dequeue an android buffer,
after egl draw,
On 2017年09月14日 10:07, zhoucm1 wrote:
On 2017年09月14日 09:52, Dave Airlie wrote:
On 14 September 2017 at 00:16, Christian König
wrote:
Am 13.09.2017 um 16:06 schrieb Marek Olšák:
On Wed, Sep 13, 2017 at 3:46 PM, Zhou, David(ChunMing)
wrote:
For android using mesa instance, egl draw will deq
>
> But think a bit more, if not by default, that isn't meaningful, we can
> continue to use seq_no base fence and Marek's syncfile fd approach.
>
At least for radv I've no interest in every CS ioctl returning an fd
that I have to
close.
Dave.
___
amd-g
On 2017年09月14日 10:33, Dave Airlie wrote:
But think a bit more, if not by default, that isn't meaningful, we can
continue to use seq_no base fence and Marek's syncfile fd approach.
At least for radv I've no interest in every CS ioctl returning an fd
that I have to
close.
OK, let's pend it now
64 matches
Mail list logo
