I'm pleased to announce the 18.0.1 release of xf86-video-ati, the Xorg
driver for ATI/AMD Radeon GPUs supported by the radeon kernel driver.
This release supports xserver versions 1.13-1.19. It also works with
xserver 1.20 RC1, so unless something unexpected happens, it should work
with xserver
and amdgpu_bo_create_reserved.
Signed-off-by: Andrey Grodzovsky
---
drivers/gpu/drm/amd/amdgpu/amdgpu_object.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c
Yeah, not really the usually flow.
Either way Reviewed-by: Christian König for both.
Christian.
Am 15.03.2018 um 17:19 schrieb Andrey Grodzovsky:
That explains the merge conflict that surprised me :)
Thanks,
Andrey
On 03/15/2018 12:18 PM, Deucher, Alexander
When CONFIG_ACPI is disabled, we never initialize the acpi_table
structure in kfd_create_crat_image_virtual:
drivers/gpu/drm/amd/amdkfd/kfd_crat.c: In function
'kfd_create_crat_image_virtual':
drivers/gpu/drm/amd/amdkfd/kfd_crat.c:888:40: error: 'acpi_table' may be used
uninitialized in this
I'm pleased to announce the 18.0.1 release of xf86-video-amdgpu, the
Xorg driver for AMD Radeon GPUs supported by the amdgpu kernel driver.
This release supports xserver versions 1.13-1.19. It also works with
xserver 1.20 RC1, so unless something unexpected happens, it should work
with xserver
On Thu, Mar 15, 2018 at 11:07 AM, Harry Wentland wrote:
> From: Clark Zheng
>
> bad case won't follow normal sense, it will not enable vga1 as usual, but
> vga2,3,4 is on.
>
> Signed-off-by: Clark Zheng
> Reviewed-by: Tony Cheng
Reusing local handle to initialize BO without resetting it to
NULL is wrong since it causes amdgpu_bo_create_reserved to skip
new BO creation and just reuse the given pointer for pinning.
v2:
Get rid of mc_addr and kaddr.
Signed-off-by: Andrey Grodzovsky
---
For the series:
Reviewed-by: Alex Deucher
I think Rex already applied the second one.
From: amd-gfx on behalf of Andrey
Grodzovsky
Sent: Thursday, March 15, 2018
On 2018-03-15 06:33 AM, Shirish S wrote:
> This patch fixes static checker warning caused by
> "36cc549d5986: "drm/amd/display: disable CRTCs with
> NULL FB on their primary plane (V2)"
>
> Reported-by: Dan Carpenter
> Signed-off-by: Shirish S
From: Mikita Lipski
By moving amdgpu_irq_disable_all earlier in the sequence
fixes an issue with disabling pflip interrupts:
*ERROR* dal_irq_service_dummy_ack: called for non-implemented irq source
Earlier patch fixed a memory corruption and revealed irq
warnings.This
Hey,
so there appears to be some progress as of today's build of
amd-staging-drm-next branch. The system I use has 2 external displays,
identical, one connected over HDMI another over DVI.
Currently:
1. Just after the boot, gamma settings on DVI-connected one are messed
up (is this HDR mode?).
I created a ticket: https://bugs.freedesktop.org/show_bug.cgi?id=105530
Let me know if there is any additional information I need to add.
Thanks!
-Dawson Dias
On Thu, Mar 15, 2018 at 8:44 PM, Harry Wentland
wrote:
> On 2018-03-15 11:01 AM, Michel Dänzer wrote:
> > On
On 2018-03-15 02:47 PM, Dawson Dias wrote:
> I created a ticket: https://bugs.freedesktop.org/show_bug.cgi?id=105530
> Let me know if there is any additional information I need to add.
>
Thanks, Dawson.
Harry
> Thanks!
>
> -Dawson Dias
>
> On Thu, Mar 15, 2018 at 8:44 PM, Harry Wentland
On Thu, Mar 15, 2018 at 5:11 PM, wrote:
> From: Mikita Lipski
>
> By moving amdgpu_irq_disable_all earlier in the sequence
> fixes an issue with disabling pflip interrupts:
>
> *ERROR* dal_irq_service_dummy_ack: called for non-implemented irq source
Add helpers for allocating GPUVM memory in kernel mode and use them
to allocate memory for the CWSR trap handler.
v2: Use dev instead of pdd->dev in kfd_process_free_gpuvm
v3:
* Cleaned up and simplified kfd_process_alloc_gpuvm
* Moved allocation for dGPU to kfd_process_device_init_vm
The events page must be accessible in user mode by the GPU and CPU
as well as in kernel mode by the CPU. On dGPUs user mode virtual
addresses are managed by the Thunk's GPU memory allocation code.
Therefore we can't allocate the memory in kernel mode like we do
on APUs. But KFD still needs to map
This adds support for allocating, mapping, unmapping and freeing
userptr BOs, and for handling MMU notifiers.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 11 +
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 566
When an MMU notifier runs in memory reclaim context, it can deadlock
trying to take locks that are already held in the thread causing the
memory reclaim. The solution is to avoid memory reclaim while holding
locks that are taken in MMU notifiers by using GFP_NOIO.
This commit fixes memory
v2:
* Fix error handling after kfd_bind_process_to_device in
kfd_ioctl_map_memory_to_gpu
v3:
* Add ioctl to acquire VM from a DRM FD
v4:
* Return number of successful map/unmap operations in failure cases
* Facilitate partial retry after failed map/unmap
* Added comments with parameter
These interfaces allow KGD to stop and resume all GPU user mode queue
access to a process address space. This is needed for handling MMU
notifiers of userptrs mapped for GPU access in KFD VMs.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_device.c
Currently the number of GPUs is limited by aperture placement options
available on GFX7 and GFX8 hardware. This limitation is not necessary.
Scratch and LDS represent per-work-item and per-work-group storage
respectively. Different work-items and work-groups use the same virtual
address to access
This commit allows amdgpu_ttm_tt_get_user_pages to work in a worker
thread rather than regular process context. This will be used when
KFD userptr BOs are restored after an MMU-notifier eviction.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c |
Also used for cleaning up on process termination.
v2: Refactored cleanup on process termination
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_priv.h| 11 +
drivers/gpu/drm/amd/amdkfd/kfd_process.c | 73
2 files
v2: Removed updating and checking of vm->vm_context
v3: Enable amdgpu_vm_clear_bo in amdgpu_vm_make_compute
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 67 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.h | 1 +
2
On GFX7 the CP does not perform a TC flush when queues are unmapped.
To avoid TC eviction from accessing an invalid VMID, flush it
explicitly before releasing a VMID.
v2: Fix unnecessary list_for_each_entry_safe
v3: Moved allocation to kfd_process_device_init_vm
Signed-off-by: Amber Lin
This commit adds the notion of MMU notifier types GFX and HSA. GFX
continues to work like MMU notifiers did before. HSA adds support for
KFD userptr BOs. The implementation of KFD userptr eviction is a stub
for now.
Signed-off-by: Felix Kuehling
---
Simulate large-BAR system by exporting only visible memory. This
limits the amount of available VRAM to the size of the BAR, but
enables CPU access to VRAM.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c | 5 +
Set up the GPUVM aperture for SVM (shared virtual memory) that allows
sharing a part of virtual address space between GPUs and CPUs.
Report the size of the the GPUVM size supported by KGD accurately.
The low part of the GPUVM aperture is reserved for kernel use. This is
for kernel-allocated
Instead of creating all VMs on process creation, create them when
a process is bound to a device. This will later allow registering
an existing VM from a DRM render node FD at runtime, before the
process is bound to the device. This way the render node VM can be
used for KFD instead of creating
This allows acquiring an existing VM from a render node FD to use it
for a compute process.
Such VMs get destroyed when the original file descriptor is released.
Added a callback from amdgpu_vm_fini to handle KFD VM destruction
correctly in this case.
v2:
* Removed vm->vm_context check in
From: Oak Zeng
Populate DRM render device minor in kfd topology
Signed-off-by: Oak Zeng
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdkfd/kfd_topology.c | 4
drivers/gpu/drm/amd/amdkfd/kfd_topology.h | 1 +
2 files
Rebased and integrated review feedback from v3:
* Removed vm->vm_context field
* Use uninterruptible waiting in initial PD validation to avoid ERESTARTSYS
* Return number of successful map/unmap operations in failure cases
* Facilitate partial retry after failed map/unmap
* Added comments with
When an MMU notifier runs in memory reclaim context, it can deadlock
trying to take locks that are already held in the thread causing the
memory reclaim. The solution is to avoid memory reclaim while holding
locks that are taken in MMU notifiers by using GFP_NOIO.
This commit fixes kmalloc while
Make sure the PD BO is valid and attach the eviction fence during VM
creation. This ensures that the pd_phys_address is actually valid
and an eviction that would invalidate it triggers a KFD process
eviction like it should.
v2: Use uninterruptible waiting in initial PD validation
Signed-off-by:
Remove struct amdkfd_vm and move the fields into struct amdgpu_vm.
This will allow turning a VM created by a DRM render node into a
KFD VM.
v2: Removed vm_context field
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h | 21 --
Sorry for jumping in late, but did you verify MST with this change ?
Just a guess but MST relies on hpd_rx interrupt for messages and
disabling to early might have impact. Maybe trying unload/reload with
MST display to be sure ?
Andrey
On 03/15/2018 06:23 PM, Alex Deucher wrote:
On Thu,
Under some heavy computing test(dgemm) environment, it may takes
the asic over 50+ seconds to finish the dispatched single job
which will trigger the timeout. It's quite annoying although it
does not seem to bring any real problems.
As a quick workround, we choose to not enfoce the timeout
setting
Change-Id: I4648ad6885f3884688d8bb90ea33be7f6ad69810
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 5 -
drivers/gpu/drm/amd/amdgpu/vi.c| 5 -
2 files changed, 10 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c
Hi Andrey,
Sorry for the confusion.
I just need to push a bunch of patches that conflict with your patch.
So apply your patch first.
Best Regards
Rex
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of
Andrey Grodzovsky
Sent: Friday, March 16, 2018 12:20 AM
To: Deucher,
That explains the merge conflict that surprised me :)
Thanks,
Andrey
On 03/15/2018 12:18 PM, Deucher, Alexander wrote:
For the series:
Reviewed-by: Alex Deucher
I think Rex already applied the second one.
Am 15.03.2018 um 10:20 schrieb Daniel Vetter:
On Tue, Mar 13, 2018 at 06:20:07PM +0100, Christian König wrote:
[SNIP]
Take a look at the DOT graphs for atomic I've done a while ago. I think we
could make a formidable competition for who's doing the worst diagrams :-)
Thanks, going to give that
This patch fixes static checker warning caused by
"36cc549d5986: "drm/amd/display: disable CRTCs with
NULL FB on their primary plane (V2)"
Reported-by: Dan Carpenter
Signed-off-by: Shirish S
---
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 3
Refine the commit f49e9bac191b066060a53d994f9c964fb71454f2
drm/amd/pp: Get and save Rv smu version
Change-Id: Iea668cc2161e5fcf339b408eb7eca1b028e26df2
Signed-off-by: Rex Zhu
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c | 5 -
Change-Id: I9e4babf1e91855fb66e65cf2f82db64a1cd6fc97
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
Users can pass in an array to decide enable/disable Zero Frame Buffer.
zfb[0] = zfb_size(MB), zfb[1] = zfb_phys_addr(MB).
If zbf_size > 0, zfb is enabled. Otherwise disabled.
Usage for example:
modprobe amdgpu zfb=256,8192
Change-Id: I711062eb86b6cdff74572cabb3df250c6708e473
Change-Id: I8253c8ff80e0cbd1f12e5ee801600e7619e6718f
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.h
> What are you trying to do? A large desktop spread across both chips?
> Use the chips independently? There is no support for crossfire on Linux. If
> you want to use a desktop across multiple chips or select which chip is used
> for rendering, you can use xrandr and the prime support in X:
>
From: Hawking Zhang
Change-Id: I09f9ddea0ad23af00fadd9af7aaccf7160e4e569
Signed-off-by: Hawking Zhang
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
Signed-off-by: Feifei Xu
Use module parameter passed from user to initialize zfb start address
and size.
Change-Id: I3d786e863114a217f89ff7c3d4ffdabf000f31a4
Signed-off-by: Feifei Xu
Signed-off-by: Hawking Zhang
Acked-by: John Bridgman
Reviewed-by: Alex
Change-Id: I866dd16548304a42298b0cb28741f27cba3a76ca
Signed-off-by: Feifei Xu
Signed-off-by: Hawking Zhang
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c |
Change-Id: I2b45d765f1f60252fa1c02aced94f8100d575ddc
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gfxhub_v1_0.c
Am 15.03.2018 um 07:12 schrieb Feifei Xu:
Use module parameter passed from user to initialize zfb start address
and size.
Change-Id: I3d786e863114a217f89ff7c3d4ffdabf000f31a4
Signed-off-by: Feifei Xu
Signed-off-by: Hawking Zhang
Acked-by: John
Am 15.03.2018 um 07:12 schrieb Feifei Xu:
Change-Id: I2b45d765f1f60252fa1c02aced94f8100d575ddc
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
Am 15.03.2018 um 08:09 schrieb Lvzhihong (ReJohn):
What are you trying to do? A large desktop spread across both chips?
Use the chips independently? There is no support for crossfire on Linux. If
you want to use a desktop across multiple chips or select which chip is used
for rendering, you
Hello Shirish S,
The patch 36cc549d5986: "drm/amd/display: disable CRTCs with NULL FB
on their primary plane (V2)" from Feb 28, 2018, leads to the
following static checker warning:
drivers/gpu/drm/amd/amdgpu/../display/amdgpu_dm/amdgpu_dm.c:4827
dm_atomic_check_plane_state_fb()
Am 15.03.2018 um 07:11 schrieb Feifei Xu:
Change-Id: I8253c8ff80e0cbd1f12e5ee801600e7619e6718f
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
Apply this patch with Christian's suggestions.
Thanks.
Best Regards
Rex
From: Christian König
Sent: Thursday, March 15, 2018 2:21 AM
To: Grodzovsky, Andrey; amd-gfx@lists.freedesktop.org
Cc: Zhu, Rex; Koenig, Christian
Am 15.03.2018 um 07:12 schrieb Feifei Xu:
From: Hawking Zhang
Change-Id: I09f9ddea0ad23af00fadd9af7aaccf7160e4e569
Signed-off-by: Hawking Zhang
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
On 2018-03-15 08:51 AM, Dawson Dias wrote:
> I managed to pinpoint the issue. The stuttering is caused by enabling
> TearFree via xrandr. When on `auto`, there is no stuttering, but when
> enabled by setting it to `on`, there is intense stuttering noticeable
> when dragging windows around, but
I managed to pinpoint the issue. The stuttering is caused by enabling
TearFree via xrandr. When on `auto`, there is no stuttering, but when
enabled by setting it to `on`, there is intense stuttering noticeable when
dragging windows around, but it's also very noticeable when just moving the
mouse
Am 15.03.2018 um 07:11 schrieb Feifei Xu:
In gmc_v9_0_vram_gtt_location(),the vram_base_offset is hardcoded
to 0 in dGPU. Fix it by reading mmMC_VM_FB_OFFSET or return
zfb_phys_addr if ZFB is enabled.
Change-Id: I585b7d4d96ebab2a5d7178fe8d1d6a746ef0c72a
Signed-off-by: Feifei Xu
Am 15.03.2018 um 07:11 schrieb Feifei Xu:
Users can pass in an array to decide enable/disable Zero Frame Buffer.
zfb[0] = zfb_size(MB), zfb[1] = zfb_phys_addr(MB).
If zbf_size > 0, zfb is enabled. Otherwise disabled.
Usage for example:
modprobe amdgpu zfb=256,8192
Change-Id:
Am 15.03.2018 um 07:12 schrieb Feifei Xu:
Change-Id: I9e4babf1e91855fb66e65cf2f82db64a1cd6fc97
Signed-off-by: Hawking Zhang
Signed-off-by: Feifei Xu
Acked-by: John Bridgman
Reviewed-by: Alex Deucher
On Tue, Mar 13, 2018 at 06:20:07PM +0100, Christian König wrote:
> Am 13.03.2018 um 17:00 schrieb Daniel Vetter:
> > On Tue, Mar 13, 2018 at 04:52:02PM +0100, Christian König wrote:
> > > Am 13.03.2018 um 16:17 schrieb Daniel Vetter:
> > > [SNIP]
> > Ok, so plan is to support fully pipeline moves
Reviewed-by: Alex Deucher
From: amd-gfx on behalf of Rex Zhu
Sent: Thursday, March 15, 2018 2:47:21 AM
To: amd-gfx@lists.freedesktop.org
Cc: Zhu, Rex
Subject: [PATCH]
On 2018-03-15 10:15 AM, Alex Deucher wrote:
On Thu, Mar 15, 2018 at 10:10 AM, wrote:
From: Mikita Lipski
Disable irq on devices before destroying them. That prevents
use-after-free memory access when unloading the driver.
Signed-off-by:
On Thu, Mar 15, 2018 at 10:23 AM, Mikita Lipski wrote:
>
>
> On 2018-03-15 10:15 AM, Alex Deucher wrote:
>>
>> On Thu, Mar 15, 2018 at 10:10 AM, wrote:
>>>
>>> From: Mikita Lipski
>>>
>>> Disable irq on devices before destroying
Acked-by: Alex Deucher
From: S, Shirish
Sent: Thursday, March 15, 2018 6:33:51 AM
To: Deucher, Alexander; Wentland, Harry; amd-gfx@lists.freedesktop.org
Cc: S, Shirish
Subject: [PATCH] drm/amd/display: fix dereferencing possible
Partial revert of 1b0ff66bc0bf1a0559255cb7b066a65d55491974
Signed-off-by: Tom St Denis
---
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/amd/display/dc/dcn10/dcn10_hw_sequencer.c
From: Mikita Lipski
Disable irq on devices before destroying them. That prevents
use-after-free memory access when unloading the driver.
Signed-off-by: Mikita Lipski
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 5 +++--
1 file changed, 3
On Thu, Mar 15, 2018 at 10:10 AM, wrote:
> From: Mikita Lipski
>
> Disable irq on devices before destroying them. That prevents
> use-after-free memory access when unloading the driver.
>
> Signed-off-by: Mikita Lipski
> ---
From: Mikita Lipski
Disable irq on devices before destroying them. That prevents
use-after-free memory access when unloading the driver.
Signed-off-by: Mikita Lipski
Reviewed-by: Alex Deucher
---
Am 15.03.2018 um 15:50 schrieb mikita.lip...@amd.com:
From: Mikita Lipski
Disable irq on devices before destroying them. That prevents
use-after-free memory access when unloading the driver.
Signed-off-by: Mikita Lipski
Reviewed-by: Alex Deucher
On 2018-03-15 03:47 PM, Dawson Dias wrote:
> Here we go! The logs are attached. I turned TearFree to `on` from
> `auto`, dragged some windows around, and captured them.
> AFAIK, I haven't made any changes via xrandr, except for toggling the
> TearFree property.
Thanks. I don't see any obvious
On 2018-03-15 10:50 AM, mikita.lip...@amd.com wrote:
> From: Mikita Lipski
>
> Disable irq on devices before destroying them. That prevents
> use-after-free memory access when unloading the driver.
>
> Signed-off-by: Mikita Lipski
> Reviewed-by:
From: Clark Zheng
bad case won't follow normal sense, it will not enable vga1 as usual, but
vga2,3,4 is on.
Signed-off-by: Clark Zheng
Reviewed-by: Tony Cheng
---
drivers/gpu/drm/amd/display/dc/dce/dce_hwseq.h | 8 +++-
On 2018-03-15 09:47 AM, Tom St Denis wrote:
> Partial revert of 1b0ff66bc0bf1a0559255cb7b066a65d55491974
>
> Signed-off-by: Tom St Denis
Thanks for your patch.
It seems like on some systems this is not enough as D1 is off and D2-4 are on,
which would still early return
On 2018-03-15 11:01 AM, Michel Dänzer wrote:
> On 2018-03-15 03:47 PM, Dawson Dias wrote:
>> Here we go! The logs are attached. I turned TearFree to `on` from
>> `auto`, dragged some windows around, and captured them.
>> AFAIK, I haven't made any changes via xrandr, except for toggling the
>>
78 matches
Mail list logo