From: Christian König
We should be able to handle BOs from other instances as well.
v2:
* Add a module option that is off-by-default
* Use new DRM helper function to check the exporting driver
Signed-off-by: Christian König
Signed-off-by:
This allows drivers to check if a DMA-buf contains a GEM object and
whether it comes from the same driver. It may be from the same or a
different device.
Signed-off-by: Felix Kuehling
---
drivers/gpu/drm/drm_prime.c | 24
include/drm/drmP.h
Am 06.07.2017 um 22:16 schrieb Felix Kuehling:
This allows drivers to check if a DMA-buf contains a GEM object and
whether it comes from the same driver. It may be from the same or a
different device.
Signed-off-by: Felix Kuehling
I think Daniel/Dave hadmore a
This patch series adds experimental P2P buffer sharing. It's disabled
by default and can be enabled with amdgpu.p2p_sharing=1.
TODO:
* Add a whitelist of chipsets that support P2P access
* Refine UVD/VCE handling on newer IP versions with GPUVM support
* Handle foreign BOs in DC path
* Test for
I split the P2P patches into a separate series for review. Patches 1-6
from this series should still be valid.
I still need Reviewed-by's for patches 1 and 4.
Patches 2 and 5 are by other authors and reviewed by me. But I'd feel
better to get at least an Acked-by from some amdgpu folks.
Thanks,
On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
> Enable SDMA context switching on CIK (copied from sdma_v3_0.c).
>
> Signed-off-by: Felix Kuehling
Reviewed-by: Alex Deucher
> ---
>
On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
> Set a configurable SDMA phase quantum when enabling SDMA context
> switching. The default value significantly reduces SDMA latency
> in page table updates when user-mode SDMA queues have concurrent
> activity,
On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
> From: Christian König
>
> Pinning them in other devices VRAM would obviously not work.
>
> Signed-off-by: Christian König
> Reviewed-by: Felix Kuehling
Chrstian,
you are probably the best person to ack this, I'd like to get the radv
code landed
and allow the GL code to get going.
Dave.
> This adds kernel semaphore support to the command submission
> interface in what should be a backwards compatible manner,
> it adds a new command submission
On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
> From: shaoyunl
>
> For GFX context, the ATC bit in SDMA*_GFX_VIRTUAL_ADDRESS can be cleared
> to perform in VM mode. For RLC context, to support ATC mode , ATC bit in
>
On Thu, Jul 6, 2017 at 5:33 PM, Yong Zhao wrote:
> Hi Alex,
>
> As far as I know, we never tested suspend/resume on the setting you
> mentioned. Theoretically it should work.
Are the kfd s/r entry points global or per GPU? If you have two GPUs
and you suspend one, will it
From: Christian König
They don't support VM mode yet.
Signed-off-by: Christian König
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
From: Christian König
This allows us to have multiple GEM objects for one BO.
Signed-off-by: Christian König
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 12 +++--
From: Amber Lin
Set the system bit for foreign BO mappings and use the remote VRAM
BAR address as the VRAM base offset.
Signed-off-by: Amber Lin
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 17
From: Christian König
Pinning them in other devices VRAM would obviously not work.
Signed-off-by: Christian König
Reviewed-by: Felix Kuehling
---
drivers/gpu/drm/amd/amdgpu/amdgpu_display.c | 6 ++
1 file
On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
> From: Yong Zhao
>
> Signed-off-by: Yong Zhao
> Reviewed-by: Felix Kuehling
Does this work properly for multiple GPUs? E.g., if one is suspended
and
On 17-07-06 05:13 PM, Alex Deucher wrote:
> On Mon, Jul 3, 2017 at 5:11 PM, Felix Kuehling wrote:
>> From: Christian König
>>
>> Pinning them in other devices VRAM would obviously not work.
>>
>> Signed-off-by: Christian König
On 2017-07-06 05:44 PM, Alex Deucher wrote:
On Thu, Jul 6, 2017 at 5:33 PM, Yong Zhao wrote:
Hi Alex,
As far as I know, we never tested suspend/resume on the setting you
mentioned. Theoretically it should work.
Are the kfd s/r entry points global or per GPU? If you have
Am 05.07.2017 um 21:51 schrieb Alex Deucher:
Not all vbios images seem to set the version appropriately.
Switch the check based on asic type instead.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_bios.c | 15 +--
1 file changed, 1
On 07/01/2017 05:32 AM, Alex Deucher wrote:
Rather than checking the CONGIG_MEMSIZE register as that may
not be reliable on some APUs.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 10 +-
1 file changed, 1 insertion(+), 9
Reviewed-by: Christian König .
Am 06.07.2017 um 04:38 schrieb Zhu, Rex:
Reviewed-by: Rex Zhu
Best Regards
Rex
-Original Message-
From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf Of Evan
Quan
Sent: Thursday, July 06,
On 5 July 2017 at 22:31, Li, Samuel wrote:
>> - above all, as-is make check will fail
> Right, I did not check that.
>
>> - keeping the radeon API symmetrical to the amdgpu one would a good idea
> The issue is Radeon does not have a struct similar to amdgpu_device_handle.
Am 06.07.2017 um 12:51 schrieb Michel Dänzer:
From: John Brooks
The BO move throttling code is designed to allow VRAM to fill quickly if it
is relatively empty. However, this does not take into account situations
where the visible VRAM is smaller than total VRAM, and total
From: John Brooks
There is no need for page faults to force BOs into visible VRAM if it's
full, and the time it takes to do so is great enough to cause noticeable
stuttering. Add GTT as a possible placement so that if visible VRAM is
full, page faults move BOs to GTT instead
From: Michel Dänzer
This is a rearrangement of some patches of John's, with some improvements
of mine, and a patch of mine, rebased on current amd-staging-4.11.
Details about changes in the individual patches.
John Brooks (4):
drm/amdgpu: Add vis_vramlimit module
From: John Brooks
When a BO is moved to VRAM, clear AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED.
This allows it to potentially later move to invisible VRAM if the CPU
does not access it again.
Setting the CPU_ACCESS flag in amdgpu_bo_fault_reserve_notify() also means
that we can
From: John Brooks
Allow specifying a limit on visible VRAM via a module parameter. This is
helpful for testing performance under visible VRAM pressure.
v2: Add cast to 64-bit (Christian König)
Signed-off-by: John Brooks
Reviewed-by: Michel Dänzer
From: Michel Dänzer
This gives BOs which haven't been accessed by the CPU since they were
moved to visible VRAM another chance to stay in VRAM when another BO
needs to go to visible VRAM.
This should allow BOs to stay in VRAM longer in some cases.
v2:
* Only do this for
Am 06.07.2017 um 12:51 schrieb Michel Dänzer:
From: John Brooks
When a BO is moved to VRAM, clear AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED.
This allows it to potentially later move to invisible VRAM if the CPU
does not access it again.
Setting the CPU_ACCESS flag in
Am 06.07.2017 um 12:51 schrieb Michel Dänzer:
From: John Brooks
There is no need for page faults to force BOs into visible VRAM if it's
full, and the time it takes to do so is great enough to cause noticeable
stuttering. Add GTT as a possible placement so that if visible
> -Original Message-
> From: Zhang, Jerry
> Sent: Thursday, July 06, 2017 2:08 AM
> To: Alex Deucher; amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander
> Subject: Re: [PATCH 4/5] drm/amdgpu: check scratch registers to see if we
> need post
>
> On 07/01/2017 05:32 AM, Alex Deucher
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Emil Velikov
> Sent: Thursday, July 06, 2017 5:21 AM
> To: Li, Samuel
> Cc: ML dri-devel; amd-gfx mailing list
> Subject: Re: [PATCH libdrm 2/2] radeon: use asic id table to get chipset name
This is meant for NPI support only as publicly released designs
can simply add to the static definitions.
This is used by using the --force (-f) option with a filename that has
a @ prefixed on, e.g.
umr -O bits -f @demo/npi/newchip -lr gfx10
Will read the file 'newchip' and import the
This is meant for NPI support only as publicly released designs
can simply add to the static definitions.
This is used by using the --force (-f) option with a filename that has
a @ prefixed on, e.g.
umr -O bits -f @demo/npi/newchip -lr gfx10
Will read the file 'newchip' and import the
The redundant fb helpers .load_lut, .gamma_set and .gamma_get are
no longer used. Remove the dead code and hook up the crtc .gamma_set
to use the crtc gamma_store directly instead of duplicating that
info locally.
Signed-off-by: Peter Rosin
---
The redundant fb helpers .load_lut, .gamma_set and .gamma_get are
no longer used. Remove the dead code and hook up the crtc .gamma_set
to use the crtc gamma_store directly instead of duplicating that
info locally.
Signed-off-by: Peter Rosin
---
Hi!
While trying to get CLUT support for the atmel_hlcdc driver, and
specifically for the emulated fbdev interface, I received some
push-back that my feeble in-driver attempts should be solved
by the core. This is my attempt to do it right.
I have obviously not tested all of this with more than
From: Colin Ian King
The arrays pctl0_data and pctl1_data do not need to be in global scope,
so them both static.
Cleans up sparse warnings:
symbol 'pctl0_data' was not declared. Should it be static?
symbol 'pctl1_data' was not declared. Should it be static?
On Thu, Jul 6, 2017 at 5:58 AM, Colin King wrote:
> From: Colin Ian King
>
> The arrays pctl0_data and pctl1_data do not need to be in global scope,
> so them both static.
>
> Cleans up sparse warnings:
> symbol 'pctl0_data' was not declared.
Am 06.07.2017 um 12:51 schrieb Michel Dänzer:
From: Michel Dänzer
This gives BOs which haven't been accessed by the CPU since they were
moved to visible VRAM another chance to stay in VRAM when another BO
needs to go to visible VRAM.
This should allow BOs to stay in
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> We want to use them as remap address space.
>
> Signed-off-by: Christian König
2x512 pages. With that updated:
Reviewed-by: Alex
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> This allows us to write the mapped PTEs into
> an IB instead of the table directly.
>
> v2: fix build with debugfs enabled, remove unused assignment
>
>
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> This way we don't need to map the full BO at a time any more.
>
> v2: use fixed windows for src/dst
>
> Signed-off-by: Christian König
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> No need to map BOs to GTT on eviction and intermediate transfers any more.
>
> Signed-off-by: Christian König
Reviewed-by: Alex
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> We can finally remove this now.
>
> v2: remove now unused max_size variable as well.
>
> Signed-off-by: Christian König
Reviewed-by:
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> Use the TTM values instead of the hardware config here.
>
> Signed-off-by: Christian König
Reviewed-by: Alex Deucher
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> No functional change.
>
> Signed-off-by: Christian König
I agree with what the patch does, but I'd prefer gart to sysvm for the
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> Instead of relying on the sysvm_size to be the same as the module parameter.
>
> Signed-off-by: Christian König
Reviewed-by: Alex
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> No intended functional change.
>
> Signed-off-by: Christian König
I think I'd prefer to keep this together with the vram_location
On Mon, Jul 3, 2017 at 5:44 AM, Christian König wrote:
> From: Christian König
>
> Limit the size of the SYSVM. This saves us a bunch of visible VRAM,
> but also limitates the maximum BO size we can swap out.
Update the description. The
Call nbio init registers on hw_init to set up any
nbio registers that need initialization at hw init time.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c
Used for nbio registers that need to be initialized. Currently
only used for a golden setting that got missed on some boards.
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/nbio_v6_1.c | 13 +
drivers/gpu/drm/amd/amdgpu/nbio_v6_1.h | 1 +
2
Change-Id: If2ac93808d1922fe3dbec64a58f674a9c342088d
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c | 4
drivers/gpu/drm/amd/include/cgs_common.h | 1 +
2 files changed, 5 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c
Change-Id: I1a62db8c764d368e6ba08440ad1761adb55d85fd
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 24
1 file changed, 24 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c
b/drivers/gpu/drm/amd/amdgpu/soc15.c
index
The late 2009, 27 inch Apple iMac10,1 has an
internal eDP display and an external Mini-
Displayport output, driven by a DCE-3.2, RV730
Radeon Mobility HD-4670.
The machine worked fine in a dual-display setup
with eDP panel + externally connected HDMI
or DVI-D digital display sink, connected via
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Evan Quan
> Sent: Thursday, July 06, 2017 9:59 PM
> To: amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander; Quan, Evan
> Subject: [PATCH 6/6] drm/amd/powerplay: added didt support for
Hi Alex,
As far as I know, we never tested suspend/resume on the setting you
mentioned. Theoretically it should work.
When I read the code now, I was wondering whether we should stop kfd
before amdgpu_bo_evict_vram() and amdgpu_fence_driver_suspend(). If
that's not needed, it may make more
Change-Id: If7088d6046c84355b8f1308d584ab6524ceab724
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/amdgpu_cgs.c | 12
drivers/gpu/drm/amd/include/cgs_common.h | 5 +
2 files changed, 17 insertions(+)
diff --git
On Thu, Jul 06, 2017 at 07:51:25PM +0900, Michel Dänzer wrote:
> From: Michel Dänzer
>
> This is a rearrangement of some patches of John's, with some improvements
> of mine, and a patch of mine, rebased on current amd-staging-4.11.
> Details about changes in the
Change-Id: I9f16ec9f5bab9abe83e9df01609893b38bb691fe
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h| 6 ++
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 1 +
2 files changed, 7 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
Change-Id: I9b925397d7ac61bcf9f0e8850f2d8b3dabd4a105
Signed-off-by: Evan Quan
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 24
1 file changed, 24 insertions(+)
diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c
b/drivers/gpu/drm/amd/amdgpu/soc15.c
index
Thanks for update.
That's fine for me.
Feel free to add my RB.
Reviewed-by: Junwei Zhang
On 07/04/2017 04:20 AM, Alex Deucher wrote:
Rather than checking the CONGIG_MEMSIZE register as that may
not be reliable on some APUs.
v2: The scratch register is only used on CIK+
62 matches
Mail list logo