As Christian and David's suggestion, submit the test ib ring debug interfaces.
It's useful for debugging with the command submission without VM case.
Signed-off-by: Huang Rui
---
V1 -> V2:
- park the scheduler thread for each ring to avoid conflict with commands from
active
On Wed, May 10, 2017 at 10:56 PM, Huang Rui wrote:
> On Thu, May 11, 2017 at 10:41:42AM +0800, Deucher, Alexander wrote:
>> > -Original Message-
>> > From: Deucher, Alexander
>> > Sent: Wednesday, May 10, 2017 10:38 PM
>> > To: Huang, Ray; amd-gfx@lists.freedesktop.org;
On 11/05/17 06:10 AM, Li, Samuel wrote:
> Also attach a sample ids file for reference. The names are from marketing, not
> related to source code and no reviews necessary here:)
Just because it's not source code doesn't mean no review is necessary. :)
> It can be put in directory
From: Xiaojie Yuan
Change-Id: I12216da14910f5e2b0970bc1fafc2a20b0ef1ba9
Signed-off-by: Samuel Li
---
amdgpu/Makefile.am | 2 +
amdgpu/Makefile.sources | 2 +-
amdgpu/amdgpu_asic_id.c | 198 +++
On Wed, May 10, 2017 at 5:20 AM, S, Shirish wrote:
> From: Shirish S
>
> amdgpu_device_resume() has a high time consuming
> call of amdgpu_late_init() which sets the clock_gating
> state of all IP blocks and is blocking.
> This patch defers only this setting
Hi,
I just tested v4.11-12441-g56868a4 on HP xw6600 with radeon graphics,
and I'm seeing the following WARNING triggered constantly.
I have not seen this earlier e.g. with the distro kernel 4.10.13-200.fc25.x86_64
$ lspci|grep -i amd
60:00.0 VGA compatible controller: Advanced Micro Devices,
> -Original Message-
> From: Christian König [mailto:deathsim...@vodafone.de]
> Sent: Wednesday, May 10, 2017 3:29 PM
> To: Alex Deucher; amd-gfx@lists.freedesktop.org
> Cc: Deucher, Alexander
> Subject: Re: [PATCH 000/117] Raven Support
>
> Am 10.05.2017 um 20:45 schrieb Alex Deucher:
>
[HK]: From Compute perspective, all validations are synchronous. PT/PD BOs are
validated when restored after evictions. However, the clearing of PT/PD BOs
during creation are still done by GPU which mandates the waiting.
[HK]: I am not clear on how to modify amdgpu_bo_kmap(). The function
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Christian König
> Sent: Wednesday, May 10, 2017 2:09 PM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH] drm/amdgpu: fix fundamental suspend/resume issue
>
> From: Christian König
From: Chunming Zhou
Add the RAVEN pci id.
Signed-off-by: Chunming Zhou
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c | 3 +++
1 file changed, 3
From: Chunming Zhou
Add the ip block and enable powerplay on raven.
Reviewed-by: Hawking Zhang
Signed-off-by: Chunming Zhou
Signed-off-by: Alex Deucher
---
From: Andrey Grodzovsky
v2: squash in some updates (Alex)
Signed-off-by: Andrey Grodzovsky
Acked-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Leo Liu
Add the ring function callbacks for the encode rings.
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Rex Zhu
smumgr provides the interface for interacting with the
smu firmware which handles power management.
v2: squash in updates (Alex)
Signed-off-by: Rex Zhu
Reviewed-by: Alex Deucher
Reviewed-by: Hawking Zhang
From: Leo Liu
Based on new vcn firmware interface changes
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 15 +--
1
From: Rex Zhu
Signed-off-by: Rex Zhu
Reviewed-by: Alex Deucher
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/include/pptable.h | 57
From: Huang Rui
Add sw controlled clockgating for VCN.
Signed-off-by: Huang Rui
Reviewed-by: Hawking Zhang
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Leo Liu
Signed-off-by: Leo Liu
Acked-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 40 -
1 file changed, 24
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
include/uapi/drm/amdgpu_drm.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
From: Leo Liu
Hope it will be generic for vcn later
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 36
From: Leo Liu
Add the ring function callbacks for the decode ring.
Signed-off-by: Leo Liu
Acked-by: Chunming Zhou
Acked-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vcn.c | 5 +
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
include/uapi/drm/amdgpu_drm.h | 3 ++-
1 file
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 21 -
1 file changed, 20 insertions(+), 1
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 34 --
1 file changed, 16
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 28
From: Leo Liu
Wire up the callback and enable them.
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Christian König
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 26 +-
From: Leo Liu
Signed-off-by: Leo Liu
Acked-by: Chunming Zhou
Acked-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 32
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.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_vcn.c | 140
1 file changed, 70
From: Leo Liu
Add encode ring and ib tests.
Signed-off-by: Leo Liu
Acked-by: Alex Deucher
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
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 | 3 +++
1 file changed, 3 insertions(+)
diff --git
From: Leo Liu
Not required on raven.
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c | 6 --
1 file changed, 4 insertions(+),
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/vcn_v1_0.c | 17 -
1 file changed, 16 insertions(+), 1 deletion(-)
From: Leo Liu
Fill in the core VCN 1.0 setup functionality.
v2: squash in fixup (Alex)
Signed-off-by: Leo Liu
Acked-by: Chunming Zhou
Acked-by: Hawking Zhang
Signed-off-by: Alex Deucher
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 4 +++-
1
From: Hawking Zhang
Used to enable/disable cp powergating.
Signed-off-by: Hawking Zhang
Reviewed-by: Alex Deucher
Reviewed-by: Huang Rui
Signed-off-by: Alex Deucher
---
From: Huang Rui
Add the psp 10.0 callbacks for PSP.
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Acked-by: Chunming Zhou
Signed-off-by: Alex Deucher
---
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Hawking Zhang
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/soc15.c | 2 +-
1 file
From: Hawking Zhang
Wire up the functions to control medium grained
powergating.
Signed-off-by: Hawking Zhang
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Chunming Zhou
Initialize the values earlier.
Signed-off-by: Chunming Zhou
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c | 11 +--
1
From: Chunming Zhou
Fetch the correct ucode for raven.
Signed-off-by: Chunming Zhou
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c | 3 +++
1 file
From: Huang Rui
Initialize sdma for powergating.
Signed-off-by: Huang Rui
Reviewed-by: Hawking Zhang
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
From: Huang Rui
Signed-off-by: Huang Rui
Reviewed-by: Alex Deucher
Reviewed-by: Hawking Zhang
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/amd/amdgpu/mmhub_v1_0.c | 50
From: Maruthi Srinivas Bayyavarapu
When audio usecase in progress and system wide suspend happens, ACP will
be powered off and when system resumes, for audio usecase to continue,
all the runtime configuration data needs to be programmed again. Added
'resume'pm call
From: Maruthi Srinivas Bayyavarapu
whenever audio data equal to I2S fifo watermark level is
produced/consumed, interrupt is generated. Amount of data is equal to
half of ALSA ring buffer size. Acknowledge the interrupt.
Signed-off-by: Maruthi Bayyavarapu
From: Maruthi Srinivas Bayyavarapu
ACP3x has a DMA controller to access system memory. This controller
transfers data from/to system memory to/from ACP internal fifo.
The patch adds PCM driver DMA operations.
Signed-off-by: Maruthi Bayyavarapu
From: Maruthi Srinivas Bayyavarapu
PCM platform driver binds to the platform device creatd by ACP3x PCI
device. PCM driver registers ALSA DMA and CPU DAI components with ASoC
framework.
Signed-off-by: Maruthi Bayyavarapu
Signed-off-by:
From: Maruthi Srinivas Bayyavarapu
ACP 3.0 is a PCI audio device. This patch adds PCI driver to bind
to this device and get PCI resources.
Signed-off-by: Maruthi Bayyavarapu
Signed-off-by: Alex Deucher
---
From: Christian König
Reinitializing the VM manager during suspend/resume is a very very bad
idea since all the VMs are still active and kicking.
This can lead to random VM faults after resume when new processes
become the same client ID assigned.
Signed-off-by:
From: Hawking Zhang
Signed-off-by: Hawking Zhang
Signed-off-by: Chunming Zhou
Reviewed-by: Junwei Zhang
Signed-off-by: Alex Deucher
---
include/drm/amdgpu_drm.h | 1 +
1 file
From: Leo Liu
Signed-off-by: Leo Liu
Reviewed-by: Alex Deucher
Signed-off-by: Alex Deucher
---
include/drm/amdgpu_drm.h | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Wednesday, May 10, 2017 5:14 AM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH xf86-video-amdgpu] Remove unused struct members from
> drmmode_display.h
>
> From:
> -Original Message-
> From: amd-gfx [mailto:amd-gfx-boun...@lists.freedesktop.org] On Behalf
> Of Michel Dänzer
> Sent: Wednesday, May 10, 2017 4:25 AM
> To: amd-gfx@lists.freedesktop.org
> Subject: [PATCH xf86-video-ati] Apply gamma correction to HW cursor
>
> From: Michel Dänzer
---
src/lib/discover_by_did.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/lib/discover_by_did.c b/src/lib/discover_by_did.c
index 2e06d5092847..7ecd7f353c92 100644
--- a/src/lib/discover_by_did.c
+++ b/src/lib/discover_by_did.c
@@ -221,7 +221,9 @@ static const struct {
{
Move the registration later to avoid a possible race
between driver initialization and a vga_switcheroo
request.
bug: https://bugzilla.redhat.com/show_bug.cgi?id=1411034
Signed-off-by: Alex Deucher
---
drivers/gpu/drm/radeon/radeon.h| 2 ++
>Because stopping all the scheduler threads takes a moment and it is entirely
>possible that the job finishes within that time.
Sounds reasonable, but instead not, because if so why not increase timed_out
value to + 0.5 sec ? that should cover the time on stop all schedulers,
So the point is we
Ping, could anybody take a look at this set?
Thanks,
Christian.
Am 27.04.2017 um 18:23 schrieb Christian König:
From: Christian König
Extern is the default for function declerations anyway and this
solves a bunch of 80char per line issues.
Signed-off-by: Christian
Am 10.05.2017 um 02:23 schrieb Michel Dänzer:
On 03/05/17 09:46 PM, Christian König wrote:
Am 02.05.2017 um 22:04 schrieb SF Markus Elfring:
From: Markus Elfring
Date: Tue, 2 May 2017 22:00:02 +0200
Three update suggestions were taken into account
from static
> Checking a second time is pointless since it can't signal any more after
> calling amd_sched_hw_job_reset().
[ML] you seems not response to me … I of cause know fence cannot signal after
hw_job_reset() ….
My question is , before you call hw_job_reset(), why you want to check the
fence ? why
Am 10.05.2017 um 12:05 schrieb Liu, Monk:
[ML] yes, but we cannot guarantee the job is 100% really hang when
entering gpu_reset(), we can only trust our amdgpu_job_timeout as a
deadline for each job.
You approach that check the fence first before charge it as
guilty/hang is incorrect looks
From: Shirish S
amdgpu_device_resume() has a high time consuming
call of amdgpu_late_init() which sets the clock_gating
state of all IP blocks and is blocking.
This patch defers only this setting of clock gating state
operation to post resume of amdgpu driver but ideally
I mean letting it signal when we force the hardware fence to complete will work
as well, but I still think that this isn't as clean as signaling it manually.
Please also see the helper function the Intel guys introduced
drm_fence_set_error(), we will run into a BUG_ON if we can't guarantee the
Am 10.05.2017 um 11:42 schrieb Chunming Zhou:
this is an improvement for previous patch, the sched_sync is to store fence
that could be skipped as scheduled, when job is executed, we didn't need
pipeline_sync if all fences in sched_sync are signalled, otherwise insert
pipeline_sync still.
v2:
[ML] yes, but we cannot guarantee the job is 100% really hang when entering
gpu_reset(), we can only trust our amdgpu_job_timeout as a deadline for each
job.
You approach that check the fence first before charge it as guilty/hang is
incorrect looks to me because why you not check it twice,
From: Michel Dänzer
Can't work currently. Fixes crash when trying to run a DRI3 client when
glamor isn't enabled.
Bugzilla: https://bugs.freedesktop.org/100968
Signed-off-by: Michel Dänzer
---
src/amdgpu_kms.c | 5 -
1 file changed, 4
before that, we have function to check if reset happens by using reset count.
v2: always update reset count after vm flush
Change-Id: I2e941dd35295d4210d57a9593d39b5ee9021be9f
Signed-off-by: Chunming Zhou
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 3 +--
1 file changed, 1
From: Michel Dänzer
Signed-off-by: Michel Dänzer
---
src/drmmode_display.c | 1 -
src/drmmode_display.h | 3 ---
2 files changed, 4 deletions(-)
diff --git a/src/drmmode_display.c b/src/drmmode_display.c
index 5b2431495..9996d2f70 100644
---
From: Michel Dänzer
Remember the shared pixmap passed to drmmode_set_scanout_pixmap for each
CRTC, and just compare against that.
Fixes leaving stale entries in ScreenRec::pixmap_dirty_list under some
circumstances, which would usually result in use-after-free and a
From: Michel Dänzer
References are held by the pixmaps corresponding to the FBs (so
the same KMS FB can be reused as long as the pixmap exists) and by the
CRTCs scanning out from them (so a KMS FB is only destroyed once it's
not being scanned out anymore, preventing
From: Michel Dänzer
This brings us in line with amdgpu and prepares for the following
change, no functional change intended.
(Ported from amdgpu commit e463b849f3e9d7b69e64a65619a22e00e78d297b)
v2:
* Be more consistent with the amdgpu code, which should make porting
On 2017年05月10日 16:50, Christian König wrote:
Am 10.05.2017 um 10:38 schrieb zhoucm1:
On 2017年05月10日 16:26, Christian König wrote:
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
this is an improvement for previous patch, the sched_sync is to
store fence
that could be skipped as scheduled,
Am 10.05.2017 um 10:38 schrieb zhoucm1:
On 2017年05月10日 16:26, Christian König wrote:
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
this is an improvement for previous patch, the sched_sync is to
store fence
that could be skipped as scheduled, when job is executed, we didn't
need
On 2017年05月10日 16:26, Christian König wrote:
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
this is an improvement for previous patch, the sched_sync is to store
fence
that could be skipped as scheduled, when job is executed, we didn't need
pipeline_sync if all fences in sched_sync are
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
Change-Id: I0ccfa0e6de0cddbcca8dd85f2862240bc5ca02b3
Signed-off-by: Chunming Zhou
---
drivers/gpu/drm/amd/amdgpu/amdgpu.h | 1 -
drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c | 6 +++---
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
this is an improvement for previous patch, the sched_sync is to store fence
that could be skipped as scheduled, when job is executed, we didn't need
pipeline_sync if all fences in sched_sync are signalled, otherwise insert
pipeline_sync still.
From: Michel Dänzer
The display hardware CLUT we're currently using for gamma correction
doesn't affect the HW cursor, so we have to apply it manually when
uploading the HW cursor data.
This currently only works in depth 24/32.
(Ported from amdgpu commit
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
before that, we have function to check if reset happens by using reset count.
Change-Id: I2e941dd35295d4210d57a9593d39b5ee9021be9f
Signed-off-by: Chunming Zhou
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 ++--
1 file
NAK, it is intentionally done like this to avoid starvation of processes.
In other words we assign a VMID only when there is at least one free.
Regards,
Christian.
Am 10.05.2017 um 09:31 schrieb Chunming Zhou:
Change-Id: If24a62b9c3097c9b040225ab0e768145b7a3db1e
Signed-off-by: Chunming Zhou
On 05/10/2017 03:31 PM, Chunming Zhou wrote:
Change-Id: If24a62b9c3097c9b040225ab0e768145b7a3db1e
Signed-off-by: Chunming Zhou
I had same idea when read this code before.
Reviewed-by: Junwei Zhang
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 92
Change-Id: If24a62b9c3097c9b040225ab0e768145b7a3db1e
Signed-off-by: Chunming Zhou
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 92 +-
1 file changed, 47 insertions(+), 45 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c
before that, we have function to check if reset happens by using reset count.
Change-Id: I2e941dd35295d4210d57a9593d39b5ee9021be9f
Signed-off-by: Chunming Zhou
---
drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
83 matches
Mail list logo