RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-25 Thread He, Jacob 
[AMD Official Use Only - Internal Distribution Only]

Yes, it’s SPM_VMID. There is VMFault with 0 if KMD doesn’t update SPM_VMID with 
current job since the default value of SPM_VMID is 0.

Thanks
Jacob

From: Koenig, Christian<mailto:christian.koe...@amd.com>
Sent: Tuesday, February 25, 2020 6:56 PM
To: He, Jacob<mailto:jacob...@amd.com>; Zhou, 
David(ChunMing)<mailto:david1.z...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Hi Jacob,

well first of all we can't rely on the correct behavior of Vulkan. In the 
kernel you have to always assume that userspace is not doing the right thing.


If there is any other UMD applications, such as OCL, and OCL driver doesn’t 
update the SPM_VMID before enable SPM, there is VMFault anyway even Vulkan 
driver restore the SPM_VMID to 0.

Question is where with which VMID is this VM fault triggered? The SPM VMID?

If yes than that would be quite problematic if an application could trigger a 
VM fault for VMID 0.

Regards,
Christian.


Am 25.02.20 um 08:20 schrieb He, Jacob:

[AMD Official Use Only - Internal Distribution Only]

Vulkan UMD makes sure that the SPM_VMID will be updated if the application 
enables the SPM. So there is no this kind of problem for Vulkan application. 
It’s kind of “atomic” operation to enable SPM and update SPM_VMID. Vulkan 
application even is not aware of SPM_VMID.
If there is any other UMD applications, such as OCL, and OCL driver doesn’t 
update the SPM_VMID before enable SPM, there is VMFault anyway even Vulkan 
driver restore the SPM_VMID to 0.

Thanks
Jacob

From: Koenig, Christian<mailto:christian.koe...@amd.com>
Sent: Tuesday, February 25, 2020 2:51 AM
To: He, Jacob<mailto:jacob...@amd.com>; Zhou, 
David(ChunMing)<mailto:david1.z...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Leave the old spm_vmid there is not a problem because other new spm enabled job 
will update it before it’s running and other spm disabled job will not access 
spm_vmid register.
Do you have a high level description of how SPM works? For example would it be 
possible to cause a problem in another application if we leave the SPM VMID 
here and the VMID is reused?

Keep in mind that we need to guarantee process isolation, e.g. we can't allow 
anything bad to happen to another application if somebody is stupid enough to 
turn on SPM tracing without setting it up properly.

Regards,
Christian.

Am 24.02.20 um 09:06 schrieb He, Jacob:
Ok, I’m glad that there is no uapi change needed.
I checked the git log, and reserve_vmid was added for shader debugger, not gpa. 
I’m fine to re-use it since the spm will not enabled for shader debug in 
general. I’ll try to change my patch.
BTW, “ring->funcs->setup_spm(ring, NULL)” to clear the vmid is not gonna work 
since the job with spm enabled execution is not done yet. Leave the old 
spm_vmid there is not a problem because other new spm enabled job will update 
it before it’s running and other spm disabled job will not access spm_vmid 
register.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Saturday, February 22, 2020 12:45 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace


[AMD Official Use Only - Internal Distribution Only]

That’s fine to me.

-David

From: Koenig, Christian 
<mailto:christian.koe...@amd.com>
Sent: Friday, February 21, 2020 11:33 PM
To: Deucher, Alexander 
<mailto:alexander.deuc...@amd.com>; Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>; 
Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; He, 
Jacob <mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just use that 
internally in the kernel to update the rlc register via the ring when we 
schedule the relevant IB?

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-25 Thread Christian König 

Hi Jacob,

well first of all we can't rely on the correct behavior of Vulkan. In 
the kernel you have to always assume that userspace is not doing the 
right thing.


If there is any other UMD applications, such as OCL, and OCL driver 
doesn’t update the SPM_VMID before enable SPM, there is VMFault anyway 
even Vulkan driver restore the SPM_VMID to 0.


Question is where with which VMID is this VM fault triggered? The SPM VMID?

If yes than that would be quite problematic if an application could 
trigger a VM fault for VMID 0.


Regards,
Christian.


Am 25.02.20 um 08:20 schrieb He, Jacob:


[AMD Official Use Only - Internal Distribution Only]


Vulkan UMD makes sure that the SPM_VMID will be updated if the 
application enables the SPM. So there is no this kind of problem for 
Vulkan application. It’s kind of “atomic” operation to enable SPM and 
update SPM_VMID. Vulkan application even is not aware of SPM_VMID.


If there is any other UMD applications, such as OCL, and OCL driver 
doesn’t update the SPM_VMID before enable SPM, there is VMFault anyway 
even Vulkan driver restore the SPM_VMID to 0.


Thanks

Jacob

*From: *Koenig, Christian <mailto:christian.koe...@amd.com>
*Sent: *Tuesday, February 25, 2020 2:51 AM
*To: *He, Jacob <mailto:jacob...@amd.com>; Zhou, David(ChunMing) 
<mailto:david1.z...@amd.com>; Deucher, Alexander 
<mailto:alexander.deuc...@amd.com>; Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

*Subject: *Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Leave the old spm_vmid there is not a problem because other new spm 
enabled job will update it before it’s running and other spm disabled 
job will not access spm_vmid register.


Do you have a high level description of how SPM works? For example 
would it be possible to cause a problem in another application if we 
leave the SPM VMID here and the VMID is reused?


Keep in mind that we need to guarantee process isolation, e.g. we 
can't allow anything bad to happen to another application if somebody 
is stupid enough to turn on SPM tracing without setting it up properly.


Regards,
Christian.

Am 24.02.20 um 09:06 schrieb He, Jacob:

Ok, I’m glad that there is no uapi change needed.

I checked the git log, and reserve_vmid was added for shader
debugger, not gpa. I’m fine to re-use it since the spm will not
enabled for shader debug in general. I’ll try to change my patch.

BTW, “ring->funcs->setup_spm(ring, NULL)” to clear the vmid is not
gonna work since the job with spm enabled execution is not done
yet. Leave the old spm_vmid there is not a problem because other
new spm enabled job will update it before it’s running and other
spm disabled job will not access spm_vmid register.

Thanks

Jacob

*From: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>
*Sent: *Saturday, February 22, 2020 12:45 AM
*To: *Koenig, Christian <mailto:christian.koe...@amd.com>;
Deucher, Alexander <mailto:alexander.deuc...@amd.com>; Christian
König <mailto:ckoenig.leichtzumer...@gmail.com>; He, Jacob
<mailto:jacob...@amd.com>; amd-gfx@lists.freedesktop.org
    <mailto:amd-gfx@lists.freedesktop.org>
*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

That’s fine to me.

-David

*From:*Koenig, Christian 
<mailto:christian.koe...@amd.com>
*Sent:* Friday, February 21, 2020 11:33 PM
*To:* Deucher, Alexander 
<mailto:alexander.deuc...@amd.com>; Christian König

<mailto:ckoenig.leichtzumer...@gmail.com>; Zhou, David(ChunMing)
 <mailto:david1.z...@amd.com>; He, Jacob
 <mailto:jacob...@amd.com>;
    amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>
*Subject:* Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes
enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we
just use that internally in the kernel to update the rlc
register via the ring when we schedule the relevant IB?  E.g.,
add a new ring callback to set SPM state and then set it to
the reserved vmid before we schedule the ib, and then reset it
to 0 after the IB in amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

index 4b2342d11520..e0db9362c6ee 100644

--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

+++ b/drivers/gpu/drm/amd

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-24 Thread He, Jacob 
[AMD Official Use Only - Internal Distribution Only]

Vulkan UMD makes sure that the SPM_VMID will be updated if the application 
enables the SPM. So there is no this kind of problem for Vulkan application. 
It’s kind of “atomic” operation to enable SPM and update SPM_VMID. Vulkan 
application even is not aware of SPM_VMID.
If there is any other UMD applications, such as OCL, and OCL driver doesn’t 
update the SPM_VMID before enable SPM, there is VMFault anyway even Vulkan 
driver restore the SPM_VMID to 0.

Thanks
Jacob

From: Koenig, Christian<mailto:christian.koe...@amd.com>
Sent: Tuesday, February 25, 2020 2:51 AM
To: He, Jacob<mailto:jacob...@amd.com>; Zhou, 
David(ChunMing)<mailto:david1.z...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Leave the old spm_vmid there is not a problem because other new spm enabled job 
will update it before it’s running and other spm disabled job will not access 
spm_vmid register.
Do you have a high level description of how SPM works? For example would it be 
possible to cause a problem in another application if we leave the SPM VMID 
here and the VMID is reused?

Keep in mind that we need to guarantee process isolation, e.g. we can't allow 
anything bad to happen to another application if somebody is stupid enough to 
turn on SPM tracing without setting it up properly.

Regards,
Christian.

Am 24.02.20 um 09:06 schrieb He, Jacob:
Ok, I’m glad that there is no uapi change needed.
I checked the git log, and reserve_vmid was added for shader debugger, not gpa. 
I’m fine to re-use it since the spm will not enabled for shader debug in 
general. I’ll try to change my patch.
BTW, “ring->funcs->setup_spm(ring, NULL)” to clear the vmid is not gonna work 
since the job with spm enabled execution is not done yet. Leave the old 
spm_vmid there is not a problem because other new spm enabled job will update 
it before it’s running and other spm disabled job will not access spm_vmid 
register.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Saturday, February 22, 2020 12:45 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace


[AMD Official Use Only - Internal Distribution Only]

That’s fine to me.

-David

From: Koenig, Christian 
<mailto:christian.koe...@amd.com>
Sent: Friday, February 21, 2020 11:33 PM
To: Deucher, Alexander 
<mailto:alexander.deuc...@amd.com>; Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>; 
Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; He, 
Jacob <mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just use that 
internally in the kernel to update the rlc register via the ring when we 
schedule the relevant IB?  E.g., add a new ring callback to set SPM state and 
then set it to the reserved vmid before we schedule the ib, and then reset it 
to 0 after the IB in amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 4b2342d11520..e0db9362c6ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
if (ring->funcs->insert_start)
ring->funcs->insert_start(ring);

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, job);
+
if (job) {
r = amdgpu_vm_flush(ring, job, need_pipe_sync);
if (r) {
@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
return r;
}

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, NULL);
+
if (ring->funcs->insert_end)
ring->funcs->insert_end(ring);



Alex

From: amd-gfx 
<mailto:amd-gfx-boun...@lists.freedesktop.org>
 on behalf of Christian König 
<mailto:ckoenig.leicht

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-24 Thread Christian König 
Leave the old spm_vmid there is not a problem because other new spm 
enabled job will update it before it’s running and other spm disabled 
job will not access spm_vmid register.
Do you have a high level description of how SPM works? For example would 
it be possible to cause a problem in another application if we leave the 
SPM VMID here and the VMID is reused?


Keep in mind that we need to guarantee process isolation, e.g. we can't 
allow anything bad to happen to another application if somebody is 
stupid enough to turn on SPM tracing without setting it up properly.


Regards,
Christian.

Am 24.02.20 um 09:06 schrieb He, Jacob:


Ok, I’m glad that there is no uapi change needed.

I checked the git log, and reserve_vmid was added for shader debugger, 
not gpa. I’m fine to re-use it since the spm will not enabled for 
shader debug in general. I’ll try to change my patch.


BTW, “ring->funcs->setup_spm(ring, NULL)” to clear the vmid is not 
gonna work since the job with spm enabled execution is not done yet. 
Leave the old spm_vmid there is not a problem because other new spm 
enabled job will update it before it’s running and other spm disabled 
job will not access spm_vmid register.


Thanks

Jacob

*From: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>
*Sent: *Saturday, February 22, 2020 12:45 AM
*To: *Koenig, Christian <mailto:christian.koe...@amd.com>; Deucher, 
Alexander <mailto:alexander.deuc...@amd.com>; Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>; He, Jacob 
<mailto:jacob...@amd.com>; amd-gfx@lists.freedesktop.org 
<mailto:amd-gfx@lists.freedesktop.org>

*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

That’s fine to me.

-David

*From:* Koenig, Christian 
*Sent:* Friday, February 21, 2020 11:33 PM
*To:* Deucher, Alexander ; Christian König 
; Zhou, David(ChunMing) 
; He, Jacob ; 
amd-gfx@lists.freedesktop.org

*Subject:* Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable 
SPM.


Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just
use that internally in the kernel to update the rlc register via
the ring when we schedule the relevant IB?  E.g., add a new ring
callback to set SPM state and then set it to the reserved vmid
before we schedule the ib, and then reset it to 0 after the IB in
amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

index 4b2342d11520..e0db9362c6ee 100644

--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring
*ring, unsigned num_ibs,

        if (ring->funcs->insert_start)

ring->funcs->insert_start(ring);

+       if (ring->funcs->setup_spm)

+ ring->funcs->setup_spm(ring, job);

+

        if (job) {

r = amdgpu_vm_flush(ring, job, need_pipe_sync);

if (r) {

@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring
*ring, unsigned num_ibs,

return r;

        }

+       if (ring->funcs->setup_spm)

+ ring->funcs->setup_spm(ring, NULL);

+

        if (ring->funcs->insert_end)

ring->funcs->insert_end(ring);

Alex

*From:*amd-gfx 
<mailto:amd-gfx-boun...@lists.freedesktop.org> on behalf of
Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>
*Sent:* Friday, February 21, 2020 5:28 AM
*To:* Zhou, David(ChunMing) 
<mailto:david1.z...@amd.com>; He, Jacob 
<mailto:jacob...@amd.com>; Koenig, Christian
 <mailto:christian.koe...@amd.com>;
amd-gfx@lists.freedesktop.org
<mailto:amd-gfx@lists.freedesktop.org>
 <mailto:amd-gfx@lists.freedesktop.org>
*Subject:* Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

That would probably be a no-go, but we could enhance the kernel
driver to update the RLC_SPM_VMID register with the reserved VMID.

Handling that in userspace is most likely not working anyway,
since the RLC registers are usually not accessible by userspace.

Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):

[AMD Official Use Only - Internal Distribution Only]

You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid
to userspace.

-David

*From:* He, Jacob  <mailto:jacob...@amd.com>
*Sent:* Thursday, February 20, 2020 10:46 PM
*To:* Zhou, David(ChunMing) 
<mailto:davi

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-24 Thread He, Jacob 
Ok, I’m glad that there is no uapi change needed.
I checked the git log, and reserve_vmid was added for shader debugger, not gpa. 
I’m fine to re-use it since the spm will not enabled for shader debug in 
general. I’ll try to change my patch.
BTW, “ring->funcs->setup_spm(ring, NULL)” to clear the vmid is not gonna work 
since the job with spm enabled execution is not done yet. Leave the old 
spm_vmid there is not a problem because other new spm enabled job will update 
it before it’s running and other spm disabled job will not access spm_vmid 
register.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Saturday, February 22, 2020 12:45 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; Deucher, 
Alexander<mailto:alexander.deuc...@amd.com>; Christian 
König<mailto:ckoenig.leichtzumer...@gmail.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace


[AMD Official Use Only - Internal Distribution Only]

That’s fine to me.

-David

From: Koenig, Christian 
Sent: Friday, February 21, 2020 11:33 PM
To: Deucher, Alexander ; Christian König 
; Zhou, David(ChunMing) 
; He, Jacob ; 
amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just use that 
internally in the kernel to update the rlc register via the ring when we 
schedule the relevant IB?  E.g., add a new ring callback to set SPM state and 
then set it to the reserved vmid before we schedule the ib, and then reset it 
to 0 after the IB in amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 4b2342d11520..e0db9362c6ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
if (ring->funcs->insert_start)
ring->funcs->insert_start(ring);

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, job);
+
if (job) {
r = amdgpu_vm_flush(ring, job, need_pipe_sync);
if (r) {
@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
return r;
}

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, NULL);
+
if (ring->funcs->insert_end)
ring->funcs->insert_end(ring);



Alex

From: amd-gfx 
<mailto:amd-gfx-boun...@lists.freedesktop.org>
 on behalf of Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>
Sent: Friday, February 21, 2020 5:28 AM
To: Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; 
He, Jacob <mailto:jacob...@amd.com>; Koenig, Christian 
<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org> 
<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

That would probably be a no-go, but we could enhance the kernel driver to 
update the RLC_SPM_VMID register with the reserved VMID.

Handling that in userspace is most likely not working anyway, since the RLC 
registers are usually not accessible by userspace.

Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):

[AMD Official Use Only - Internal Distribution Only]



You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to userspace.



-David





From: He, Jacob <mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 10:46 PM
To: Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; 
Koenig, Christian <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



amdgpu_vm_reserve_vmid doesn’t return the reserved vmid back to user space. 
There is no chance for user mode driver to update RLC_SPM_VMID.



Thanks

Jacob



From: He, Jacob<mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 6:20 PM
To: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>; Koenig, 
Christian<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.



Th

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-21 Thread Zhou, David(ChunMing) 
[AMD Official Use Only - Internal Distribution Only]

That's fine to me.

-David

From: Koenig, Christian 
Sent: Friday, February 21, 2020 11:33 PM
To: Deucher, Alexander ; Christian König 
; Zhou, David(ChunMing) 
; He, Jacob ; 
amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:

[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just use that 
internally in the kernel to update the rlc register via the ring when we 
schedule the relevant IB?  E.g., add a new ring callback to set SPM state and 
then set it to the reserved vmid before we schedule the ib, and then reset it 
to 0 after the IB in amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 4b2342d11520..e0db9362c6ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
if (ring->funcs->insert_start)
ring->funcs->insert_start(ring);

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, job);
+
if (job) {
r = amdgpu_vm_flush(ring, job, need_pipe_sync);
if (r) {
@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
return r;
}

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, NULL);
+
if (ring->funcs->insert_end)
ring->funcs->insert_end(ring);



Alex

From: amd-gfx 
<mailto:amd-gfx-boun...@lists.freedesktop.org>
 on behalf of Christian König 
<mailto:ckoenig.leichtzumer...@gmail.com>
Sent: Friday, February 21, 2020 5:28 AM
To: Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; 
He, Jacob <mailto:jacob...@amd.com>; Koenig, Christian 
<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org> 
<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

That would probably be a no-go, but we could enhance the kernel driver to 
update the RLC_SPM_VMID register with the reserved VMID.

Handling that in userspace is most likely not working anyway, since the RLC 
registers are usually not accessible by userspace.

Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):

[AMD Official Use Only - Internal Distribution Only]



You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to userspace.



-David





From: He, Jacob <mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 10:46 PM
To: Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; 
Koenig, Christian <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



amdgpu_vm_reserve_vmid doesn't return the reserved vmid back to user space. 
There is no chance for user mode driver to update RLC_SPM_VMID.



Thanks

Jacob



From: He, Jacob<mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 6:20 PM
To: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>; Koenig, 
Christian<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.



Thanks

Jacob



From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Thursday, February 20, 2020 10:13 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx 
mailto:amd-gfx-boun...@lists.freedesktop.org>>
 On Behalf Of Christian König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob mailto:jacob...@amd.com>>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-21 Thread Christian König 

I would just do this as part of the vm_flush() callback on the ring.

E.g. check if the VMID you want to flush is reserved and if yes enable SPM.

Maybe pass along a flag or something in the job to make things easier.

Christian.

Am 21.02.20 um 16:31 schrieb Deucher, Alexander:


[AMD Public Use]


We already have the RESERVE_VMID ioctl interface, can't we just use 
that internally in the kernel to update the rlc register via the ring 
when we schedule the relevant IB?  E.g., add a new ring callback to 
set SPM state and then set it to the reserved vmid before we schedule 
the ib, and then reset it to 0 after the IB in amdgpu_ib_schedule().


diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c

index 4b2342d11520..e0db9362c6ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, 
unsigned num_ibs,

        if (ring->funcs->insert_start)
                ring->funcs->insert_start(ring);

+       if (ring->funcs->setup_spm)
+               ring->funcs->setup_spm(ring, job);
+
        if (job) {
                r = amdgpu_vm_flush(ring, job, need_pipe_sync);
                if (r) {
@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, 
unsigned num_ibs,

                return r;
        }

+       if (ring->funcs->setup_spm)
+               ring->funcs->setup_spm(ring, NULL);
+
        if (ring->funcs->insert_end)
                ring->funcs->insert_end(ring);



Alex

*From:* amd-gfx  on behalf of 
Christian König 

*Sent:* Friday, February 21, 2020 5:28 AM
*To:* Zhou, David(ChunMing) ; He, Jacob 
; Koenig, Christian ; 
amd-gfx@lists.freedesktop.org 

*Subject:* Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace
That would probably be a no-go, but we could enhance the kernel driver 
to update the RLC_SPM_VMID register with the reserved VMID.


Handling that in userspace is most likely not working anyway, since 
the RLC registers are usually not accessible by userspace.


Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):


[AMD Official Use Only - Internal Distribution Only]

You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to 
userspace.


-David

*From:* He, Jacob  <mailto:jacob...@amd.com>
*Sent:* Thursday, February 20, 2020 10:46 PM
*To:* Zhou, David(ChunMing)  
<mailto:david1.z...@amd.com>; Koenig, Christian 
 <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

*Subject:* RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

amdgpu_vm_reserve_vmid doesn’t return the reserved vmid back to user 
space. There is no chance for user mode driver to update RLC_SPM_VMID.


Thanks

Jacob

*From: *He, Jacob <mailto:jacob...@amd.com>
*Sent: *Thursday, February 20, 2020 6:20 PM
*To: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; Koenig, 
Christian <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Looks like amdgpu_vm_reserve_vmid could work, let me have a try to 
update the RLC_SPM_VMID with pm4 packets in UMD.


Thanks

Jacob

*From: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>
*Sent: *Thursday, February 20, 2020 10:13 AM
*To: *Koenig, Christian <mailto:christian.koe...@amd.com>; He, Jacob 
<mailto:jacob...@amd.com>; amd-gfx@lists.freedesktop.org 
<mailto:amd-gfx@lists.freedesktop.org>

*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply 
using VMID in kernel.
We already have an pair interface for RGP, I think you can use it 
instead of involving additional kernel change.

amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx <mailto:amd-gfx-boun...@lists.freedesktop.org>> On Behalf Of 
Christian König

Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob mailto:jacob...@amd.com>>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-21 Thread Deucher, Alexander 
[AMD Public Use]

We already have the RESERVE_VMID ioctl interface, can't we just use that 
internally in the kernel to update the rlc register via the ring when we 
schedule the relevant IB?  E.g., add a new ring callback to set SPM state and 
then set it to the reserved vmid before we schedule the ib, and then reset it 
to 0 after the IB in amdgpu_ib_schedule().

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index 4b2342d11520..e0db9362c6ee 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -185,6 +185,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
if (ring->funcs->insert_start)
ring->funcs->insert_start(ring);

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, job);
+
if (job) {
r = amdgpu_vm_flush(ring, job, need_pipe_sync);
if (r) {
@@ -273,6 +276,9 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
return r;
}

+   if (ring->funcs->setup_spm)
+   ring->funcs->setup_spm(ring, NULL);
+
if (ring->funcs->insert_end)
ring->funcs->insert_end(ring);



Alex

From: amd-gfx  on behalf of Christian 
König 
Sent: Friday, February 21, 2020 5:28 AM
To: Zhou, David(ChunMing) ; He, Jacob ; 
Koenig, Christian ; amd-gfx@lists.freedesktop.org 

Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

That would probably be a no-go, but we could enhance the kernel driver to 
update the RLC_SPM_VMID register with the reserved VMID.

Handling that in userspace is most likely not working anyway, since the RLC 
registers are usually not accessible by userspace.

Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):

[AMD Official Use Only - Internal Distribution Only]



You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to userspace.



-David





From: He, Jacob <mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 10:46 PM
To: Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; 
Koenig, Christian <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



amdgpu_vm_reserve_vmid doesn’t return the reserved vmid back to user space. 
There is no chance for user mode driver to update RLC_SPM_VMID.



Thanks

Jacob



From: He, Jacob<mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 6:20 PM
To: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>; Koenig, 
Christian<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.



Thanks

Jacob



From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Thursday, February 20, 2020 10:13 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace



[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx 
mailto:amd-gfx-boun...@lists.freedesktop.org>>
 On Behalf Of Christian König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob mailto:jacob...@amd.com>>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the commit 
message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously on the 
ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise we run 
into a bunch of trouble when the VMID is reused.

You also need to make sur

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-21 Thread Christian König 
That would probably be a no-go, but we could enhance the kernel driver 
to update the RLC_SPM_VMID register with the reserved VMID.


Handling that in userspace is most likely not working anyway, since the 
RLC registers are usually not accessible by userspace.


Regards,
Christian.

Am 20.02.20 um 16:15 schrieb Zhou, David(ChunMing):


[AMD Official Use Only - Internal Distribution Only]

You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to 
userspace.


-David

*From:* He, Jacob 
*Sent:* Thursday, February 20, 2020 10:46 PM
*To:* Zhou, David(ChunMing) ; Koenig, Christian 
; amd-gfx@lists.freedesktop.org

*Subject:* RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

amdgpu_vm_reserve_vmid doesn’t return the reserved vmid back to user 
space. There is no chance for user mode driver to update RLC_SPM_VMID.


Thanks

Jacob

*From: *He, Jacob <mailto:jacob...@amd.com>
*Sent: *Thursday, February 20, 2020 6:20 PM
*To: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>; Koenig, 
Christian <mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Looks like amdgpu_vm_reserve_vmid could work, let me have a try to 
update the RLC_SPM_VMID with pm4 packets in UMD.


Thanks

Jacob

*From: *Zhou, David(ChunMing) <mailto:david1.z...@amd.com>
*Sent: *Thursday, February 20, 2020 10:13 AM
*To: *Koenig, Christian <mailto:christian.koe...@amd.com>; He, Jacob 
<mailto:jacob...@amd.com>; amd-gfx@lists.freedesktop.org 
<mailto:amd-gfx@lists.freedesktop.org>

*Subject: *RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply 
using VMID in kernel.
We already have an pair interface for RGP, I think you can use it 
instead of involving additional kernel change.

amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx <mailto:amd-gfx-boun...@lists.freedesktop.org>> On Behalf Of Christian 
König

Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob mailto:jacob...@amd.com>>; 
amd-gfx@lists.freedesktop.org <mailto:amd-gfx@lists.freedesktop.org>

Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the 
commit message.


Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously 
on the ring. Otherwise we might corrupt an already executing SPM trace.


And you also need to make sure to disable the tracing again or 
otherwise we run into a bunch of trouble when the VMID is reused.


You also need to make sure that IBs using the SPM trace are serialized 
with each other, e.g. hack into amdgpu_ids.c file and make sure that 
only one VMID at a time can have that attribute.


Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He mailto:jacob...@amd.com>>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct 
amdgpu_cs_parser *p, union drm_amdgpu_cs

>    uint32_t uf_offset = 0;
>    int i;
>    int ret;
> + bool update_spm_vmid = false;
>
>    if (cs->in.num_chunks == 0)
>    return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct 
amdgpu_cs_parser *p, union drm_amdgpu_cs

>    case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>    break;
>
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-20 Thread Zhou, David(ChunMing) 
[AMD Official Use Only - Internal Distribution Only]

You can enhance amdgpu_vm_ioctl In amdgpu_vm.c to return vmid to userspace.

-David


From: He, Jacob 
Sent: Thursday, February 20, 2020 10:46 PM
To: Zhou, David(ChunMing) ; Koenig, Christian 
; amd-gfx@lists.freedesktop.org
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

amdgpu_vm_reserve_vmid doesn't return the reserved vmid back to user space. 
There is no chance for user mode driver to update RLC_SPM_VMID.

Thanks
Jacob

From: He, Jacob<mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 6:20 PM
To: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>; Koenig, 
Christian<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Thursday, February 20, 2020 10:13 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx 
mailto:amd-gfx-boun...@lists.freedesktop.org>>
 On Behalf Of Christian König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob mailto:jacob...@amd.com>>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the commit 
message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously on the 
ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise we run 
into a bunch of trouble when the VMID is reused.

You also need to make sure that IBs using the SPM trace are serialized with 
each other, e.g. hack into amdgpu_ids.c file and make sure that only one VMID 
at a time can have that attribute.

Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He mailto:jacob...@amd.com>>
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>uint32_t uf_offset = 0;
>int i;
>int ret;
> + bool update_spm_vmid = false;
>
>if (cs->in.num_chunks == 0)
>return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>break;
>
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid = true;
> + break;
> +
>default:
>ret = -EINVAL;
>goto free_partial_kdata;
> @@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>if (ret)
>goto free_all_kdata;
>
> + p->job->ne

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-20 Thread He, Jacob 
[AMD Official Use Only - Internal Distribution Only]

Hi Christian,
  This patch is just for a simple scenario. Multiple SPM processes is not 
supported. That is saying, it’s not necessary to take MCBP or other complicated 
cases into account. UMD waits GPU idle right before enabling SPM trace. 
Serialization is made by UMD.
  It’s not necessary to clear mmRLC_SPM_MC_CNTL also. For those spm disabled 
submission, mmRLC_SPM_MC_CNTL is useless. For those spm enabled submission, it 
will require KMD to update mmRLC_SPM_MC_CNTL.

Thanks
Jacob

From: Christian König<mailto:ckoenig.leichtzumer...@gmail.com>
Sent: Wednesday, February 19, 2020 7:02 PM
To: He, Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the
commit message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously
on the ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise
we run into a bunch of trouble when the VMID is reused.

You also need to make sure that IBs using the SPM trace are serialized
with each other, e.g. hack into amdgpu_ids.c file and make sure that
only one VMID at a time can have that attribute.

Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He 
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>uint32_t uf_offset = 0;
>int i;
>int ret;
> + bool update_spm_vmid = false;
>
>if (cs->in.num_chunks == 0)
>return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>break;
>
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid = true;
> + break;
> +
>default:
>ret = -EINVAL;
>goto free_partial_kdata;
> @@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>if (ret)
>goto free_all_kdata;
>
> + p->job->need_update_spm_vmid = update_spm_vmid;
> +
>if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
>ret = -ECANCELED;
>goto free_all_kdata;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> index cae81914c821..36faab12b585 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> @@ -156,9 +156,13 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, 
> unsigned num_ibs,
>return -EINVAL;
>}
>
> - if (vm && !job->vmid) {
> - dev_err(adev->dev, "VM IB without ID\n");
> - return -EINVAL;
> + if (vm) {
> + if (!job->vmid) {
> + dev_err(adev->dev, "VM IB without ID\n");
> + return -EINVAL;
> + } else if (adev->gfx.rlc.funcs->update_spm_vmid && 
> job->need_update_spm_vmid) {
> + adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid);
> +

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-20 Thread He, Jacob 
amdgpu_vm_reserve_vmid doesn’t return the reserved vmid back to user space. 
There is no chance for user mode driver to update RLC_SPM_VMID.

Thanks
Jacob

From: He, Jacob<mailto:jacob...@amd.com>
Sent: Thursday, February 20, 2020 6:20 PM
To: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>; Koenig, 
Christian<mailto:christian.koe...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Thursday, February 20, 2020 10:13 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx  On Behalf Of Christian 
König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob ; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the commit 
message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously on the 
ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise we run 
into a bunch of trouble when the VMID is reused.

You also need to make sure that IBs using the SPM trace are serialized with 
each other, e.g. hack into amdgpu_ids.c file and make sure that only one VMID 
at a time can have that attribute.

Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He 
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>uint32_t uf_offset = 0;
>int i;
>int ret;
> + bool update_spm_vmid = false;
>
>if (cs->in.num_chunks == 0)
>return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>break;
>
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid = true;
> + break;
> +
>default:
>ret = -EINVAL;
>goto free_partial_kdata;
> @@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>if (ret)
>goto free_all_kdata;
>
> + p->job->need_update_spm_vmid = update_spm_vmid;
> +
>if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
>ret = -ECANCELED;
>goto free_all_kdata;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> index cae81914c821..36faab12b585 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> @@ -156,9 +156,13 @@ int amdgpu_ib

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-20 Thread He, Jacob 
Looks like amdgpu_vm_reserve_vmid could work, let me have a try to update the 
RLC_SPM_VMID with pm4 packets in UMD.

Thanks
Jacob

From: Zhou, David(ChunMing)<mailto:david1.z...@amd.com>
Sent: Thursday, February 20, 2020 10:13 AM
To: Koenig, Christian<mailto:christian.koe...@amd.com>; He, 
Jacob<mailto:jacob...@amd.com>; 
amd-gfx@lists.freedesktop.org<mailto:amd-gfx@lists.freedesktop.org>
Subject: RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx  On Behalf Of Christian 
König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob ; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the commit 
message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously on the 
ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise we run 
into a bunch of trouble when the VMID is reused.

You also need to make sure that IBs using the SPM trace are serialized with 
each other, e.g. hack into amdgpu_ids.c file and make sure that only one VMID 
at a time can have that attribute.

Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He 
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>uint32_t uf_offset = 0;
>int i;
>int ret;
> + bool update_spm_vmid = false;
>
>if (cs->in.num_chunks == 0)
>return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>break;
>
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid = true;
> + break;
> +
>default:
>ret = -EINVAL;
>goto free_partial_kdata;
> @@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>if (ret)
>goto free_all_kdata;
>
> + p->job->need_update_spm_vmid = update_spm_vmid;
> +
>if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
>ret = -ECANCELED;
>goto free_all_kdata;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> index cae81914c821..36faab12b585 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> @@ -156,9 +156,13 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, 
> unsigned num_ibs,
>return -EINVAL;
>}
>
> - if (vm && !job->vmid) {
> - dev_err(adev->dev, "VM IB without ID\n");
> - return -EINVAL;
> + if (vm) {
> + if (!job->vmid) {
> + dev_err(adev->dev, "VM IB without ID\n");
> + return -EINVAL;
> + } else if (adev->gfx.rlc.f

RE: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-19 Thread Zhou, David(ChunMing) 
[AMD Official Use Only - Internal Distribution Only]

Christian is right here, that will cause many problems for simply using VMID in 
kernel.
We already have an pair interface for RGP, I think you can use it instead of 
involving additional kernel change.
amdgpu_vm_reserve_vmid/ amdgpu_vm_unreserve_vmid.

-David

-Original Message-
From: amd-gfx  On Behalf Of Christian 
König
Sent: Wednesday, February 19, 2020 7:03 PM
To: He, Jacob ; amd-gfx@lists.freedesktop.org
Subject: Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

Am 19.02.20 um 11:15 schrieb Jacob He:
> [WHY]
> When SPM trace enabled, SPM_VMID should be updated with the current 
> vmid.
>
> [HOW]
> Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us 
> which job should update SPM_VMID.
> Right before a job is submitted to GPU, set the SPM_VMID accordingly.
>
> [Limitation]
> Running more than one SPM trace enabled processes simultaneously is 
> not supported.

Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the commit 
message.

Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously on the 
ring. Otherwise we might corrupt an already executing SPM trace.

And you also need to make sure to disable the tracing again or otherwise we run 
into a bunch of trouble when the VMID is reused.

You also need to make sure that IBs using the SPM trace are serialized with 
each other, e.g. hack into amdgpu_ids.c file and make sure that only one VMID 
at a time can have that attribute.

Regards,
Christian.

>
> Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
> Signed-off-by: Jacob He 
> ---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
>   drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
>   drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
>   drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
>   drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
>   8 files changed, 48 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> index f9fa6e104fef..3f32c4db5232 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
> @@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>   uint32_t uf_offset = 0;
>   int i;
>   int ret;
> + bool update_spm_vmid = false;
>   
>   if (cs->in.num_chunks == 0)
>   return 0;
> @@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>   case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
>   break;
>   
> + case AMDGPU_CHUNK_ID_SPM_TRACE:
> + update_spm_vmid = true;
> + break;
> +
>   default:
>   ret = -EINVAL;
>   goto free_partial_kdata;
> @@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
> *p, union drm_amdgpu_cs
>   if (ret)
>   goto free_all_kdata;
>   
> + p->job->need_update_spm_vmid = update_spm_vmid;
> +
>   if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
>   ret = -ECANCELED;
>   goto free_all_kdata;
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> index cae81914c821..36faab12b585 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
> @@ -156,9 +156,13 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, 
> unsigned num_ibs,
>   return -EINVAL;
>   }
>   
> - if (vm && !job->vmid) {
> - dev_err(adev->dev, "VM IB without ID\n");
> - return -EINVAL;
> + if (vm) {
> + if (!job->vmid) {
> + dev_err(adev->dev, "VM IB without ID\n");
> + return -EINVAL;
> + } else if (adev->gfx.rlc.funcs->update_spm_vmid && 
> job->need_update_spm_vmid) {
> + adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid);
> + }
>   }
>   
>   alloc_size = ring->funcs->emit_frame_size + num_ibs * diff --git 
> a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h 
> b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
> index 2e2110dddb76..4582536961c7 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.

Re: [PATCH] drm/amdgpu: Add a chunk ID for spm trace

2020-02-19 Thread Christian König 

Am 19.02.20 um 11:15 schrieb Jacob He:

[WHY]
When SPM trace enabled, SPM_VMID should be updated with the current
vmid.

[HOW]
Add a chunk id, AMDGPU_CHUNK_ID_SPM_TRACE, so that UMD can tell us
which job should update SPM_VMID.
Right before a job is submitted to GPU, set the SPM_VMID accordingly.

[Limitation]
Running more than one SPM trace enabled processes simultaneously is
not supported.


Well there are multiple problems with that patch.

First of all you need to better describe what SPM tracing is in the 
commit message.


Then the updating of mmRLC_SPM_MC_CNTL must be executed asynchronously 
on the ring. Otherwise we might corrupt an already executing SPM trace.


And you also need to make sure to disable the tracing again or otherwise 
we run into a bunch of trouble when the VMID is reused.


You also need to make sure that IBs using the SPM trace are serialized 
with each other, e.g. hack into amdgpu_ids.c file and make sure that 
only one VMID at a time can have that attribute.


Regards,
Christian.



Change-Id: Ic932ef6ac9dbf244f03aaee90550e8ff3a675666
Signed-off-by: Jacob He 
---
  drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c  |  7 +++
  drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c  | 10 +++---
  drivers/gpu/drm/amd/amdgpu/amdgpu_job.h |  1 +
  drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h |  1 +
  drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c  | 15 ++-
  drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c   |  3 ++-
  drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c   |  3 ++-
  drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c   | 15 ++-
  8 files changed, 48 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index f9fa6e104fef..3f32c4db5232 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -113,6 +113,7 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
*p, union drm_amdgpu_cs
uint32_t uf_offset = 0;
int i;
int ret;
+   bool update_spm_vmid = false;
  
  	if (cs->in.num_chunks == 0)

return 0;
@@ -221,6 +222,10 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
*p, union drm_amdgpu_cs
case AMDGPU_CHUNK_ID_SYNCOBJ_TIMELINE_SIGNAL:
break;
  
+		case AMDGPU_CHUNK_ID_SPM_TRACE:

+   update_spm_vmid = true;
+   break;
+
default:
ret = -EINVAL;
goto free_partial_kdata;
@@ -231,6 +236,8 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser 
*p, union drm_amdgpu_cs
if (ret)
goto free_all_kdata;
  
+	p->job->need_update_spm_vmid = update_spm_vmid;

+
if (p->ctx->vram_lost_counter != p->job->vram_lost_counter) {
ret = -ECANCELED;
goto free_all_kdata;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
index cae81914c821..36faab12b585 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c
@@ -156,9 +156,13 @@ int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned 
num_ibs,
return -EINVAL;
}
  
-	if (vm && !job->vmid) {

-   dev_err(adev->dev, "VM IB without ID\n");
-   return -EINVAL;
+   if (vm) {
+   if (!job->vmid) {
+   dev_err(adev->dev, "VM IB without ID\n");
+   return -EINVAL;
+   } else if (adev->gfx.rlc.funcs->update_spm_vmid && 
job->need_update_spm_vmid) {
+   adev->gfx.rlc.funcs->update_spm_vmid(adev, job->vmid);
+   }
}
  
  	alloc_size = ring->funcs->emit_frame_size + num_ibs *

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
index 2e2110dddb76..4582536961c7 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.h
@@ -52,6 +52,7 @@ struct amdgpu_job {
boolvm_needs_flush;
uint64_tvm_pd_addr;
unsignedvmid;
+   boolneed_update_spm_vmid;
unsignedpasid;
uint32_tgds_base, gds_size;
uint32_tgws_base, gws_size;
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h 
b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
index d3d4707f2168..52509c254cbd 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_rlc.h
@@ -126,6 +126,7 @@ struct amdgpu_rlc_funcs {
void (*stop)(struct amdgpu_device *adev);
void (*reset)(struct amdgpu_device *adev);
void (*start)(struct amdgpu_device *adev);
+   void (*update_spm_vmid)(struct amdgpu_device *adev, unsigned vmid);
  };
  
  struct amdgpu_rlc {

diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c 
b/drivers/gpu/drm/amd/amdgpu/gfx_v10_0.c
index