Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
On 24/07/2024 12:16, Christian König wrote: Am 24.07.24 um 10:16 schrieb Tvrtko Ursulin: [SNIP] Absolutely. Absolutely good and absolutely me, or absolutely you? :) You, I don't even have time to finish all the stuff I already started :/ Okay, I think I can squeeze it in. These are the TODO points and their opens: - Adjust amdgpu_ctx_set_entity_priority() to call drm_sched_entity_modify_sched() regardless of the hw_type - to fix priority changes on a single sched other than gfx or compute. Either that or to stop using the scheduler priority to implement userspace priorities and always use different HW queues for that. - Document sched_list array lifetime must align with the entity and adjust the callers. Open: Do you still oppose keeping sched_list for num_scheds == 1? Not if you can fix up all callers. If so do you propose drm_sched_entity_modify_sched() keeps disagreeing with drm_sched_entity_init() on this detail? And keep the "one shot single sched_list" quirk in? Why is that nicer than simply keeping the list and remove that quirk? Once lifetime rules are clear it IMO is okay to always keep the list. Yeah if every caller of drm_sched_entity_init() can be fixed I'm fine with that as well. Okay so I will tackle the above few first. - Remove drm_sched_entity_set_priority(). Open: Should we at this point also modify amdgpu_device_init_schedulers() to stop initialising schedulers with DRM_SCHED_PRIORITY_COUNT run queues? One step at a time. And leave this for later. Regards, Tvrtko
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
Am 24.07.24 um 10:16 schrieb Tvrtko Ursulin:
[SNIP]
Absolutely.
Absolutely good and absolutely me, or absolutely you? :)
You, I don't even have time to finish all the stuff I already started :/
These are the TODO points and their opens:
- Adjust amdgpu_ctx_set_entity_priority() to call
drm_sched_entity_modify_sched() regardless of the hw_type - to fix
priority changes on a single sched other than gfx or compute.
Either that or to stop using the scheduler priority to implement
userspace priorities and always use different HW queues for that.
- Document sched_list array lifetime must align with the entity and
adjust the callers.
Open:
Do you still oppose keeping sched_list for num_scheds == 1?
Not if you can fix up all callers.
If so do you propose drm_sched_entity_modify_sched() keeps disagreeing
with drm_sched_entity_init() on this detail? And keep the "one shot
single sched_list" quirk in? Why is that nicer than simply keeping the
list and remove that quirk? Once lifetime rules are clear it IMO is
okay to always keep the list.
Yeah if every caller of drm_sched_entity_init() can be fixed I'm fine
with that as well.
- Remove drm_sched_entity_set_priority().
Open:
Should we at this point also modify amdgpu_device_init_schedulers() to
stop initialising schedulers with DRM_SCHED_PRIORITY_COUNT run queues?
One step at a time.
Once drm_sched_entity_set_priority() is gone there is little point.
Unless there are some non-obvious games with the kernel priority or
something.
In general scheduler priorities were meant to be used for things
like kernel queues which would always have higher priority than
user space submissions and using them for userspace turned out to
be not such a good idea.
Out of curiousity what were the problems? I cannot think of
anything fundamental since there are priorities at the backend
level after all, just fewer levels.
A higher level queue can starve lower level queues to the point
that they never get a chance to get anything running.
Oh that.. well I call that implementation details. :) Because
nowhere in the uapi is I think guaranteed execution ordering needs
to be strictly in descending priority. This potentially goes back to
what you said above that a potential larger rewrite might be
beneficial. Implementing some smarter scheduling. Although the issue
will be it is just frontend scheduling after all. So a bit
questionable to invest in making it too smart.
+1 and we have a bug report complaining that RR is in at least one
situation better than FIFO. So it is actually quite hard to remove.
On the other hand FIFO has some really nice properties as well. E.g.
try to run >100 glxgears instances (on weaker hw) and just visually
compare the result differences between RR and FIFO. FIFO is baby
smooth and RR is basically stuttering all the time.
I think this goes more back to what I was suggesting during early xe
days, that potentially drm scheduler should be split into dependency
handling part and the scheduler part. Drivers with 1:1
entity:scheduler and full hardware/firmware scheduling do not really
need neither fifo or rr.
Yeah that's my thinking as well and I also suggested that multiple
times in discussions with Sima and others.
This basically means that userspace gets a chance to submit
infinity fences with all the bad consequences.
I mean one problem unrelated to this discussion is this:
void drm_sched_entity_select_rq(struct drm_sched_entity *entity)
{
struct dma_fence *fence;
struct drm_gpu_scheduler *sched;
struct drm_sched_rq *rq;
/* queue non-empty, stay on the same engine */
if (spsc_queue_count(&entity->job_queue))
return;
Which makes it look like the entity with a constant trickle of
jobs will never actually manage to change it's run queue. Neither
today, nor after the potential drm_sched_entity_set_priority()
removal.
That's intentional and based on how the scheduler load balancing
works.
I see that it is intentional but if it can silently prevent priority
changes (even hw priority) it is not very solid. Unless I am missing
something here.
IIRC the GSoC student who implemented this (with me as mentor)
actually documented that behavior somewhere.
And to be honest it kind of makes sense because switching priorities
would otherwise be disruptive, e.g. you have a moment were you need
to drain all previous submissions with the old priority before you
can do new ones with the new priority.
Hmmm I don't see how it makes sense. Perhaps a test case for
AMDGPU_SCHED_OP_*_PRIORITY_OVERRIDE is missing to show how it doesn't
work. Or at least how easy it can be defeated with callers none the
wiser.
Ok, that needs a bit longer explanation. You don't by any chance have teams?
For context I am kind of interested because I wired up amdgpu to the
DRM cgroup controller and use priority override to de-prioritize
certain cgr
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
On 22/07/2024 16:13, Christian König wrote:
Am 22.07.24 um 16:43 schrieb Tvrtko Ursulin:
On 22/07/2024 15:06, Christian König wrote:
Am 22.07.24 um 15:52 schrieb Tvrtko Ursulin:
On 19/07/2024 16:18, Christian König wrote:
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply
update the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver
which can do
dynamic priority changes. And that appears was attempted to be
rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx
priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if*
drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is
actually
called, which in amdgpu's case today is true only for gfx and
compute.
Priority changes for other engines with only one scheduler
assigned, such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched:
improve docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of
schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of
sched
list")), drm_sched_entity_modify_sched() was disagreeing with
that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for
many
callers with just one scheduler. With more than one scheduler
amdgpu is
the only caller, and there the container is not on the stack.
Keeping a
stack backed list in the entity would obviously be undefined
behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container
for the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally
prevent the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched()
messed this up.
To give more background: Amdgpu has two different ways of handling
priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially
dropped the scheduler list to avoid using a stack allocated list,
and that functionality is still used in amdgpu_ctx_init_entity()
for example.
Setting the scheduler priority was basically just a workaround
because we didn't had the hw priorities at that time. Since that is
no longer the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
Removing drm_sched_entity_set_priority() is one thing, but we also
need to clear up the sched_list container ownership issue. It is
neither documented, nor robustly handled in the code. The
"num_scheds == 1" special casing throughout IMHO has to go too.
I disagree. Keeping the scheduler list in the entity is only useful
for load balancing.
As long as only one scheduler is provided and we don't load balance
the entity doesn't needs the scheduler list in the first place.
Once set_priority is removed then it indeed it doesn't. But even when
it is removed it needs documenting who owns the passed in container.
Today drivers are okay to pass a stack array when it is one element,
but if they did it with more than one they would be in for a nasty
surprise.
Yes, completely agree. But instead of copying the array I would rather
go into the direction to cleanup all callers and make the scheduler list
mandatory to stay around as long as the scheduler lives.
The whole thing of one calling convention there and another one at a
different place really sucks.
Ok, lets scroll a bit down to formulate a plan.
Another thing if you want to get rid of frontend priority handling
is to stop configuring scheduler instances with
DRM
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
Am 22.07.24 um 16:43 schrieb Tvrtko Ursulin:
On 22/07/2024 15:06, Christian König wrote:
Am 22.07.24 um 15:52 schrieb Tvrtko Ursulin:
On 19/07/2024 16:18, Christian König wrote:
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply
update the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver
which can do
dynamic priority changes. And that appears was attempted to be
rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx
priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if*
drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is
actually
called, which in amdgpu's case today is true only for gfx and
compute.
Priority changes for other engines with only one scheduler
assigned, such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched:
improve docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of
schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of
sched
list")), drm_sched_entity_modify_sched() was disagreeing with
that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for
many
callers with just one scheduler. With more than one scheduler
amdgpu is
the only caller, and there the container is not on the stack.
Keeping a
stack backed list in the entity would obviously be undefined
behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container
for the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally
prevent the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched()
messed this up.
To give more background: Amdgpu has two different ways of handling
priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially
dropped the scheduler list to avoid using a stack allocated list,
and that functionality is still used in amdgpu_ctx_init_entity()
for example.
Setting the scheduler priority was basically just a workaround
because we didn't had the hw priorities at that time. Since that is
no longer the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
Removing drm_sched_entity_set_priority() is one thing, but we also
need to clear up the sched_list container ownership issue. It is
neither documented, nor robustly handled in the code. The
"num_scheds == 1" special casing throughout IMHO has to go too.
I disagree. Keeping the scheduler list in the entity is only useful
for load balancing.
As long as only one scheduler is provided and we don't load balance
the entity doesn't needs the scheduler list in the first place.
Once set_priority is removed then it indeed it doesn't. But even when
it is removed it needs documenting who owns the passed in container.
Today drivers are okay to pass a stack array when it is one element,
but if they did it with more than one they would be in for a nasty
surprise.
Yes, completely agree. But instead of copying the array I would rather
go into the direction to cleanup all callers and make the scheduler list
mandatory to stay around as long as the scheduler lives.
The whole thing of one calling convention there and another one at a
different place really sucks.
Another thing if you want to get rid of frontend priority handling
is to stop configuring scheduler instances with
DRM_SCHED_PRIORITY_COUNT priority levels, to avoid wasting memory on
pointless run queues.
I wo
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
On 22/07/2024 15:06, Christian König wrote:
Am 22.07.24 um 15:52 schrieb Tvrtko Ursulin:
On 19/07/2024 16:18, Christian König wrote:
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply update
the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver which
can do
dynamic priority changes. And that appears was attempted to be
rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx
priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if*
drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is
actually
called, which in amdgpu's case today is true only for gfx and compute.
Priority changes for other engines with only one scheduler
assigned, such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched:
improve docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of
schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")), drm_sched_entity_modify_sched() was disagreeing with that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for many
callers with just one scheduler. With more than one scheduler
amdgpu is
the only caller, and there the container is not on the stack.
Keeping a
stack backed list in the entity would obviously be undefined behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container for
the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally prevent
the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched() messed
this up.
To give more background: Amdgpu has two different ways of handling
priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially
dropped the scheduler list to avoid using a stack allocated list, and
that functionality is still used in amdgpu_ctx_init_entity() for
example.
Setting the scheduler priority was basically just a workaround
because we didn't had the hw priorities at that time. Since that is
no longer the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
Removing drm_sched_entity_set_priority() is one thing, but we also
need to clear up the sched_list container ownership issue. It is
neither documented, nor robustly handled in the code. The "num_scheds
== 1" special casing throughout IMHO has to go too.
I disagree. Keeping the scheduler list in the entity is only useful for
load balancing.
As long as only one scheduler is provided and we don't load balance the
entity doesn't needs the scheduler list in the first place.
Once set_priority is removed then it indeed it doesn't. But even when it
is removed it needs documenting who owns the passed in container. Today
drivers are okay to pass a stack array when it is one element, but if
they did it with more than one they would be in for a nasty surprise.
Another thing if you want to get rid of frontend priority handling is
to stop configuring scheduler instances with DRM_SCHED_PRIORITY_COUNT
priority levels, to avoid wasting memory on pointless run queues.
I would rather like to completely drop the RR with the runlists
altogether and keep only the FIFO approach around. This way priority can
be implemented by boosting the score of submissions by a certain degree.
You mean larger refactoring of the scheduler removing the 1:N between
drm_sched and drm_sched_rq?
And final thing is to check whether the locking in
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
Am 22.07.24 um 15:52 schrieb Tvrtko Ursulin:
On 19/07/2024 16:18, Christian König wrote:
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply update
the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver which
can do
dynamic priority changes. And that appears was attempted to be
rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx
priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if*
drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is
actually
called, which in amdgpu's case today is true only for gfx and compute.
Priority changes for other engines with only one scheduler
assigned, such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched:
improve docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of
schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")), drm_sched_entity_modify_sched() was disagreeing with that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for many
callers with just one scheduler. With more than one scheduler
amdgpu is
the only caller, and there the container is not on the stack.
Keeping a
stack backed list in the entity would obviously be undefined behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container for
the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally prevent
the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched() messed
this up.
To give more background: Amdgpu has two different ways of handling
priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially
dropped the scheduler list to avoid using a stack allocated list, and
that functionality is still used in amdgpu_ctx_init_entity() for
example.
Setting the scheduler priority was basically just a workaround
because we didn't had the hw priorities at that time. Since that is
no longer the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
Removing drm_sched_entity_set_priority() is one thing, but we also
need to clear up the sched_list container ownership issue. It is
neither documented, nor robustly handled in the code. The "num_scheds
== 1" special casing throughout IMHO has to go too.
I disagree. Keeping the scheduler list in the entity is only useful for
load balancing.
As long as only one scheduler is provided and we don't load balance the
entity doesn't needs the scheduler list in the first place.
Another thing if you want to get rid of frontend priority handling is
to stop configuring scheduler instances with DRM_SCHED_PRIORITY_COUNT
priority levels, to avoid wasting memory on pointless run queues.
I would rather like to completely drop the RR with the runlists
altogether and keep only the FIFO approach around. This way priority can
be implemented by boosting the score of submissions by a certain degree.
And final thing is to check whether the locking in
drm_sched_entity_modify_sched() is okay. Because according to kerneldoc:
* Note that this must be called under the same common lock for
@entity as
* drm_sched_job_arm() and drm_sched_entity_push_job(), or the driver
needs to
* guarantee through some other means that this is never called while
new jobs
* can be pushed to @entity.
I don't see that is the case. Priority override is under
amdgpu_ctx_mgr->lock, while job ar
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
On 19/07/2024 16:18, Christian König wrote:
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply update the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver which
can do
dynamic priority changes. And that appears was attempted to be rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if* drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is actually
called, which in amdgpu's case today is true only for gfx and compute.
Priority changes for other engines with only one scheduler assigned,
such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched: improve
docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")), drm_sched_entity_modify_sched() was disagreeing with that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for many
callers with just one scheduler. With more than one scheduler amdgpu is
the only caller, and there the container is not on the stack. Keeping a
stack backed list in the entity would obviously be undefined behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container for
the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally prevent
the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched() messed
this up.
To give more background: Amdgpu has two different ways of handling
priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially dropped
the scheduler list to avoid using a stack allocated list, and that
functionality is still used in amdgpu_ctx_init_entity() for example.
Setting the scheduler priority was basically just a workaround because
we didn't had the hw priorities at that time. Since that is no longer
the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
Removing drm_sched_entity_set_priority() is one thing, but we also need
to clear up the sched_list container ownership issue. It is neither
documented, nor robustly handled in the code. The "num_scheds == 1"
special casing throughout IMHO has to go too.
Another thing if you want to get rid of frontend priority handling is to
stop configuring scheduler instances with DRM_SCHED_PRIORITY_COUNT
priority levels, to avoid wasting memory on pointless run queues.
And final thing is to check whether the locking in
drm_sched_entity_modify_sched() is okay. Because according to kerneldoc:
* Note that this must be called under the same common lock for @entity as
* drm_sched_job_arm() and drm_sched_entity_push_job(), or the driver
needs to
* guarantee through some other means that this is never called while
new jobs
* can be pushed to @entity.
I don't see that is the case. Priority override is under
amdgpu_ctx_mgr->lock, while job arm and push appear not. I also cannot
spot anything else preventing amdgpu_sched_ioctl() running in parallel
to everything else.
In general scheduler priorities were meant to be used for things like
kernel queues which would always have higher priority than user space
submissions and using them for userspace turned out to be not such a
good idea.
Out of curiousity what were the problems? I cannot think of anything
fundamental since there are priorities at the backend level after all,
jus
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
On Fri, Jul 19, 2024 at 05:18:05PM +0200, Christian König wrote:
> Am 19.07.24 um 15:02 schrieb Christian König:
> > Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
> > > From: Tvrtko Ursulin
> > >
> > > Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
> > > creation") a change was made which prevented priority changes for
> > > entities
> > > with only one assigned scheduler.
> > >
> > > The commit reduced drm_sched_entity_set_priority() to simply update the
> > > entities priority, but the run queue selection logic in
> > > drm_sched_entity_select_rq() was never able to actually change the
> > > originally assigned run queue.
> > >
> > > In practice that only affected amdgpu, being the only driver which
> > > can do
> > > dynamic priority changes. And that appears was attempted to be rectified
> > > there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx priority
> > > override").
> > >
> > > A few unresolved problems however were that this only fixed
> > > drm_sched_entity_set_priority() *if* drm_sched_entity_modify_sched() was
> > > called first. That was not documented anywhere.
> > >
> > > Secondly, this only works if drm_sched_entity_modify_sched() is actually
> > > called, which in amdgpu's case today is true only for gfx and compute.
> > > Priority changes for other engines with only one scheduler assigned,
> > > such
> > > as jpeg and video decode will still not work.
> > >
> > > Note that this was also noticed in 981b04d96856 ("drm/sched: improve
> > > docs
> > > around drm_sched_entity").
> > >
> > > Completely different set of non-obvious confusion was that whereas
> > > drm_sched_entity_init() was not keeping the passed in list of schedulers
> > > (courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
> > > list")), drm_sched_entity_modify_sched() was disagreeing with that and
> > > would simply assign the single item list.
> > >
> > > That incosistency appears to be semi-silently fixed in ac4eb83ab255
> > > ("drm/sched: select new rq even if there is only one v3").
> > >
> > > What was also not documented is why it was important to not keep the
> > > list of schedulers when there is only one. I suspect it could have
> > > something to do with the fact the passed in array is on stack for many
> > > callers with just one scheduler. With more than one scheduler amdgpu is
> > > the only caller, and there the container is not on the stack. Keeping a
> > > stack backed list in the entity would obviously be undefined behaviour
> > > *if* the list was kept.
> > >
> > > Amdgpu however did only stop passing in stack backed container for
> > > the more
> > > than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
> > > entities on
> > > demand"). Until then I suspect dereferencing freed stack from
> > > drm_sched_entity_select_rq() was still present.
> > >
> > > In order to untangle all that and fix priority changes this patch is
> > > bringing back the entity owned container for storing the passed in
> > > scheduler list.
> >
> > Please don't. That makes the mess just more horrible.
> >
> > The background of not keeping the array is to intentionally prevent the
> > priority override from working.
> >
> > The bug is rather that adding drm_sched_entity_modify_sched() messed
> > this up.
>
> To give more background: Amdgpu has two different ways of handling priority:
> 1. The priority in the DRM scheduler.
> 2. Different HW rings with different priorities.
>
> Your analysis is correct that drm_sched_entity_init() initially dropped the
> scheduler list to avoid using a stack allocated list, and that functionality
> is still used in amdgpu_ctx_init_entity() for example.
>
> Setting the scheduler priority was basically just a workaround because we
> didn't had the hw priorities at that time. Since that is no longer the case
> I suggest to just completely drop the drm_sched_entity_set_priority()
> function instead.
>
+1 on this idea of dropping drm_sched_entity_set_priority if it doesn't
really work and unused.
It certainly unused in Xe and Xe has HW rings with different priorities
via the GuC interface. I belive this is also true for all new drivers
based on my interactions.
We should not be adding complexity the scheduler without a use case.
Matt
> In general scheduler priorities were meant to be used for things like kernel
> queues which would always have higher priority than user space submissions
> and using them for userspace turned out to be not such a good idea.
>
> Regards,
> Christian.
>
> >
> > Regards,
> > Christian.
> >
> >
> > > Container is now owned by the entity and the pointers are
> > > owned by the drivers. List of schedulers is always kept including
> > > for the
> > > one scheduler case.
> > >
> > > The patch can therefore also removes the single scheduler special case,
> > > which means that priority changes should now work (be able to change the
> > > selected run-queue) for all drivers and engines. In other w
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
Am 19.07.24 um 15:02 schrieb Christian König:
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for
entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply update the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver which
can do
dynamic priority changes. And that appears was attempted to be rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if* drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is actually
called, which in amdgpu's case today is true only for gfx and compute.
Priority changes for other engines with only one scheduler assigned,
such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched: improve
docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")), drm_sched_entity_modify_sched() was disagreeing with that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for many
callers with just one scheduler. With more than one scheduler amdgpu is
the only caller, and there the container is not on the stack. Keeping a
stack backed list in the entity would obviously be undefined behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container for
the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate
entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally prevent
the priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched() messed
this up.
To give more background: Amdgpu has two different ways of handling priority:
1. The priority in the DRM scheduler.
2. Different HW rings with different priorities.
Your analysis is correct that drm_sched_entity_init() initially dropped
the scheduler list to avoid using a stack allocated list, and that
functionality is still used in amdgpu_ctx_init_entity() for example.
Setting the scheduler priority was basically just a workaround because
we didn't had the hw priorities at that time. Since that is no longer
the case I suggest to just completely drop the
drm_sched_entity_set_priority() function instead.
In general scheduler priorities were meant to be used for things like
kernel queues which would always have higher priority than user space
submissions and using them for userspace turned out to be not such a
good idea.
Regards,
Christian.
Regards,
Christian.
Container is now owned by the entity and the pointers are
owned by the drivers. List of schedulers is always kept including for
the
one scheduler case.
The patch can therefore also removes the single scheduler special case,
which means that priority changes should now work (be able to change the
selected run-queue) for all drivers and engines. In other words
drm_sched_entity_set_priority() should now just work for all cases.
To enable maintaining its own container some API calls needed to grow a
capability for returning success/failure, which is a change which
percolates mostly through amdgpu source.
Signed-off-by: Tvrtko Ursulin
Fixes: b3ac17667f11 ("drm/scheduler: rework entity creation")
References: 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")
References: 977f7e1068be ("drm/amdgpu: allocate entities on demand")
References: 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx
priority override")
References: ac4eb83ab255 ("drm/sched: select new rq even if there is
only one v3")
References: 981b04d96856 ("drm/sched: improve docs around
drm_sched_entity")
Cc: Christian König
Cc: Alex Deucher
Cc: Luben Tuikov
Cc: Matthew Brost
Cc: Daniel Vetter
Cc: [email protected]
Re: [PATCH] drm/scheduler: Fix drm_sched_entity_set_priority()
Am 19.07.24 um 11:47 schrieb Tvrtko Ursulin:
From: Tvrtko Ursulin
Long time ago in commit b3ac17667f11 ("drm/scheduler: rework entity
creation") a change was made which prevented priority changes for entities
with only one assigned scheduler.
The commit reduced drm_sched_entity_set_priority() to simply update the
entities priority, but the run queue selection logic in
drm_sched_entity_select_rq() was never able to actually change the
originally assigned run queue.
In practice that only affected amdgpu, being the only driver which can do
dynamic priority changes. And that appears was attempted to be rectified
there in 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx priority
override").
A few unresolved problems however were that this only fixed
drm_sched_entity_set_priority() *if* drm_sched_entity_modify_sched() was
called first. That was not documented anywhere.
Secondly, this only works if drm_sched_entity_modify_sched() is actually
called, which in amdgpu's case today is true only for gfx and compute.
Priority changes for other engines with only one scheduler assigned, such
as jpeg and video decode will still not work.
Note that this was also noticed in 981b04d96856 ("drm/sched: improve docs
around drm_sched_entity").
Completely different set of non-obvious confusion was that whereas
drm_sched_entity_init() was not keeping the passed in list of schedulers
(courtesy of 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched
list")), drm_sched_entity_modify_sched() was disagreeing with that and
would simply assign the single item list.
That incosistency appears to be semi-silently fixed in ac4eb83ab255
("drm/sched: select new rq even if there is only one v3").
What was also not documented is why it was important to not keep the
list of schedulers when there is only one. I suspect it could have
something to do with the fact the passed in array is on stack for many
callers with just one scheduler. With more than one scheduler amdgpu is
the only caller, and there the container is not on the stack. Keeping a
stack backed list in the entity would obviously be undefined behaviour
*if* the list was kept.
Amdgpu however did only stop passing in stack backed container for the more
than one scheduler case in 977f7e1068be ("drm/amdgpu: allocate entities on
demand"). Until then I suspect dereferencing freed stack from
drm_sched_entity_select_rq() was still present.
In order to untangle all that and fix priority changes this patch is
bringing back the entity owned container for storing the passed in
scheduler list.
Please don't. That makes the mess just more horrible.
The background of not keeping the array is to intentionally prevent the
priority override from working.
The bug is rather that adding drm_sched_entity_modify_sched() messed
this up.
Regards,
Christian.
Container is now owned by the entity and the pointers are
owned by the drivers. List of schedulers is always kept including for the
one scheduler case.
The patch can therefore also removes the single scheduler special case,
which means that priority changes should now work (be able to change the
selected run-queue) for all drivers and engines. In other words
drm_sched_entity_set_priority() should now just work for all cases.
To enable maintaining its own container some API calls needed to grow a
capability for returning success/failure, which is a change which
percolates mostly through amdgpu source.
Signed-off-by: Tvrtko Ursulin
Fixes: b3ac17667f11 ("drm/scheduler: rework entity creation")
References: 8c23056bdc7a ("drm/scheduler: do not keep a copy of sched list")
References: 977f7e1068be ("drm/amdgpu: allocate entities on demand")
References: 2316a86bde49 ("drm/amdgpu: change hw sched list on ctx priority
override")
References: ac4eb83ab255 ("drm/sched: select new rq even if there is only one
v3")
References: 981b04d96856 ("drm/sched: improve docs around drm_sched_entity")
Cc: Christian König
Cc: Alex Deucher
Cc: Luben Tuikov
Cc: Matthew Brost
Cc: Daniel Vetter
Cc: [email protected]
Cc: [email protected]
Cc: # v5.6+
---
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c | 31 +---
drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.h | 2 +-
drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c | 13 +--
drivers/gpu/drm/amd/amdgpu/vcn_v4_0_5.c | 3 +-
drivers/gpu/drm/scheduler/sched_entity.c | 96 ---
include/drm/gpu_scheduler.h | 16 ++--
6 files changed, 100 insertions(+), 61 deletions(-)
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
index 5cb33ac99f70..387247f8307e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c
@@ -802,15 +802,15 @@ struct dma_fence *amdgpu_ctx_get_fence(struct amdgpu_ctx
*ctx,
return fence;
}
-static void amdgpu_ctx_set_entity_priority(struct amdgpu_ctx *ctx,
- struct amdgpu_ctx_entity *a
