public inbox for intel-gfx@lists.freedesktop.org
 help / color / mirror / Atom feed
From: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH v3 13/13] drm/i915/scheduler: Support user-defined priorities
Date: Fri, 29 Sep 2017 10:32:03 +0300	[thread overview]
Message-ID: <1506670323.4729.10.camel@linux.intel.com> (raw)
In-Reply-To: <20170928193910.17988-13-chris@chris-wilson.co.uk>

On Thu, 2017-09-28 at 20:39 +0100, Chris Wilson wrote:
> Use a priority stored in the context as the initial value when
> submitting a request. This allows us to change the default priority on a
> per-context basis, allowing different contexts to be favoured with GPU
> time at the expense of lower importance work. The user can adjust the
> context's priority via I915_CONTEXT_PARAM_PRIORITY, with more positive
> values being higher priority (they will be serviced earlier, after their
> dependencies have been resolved). Any prerequisite work for an execbuf
> will have its priority raised to match the new request as required.
> 
> Normal users can specify any value in the range of -1023 to 0 [default],
> i.e. they can reduce the priority of their workloads (and temporarily
> boost it back to normal if so desired).
> 
> Privileged users can specify any value in the range of -1023 to 1023,
> [default is 0], i.e. they can raise their priority above all overs and
> so potentially starve the system.
> 
> Note that the existing schedulers are not fair, nor load balancing, the
> execution is strictly by priority on a first-come, first-served basis,
> and the driver may choose to boost some requests above the range
> available to users.
> 
> This priority was originally based around nice(2), but evolved to allow
> clients to adjust their priority within a small range, and allow for a
> privileged high priority range.
> 
> For example, this can be used to implement EGL_IMG_context_priority
> https://www.khronos.org/registry/egl/extensions/IMG/EGL_IMG_context_priority.txt
> 
> 	EGL_CONTEXT_PRIORITY_LEVEL_IMG determines the priority level of
>         the context to be created. This attribute is a hint, as an
>         implementation may not support multiple contexts at some
>         priority levels and system policy may limit access to high
>         priority contexts to appropriate system privilege level. The
>         default value for EGL_CONTEXT_PRIORITY_LEVEL_IMG is
>         EGL_CONTEXT_PRIORITY_MEDIUM_IMG."
> 
> so we can map
> 
> 	PRIORITY_HIGH -> 1023 [privileged, will failback to 0]
> 	PRIORITY_MED -> 0 [default]
> 	PRIORITY_LOW -> -1023
> 
> They also map onto the priorities used by VkQueue (and a VkQueue is
> essentially a timeline, our i915_gem_context under full-ppgtt).
> 
> v2: s/CAP_SYS_ADMIN/CAP_SYS_NICE/
> v3: Report min/max user priorities as defines in the uapi, and rebase
> internal priorities on the exposed values.
> 
> Testcase: igt/gem_exec_schedule
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

Didn't seem to stick last time :P

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2017-09-29  7:32 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-09-28 19:38 [PATCH v3 01/13] drm/i915: Inherit Kabylake platform features from Skylake Chris Wilson
2017-09-28 19:38 ` [PATCH v3 02/13] drm/i915/execlists: Move request unwinding to a separate function Chris Wilson
2017-09-28 19:39 ` [PATCH v3 03/13] drm/i915: Give the invalid priority a magic name Chris Wilson
2017-09-29  6:39   ` Mika Kuoppala
2017-09-29  7:01   ` Joonas Lahtinen
2017-09-28 19:39 ` [PATCH v3 04/13] drm/i915/execlists: Cache the last priolist lookup Chris Wilson
2017-09-29 11:45   ` Chris Wilson
2017-09-28 19:39 ` [PATCH v3 05/13] drm/i915/preempt: Fix WaEnablePreemptionGranularityControlByUMD Chris Wilson
2017-09-28 19:39 ` [PATCH v3 06/13] drm/i915/preempt: Default to disabled mid-command preemption levels Chris Wilson
2017-09-29  7:18   ` Joonas Lahtinen
2017-09-28 19:39 ` [PATCH v3 07/13] drm/i915/execlists: Distinguish the incomplete context notifies Chris Wilson
2017-09-28 19:39 ` [PATCH v3 08/13] drm/i915: Introduce a preempt context Chris Wilson
2017-09-29  7:23   ` Joonas Lahtinen
2017-09-28 19:39 ` [PATCH v3 09/13] drm/i915/execlists: Move bdw GPGPU w/a to emit_bb Chris Wilson
2017-09-28 19:39 ` [PATCH v3 10/13] drm/i915/execlists: Keep request->priority for its lifetime Chris Wilson
2017-09-28 19:39 ` [PATCH v3 11/13] drm/i915: Expand I915_PARAM_HAS_SCHEDULER into a capability bitmask Chris Wilson
2017-09-28 19:39 ` [PATCH v3 12/13] drm/i915/execlists: Preemption! Chris Wilson
2017-09-29  6:55   ` Mika Kuoppala
2017-09-29 10:07     ` Chris Wilson
2017-09-29  7:29   ` Joonas Lahtinen
2017-09-29 10:12     ` Chris Wilson
2017-09-29 13:14       ` Joonas Lahtinen
2017-09-28 19:39 ` [PATCH v3 13/13] drm/i915/scheduler: Support user-defined priorities Chris Wilson
2017-09-29  7:32   ` Joonas Lahtinen [this message]
2017-09-28 19:58 ` [PATCH v3 01/13] drm/i915: Inherit Kabylake platform features from Skylake Rodrigo Vivi
2017-09-28 20:05   ` Chris Wilson
2017-09-28 21:04 ` ✗ Fi.CI.BAT: failure for series starting with [v3,01/13] " Patchwork
2017-09-29 10:29 ` ✓ Fi.CI.BAT: success " Patchwork
2017-09-29 11:20 ` [PATCH v3 01/13] " David Weinehall
2017-09-29 11:24 ` ✗ Fi.CI.IGT: failure for series starting with [v3,01/13] " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1506670323.4729.10.camel@linux.intel.com \
    --to=joonas.lahtinen@linux.intel.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox