All of lore.kernel.org
 help / color / mirror / Atom feed
From: Ben Widawsky <ben@bwidawsk.net>
To: Daniel Vetter <daniel@ffwll.ch>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 14/18] drm/i915/context: switch contexts with execbuf2
Date: Fri, 30 Mar 2012 11:58:20 -0700	[thread overview]
Message-ID: <20120330115820.6eafe8e3@bwidawsk.net> (raw)
In-Reply-To: <20120329193820.GH27737@phenom.ffwll.local>

On Thu, 29 Mar 2012 21:38:20 +0200
Daniel Vetter <daniel@ffwll.ch> wrote:

> On Sun, Mar 18, 2012 at 01:39:54PM -0700, Ben Widawsky wrote:
> > Use the rsvd1 field in execbuf2 to specify the context ID associated
> > with the workload. This will allow the driver to do the proper context
> > switch when/if needed.
> > 
> > Signed-off-by: Ben Widawsky <ben@bwidawsk.net>
> > ---
> >  drivers/gpu/drm/i915/i915_gem_execbuffer.c |    6 ++++++
> >  include/drm/i915_drm.h                     |    4 +++-
> >  2 files changed, 9 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> > index 81687af..c365e12 100644
> > --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> > +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
> > @@ -1058,6 +1058,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
> >  	struct drm_i915_gem_object *batch_obj;
> >  	struct drm_clip_rect *cliprects = NULL;
> >  	struct intel_ring_buffer *ring;
> > +	u32 ctx_id = args->context_info & I915_EXEC_CONTEXT_ID_MASK;
> >  	u32 exec_start, exec_len;
> >  	u32 seqno;
> >  	u32 mask;
> > @@ -1266,6 +1267,10 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
> >  			goto err;
> >  	}
> >  
> > +	ret = i915_switch_context(ring, file, ctx_id, seqno, 0);
> > +	if (ret)
> > +		goto err;
> > +
> 
> Already complained a bit about these:
> - Imo the seqno and hw_flags param can/should go away.

Responding to the other checks for this.

> - You need to add some sanity checks somewhere so that we correctly bail
>   out of the do_execbuffer stuff without launching the batch. Obviously
>   also a prime candidate for an i-g-t tests (because it'll nicely exercise
>   a piece of code which usually is rather hard to tests).

I'm not sure what you're asking for, aside from - this is hairy code; be
careful. What was your idea?

> 
> Aside: I haven't yet looked at your testcases, so maybe I'll come up with
> a crazy idea for another testcase ;-)

Testcases are pretty sparse in i-g-t. Most of it is done implicitly with
mesa. If you have ideas, I'll be happy to write them

> 
> >  	trace_i915_gem_ring_dispatch(ring, seqno);
> >  
> >  	exec_start = batch_obj->gtt_offset + args->batch_start_offset;
> > @@ -1372,6 +1377,7 @@ i915_gem_execbuffer(struct drm_device *dev, void *data,
> >  	exec2.num_cliprects = args->num_cliprects;
> >  	exec2.cliprects_ptr = args->cliprects_ptr;
> >  	exec2.flags = I915_EXEC_RENDER;
> > +	exec2.context_info = 0;
> >  
> >  	ret = i915_gem_do_execbuffer(dev, data, file, &exec2, exec2_list);
> >  	if (!ret) {
> > diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h
> > index bead13e..03d159f 100644
> > --- a/include/drm/i915_drm.h
> > +++ b/include/drm/i915_drm.h
> > @@ -660,13 +660,15 @@ struct drm_i915_gem_execbuffer2 {
> >  #define I915_EXEC_CONSTANTS_ABSOLUTE 	(1<<6)
> >  #define I915_EXEC_CONSTANTS_REL_SURFACE (2<<6) /* gen4/5 only */
> >  	__u64 flags;
> > -	__u64 rsvd1;
> > +	__u64 context_info;
> >  	__u64 rsvd2;
> >  };
> >  
> >  /** Resets the SO write offset registers for transform feedback on gen7. */
> >  #define I915_EXEC_GEN7_SOL_RESET	(1<<8)
> >  
> > +#define I915_EXEC_CONTEXT_ID_MASK	(0xffffffff)
> > +
> >  struct drm_i915_gem_pin {
> >  	/** Handle of the buffer to be pinned. */
> >  	__u32 handle;
> > -- 
> > 1.7.9.4
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 

  reply	other threads:[~2012-03-30 18:58 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-03-18 20:39 [PATCH 00/18] i915 HW Context Support Ben Widawsky
2012-03-18 20:39 ` [PATCH 01/18] drm/i915: CXT_SIZE register offsets added Ben Widawsky
2012-03-18 20:39 ` [PATCH 02/18] drm/i915: preliminary context support Ben Widawsky
2012-03-28 22:43   ` Daniel Vetter
2012-03-28 22:59     ` Ben Widawsky
2012-03-29  8:43       ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 03/18] drm/i915: context basic create & destroy Ben Widawsky
2012-03-18 20:39 ` [PATCH 04/18] drm/i915: add context information to objects Ben Widawsky
2012-03-28 22:36   ` Daniel Vetter
2012-03-29  0:20     ` Ben Widawsky
2012-03-29  8:47       ` Daniel Vetter
2012-03-29 15:57         ` Ben Widawsky
2012-03-18 20:39 ` [PATCH 05/18] drm/i915: context switch implementation Ben Widawsky
2012-03-29 18:24   ` Daniel Vetter
2012-03-29 18:43     ` Ben Widawsky
2012-03-29 18:49       ` Daniel Vetter
2012-03-30 18:11         ` Ben Widawsky
2012-03-29 18:47   ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 06/18] drm/i915: trace events for contexts Ben Widawsky
2012-03-18 20:39 ` [PATCH 07/18] drm/i915: Ivybridge MI_ARB_ON_OFF context w/a Ben Widawsky
2012-03-18 20:39 ` [PATCH 08/18] drm/i915: PIPE_CONTROL_TLB_INVALIDATE Ben Widawsky
2012-03-18 20:39 ` [PATCH 09/18] drm/i915: possibly invalidate TLB before context switch Ben Widawsky
2012-03-29 19:25   ` Daniel Vetter
2012-03-30 18:39     ` Ben Widawsky
2012-03-30 19:01       ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 10/18] drm/i915: use the default context Ben Widawsky
2012-03-18 20:39 ` [PATCH 11/18] drm/i915: switch to default context on idle Ben Widawsky
2012-03-29 19:29   ` Daniel Vetter
2012-03-29 20:28     ` Chris Wilson
2012-03-30 21:17     ` Ben Widawsky
2012-03-30 21:30       ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 12/18] drm/i915: try to reset the gpu before unload Ben Widawsky
2012-03-29 19:31   ` Daniel Vetter
2012-03-30 18:50     ` Ben Widawsky
2012-03-30 19:05       ` Daniel Vetter
2012-03-30 16:54   ` Jesse Barnes
2012-03-30 17:30     ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 13/18] drm/i915/context: create & destroy ioctls Ben Widawsky
2012-03-29 19:35   ` Daniel Vetter
2012-03-30 18:55     ` Ben Widawsky
2012-03-30 19:16       ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 14/18] drm/i915/context: switch contexts with execbuf2 Ben Widawsky
2012-03-29 19:38   ` Daniel Vetter
2012-03-30 18:58     ` Ben Widawsky [this message]
2012-03-30 19:20       ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 15/18] drm/i915/context: add params Ben Widawsky
2012-03-18 20:39 ` [PATCH 16/18] drm/i915/context: anonymous context interfaces Ben Widawsky
2012-03-29 19:42   ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 17/18] drm/i915: Ironlake rc6 can use " Ben Widawsky
2012-03-29 19:47   ` Daniel Vetter
2012-03-18 20:39 ` [PATCH 18/18] drm/i915: try to enable rc6 on Ironlake... again Ben Widawsky
2012-03-19  3:47 ` [PATCH 00/18] i915 HW Context Support Ben Widawsky
2012-03-19 10:14 ` Daniel Vetter
2012-03-29 19:51   ` Daniel Vetter

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=20120330115820.6eafe8e3@bwidawsk.net \
    --to=ben@bwidawsk.net \
    --cc=daniel@ffwll.ch \
    --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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.