Re: [PATCH 1/4] drm/v3d: Delay the scheduler timeout if we're still making progress.

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Tue, Jul 03, 2018 at 10:05:12AM -0700, Eric Anholt wrote:
> GTF-GLES2.gtf.GL.acos.acos_float_vert_xvary submits jobs that take 4
> seconds at maximum resolution, but we still want to reset quickly if a
> job is really hung.  Sample the CL's current address and the return
> address (since we call into tile lists repeatedly) and if either has
> changed then assume we've made progress.
> 
> Signed-off-by: Eric Anholt <eric@xxxxxxxxxx>
> Cc: Lucas Stach <l.stach@xxxxxxxxxxxxxx>
> ---
>  drivers/gpu/drm/v3d/v3d_drv.h   |  2 ++
>  drivers/gpu/drm/v3d/v3d_regs.h  |  1 +
>  drivers/gpu/drm/v3d/v3d_sched.c | 18 ++++++++++++++++++
>  3 files changed, 21 insertions(+)
> 
> diff --git a/drivers/gpu/drm/v3d/v3d_drv.h b/drivers/gpu/drm/v3d/v3d_drv.h
> index f546e0ab9562..a5d96d823416 100644
> --- a/drivers/gpu/drm/v3d/v3d_drv.h
> +++ b/drivers/gpu/drm/v3d/v3d_drv.h
> @@ -189,6 +189,8 @@ struct v3d_job {
>  
>  	/* GPU virtual addresses of the start/end of the CL job. */
>  	u32 start, end;
> +
> +	u32 timedout_ctca, timedout_ctra;
>  };
>  
>  struct v3d_exec_info {
> diff --git a/drivers/gpu/drm/v3d/v3d_regs.h b/drivers/gpu/drm/v3d/v3d_regs.h
> index fc13282dfc2f..854046565989 100644
> --- a/drivers/gpu/drm/v3d/v3d_regs.h
> +++ b/drivers/gpu/drm/v3d/v3d_regs.h
> @@ -222,6 +222,7 @@
>  #define V3D_CLE_CTNCA(n) (V3D_CLE_CT0CA + 4 * n)
>  #define V3D_CLE_CT0RA                                  0x00118
>  #define V3D_CLE_CT1RA                                  0x0011c
> +#define V3D_CLE_CTNRA(n) (V3D_CLE_CT0RA + 4 * n)
>  #define V3D_CLE_CT0LC                                  0x00120
>  #define V3D_CLE_CT1LC                                  0x00124
>  #define V3D_CLE_CT0PC                                  0x00128
> diff --git a/drivers/gpu/drm/v3d/v3d_sched.c b/drivers/gpu/drm/v3d/v3d_sched.c
> index 808bc901f567..00667c733dca 100644
> --- a/drivers/gpu/drm/v3d/v3d_sched.c
> +++ b/drivers/gpu/drm/v3d/v3d_sched.c
> @@ -153,7 +153,25 @@ v3d_job_timedout(struct drm_sched_job *sched_job)
>  	struct v3d_job *job = to_v3d_job(sched_job);
>  	struct v3d_exec_info *exec = job->exec;
>  	struct v3d_dev *v3d = exec->v3d;
> +	enum v3d_queue job_q = job == &exec->bin ? V3D_BIN : V3D_RENDER;
>  	enum v3d_queue q;
> +	u32 ctca = V3D_CORE_READ(0, V3D_CLE_CTNCA(job_q));
> +	u32 ctra = V3D_CORE_READ(0, V3D_CLE_CTNRA(job_q));
> +
> +	/* If the current address or return address have changed, then
> +	 * the GPU has probably made progress and we should delay the
> +	 * reset.  This could fail if the GPU got in an infinite loop
> +	 * in the CL, but that is pretty unlikely outside of an i-g-t
> +	 * testcase.
> +	 */

In i915 we have a credit system that allows only a limited amount of
postponing the timeout. That would block the fairly easy DOS. Otoh I have
no idea how robust v3d is against DOS attacks at the ioctl level and
whether that's worth it.

>From your description I'm assuming that a loop in the shaders won't trick
this, so ARB_robustness/WebGL should still be happy.

Anyway, Ack on the entire pile.
-Daniel
> +	if (job->timedout_ctca != ctca || job->timedout_ctra != ctra) {
> +		job->timedout_ctca = ctca;
> +		job->timedout_ctra = ctra;
> +
> +		schedule_delayed_work(&job->base.work_tdr,
> +				      job->base.sched->timeout);
> +		return;
> +	}
>  
>  	mutex_lock(&v3d->reset_lock);
>  
> -- 
> 2.18.0
> 
> _______________________________________________
> dri-devel mailing list
> dri-devel@xxxxxxxxxxxxxxxxxxxxx
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel




[Index of Archives]     [Linux DRI Users]     [Linux Intel Graphics]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]
  Powered by Linux