All of lore.kernel.org
 help / color / mirror / Atom feed
From: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
To: "Chris Wilson" <chris@chris-wilson.co.uk>,
	"Michał Winiarski" <michal.winiarski@intel.com>,
	intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 13/14] drm/i915/guc: Workaround the missing user interrupt after preemption
Date: Fri, 20 Oct 2017 10:42:55 -0700	[thread overview]
Message-ID: <07d36f51-5908-c7b5-1b33-69160f2de4e4@intel.com> (raw)
In-Reply-To: <150852014424.21927.10081115525041712021@mail.alporthouse.com>



On 20/10/17 10:22, Chris Wilson wrote:
> Quoting Chris Wilson (2017-10-19 20:44:41)
>> Quoting Michał Winiarski (2017-10-19 19:36:18)
>>> With GuC, we're scheduling tasklet on USER_INTERRUPT - since some work
>>> has completed, perhaps we're able to submit some more work. We're doing
>>> similar thing for preemption - after preemption has completed, it's time
>>> to schedule the tasklet and submit more work (since the engine is now
>>> idle). Unfortunately, we can hit the scenarios where the preemption is
>>> done, but the interrupt is nowhere to be seen. To work around the
>>> problem, let's use a delayed work that's kicking the tasklet if
>>> preemption is done, and queueing itself otherwise.
>>
>> I'm not buying it yet.
> 
> I think I know what it is. I think it is as simple as us turning off the
> irq when there's no more requests being signaled on the engine.
> -Chris

I can confirm this, I was just about to reply with the same findings ;)
By hackily adding USER_INTERRUPT to irq_keep_mask we've been able to run 
exec_whisper@fds-priority several time in a row with no problems.

Daniele
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

  reply	other threads:[~2017-10-20 17:42 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-19 18:36 [PATCH 00/14] Preemption with GuC, third attempt Michał Winiarski
2017-10-19 18:36 ` [PATCH 01/14] drm/i915/guc: Do not use 0 for GuC doorbell cookie Michał Winiarski
2017-10-21  6:28   ` Sagar Arun Kamble
2017-10-19 18:36 ` [PATCH 02/14] drm/i915/guc: Extract GuC stage desc pool creation into a helper Michał Winiarski
2017-10-19 18:36 ` [PATCH v2 03/14] drm/i915/guc: Allocate separate shared data object for GuC communication Michał Winiarski
2017-10-19 18:36 ` [PATCH 04/14] drm/i915/guc: Initialize GuC before restarting engines Michał Winiarski
2017-10-25 18:44   ` Chris Wilson
2017-10-19 18:36 ` [PATCH 05/14] drm/i915/guc: Add preemption action to GuC firmware interface Michał Winiarski
2017-10-19 18:36 ` [PATCH v2 06/14] drm/i915/guc: Add a second client, to be used for preemption Michał Winiarski
2017-10-19 19:36   ` Michal Wajdeczko
2017-10-19 18:36 ` [PATCH 07/14] drm/i915/guc: Split guc_wq_item_append Michał Winiarski
2017-10-19 18:36 ` [PATCH v2 08/14] drm/i915: Extract "emit write" part of emit breadcrumb functions Michał Winiarski
2017-10-19 18:36 ` [PATCH 09/14] drm/i915: Add information needed to track engine preempt state Michał Winiarski
2017-10-19 18:36 ` [PATCH 10/14] drm/i915/guc: Keep request->priority for its lifetime Michał Winiarski
2017-10-19 18:36 ` [PATCH v2 11/14] drm/i915: Rename helpers used for unwinding, use macro for can_preempt Michał Winiarski
2017-10-19 19:18   ` Chris Wilson
2017-10-19 19:20   ` Chris Wilson
2017-10-19 18:36 ` [PATCH v3 12/14] drm/i915/guc: Preemption! With GuC Michał Winiarski
2017-10-19 19:14   ` Chris Wilson
2017-10-19 19:16   ` Chris Wilson
2017-10-20  9:00   ` Chris Wilson
2017-10-20 17:00   ` Chris Wilson
2017-10-23 23:14   ` Daniele Ceraolo Spurio
2017-10-25 18:15   ` Chris Wilson
2017-10-19 18:36 ` [PATCH 13/14] drm/i915/guc: Workaround the missing user interrupt after preemption Michał Winiarski
2017-10-19 19:44   ` Chris Wilson
2017-10-20 17:22     ` Chris Wilson
2017-10-20 17:42       ` Daniele Ceraolo Spurio [this message]
2017-10-20 16:08   ` Chris Wilson
2017-10-19 18:36 ` [PATCH 14/14] HAX Enable GuC Submission for CI Michał Winiarski
2017-10-19 19:11 ` ✗ Fi.CI.BAT: failure for Preemption with GuC, third attempt Patchwork
2017-10-26 21:21 ` 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=07d36f51-5908-c7b5-1b33-69160f2de4e4@intel.com \
    --to=daniele.ceraolospurio@intel.com \
    --cc=chris@chris-wilson.co.uk \
    --cc=intel-gfx@lists.freedesktop.org \
    --cc=michal.winiarski@intel.com \
    /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.