From: John Harrison <John.C.Harrison@Intel.com>
To: Chris Wilson <chris@chris-wilson.co.uk>, Intel-GFX@Lists.FreeDesktop.Org
Subject: Re: [PATCH 1/7] drm/i915: Convert requests to use struct fence
Date: Mon, 11 Jan 2016 19:03:39 +0000 [thread overview]
Message-ID: <5693FC8B.9000102@Intel.com> (raw)
In-Reply-To: <20160108215907.GB28013@nuc-i3427.alporthouse.com>
On 08/01/2016 21:59, Chris Wilson wrote:
> On Fri, Jan 08, 2016 at 06:47:22PM +0000, John.C.Harrison@Intel.com wrote:
>> From: John Harrison <John.C.Harrison@Intel.com>
>>
>> There is a construct in the linux kernel called 'struct fence' that is
>> intended to keep track of work that is executed on hardware. I.e. it
>> solves the basic problem that the drivers 'struct
>> drm_i915_gem_request' is trying to address. The request structure does
>> quite a lot more than simply track the execution progress so is very
>> definitely still required. However, the basic completion status side
>> could be updated to use the ready made fence implementation and gain
>> all the advantages that provides.
>>
>> This patch makes the first step of integrating a struct fence into the
>> request. It replaces the explicit reference count with that of the
>> fence. It also replaces the 'is completed' test with the fence's
>> equivalent. Currently, that simply chains on to the original request
>> implementation. A future patch will improve this.
> But this forces everyone to do the heavyweight polling until the request
> is completed?
Not sure what you mean by heavy weight polling. And as described, this
is only an intermediate step.
> The seqno is already CPU cacheable and with the exception
> of interrupt polling, the question of whether a fence is complete can be
> determined by just inspecting that value. Only one place (the
> interrupt/signalling path) should ever be concerned about the
> complication of how we emit the breadcrumb and interrupt from the ring.
There is still only one piece of code that needs to worry about the
internal details. This is change is simply moving that around to make it
easier to be interrupt driven later on (which is required for the
scheduler).
> -Chris
>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2016-01-11 19:03 UTC|newest]
Thread overview: 74+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-12-11 13:11 [PATCH 00/13] Convert requests to use struct fence John.C.Harrison
2015-12-11 13:11 ` [PATCH 01/13] staging/android/sync: Support sync points created from dma-fences John.C.Harrison
2015-12-17 17:32 ` [Intel-gfx] " Jesse Barnes
2015-12-11 13:11 ` [PATCH 02/13] staging/android/sync: add sync_fence_create_dma John.C.Harrison
2015-12-17 17:29 ` Jesse Barnes
2015-12-11 13:11 ` [PATCH 03/13] staging/android/sync: Move sync framework out of staging John.C.Harrison
2015-12-17 17:35 ` Jesse Barnes
2015-12-21 10:03 ` Daniel Vetter
2015-12-21 14:20 ` John Harrison
2015-12-21 15:46 ` Daniel Vetter
2015-12-22 12:14 ` John Harrison
2015-12-11 13:11 ` [PATCH 04/13] android/sync: Improved debug dump to dmesg John.C.Harrison
2015-12-17 17:36 ` Jesse Barnes
2015-12-11 13:11 ` [PATCH 05/13] drm/i915: Convert requests to use struct fence John.C.Harrison
2015-12-17 17:43 ` Jesse Barnes
2016-01-04 17:20 ` Jesse Barnes
2016-01-04 20:57 ` Chris Wilson
2016-01-04 21:16 ` Jesse Barnes
2016-01-08 21:47 ` Chris Wilson
2016-01-08 21:55 ` Jesse Barnes
2015-12-11 13:11 ` [PATCH 06/13] drm/i915: Removed now redudant parameter to i915_gem_request_completed() John.C.Harrison
2015-12-11 13:11 ` [PATCH 07/13] drm/i915: Add per context timelines to fence object John.C.Harrison
2015-12-17 17:49 ` Jesse Barnes
2015-12-21 10:16 ` Chris Wilson
2015-12-11 13:11 ` [PATCH 08/13] drm/i915: Delay the freeing of requests until retire time John.C.Harrison
2015-12-11 13:11 ` [PATCH 09/13] drm/i915: Interrupt driven fences John.C.Harrison
2015-12-11 15:30 ` John Harrison
2015-12-11 16:07 ` Tvrtko Ursulin
2015-12-11 13:11 ` [PATCH 10/13] drm/i915: Updated request structure tracing John.C.Harrison
2015-12-11 13:11 ` [PATCH 11/13] android/sync: Fix reversed sense of signaled fence John.C.Harrison
2015-12-11 15:57 ` Tvrtko Ursulin
2015-12-14 11:22 ` John Harrison
2015-12-14 12:37 ` Tvrtko Ursulin
2015-12-11 13:12 ` [PATCH 12/13] drm/i915: Add sync framework support to execbuff IOCTL John.C.Harrison
2015-12-11 15:29 ` Tvrtko Ursulin
2015-12-14 11:46 ` John Harrison
2015-12-14 12:23 ` Chris Wilson
2015-12-11 13:12 ` [PATCH 13/13] drm/i915: Cache last IRQ seqno to reduce IRQ overhead John.C.Harrison
2015-12-11 14:28 ` Tvrtko Ursulin
2015-12-14 11:58 ` John Harrison
2015-12-14 12:52 ` Tvrtko Ursulin
2015-12-11 14:55 ` Chris Wilson
2015-12-11 15:35 ` John Harrison
2015-12-11 16:07 ` Chris Wilson
2016-01-08 18:47 ` [PATCH 0/7] Convert requests to use struct fence John.C.Harrison
2016-01-08 18:47 ` [PATCH 1/7] drm/i915: " John.C.Harrison
2016-01-08 21:59 ` Chris Wilson
2016-01-11 19:03 ` John Harrison [this message]
2016-01-11 22:41 ` Jesse Barnes
2016-01-08 18:47 ` [PATCH 2/7] drm/i915: Removed now redudant parameter to i915_gem_request_completed() John.C.Harrison
2016-01-11 22:43 ` Jesse Barnes
2016-01-08 18:47 ` [PATCH 3/7] drm/i915: Add per context timelines to fence object John.C.Harrison
2016-01-08 22:05 ` Chris Wilson
2016-01-11 19:03 ` John Harrison
2016-01-11 22:47 ` Jesse Barnes
2016-01-11 22:58 ` Chris Wilson
2016-01-12 11:03 ` John Harrison
2016-01-12 11:26 ` Chris Wilson
2016-01-08 18:47 ` [PATCH 4/7] drm/i915: Delay the freeing of requests until retire time John.C.Harrison
2016-01-08 22:08 ` Chris Wilson
2016-01-11 19:06 ` John Harrison
2016-01-25 11:52 ` Maarten Lankhorst
2016-01-25 12:11 ` Chris Wilson
2016-01-08 18:47 ` [PATCH 5/7] drm/i915: Interrupt driven fences John.C.Harrison
2016-01-08 22:14 ` Chris Wilson
2016-01-09 0:30 ` Chris Wilson
2016-01-08 22:46 ` Chris Wilson
2016-01-11 19:10 ` John Harrison
2016-01-11 23:01 ` Jesse Barnes
2016-01-08 18:47 ` [PATCH 6/7] drm/i915: Updated request structure tracing John.C.Harrison
2016-01-08 22:16 ` Chris Wilson
2016-01-08 18:47 ` [PATCH 7/7] drm/i915: Cache last IRQ seqno to reduce IRQ overhead John.C.Harrison
2016-01-08 22:47 ` [PATCH 0/7] Convert requests to use struct fence Chris Wilson
2016-01-11 19:15 ` John Harrison
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=5693FC8B.9000102@Intel.com \
--to=john.c.harrison@intel.com \
--cc=Intel-GFX@Lists.FreeDesktop.Org \
--cc=chris@chris-wilson.co.uk \
/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.