From: Daniel Vetter <daniel.vetter@ffwll.ch>
To: intel-gfx <intel-gfx@lists.freedesktop.org>
Cc: linux-kernel@vger.kernel.org, dri-devel@lists.freedesktop.org,
Daniel Vetter <daniel.vetter@ffwll.ch>
Subject: [PATCH 00/13] pwrite/pread rework
Date: Sun, 6 Nov 2011 20:13:47 +0100 [thread overview]
Message-ID: <1320606840-21132-1-git-send-email-daniel.vetter@ffwll.ch> (raw)
Hi all,
I've finally managed to clean up my pwrite/pread rework. Major changes since the
first submission:
- fixed a bunch of bugs, some of them discovered by the intel-gpu-tools gem test
suite
- extracted the clflush helper into common drm code as suggested by Chris
Wilson.
- extended the prefault helpers in pagemap.h instead of rolling our own. Also
add a patch to not prefault for writes (which change the userspace memory)
before we're committed to the write. Both address issues raised by Keith
Packard.
Why this is cool stuff:
- fixes the spurious -EFAULTS when handing in a pointer to a gem gtt mappings.
See https://bugs.freedesktop.org/show_bug.cgi?id=38115
- Kills a bunch of code by dropping the practically useless ranged cpu read
domain tracking and integrating the slow paths into the fast paths.
- Fixes pwrite/pread to use snooped cpu access on llc machines, hereby decently
speeding up microbenchmarks.
- Paves the way for further clever tricks to reduce pressure on the mappable gtt
area.
Reviews and comments highly welcome.
Yours, Daniel
Daniel Vetter (13):
drm/i915: fall through pwrite_gtt_slow to the shmem slow path
drm/i915: rewrite shmem_pwrite_slow to use copy_from_user
drm/i915: rewrite shmem_pread_slow to use copy_to_user
drm/i915: merge shmem_pwrite slow&fast-path
drm/i915: merge shmem_pread slow&fast-path
drm: add helper to clflush a virtual address range
drm/i915: move clflushing into shmem_pread
drm/i915: kill ranged cpu read domain support
drm/i915: don't use gtt_pwrite on LLC cached objects
drm/i915: don't call shmem_read_mapping unnecessarily
mm: extend prefault helpers to fault in more than PAGE_SIZE
drm/i915: drop gtt slowpath
drm/i915: don't clobber userspace memory before commiting to the
pread
drivers/gpu/drm/drm_cache.c | 23 ++
drivers/gpu/drm/i915/i915_drv.h | 7 -
drivers/gpu/drm/i915/i915_gem.c | 824 +++++++++++----------------------------
include/drm/drmP.h | 1 +
include/linux/pagemap.h | 28 +-
5 files changed, 273 insertions(+), 610 deletions(-)
--
1.7.6.4
next reply other threads:[~2011-11-06 19:15 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-11-06 19:13 Daniel Vetter [this message]
2011-11-06 19:13 ` [PATCH 01/13] drm/i915: fall through pwrite_gtt_slow to the shmem slow path Daniel Vetter
2011-11-21 3:09 ` [Intel-gfx] " Ben Widawsky
2011-11-21 10:20 ` Chris Wilson
2011-11-06 19:13 ` [PATCH 02/13] drm/i915: rewrite shmem_pwrite_slow to use copy_from_user Daniel Vetter
2011-11-21 5:56 ` [Intel-gfx] " Ben Widawsky
2011-11-21 16:02 ` Daniel Vetter
2011-11-21 17:55 ` Ben Widawsky
2011-11-21 18:43 ` Ben Widawsky
2011-11-06 19:13 ` [PATCH 03/13] drm/i915: rewrite shmem_pread_slow to use copy_to_user Daniel Vetter
2011-11-06 19:13 ` [PATCH 04/13] drm/i915: merge shmem_pwrite slow&fast-path Daniel Vetter
2011-11-06 19:13 ` [PATCH 05/13] drm/i915: merge shmem_pread slow&fast-path Daniel Vetter
2011-11-06 19:13 ` [PATCH 06/13] drm: add helper to clflush a virtual address range Daniel Vetter
2011-11-21 19:46 ` [Intel-gfx] " Ben Widawsky
2011-11-06 19:13 ` [PATCH 07/13] drm/i915: move clflushing into shmem_pread Daniel Vetter
2011-11-06 19:13 ` [PATCH 08/13] drm/i915: kill ranged cpu read domain support Daniel Vetter
2011-11-06 19:13 ` [PATCH 09/13] drm/i915: don't use gtt_pwrite on LLC cached objects Daniel Vetter
2011-11-06 21:16 ` Chris Wilson
2011-11-06 22:19 ` Daniel Vetter
2011-11-06 19:13 ` [PATCH 10/13] drm/i915: don't call shmem_read_mapping unnecessarily Daniel Vetter
2011-11-06 19:13 ` [PATCH 11/13] mm: extend prefault helpers to fault in more than PAGE_SIZE Daniel Vetter
2011-11-06 22:24 ` Chris Wilson
2011-11-06 19:13 ` [PATCH 12/13] drm/i915: drop gtt slowpath Daniel Vetter
2011-11-06 19:14 ` [PATCH 13/13] drm/i915: don't clobber userspace memory before commiting to the pread 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=1320606840-21132-1-git-send-email-daniel.vetter@ffwll.ch \
--to=daniel.vetter@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=linux-kernel@vger.kernel.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
all inboxes | Powered by JetHome®