From: Dave Airlie <airlied@gmail.com>
To: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Intel Graphics Development <intel-gfx@lists.freedesktop.org>
Subject: Re: [Intel-gfx] [CI] drm/i915: Filter wake_flags passed to default_wake_function
Date: Mon, 3 Aug 2020 17:16:40 +1000 [thread overview]
Message-ID: <CAPM=9twnbntmtCqoKa6RsDeKEwrjjJBi=St7mDtMXtP9Uy+EZA@mail.gmail.com> (raw)
In-Reply-To: <20200728152144.1100-1-chris@chris-wilson.co.uk>
On Wed, 29 Jul 2020 at 01:21, Chris Wilson <chris@chris-wilson.co.uk> wrote:
>
> The flags passed to the wait_entry.func are passed onwards to
> try_to_wake_up(), which has a very particular interpretation for its
> wake_flags. In particular, beyond the published WF_SYNC, it has a few
> internal flags as well. Since we passed the fence->error down the chain
> via the flags argument, these ended up in the default_wake_function
> confusing the kernel/sched.
>
> Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/2110
> Fixes: ef4688497512 ("drm/i915: Propagate fence errors")
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Cc: <stable@vger.kernel.org> # v5.4+
> Reviewed-by: Matthew Auld <matthew.auld@intel.com>
> ---
> drivers/gpu/drm/i915/i915_sw_fence.c | 10 +++++++---
> 1 file changed, 7 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c
> index 295b9829e2da..4cd2038cbe35 100644
> --- a/drivers/gpu/drm/i915/i915_sw_fence.c
> +++ b/drivers/gpu/drm/i915/i915_sw_fence.c
> @@ -164,9 +164,13 @@ static void __i915_sw_fence_wake_up_all(struct i915_sw_fence *fence,
>
> do {
> list_for_each_entry_safe(pos, next, &x->head, entry) {
> - pos->func(pos,
> - TASK_NORMAL, fence->error,
> - &extra);
> + int wake_flags;
> +
> + wake_flags = fence->error;
> + if (pos->func == autoremove_wake_function)
> + wake_flags = 0;
> +
> + pos->func(pos, TASK_NORMAL, wake_flags, &extra);
> }
>
> if (list_empty(&extra))
This seems to be heading for my tree at the moment, there is only one
place in the kernel where someone compares pos->func with
autoremove_wake_function, and it's in this file.
This seems horribly brittle, can we at least make this better in -next
even if we have to have this fix in fixes?
I also have to question the whole raison d'etre for i915_sw_fence,
it's initial commit says it was meant to be a core kernel struct, but
I haven't seen any effort on behalf of i915 team to make that happen,
I expect when that is attempted the whole thing will get shredded for
layering violations like the above.
Dave.
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
prev parent reply other threads:[~2020-08-03 7:16 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-28 15:21 [Intel-gfx] [CI] drm/i915: Filter wake_flags passed to default_wake_function Chris Wilson
2020-07-28 18:26 ` [Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915: Filter wake_flags passed to default_wake_function (rev2) Patchwork
2020-08-03 7:16 ` Dave Airlie [this message]
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='CAPM=9twnbntmtCqoKa6RsDeKEwrjjJBi=St7mDtMXtP9Uy+EZA@mail.gmail.com' \
--to=airlied@gmail.com \
--cc=chris@chris-wilson.co.uk \
--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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).