From: Daniel Vetter <daniel@ffwll.ch>
To: Matthew Auld <matthew.auld@intel.com>
Cc: intel-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Subject: Re: [Intel-gfx] [PATCH] drm/i915: return DRIVER_NAME for the fence driver name
Date: Wed, 16 Jun 2021 15:32:00 +0200 [thread overview]
Message-ID: <YMn9UDHpMA6MKiSL@phenom.ffwll.local> (raw)
In-Reply-To: <20210616122833.332954-1-matthew.auld@intel.com>
On Wed, Jun 16, 2021 at 01:28:33PM +0100, Matthew Auld wrote:
> The first tracepoint for a request is trace_dma_fence_init which is
> called in the ctor before we have properly setup the request->engine. So
> if it's a non-recycled request the rq->engine might be NULL, or some
> garbage value, which leads to a crash.
I'd hammer it more in here that we cannot move the dma_fence_init out of
the slab constructor (aka ctor), because that breaks the magic request
recycling through SLAB_TYPESAFE_BY_RCU. You do explain it all, but in a
rather dense fashion. More verbosity here would be good.
> Since we are not permitted to use kmem_cache_zalloc() here with
This part I'm not really clear on ... what would zalloc break? Ideally
references to the commits/bugs/functions in the code that would blow up
would be good here.
> SLAB_TYPESAFE_BY_RCU, one approach is simply to return DRIVER_NAME. We
> can then revisit this later if we decide to get rid of
> SLAB_TYPESAFE_BY_RCU.
>
> Fixes: 855e39e65cfc ("drm/i915: Initialise basic fence before acquiring seqno")
> Signed-off-by: Matthew Auld <matthew.auld@intel.com>
> Cc: Michael Mason <michael.w.mason@intel.com>
> Cc: Daniel Vetter <daniel@ffwll.ch>
As a stop-gab the code change looks reasonable, but I think the commit
message needs work and much more of the history/background we're hitting
here.
Thanks, Daniel
> ---
> drivers/gpu/drm/i915/i915_request.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
> index 1014c71cf7f5..55fa94bde22e 100644
> --- a/drivers/gpu/drm/i915/i915_request.c
> +++ b/drivers/gpu/drm/i915/i915_request.c
> @@ -61,7 +61,7 @@ static struct i915_global_request {
>
> static const char *i915_fence_get_driver_name(struct dma_fence *fence)
> {
> - return dev_name(to_request(fence)->engine->i915->drm.dev);
> + return DRIVER_NAME;
> }
>
> static const char *i915_fence_get_timeline_name(struct dma_fence *fence)
> --
> 2.26.3
>
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2021-06-16 13:32 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-16 12:28 [Intel-gfx] [PATCH] drm/i915: return DRIVER_NAME for the fence driver name Matthew Auld
2021-06-16 13:32 ` Daniel Vetter [this message]
2021-06-16 19:19 ` [Intel-gfx] ✓ Fi.CI.BAT: success for " Patchwork
2021-06-16 22:40 ` [Intel-gfx] ✓ Fi.CI.IGT: " Patchwork
2021-06-22 21:32 ` [Intel-gfx] [PATCH] " Mason, Michael W
2021-06-23 10:12 ` Thomas Zimmermann
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=YMn9UDHpMA6MKiSL@phenom.ffwll.local \
--to=daniel@ffwll.ch \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=matthew.auld@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox