From: Matthew Brost <matthew.brost@intel.com>
To: Matthew Auld <matthew.auld@intel.com>
Cc: intel-xe@lists.freedesktop.org,
"Himal Prasad Ghimiray" <himal.prasad.ghimiray@intel.com>,
"Tejas Upadhyay" <tejas.upadhyay@intel.com>,
"Thomas Hellström" <thomas.hellstrom@linux.intel.com>
Subject: Re: [PATCH 4/4] drm/xe/bo: add some annotations in bo_put()
Date: Tue, 10 Sep 2024 15:59:32 +0000 [thread overview]
Message-ID: <ZuBs5DazbyngXiwt@DUT025-TGLU.fm.intel.com> (raw)
In-Reply-To: <62149100-98df-46ea-87d3-9aefb060b655@intel.com>
On Tue, Sep 10, 2024 at 03:52:26PM +0100, Matthew Auld wrote:
> Hi,
>
> On 10/09/2024 14:59, Matthew Brost wrote:
> > On Tue, Sep 10, 2024 at 02:11:49PM +0100, Matthew Auld wrote:
> > > If the put() triggers bo destroy then there is at least one potential
> > > sleeping lock. Also annotate bos_lock and ggtt lock.
> > >
> > > Signed-off-by: Matthew Auld <matthew.auld@intel.com>
> > > Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
> > > Cc: Tejas Upadhyay <tejas.upadhyay@intel.com>
> > > Cc: "Thomas Hellström" <thomas.hellstrom@linux.intel.com>
> > > ---
> > > drivers/gpu/drm/xe/xe_bo.c | 14 ++++++++++++++
> > > drivers/gpu/drm/xe/xe_bo.h | 6 +-----
> > > 2 files changed, 15 insertions(+), 5 deletions(-)
> > >
> > > diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> > > index 53948a965404..4176b302c38d 100644
> > > --- a/drivers/gpu/drm/xe/xe_bo.c
> > > +++ b/drivers/gpu/drm/xe/xe_bo.c
> > > @@ -2320,6 +2320,20 @@ void xe_bo_put_commit(struct llist_head *deferred)
> > > drm_gem_object_free(&bo->ttm.base.refcount);
> > > }
> > > +void xe_bo_put(struct xe_bo *bo)
> >
> > These are helpful but thinking we might need a version without asserts
> > that can be used in code paths where we know that we have another ref to
> > the BO. I can already think of a case in the SVM code where this would
> > be needed. In that version maybe we assert the ref count > 1, granted
> > that races but probably the best we can there.
>
> Yeah, adding another put() variant or similar for that case should be
> totally fine I think.
>
CI seems to indicate this change is good with the current code base,
will add another variant if / when it is needed.
With that:
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
> >
> > What do you think?
> >
> > Asserts look good though.
> >
> > Matt
> >
> > > +{
> > > + might_sleep();
> > > + if (bo) {
> > > +#ifdef CONFIG_PROC_FS
> > > + if (bo->client)
> > > + might_lock(&bo->client->bos_lock);
> > > +#endif
> > > + if (bo->ggtt_node && bo->ggtt_node->ggtt)
> > > + might_lock(&bo->ggtt_node->ggtt->lock);
> > > + drm_gem_object_put(&bo->ttm.base);
> > > + }
> > > +}
> > > +
> > > /**
> > > * xe_bo_dumb_create - Create a dumb bo as backing for a fb
> > > * @file_priv: ...
> > > diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h
> > > index dbfb3209615d..6e4be52306df 100644
> > > --- a/drivers/gpu/drm/xe/xe_bo.h
> > > +++ b/drivers/gpu/drm/xe/xe_bo.h
> > > @@ -126,11 +126,7 @@ static inline struct xe_bo *xe_bo_get(struct xe_bo *bo)
> > > return bo;
> > > }
> > > -static inline void xe_bo_put(struct xe_bo *bo)
> > > -{
> > > - if (bo)
> > > - drm_gem_object_put(&bo->ttm.base);
> > > -}
> > > +void xe_bo_put(struct xe_bo *bo);
> > > static inline void __xe_bo_unset_bulk_move(struct xe_bo *bo)
> > > {
> > > --
> > > 2.46.0
> > >
next prev parent reply other threads:[~2024-09-10 16:01 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-10 13:11 [PATCH 1/4] drm/xe/client: fix deadlock in show_meminfo() Matthew Auld
2024-09-10 13:11 ` [PATCH 2/4] drm/xe/client: add missing bo locking " Matthew Auld
2024-09-10 14:16 ` Matthew Brost
2024-09-11 5:39 ` Upadhyay, Tejas
2024-09-11 8:35 ` Matthew Auld
2024-09-11 9:40 ` Upadhyay, Tejas
2024-09-10 13:11 ` [PATCH 3/4] drm/xe/client: use mem_type from the current resource Matthew Auld
2024-09-10 14:18 ` Matthew Brost
2024-09-11 5:45 ` Upadhyay, Tejas
2024-09-10 13:11 ` [PATCH 4/4] drm/xe/bo: add some annotations in bo_put() Matthew Auld
2024-09-10 13:59 ` Matthew Brost
2024-09-10 14:52 ` Matthew Auld
2024-09-10 15:59 ` Matthew Brost [this message]
2024-09-10 14:49 ` Matthew Brost
2024-09-10 15:03 ` Matthew Auld
2024-09-10 15:26 ` Matthew Brost
2024-09-10 15:29 ` Matthew Brost
2024-09-11 5:40 ` Upadhyay, Tejas
2024-09-10 13:29 ` ✓ CI.Patch_applied: success for series starting with [1/4] drm/xe/client: fix deadlock in show_meminfo() Patchwork
2024-09-10 13:30 ` ✓ CI.checkpatch: " Patchwork
2024-09-10 13:31 ` ✓ CI.KUnit: " Patchwork
2024-09-10 13:48 ` ✓ CI.Build: " Patchwork
2024-09-10 13:53 ` ✓ CI.Hooks: " Patchwork
2024-09-10 13:55 ` [PATCH 1/4] " Matthew Brost
2024-09-10 13:56 ` ✓ CI.checksparse: success for series starting with [1/4] " Patchwork
2024-09-10 14:39 ` ✓ CI.BAT: " Patchwork
2024-09-10 15:45 ` ✗ CI.FULL: failure " Patchwork
2024-09-11 5:19 ` [PATCH 1/4] " Upadhyay, Tejas
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=ZuBs5DazbyngXiwt@DUT025-TGLU.fm.intel.com \
--to=matthew.brost@intel.com \
--cc=himal.prasad.ghimiray@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.auld@intel.com \
--cc=tejas.upadhyay@intel.com \
--cc=thomas.hellstrom@linux.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