All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dan Carpenter <dan.carpenter@oracle.com>
To: "Christian König" <christian.koenig@amd.com>
Cc: dri-devel@lists.freedesktop.org
Subject: Re: [bug report] drm/ttm: fix re-init of global structures
Date: Tue, 4 Feb 2020 17:24:58 +0300	[thread overview]
Message-ID: <20200204142457.GP11068@kadam> (raw)
In-Reply-To: <c83eb759-3ca9-64a6-d40c-32786304d82c@amd.com>

On Tue, Feb 04, 2020 at 03:03:43PM +0100, Christian König wrote:
> Am 04.02.20 um 13:57 schrieb Dan Carpenter:
> > Hello Christian König,
> > 
> > The patch bd4264112f93: "drm/ttm: fix re-init of global structures"
> > from Apr 16, 2019, leads to the following static checker warning:
> > 
> > 	drivers/gpu/drm/ttm/ttm_bo.c:1610 ttm_bo_global_release()
> > 	warn: passing freed memory 'glob'
> > 
> > drivers/gpu/drm/ttm/ttm_bo.c
> >    1591  static void ttm_bo_global_kobj_release(struct kobject *kobj)
> >    1592  {
> >    1593          struct ttm_bo_global *glob =
> >    1594                  container_of(kobj, struct ttm_bo_global, kobj);
> >    1595
> >    1596          __free_page(glob->dummy_read_page);
> >    1597  }
> >    1598
> >    1599  static void ttm_bo_global_release(void)
> >    1600  {
> >    1601          struct ttm_bo_global *glob = &ttm_bo_glob;
> >    1602
> >    1603          mutex_lock(&ttm_global_mutex);
> >    1604          if (--ttm_bo_glob_use_count > 0)
> >    1605                  goto out;
> >    1606
> >    1607          kobject_del(&glob->kobj);
> >    1608          kobject_put(&glob->kobj);
> >    1609          ttm_mem_global_release(&ttm_mem_glob);
> >    1610          memset(glob, 0, sizeof(*glob));
> >                         ^^^^^^^^^^^^^^^^^^^^^^
> > Depending on the config kobject_release() might call ttm_bo_global_kobj_release()
> > a few seconds after this memset.  Maybe put the memset into
> > ttm_bo_global_kobj_release()?
> 
> That's not possible. The object might be re-used directly after we drop the
> ttm_global_mutex.
> 

Hm...  That sucks.  If we reallocate glob->dummy_read_page before the
ttm_bo_global_kobj_release() gets called then we're toasted.

> How can we wait for the ttm_mem_global_release() to have finished?
> 

A bunch of these release functions use a completion.  But you probably
don't want a four second delay before we can re-use the struct.

regards,
dan carpenter
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

  reply	other threads:[~2020-02-04 14:25 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-04 12:57 [bug report] drm/ttm: fix re-init of global structures Dan Carpenter
2020-02-04 14:03 ` Christian König
2020-02-04 14:24   ` Dan Carpenter [this message]
2020-02-04 14:27     ` Christian König

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=20200204142457.GP11068@kadam \
    --to=dan.carpenter@oracle.com \
    --cc=christian.koenig@amd.com \
    --cc=dri-devel@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 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.