[PATCH] drm/vmwgfx: Fix dumb buffer leak

Zack Rusin zack.rusin at broadcom.com
Tue Jan 21 19:42:11 UTC 2025


On Tue, Jan 21, 2025 at 2:11 PM Ian Forbes <ian.forbes at broadcom.com> wrote:
>
> On Fri, Jan 17, 2025 at 1:20 PM Zack Rusin <zack.rusin at broadcom.com> wrote:
> >
> > You're going to have to explain that one in the commit message a lot
> > better because as is it doesn't make sense to me. Especially the
> > !vbo->is_dumb in vmw_bo_free.
> >
> > z
>
> The dirty tracker is freed later in vmw_bo_release when it's a
> coherent dumb_buffer.

Ah, that makes sense. If you could improve the commit message that'd
be great (just describe both issues. i.e. the buffer handle being
allocated via vmw_gem_object_create_with_handle which create a gem
handle that holds the reference and the dirty being explicitly freed
for dumb buffers in vmw_bo_release).

If you don't have anything too important on your plate right now maybe
you could try to come up with some way of testing the
creation/destruction of buffers matches. IGT could technically use the
debugfs file that Maaz added that should contain all the buffer stats.
Even for the above change it'd be a lot nicer if you could just add to
the commit message the debugfs output before and after this change
while running gnome or kde that shows the leak has been fixed.

z
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 5427 bytes
Desc: S/MIME Cryptographic Signature
URL: <https://lists.freedesktop.org/archives/dri-devel/attachments/20250121/5855bf8a/attachment-0001.bin>


More information about the dri-devel mailing list