On Fri, May 06, 2022 at 03:22:25PM +0200, Javier Martinez Canillas wrote:
Commit d258d00fb9c7 ("fbdev: efifb: Cleanup fb_info in .fb_destroy rather
than .remove") attempted to fix a use-after-free error due driver freeing
the fb_info in the .remove handler instead of doing it in .fb_destroy.
But ironically that change introduced yet another use-after-free since the
fb_info was still used after the free.
This should fix for good by freeing the fb_info at the end of the handler.
Fixes: d258d00fb9c7 ("fbdev: efifb: Cleanup fb_info in .fb_destroy rather than .remove")
are these patches going through any CI before being applied? Maybe would
be a good idea to cc intel-gfx mailing list on these fixes to have Intel
CI to pick them up for some tests?
pushed to drm-misc-fixes where the previous patch was applied.
thanks
LUcas De Marchi
Reported-by: Ville Syrjälä <ville.syrjala@xxxxxxxxxxxxxxx>
Reported-by: Andrzej Hajda <andrzej.hajda@xxxxxxxxx>
Signed-off-by: Javier Martinez Canillas <javierm@xxxxxxxxxx>
---
drivers/video/fbdev/efifb.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/video/fbdev/efifb.c b/drivers/video/fbdev/efifb.c
index cfa3dc0b4eee..b3d5f884c544 100644
--- a/drivers/video/fbdev/efifb.c
+++ b/drivers/video/fbdev/efifb.c
@@ -259,12 +259,12 @@ static void efifb_destroy(struct fb_info *info)
memunmap(info->screen_base);
}
- framebuffer_release(info);
-
if (request_mem_succeeded)
release_mem_region(info->apertures->ranges[0].base,
info->apertures->ranges[0].size);
fb_dealloc_cmap(&info->cmap);
+
+ framebuffer_release(info);
}
static const struct fb_ops efifb_ops = {
--
2.35.1
