[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <aIHBO_2-hKWgb8Dq@sultan-box>
Date: Wed, 23 Jul 2025 22:14:35 -0700
From: Sultan Alsawaf <sultan@...neltoast.com>
To: Bin Du <Bin.Du@....com>
Cc: mchehab@...nel.org, hverkuil@...all.nl,
laurent.pinchart+renesas@...asonboard.com,
bryan.odonoghue@...aro.org, sakari.ailus@...ux.intel.com,
prabhakar.mahadev-lad.rj@...renesas.com,
linux-media@...r.kernel.org, linux-kernel@...r.kernel.org,
pratap.nirujogi@....com, benjamin.chan@....com, king.li@....com,
gjorgji.rosikopulos@....com, Phil.Jawich@....com,
Dominic.Antony@....com, Svetoslav.Stoilov@....com
Subject: Re: [PATCH v2 6/8] media: platform: amd: isp4 video node and buffers
handling added
On Wed, Jul 23, 2025 at 10:55:48AM -0700, Sultan Alsawaf wrote:
> On Wed, Jun 18, 2025 at 05:19:57PM +0800, Bin Du wrote:
> > +
> > + buf->bo = (void *)bo;
> > + buf->gpu_addr = gpu_addr;
> > +
> > + refcount_set(&buf->refcount, 1);
>
> This discards the refcount inc triggered from amdgpu_bo_create_isp_user() when
> it calls get_dma_buf(), leading to a use-after-free. Move this refcount_set()
> up, preferably right after vmalloc_user() or right after `buf` is allocated so
> there's no risk of this issue occurring again in the future.
Following up to correct myself: please disregard this one comment from my
previous email since it's incorrect. Sorry for any confusion.
Sultan
Powered by blists - more mailing lists