Re: [PATCH v3] misc: fastrpc: possible double-free of cctx->remote_heap

[Xingjing Deng]

This issue was also identified through static program analysis and
subsequently verified via manual inspection. I believe I have
uncovered a potential risk of abnormal execution here, hence I’m
reporting this problem.

Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> 于2026年1月26日周一 23:24写道：
>
> On Sat, Jan 17, 2026 at 10:09:59PM +0800, Xingjing Deng wrote:
> > fastrpc_init_create_static_process() may free cctx->remote_heap on the
> > err_map path but does not clear the pointer. Later, fastrpc_rpmsg_remove()
> > frees cctx->remote_heap again if it is non-NULL, which can lead to a
> > double-free if the INIT_CREATE_STATIC ioctl hits the error path and the rpmsg
> > device is subsequently removed/unbound.
> > Clear cctx->remote_heap after freeing it in the error path to prevent the
> > later cleanup from freeing it again.
> >
> > Fixes: 0871561055e66 ("misc: fastrpc: Add support for audiopd")
> > Cc: stable@xxxxxxxxxxxxxxx # 6.2+
> > Signed-off-by: Xingjing Deng <xjdeng@xxxxxxxxxxx>
> >
> > ---
> >
> > v3:
> > - Adjust the email format.
> > - Link to v2: https://lore.kernel.org/linux-arm-msm/2026011650-gravitate-happily-5d0c@gregkh/T/#t
> >
> > v2:
> > - Add Fixes: and Cc: stable@xxxxxxxxxxxxxxx.
> > - Link to v1: https://lore.kernel.org/linux-arm-msm/2026011227-casualty-rephrase-9381@gregkh/T/#t
> >
> >  drivers/misc/fastrpc.c | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/drivers/misc/fastrpc.c b/drivers/misc/fastrpc.c
> > index ee652ef01534..fb3b54e05928 100644
> > --- a/drivers/misc/fastrpc.c
> > +++ b/drivers/misc/fastrpc.c
> > @@ -1370,6 +1370,7 @@ static int fastrpc_init_create_static_process(struct fastrpc_user *fl,
> >       }
> >  err_map:
> >       fastrpc_buf_free(fl->cctx->remote_heap);
> > +     fl->cctx->remote_heap = NULL;
> >  err_name:
> >       kfree(name);
> >  err:
> > --
> > 2.25.1
> >
> >
>
> How was this found and tested?
>
> And randomly setting a pointer to null doesn't really document what is
> happening here, what would you want to see here if you were to look at
> this code in 5 years?
>
> thanks,
>
> greg k-h