Linux Framebuffer Layer development
 help / color / mirror / Atom feed
From: Helge Deller <deller@gmx.de>
To: YuanZhe Liu <25031212351@stu.xidian.edu.cn>,
	Steve Glendinning <steve.glendinning@shawell.net>
Cc: linux-fbdev@vger.kernel.org, dri-devel@lists.freedesktop.org,
	linux-kernel@vger.kernel.org, Bin Yu <byu@xidian.edu.cn>,
	Cheng Wen <wencheng@xidian.edu.cn>
Subject: Re: [PATCH] fbdev: smscufx: don't free registered framebuffer on probe error
Date: Wed, 18 Nov 2026 20:15:28 +0100	[thread overview]
Message-ID: <4a5d5e4b-e376-463e-a919-2c16a625be1d@gmx.de> (raw)
In-Reply-To: <20261102084830.9890-1-25031212351@stu.xidian.edu.cn>

On 11/2/26 09:48, YuanZhe Liu wrote:
> register_framebuffer() can return an error after the framebuffer has
> already been installed: do_register_framebuffer() bumps info->count,
> creates /dev/fbN and stores the fb_info in registered_fb[] before it
> returns the result of the console takeover, and that takeover may fail.
> ufx_usb_probe() treats any error as "nothing was registered" and unwinds
> through reset_active, which releases the fb_info and drops both
> references to struct ufx_data, freeing @dev.
> 
> The framebuffer stays registered and can still be opened, so the next
> open() of /dev/fbN reaches ufx_ops_open() with info->par pointing at
> freed memory:
> 
>    refcount_t: addition on 0; use-after-free.
>    WARNING: lib/refcount.c:25 at refcount_warn_saturate
>    Call Trace:
>     kref_get
>     ufx_ops_open    drivers/video/fbdev/smscufx.c:1033
>     fb_open         drivers/video/fbdev/core/fb_chrdev.c:351
>     chrdev_open
>     ...
>    refcount_t: underflow; use-after-free.
>    WARNING: lib/refcount.c:28 at refcount_warn_saturate
>    Call Trace:
>     ufx_ops_release drivers/video/fbdev/smscufx.c:1135
>     fb_release       drivers/video/fbdev/core/fb_chrdev.c:378
>     __fput
> 
> Reuse the "has this fb_info been registered" test that ufx_setup_modes()
> already uses: take the error path only when refcount_read(&info->count)
> says the core did not register the framebuffer.  A failed console
> takeover is not fatal for the driver, so report it and keep the
> framebuffer registered; the disconnect callback will unregister it
> normally.
> 
> Fixes: 3c8a63e22a08 ("Add support for SMSC UFX6000/7000 USB display adapters")
> Cc: stable@vger.kernel.org
> ---
>   drivers/video/fbdev/smscufx.c | 12 ++++++++++--
>   1 file changed, 10 insertions(+), 2 deletions(-)

applied.

Thanks!
Helge

      reply	other threads:[~2026-11-18 19:16 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-02  8:48 [PATCH] fbdev: smscufx: don't free registered framebuffer on probe error YuanZhe Liu
2026-11-18 19:15 ` Helge Deller [this message]

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=4a5d5e4b-e376-463e-a919-2c16a625be1d@gmx.de \
    --to=deller@gmx.de \
    --cc=25031212351@stu.xidian.edu.cn \
    --cc=byu@xidian.edu.cn \
    --cc=dri-devel@lists.freedesktop.org \
    --cc=linux-fbdev@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=steve.glendinning@shawell.net \
    --cc=wencheng@xidian.edu.cn \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox