Linux USB
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Xinsheng Zhu" <xinsheng.zhu.ny@gmail.com>, linux-usb@vger.kernel.org
Subject: Re: [PATCH usb-next] usb: gadget: f_ncm: prevent double-free of OS descriptor table
Date: Mon, 09 Nov 2026 05:41:08 +0000	[thread overview]
Message-ID: <sashiko-outbox-204450@kernel.org> (raw)
In-Reply-To: <20261109053440.49156-1-xinsheng.zhu.ny@gmail.com>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Low] usb: gadget: f_ncm: missing Fixes trailer for ncm_unbind()
  double-free fix
It also noticed 1 related pre-existing issue this patch does not
introduce:
- [Medium] usb: gadget: f_ncm: memory leak in ncm_unbind() [1]
--

commit 1d3d70b425a5f8ec0afdd69a5b560a6e1867217a
Author: Xinsheng Zhu <xinsheng.zhu.ny@gmail.com>

> usb: gadget: f_ncm: prevent double-free of OS descriptor table
>
> ncm_bind() allocates an OS descriptor table only when
> cdev->use_os_string is enabled. However, ncm_unbind() frees the
> table and resets f->os_desc_n without clearing f->os_desc_table,
> leaving a dangling pointer.
>
> When a ConfigFS gadget is unbound, the function object can be reused
> for a subsequent bind. The OS descriptor support can be toggled
> through the ConfigFS os_desc/use attribute, which determines
> cdev->use_os_string on the next bind. If this flag changes from true
> to false between binds, ncm_bind() skips the allocation and leaves
> f->os_desc_table pointing to the previously freed memory.
>
> A subsequent unbind may therefore free the stale pointer again,
> resulting in a double-free.
>
> Set f->os_desc_table to NULL after freeing it to prevent the stale
> pointer from being reused across bind/unbind cycles.
>
> Signed-off-by: Xinsheng Zhu <xinsheng.zhu.ny@gmail.com>

[Severity: Low]
This is a bug fix for OS descriptor handling in ncm_unbind(). Should the
commit message include:

  Fixes: 793409292382 ("usb: gadget: f_ncm: Add OS descriptor support")

> diff --git a/drivers/usb/gadget/function/f_ncm.c b/drivers/usb/gadget/function/f_ncm.c
> index 500c9c19c78b4..685152ec9a5d0 100644
> --- a/drivers/usb/gadget/function/f_ncm.c
> +++ b/drivers/usb/gadget/function/f_ncm.c
> @@ -1741,6 +1741,7 @@ static void ncm_unbind(struct usb_configuration *c, struct usb_function *f)
>  	hrtimer_cancel(&ncm->task_timer);
>  
>  	kfree(f->os_desc_table);
> +	f->os_desc_table = NULL;
>  	f->os_desc_n = 0;

[Severity: Medium]
This problem wasn't introduced by this patch, but does ncm_unbind() leak
ncm->skb_tx_data and ncm->skb_tx_ndp [1]?

ncm_wrap_ntb() allocates skb_tx_data and skb_tx_ndp to aggregate outgoing
packets and arms task_timer to flush them. If the gadget is unbound before
task_timer fires or the NTB fills up, hrtimer_cancel(&ncm->task_timer) stops
the pending flush without freeing those skbs, and ncm_free() later frees ncm
directly.

[1] https://sashiko.dev/#/bug/linux-c8af1819-7ebd-4e08-89f6-265e1b0462fd

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261109053440.49156-1-xinsheng.zhu.ny@gmail.com?part=1

      reply	other threads:[~2026-11-09  5:41 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-09  5:34 [PATCH usb-next] usb: gadget: f_ncm: prevent double-free of OS descriptor table Xinsheng Zhu
2026-11-09  5:41 ` sashiko-bot [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=sashiko-outbox-204450@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=linux-usb@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=xinsheng.zhu.ny@gmail.com \
    /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