[PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Xu Yang]

From: Xu Yang <xu.yang_2@nxp.com>

kstrndup() returns NULL on allocation failure. The condition was
checking 'if (tmp)' to detect failure, but this is inverted — it
would treat a successful allocation as an error and return -ENOMEM
while leaking the string, and proceed with a NULL pointer on failure.

Fix by changing the condition to 'if (!tmp)'.

Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
Cc: stable@vger.kernel.org
Assisted-by: Claude:claude-sonnet-4-6
Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
---
 drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
index 5d201a2e30e7..e9f2632ce785 100644
--- a/drivers/usb/gadget/function/f_uac1_legacy.c
+++ b/drivers/usb/gadget/function/f_uac1_legacy.c
@@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
 		goto end;						\
 									\
 	tmp = kstrndup(page, len, GFP_KERNEL);				\
-	if (tmp) {							\
+	if (!tmp) {							\
 		ret = -ENOMEM;						\
 		goto end;						\
 	}								\
-- 
2.34.1

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[sashiko-bot]

Thank you for your contribution! Sashiko AI review found 6 potential issue(s) to consider:

Pre-existing issues:
- [High] Missing workqueue cancellation and queue cleanup before freeing `audio` context.
- [High] Unbounded memory leak of `usb_request` objects and buffers upon alternate setting changes.
- [High] Improper handling of `f_audio_buffer_alloc` failure leads to subsequent UAF or ERR_PTR dereference.
- [High] `audio->copy_buf` is not cleared when queued on `alt=0`, causing double list additions.
- [High] The playback workqueue only processes one buffer per run, stranding/leaking buffers.
- [High] A 1-byte out-of-bounds write of a NUL byte occurs in `f_uac1_opts_##name##_show` when reading a string attribute of maximum length.
--

commit 1d07b1a645b0a1199b66ed32d3217e953cef8087
Author: Xu Yang <xu.yang_2@nxp.com>
Date:   Thu Jun 25 11:31:54 2026 +0000

usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()
    
This patch fixes an inverted NULL check on the result of kstrndup() in
f_uac1_opts_##name##_store(). The original code treated a successful
allocation as an error and would return -ENOMEM while proceeding with a
NULL pointer on failure.

> diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> index 5d201a2e30e7f..e9f2632ce7852 100644
> --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
>  		goto end;						\
>  									\
>  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> -	if (tmp) {							\
> +	if (!tmp) {							\
>  		ret = -ENOMEM;						\
>  		goto end;						\
>  	}								\

[Severity: High]
This is a pre-existing issue, but does f_uac1_opts_##name##_show() have an
out-of-bounds write here?

drivers/usb/gadget/function/f_uac1_legacy.c:UAC1_STR_ATTRIBUTE() {
    ...
static ssize_t f_uac1_opts_##name##_show(struct config_item *item,	\
					 char *page)			\
{									\
...
	result = sprintf(page, "%s\n", opts->name);			\
    ...
}

If a user writes an attribute of exactly 4095 bytes, kstrndup stores it with
a length of 4095. During a subsequent read, sprintf() writes the 4095
characters, appends a newline (byte 4096), and then writes the NUL terminator
at index 4096, which overflows the 4096-byte page allocation by 1 byte.
Could this use sysfs_emit() or scnprintf() instead?

[Severity: High]
This is a pre-existing issue, but does f_audio_free() leak resources or cause
a use-after-free?

drivers/usb/gadget/function/f_uac1_legacy.c:f_audio_free() {
    ...
static void f_audio_free(struct usb_function *f)
{
	struct f_audio *audio = func_to_audio(f);
	struct f_uac1_legacy_opts *opts;

	gaudio_cleanup(&audio->card);
	opts = container_of(f->fi, struct f_uac1_legacy_opts, func_inst);
	kfree(audio);
    ...
}

The audio structure is freed without cancelling the asynchronous
audio->playback_work or draining audio->play_queue. If the workqueue is
pending or executing, it will dereference the freed audio object. Any
elements left in the queue are also leaked.

[Severity: High]
This is a pre-existing issue, but does f_audio_set_alt() leak usb_request
objects and buffers?

drivers/usb/gadget/function/f_uac1_legacy.c:f_audio_set_alt() {
    ...
			for (i = 0; i < req_count && err == 0; i++) {
				req = usb_ep_alloc_request(out_ep, GFP_ATOMIC);
				if (req) {
					req->buf = kzalloc(req_buf_size,
							GFP_ATOMIC);
    ...
}

When the interface is set to alt == 1, requests and buffers are dynamically
allocated. However, when the host stops the stream by switching to alt == 0,
these are never freed. Repeated toggling of the alternate setting could
quickly exhaust system memory.

[Severity: High]
This is a pre-existing issue, but does f_audio_out_ep_complete() improperly
handle f_audio_buffer_alloc() failures?

drivers/usb/gadget/function/f_uac1_legacy.c:f_audio_out_ep_complete() {
    ...
		schedule_work(&audio->playback_work);
		copy_buf = f_audio_buffer_alloc(audio_buf_size);
		if (IS_ERR(copy_buf))
			return -ENOMEM;
	}

	memcpy(copy_buf->buf + copy_buf->actual, req->buf, req->actual);
    ...
}

If f_audio_buffer_alloc() fails, it returns -ENOMEM immediately, but
audio->copy_buf is left pointing to the old buffer that was just queued for
playback. A subsequent completion would write to this old buffer, potentially
causing a use-after-free.

[Severity: High]
This is a pre-existing issue, but does f_audio_set_alt() incorrectly leave
audio->copy_buf populated when queued?

drivers/usb/gadget/function/f_uac1_legacy.c:f_audio_set_alt() {
    ...
		} else {
			struct f_audio_buf *copy_buf = audio->copy_buf;
			if (copy_buf) {
				list_add_tail(&copy_buf->list,
						&audio->play_queue);
				schedule_work(&audio->playback_work);
			}
		}
    ...
}

The copy_buf is added to the play_queue, but audio->copy_buf is not cleared.
If the host issues consecutive SET_INTERFACE 0 requests, the driver will
attempt to add the same buffer to the list again, which corrupts the
doubly-linked list.

[Severity: High]
This is a pre-existing issue, but does the playback workqueue strand buffers in
the play queue?

drivers/usb/gadget/function/f_uac1_legacy.c:f_audio_playback_work() {
    ...
	spin_lock_irq(&audio->lock);
	if (list_empty(&audio->play_queue)) {
		spin_unlock_irq(&audio->lock);
		return;
	}
	play_buf = list_first_entry(&audio->play_queue,
			struct f_audio_buf, list);
	list_del(&play_buf->list);
	spin_unlock_irq(&audio->lock);
    ...
}

The work item pulls and processes exactly one buffer per execution. If
f_audio_out_ep_complete() adds multiple buffers quickly, schedule_work() may
coalesce the work items. Since there is no while(!list_empty()) loop, any
coalesced buffers are left stranded in the queue, leaking them when the stream
stops.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260625113154.1954813-1-xu.yang_2@oss.nxp.com?part=1

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Frank Li]

On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
>
> From: Xu Yang <xu.yang_2@nxp.com>
>
> kstrndup() returns NULL on allocation failure. The condition was
> checking 'if (tmp)' to detect failure, but this is inverted — it
> would treat a successful allocation as an error and return -ENOMEM
> while leaking the string, and proceed with a NULL pointer on failure.
>
> Fix by changing the condition to 'if (!tmp)'.
>
> Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> Cc: stable@vger.kernel.org
> Assisted-by: Claude:claude-sonnet-4-6
> Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> ---

Reviewed-by: Frank Li <Frank.Li@nxp.com>

>  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> index 5d201a2e30e7..e9f2632ce785 100644
> --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item, \
>                 goto end;                                               \
>                                                                         \
>         tmp = kstrndup(page, len, GFP_KERNEL);                          \
> -       if (tmp) {                                                      \
> +       if (!tmp) {                                                     \
>                 ret = -ENOMEM;                                          \
>                 goto end;                                               \
>         }                                                               \
> --
> 2.34.1
>
>

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Greg KH]

On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> From: Xu Yang <xu.yang_2@nxp.com>
> 
> kstrndup() returns NULL on allocation failure. The condition was
> checking 'if (tmp)' to detect failure, but this is inverted — it
> would treat a successful allocation as an error and return -ENOMEM
> while leaking the string, and proceed with a NULL pointer on failure.
> 
> Fix by changing the condition to 'if (!tmp)'.
> 
> Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> Cc: stable@vger.kernel.org
> Assisted-by: Claude:claude-sonnet-4-6
> Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> ---
>  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> index 5d201a2e30e7..e9f2632ce785 100644
> --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
>  		goto end;						\
>  									\
>  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> -	if (tmp) {							\
> +	if (!tmp) {							\
>  		ret = -ENOMEM;						\
>  		goto end;						\
>  	}								\
> -- 
> 2.34.1
> 
> 

So this has never worked?  How has no one ever noticed that in the
decade this has been in the kernel?  Does that mean that no one uses
this file/driver and we can just remove it?

thanks,

greg k-h

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Frank Li]

On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > From: Xu Yang <xu.yang_2@nxp.com>
> >
> > kstrndup() returns NULL on allocation failure. The condition was
> > checking 'if (tmp)' to detect failure, but this is inverted — it
> > would treat a successful allocation as an error and return -ENOMEM
> > while leaking the string, and proceed with a NULL pointer on failure.
> >
> > Fix by changing the condition to 'if (!tmp)'.
> >
> > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > Cc: stable@vger.kernel.org
> > Assisted-by: Claude:claude-sonnet-4-6
> > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > ---
> >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > index 5d201a2e30e7..e9f2632ce785 100644
> > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> >  		goto end;						\
> >  									\
> >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > -	if (tmp) {							\
> > +	if (!tmp) {							\
> >  		ret = -ENOMEM;						\
> >  		goto end;						\
> >  	}								\
> > --
> > 2.34.1
> >
> >
>
> So this has never worked?  How has no one ever noticed that in the
> decade this has been in the kernel?  Does that mean that no one uses
> this file/driver and we can just remove it?

I think just no one use sys interface to change name.

Frank

>
> thanks,
>
> greg k-h
>

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Greg KH]

On Wed, Jul 08, 2026 at 03:54:03PM -0500, Frank Li wrote:
> On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > > From: Xu Yang <xu.yang_2@nxp.com>
> > >
> > > kstrndup() returns NULL on allocation failure. The condition was
> > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > would treat a successful allocation as an error and return -ENOMEM
> > > while leaking the string, and proceed with a NULL pointer on failure.
> > >
> > > Fix by changing the condition to 'if (!tmp)'.
> > >
> > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > Cc: stable@vger.kernel.org
> > > Assisted-by: Claude:claude-sonnet-4-6
> > > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > > ---
> > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > index 5d201a2e30e7..e9f2632ce785 100644
> > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > >  		goto end;						\
> > >  									\
> > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > -	if (tmp) {							\
> > > +	if (!tmp) {							\
> > >  		ret = -ENOMEM;						\
> > >  		goto end;						\
> > >  	}								\
> > > --
> > > 2.34.1
> > >
> > >
> >
> > So this has never worked?  How has no one ever noticed that in the
> > decade this has been in the kernel?  Does that mean that no one uses
> > this file/driver and we can just remove it?
> 
> I think just no one use sys interface to change name.

So it can just be removed?  Please submit a patch to do that instead.

Ideally this whole file can be removed, does anyone still use it?

thanks,

greg k-h

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Xu Yang]

On Thu, Jul 09, 2026 at 07:33:03AM +0200, Greg KH wrote:
> On Wed, Jul 08, 2026 at 03:54:03PM -0500, Frank Li wrote:
> > On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > > > From: Xu Yang <xu.yang_2@nxp.com>
> > > >
> > > > kstrndup() returns NULL on allocation failure. The condition was
> > > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > > would treat a successful allocation as an error and return -ENOMEM
> > > > while leaking the string, and proceed with a NULL pointer on failure.
> > > >
> > > > Fix by changing the condition to 'if (!tmp)'.
> > > >
> > > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > > Cc: stable@vger.kernel.org
> > > > Assisted-by: Claude:claude-sonnet-4-6
> > > > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > > > ---
> > > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > index 5d201a2e30e7..e9f2632ce785 100644
> > > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > > >  		goto end;						\
> > > >  									\
> > > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > > -	if (tmp) {							\
> > > > +	if (!tmp) {							\
> > > >  		ret = -ENOMEM;						\
> > > >  		goto end;						\
> > > >  	}								\
> > > > --
> > > > 2.34.1
> > > >
> > > >
> > >
> > > So this has never worked?  How has no one ever noticed that in the
> > > decade this has been in the kernel?  Does that mean that no one uses
> > > this file/driver and we can just remove it?
> > 
> > I think just no one use sys interface to change name.
> 
> So it can just be removed?  Please submit a patch to do that instead.
> 
> Ideally this whole file can be removed, does anyone still use it?

This legacy driver is still built in kernel in multi_v7_defconfig: 

~/work/usb$ grep -nr CONFIG_USB_CONFIGFS_F_UAC1_LEGACY arch/*/configs/
arch/arm/configs/multi_v7_defconfig:939:CONFIG_USB_CONFIGFS_F_UAC1_LEGACY=y

It seems we can't simply remove it. 

BTW, do we have other ways to figure out if it can be removed?

Thanks,
Xu Yang

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Greg KH]

On Thu, Jul 09, 2026 at 02:23:51PM +0800, Xu Yang wrote:
> On Thu, Jul 09, 2026 at 07:33:03AM +0200, Greg KH wrote:
> > On Wed, Jul 08, 2026 at 03:54:03PM -0500, Frank Li wrote:
> > > On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > > > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > > > > From: Xu Yang <xu.yang_2@nxp.com>
> > > > >
> > > > > kstrndup() returns NULL on allocation failure. The condition was
> > > > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > > > would treat a successful allocation as an error and return -ENOMEM
> > > > > while leaking the string, and proceed with a NULL pointer on failure.
> > > > >
> > > > > Fix by changing the condition to 'if (!tmp)'.
> > > > >
> > > > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > > > Cc: stable@vger.kernel.org
> > > > > Assisted-by: Claude:claude-sonnet-4-6
> > > > > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > > > > ---
> > > > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > >
> > > > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > index 5d201a2e30e7..e9f2632ce785 100644
> > > > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > > > >  		goto end;						\
> > > > >  									\
> > > > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > > > -	if (tmp) {							\
> > > > > +	if (!tmp) {							\
> > > > >  		ret = -ENOMEM;						\
> > > > >  		goto end;						\
> > > > >  	}								\
> > > > > --
> > > > > 2.34.1
> > > > >
> > > > >
> > > >
> > > > So this has never worked?  How has no one ever noticed that in the
> > > > decade this has been in the kernel?  Does that mean that no one uses
> > > > this file/driver and we can just remove it?
> > > 
> > > I think just no one use sys interface to change name.
> > 
> > So it can just be removed?  Please submit a patch to do that instead.
> > 
> > Ideally this whole file can be removed, does anyone still use it?
> 
> This legacy driver is still built in kernel in multi_v7_defconfig: 
> 
> ~/work/usb$ grep -nr CONFIG_USB_CONFIGFS_F_UAC1_LEGACY arch/*/configs/
> arch/arm/configs/multi_v7_defconfig:939:CONFIG_USB_CONFIGFS_F_UAC1_LEGACY=y

defconfig doesn't mean much.  Perhaps look at some of the Android
configs?

> It seems we can't simply remove it. 

Why?

> BTW, do we have other ways to figure out if it can be removed?

Remove it and see who complains?  :)

thanks,

greg k-h

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[David Laight]

On Wed, 8 Jul 2026 15:54:03 -0500
Frank Li <Frank.li@oss.nxp.com> wrote:

> On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:  
> > > From: Xu Yang <xu.yang_2@nxp.com>
> > >
> > > kstrndup() returns NULL on allocation failure. The condition was
> > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > would treat a successful allocation as an error and return -ENOMEM
> > > while leaking the string, and proceed with a NULL pointer on failure.
> > >
> > > Fix by changing the condition to 'if (!tmp)'.
> > >
> > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > Cc: stable@vger.kernel.org
> > > Assisted-by: Claude:claude-sonnet-4-6
> > > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > > ---
> > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > index 5d201a2e30e7..e9f2632ce785 100644
> > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > >  		goto end;						\
> > >  									\
> > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > -	if (tmp) {							\
> > > +	if (!tmp) {							\
> > >  		ret = -ENOMEM;						\
> > >  		goto end;						\
> > >  	}								\
> > > --
> > > 2.34.1
> > >
> > >  
> >
> > So this has never worked?  How has no one ever noticed that in the
> > decade this has been in the kernel?  Does that mean that no one uses
> > this file/driver and we can just remove it?  
> 
> I think just no one use sys interface to change name.

Or doesn't notice/care that is didn't work and didn't try often
enough to leak enough memory to break things.

	David

> 
> Frank
> 
> >
> > thanks,
> >
> > greg k-h
> >  
> 

Re: [PATCH] usb: gadget: f_uac1_legacy: fix inverted NULL check after kstrndup()

[Greg KH]

On Thu, Jul 09, 2026 at 08:48:08AM +0200, Greg KH wrote:
> On Thu, Jul 09, 2026 at 02:23:51PM +0800, Xu Yang wrote:
> > On Thu, Jul 09, 2026 at 07:33:03AM +0200, Greg KH wrote:
> > > On Wed, Jul 08, 2026 at 03:54:03PM -0500, Frank Li wrote:
> > > > On Wed, Jul 08, 2026 at 05:05:03PM +0200, Greg KH wrote:
> > > > > On Thu, Jun 25, 2026 at 07:31:54PM +0800, Xu Yang wrote:
> > > > > > From: Xu Yang <xu.yang_2@nxp.com>
> > > > > >
> > > > > > kstrndup() returns NULL on allocation failure. The condition was
> > > > > > checking 'if (tmp)' to detect failure, but this is inverted — it
> > > > > > would treat a successful allocation as an error and return -ENOMEM
> > > > > > while leaking the string, and proceed with a NULL pointer on failure.
> > > > > >
> > > > > > Fix by changing the condition to 'if (!tmp)'.
> > > > > >
> > > > > > Fixes: 0854611a19ae ("usb: gadget: f_uac1: add configfs support")
> > > > > > Cc: stable@vger.kernel.org
> > > > > > Assisted-by: Claude:claude-sonnet-4-6
> > > > > > Signed-off-by: Xu Yang <xu.yang_2@nxp.com>
> > > > > > ---
> > > > > >  drivers/usb/gadget/function/f_uac1_legacy.c | 2 +-
> > > > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > > > >
> > > > > > diff --git a/drivers/usb/gadget/function/f_uac1_legacy.c b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > > index 5d201a2e30e7..e9f2632ce785 100644
> > > > > > --- a/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > > +++ b/drivers/usb/gadget/function/f_uac1_legacy.c
> > > > > > @@ -914,7 +914,7 @@ static ssize_t f_uac1_opts_##name##_store(struct config_item *item,	\
> > > > > >  		goto end;						\
> > > > > >  									\
> > > > > >  	tmp = kstrndup(page, len, GFP_KERNEL);				\
> > > > > > -	if (tmp) {							\
> > > > > > +	if (!tmp) {							\
> > > > > >  		ret = -ENOMEM;						\
> > > > > >  		goto end;						\
> > > > > >  	}								\
> > > > > > --
> > > > > > 2.34.1
> > > > > >
> > > > > >
> > > > >
> > > > > So this has never worked?  How has no one ever noticed that in the
> > > > > decade this has been in the kernel?  Does that mean that no one uses
> > > > > this file/driver and we can just remove it?
> > > > 
> > > > I think just no one use sys interface to change name.
> > > 
> > > So it can just be removed?  Please submit a patch to do that instead.
> > > 
> > > Ideally this whole file can be removed, does anyone still use it?
> > 
> > This legacy driver is still built in kernel in multi_v7_defconfig: 
> > 
> > ~/work/usb$ grep -nr CONFIG_USB_CONFIGFS_F_UAC1_LEGACY arch/*/configs/
> > arch/arm/configs/multi_v7_defconfig:939:CONFIG_USB_CONFIGFS_F_UAC1_LEGACY=y
> 
> defconfig doesn't mean much.  Perhaps look at some of the Android
> configs?
> 
> > It seems we can't simply remove it. 
> 
> Why?
> 
> > BTW, do we have other ways to figure out if it can be removed?
> 
> Remove it and see who complains?  :)

Seriously, let's remove the ability to set the name as that has never
worked.  If anyone wants it to work, they can make a new patch to add
that feature as a new one.

thanks,

greg k-h