[RESEND PATCH 0/1] usb: gadget: f_fs: fix NULL pointer dereference in FUNCTIONFS_ENDPOINT_DESC

[RESEND PATCH 0/1] usb: gadget: f_fs: fix NULL pointer dereference in FUNCTIONFS_ENDPOINT_DESC

[zjamg]

From: Yuchao Zhang <ndaugoing@gmail.com>

Hi,

This patch addresses a kernel NULL pointer dereference panic in
FunctionFS when calling the FUNCTIONFS_ENDPOINT_DESC ioctl.

In FunctionFS, user space can legitimately supply only a subset of
speed descriptors (e.g., only full-speed descriptors via
FUNCTIONFS_HAS_FS_DESC). When the USB gadget is subsequently connected
to a high-speed or super-speed host:
1. composite.c:config_ep_by_speed() detects that the function lacks
   descriptors for the current speed and gracefully falls back to
   full-speed descriptors (f->fs_descriptors).
2. The endpoint is enabled via usb_ep_enable(), and epfile->ep becomes
   valid.
3. However, when user space calls ioctl(fd, FUNCTIONFS_ENDPOINT_DESC),
   ffs_epfile_ioctl() selects the descriptor index solely based on
   epfile->ffs->gadget->speed (desc_idx = 1 for HIGH, 2 for
   SUPER/SUPER_PLUS).
4. Because user space only supplied full-speed descriptors,
   epfile->ep->descs[desc_idx] is NULL.
5. The subsequent memcpy(&desc1, desc, desc->bLength) immediately
   dereferences the NULL pointer, causing an instant kernel crash:

[    2.154836] Call trace:
[    2.154836]  ffs_epfile_ioctl+0xe14/0x2c40 [usb_f_fs] (P)
[    2.155003]  __arm64_sys_ioctl+0x128/0x1a4
[    2.155298]  invoke_syscall.constprop.0+0x64/0x260
[    2.155374]  el0_svc_common.constprop.0+0xa0/0x240
[    2.155441]  do_el0_svc+0x3c/0x60
[    2.155494]  el0_svc+0x38/0x98
[    2.155547]  el0t_64_sync_handler+0xa0/0xe4
[    2.155606]  el0t_64_sync+0x198/0x19c
[    2.156463] Kernel panic - not syncing: Oops: Fatal exception

The fix restores the speed fallback loop mirroring config_ep_by_speed()
and historical f_fs behavior, defensively checks epfile->ep->ep->desc
as a fallback, and returns -EINVAL safely if no valid descriptor is
found.

Testing with QEMU ARM64 using dummy_hcd confirms that after applying
this patch, the ioctl succeeds and returns the fallback descriptor
cleanly without any crash.

zjamg (1):
  usb: gadget: f_fs: Fix NULL pointer dereference in
    FUNCTIONFS_ENDPOINT_DESC

 drivers/usb/gadget/function/f_fs.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

-- 
2.53.0

[RESEND PATCH 1/1] usb: gadget: f_fs: Fix NULL pointer dereference in FUNCTIONFS_ENDPOINT_DESC

[zjamg]

From: Yuchao Zhang <ndaugoing@gmail.com>

When user space sets up FunctionFS with only full-speed descriptors
(e.g., flags = FUNCTIONFS_HAS_FS_DESC) and the gadget operates at a
higher speed (such as USB_SPEED_HIGH or USB_SPEED_SUPER),
config_ep_by_speed() detects that the function lacks descriptors for
the current speed, logs a warning, and falls back to full-speed
descriptors. The endpoint is then successfully enabled via
usb_ep_enable(), and epfile->ep becomes valid.

However, when user space subsequently issues the FUNCTIONFS_ENDPOINT_DESC
ioctl to query the endpoint descriptor, ffs_epfile_ioctl() selects the
descriptor index solely based on gadget->speed:

	switch (epfile->ffs->gadget->speed) {
	case USB_SPEED_SUPER:
	case USB_SPEED_SUPER_PLUS:
		desc_idx = 2;
		break;
	case USB_SPEED_HIGH:
		desc_idx = 1;
		break;
	default:
		desc_idx = 0;
	}
	desc = epfile->ep->descs[desc_idx];
	memcpy(&desc1, desc, desc->bLength);

Because user space only supplied full-speed descriptors,
epfile->ep->descs[1] is NULL. Dereferencing desc->bLength triggers an
immediate kernel NULL pointer dereference panic.

Fix this by falling back through lower speed descriptors if the current
speed descriptor was not supplied, matching the fallback logic in
config_ep_by_speed() and historical f_fs behavior. If still missing,
fall back to epfile->ep->ep->desc (the active descriptor), and return
-EINVAL safely if no valid descriptor is found. In addition, mark desc
as const to match struct usb_ep::desc and avoid discarding qualifiers.

Fixes: c559a3534109 ("usb: gadget: f_fs: add ioctl returning ep descriptor")
Cc: stable@vger.kernel.org
Signed-off-by: Yuchao Zhang <ndaugoing@gmail.com>
---
 drivers/usb/gadget/function/f_fs.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)

diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c
index 43962e05eacf..c64a268e98a4 100644
--- a/drivers/usb/gadget/function/f_fs.c
+++ b/drivers/usb/gadget/function/f_fs.c
@@ -1877,8 +1877,9 @@ static long ffs_epfile_ioctl(struct file *file, unsigned code,
 		break;
 	case FUNCTIONFS_ENDPOINT_DESC:
 	{
+		const struct usb_endpoint_descriptor *desc;
+		struct usb_endpoint_descriptor desc1;
 		int desc_idx;
-		struct usb_endpoint_descriptor desc1, *desc;
 
 		switch (epfile->ffs->gadget->speed) {
 		case USB_SPEED_SUPER:
@@ -1892,7 +1893,17 @@ static long ffs_epfile_ioctl(struct file *file, unsigned code,
 			desc_idx = 0;
 		}
 
-		desc = epfile->ep->descs[desc_idx];
+		do {
+			desc = epfile->ep->descs[desc_idx];
+		} while (!desc && --desc_idx >= 0);
+
+		if (!desc)
+			desc = epfile->ep->ep->desc;
+		if (!desc) {
+			ret = -EINVAL;
+			break;
+		}
+
 		memcpy(&desc1, desc, desc->bLength);
 
 		spin_unlock_irq(&epfile->ffs->eps_lock);
-- 
2.53.0

Re: [RESEND PATCH 1/1] usb: gadget: f_fs: Fix NULL pointer dereference in FUNCTIONFS_ENDPOINT_DESC

[Michał Nazarewicz]

On Fri, Sep 18 2026, zjamg wrote:
> From: Yuchao Zhang <ndaugoing@gmail.com>
>
> Fixes: c559a3534109 ("usb: gadget: f_fs: add ioctl returning ep descriptor")
> Cc: stable@vger.kernel.org
> Signed-off-by: Yuchao Zhang <ndaugoing@gmail.com>

Acked-by: Michał Nazarewicz <mina86@mina86.com>

> ---
>  drivers/usb/gadget/function/f_fs.c | 15 +++++++++++++--
>  1 file changed, 13 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/usb/gadget/function/f_fs.c b/drivers/usb/gadget/function/f_fs.c
> index 43962e05eacf..c64a268e98a4 100644
> --- a/drivers/usb/gadget/function/f_fs.c
> +++ b/drivers/usb/gadget/function/f_fs.c
> @@ -1877,8 +1877,9 @@ static long ffs_epfile_ioctl(struct file *file, unsigned code,
>  		break;
>  	case FUNCTIONFS_ENDPOINT_DESC:
>  	{
> +		const struct usb_endpoint_descriptor *desc;
> +		struct usb_endpoint_descriptor desc1;
>  		int desc_idx;
> -		struct usb_endpoint_descriptor desc1, *desc;
>  
>  		switch (epfile->ffs->gadget->speed) {
>  		case USB_SPEED_SUPER:
> @@ -1892,7 +1893,17 @@ static long ffs_epfile_ioctl(struct file *file, unsigned code,
>  			desc_idx = 0;
>  		}
>  
> -		desc = epfile->ep->descs[desc_idx];
> +		do {
> +			desc = epfile->ep->descs[desc_idx];
> +		} while (!desc && --desc_idx >= 0);
> +
> +		if (!desc)
> +			desc = epfile->ep->ep->desc;

IIUC, this will never be non-NULL if all the descriptors are NULL.  Then
again, it probably won’t hurt, so dosen’t matter one way or another.

> +		if (!desc) {
> +			ret = -EINVAL;
> +			break;
> +		}
> +
>  		memcpy(&desc1, desc, desc->bLength);
>  
>  		spin_unlock_irq(&epfile->ffs->eps_lock);

-- 
Best regards
ミハウ “𝓶𝓲𝓷𝓪86” ナザレヴィツ
«If at first you don’t succeed, give up skydiving»