Re: [PATCH] HID: core: fix device cleanup on allocation failure

On Wed, Aug 19, 2026 at 01:31:49AM +0000, Tristan Madani wrote:
> From: Tristan Madani <[email protected]>
> 
> hid_allocate_device() calls hid_destroy_device() in its error path when
> hid_bpf_device_init() fails.  hid_destroy_device() in turn calls
> hid_bpf_destroy_device() which invokes synchronize_srcu() and
> cleanup_srcu_struct() on the SRCU structure.  However, at this point
> init_srcu_struct() has not been called (or has failed), so the SRCU
> internal pointers are all zeroed from kzalloc.  Dereferencing them
> causes a general protection fault:
> 
>   general protection fault in check_init_srcu_struct
>   KASAN: null-ptr-deref in range [0x00000000000001c0-0x00000000000001c7]
>   Call Trace:
>    synchronize_srcu+0x1c/0x2b0
>    hid_bpf_destroy_device+0x64/0x90
>    hid_destroy_device+0x29/0x240
>    hid_allocate_device+0x420/0x4e0
>    uhid_dev_create2+0x161/0x890
> 
> Fix this by replacing hid_destroy_device() with put_device() in the
> error path.  At this point in hid_allocate_device(), the device has
> been initialized (device_initialize) but not yet added (device_add),
> so put_device() is the correct cleanup call.  It triggers the release
> callback which frees the hid_device without attempting to clean up
> resources that were never set up.
> 
> Found by syzkaller on Ubuntu 6.14.11 KASAN.
> 
> Fixes: 6cd735f0e57a ("HID: bpf: protect HID-BPF prog_list access by a SRCU")
> Cc: [email protected]
> Signed-off-by: Tristan Madani <[email protected]>
> ---
>  drivers/hid/hid-core.c | 7 ++++++-
>  1 file changed, 6 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/hid/hid-core.c b/drivers/hid/hid-core.c
> index cf123347a2af7..9534de5317783 100644
> --- a/drivers/hid/hid-core.c
> +++ b/drivers/hid/hid-core.c
> @@ -3060,7 +3060,12 @@ struct hid_device *hid_allocate_device(void)
>  	return hdev;
>  
>  out_err:
> -	hid_destroy_device(hdev);
> +	/*
> +	 * hid_destroy_device() cannot be used here because
> +	 * hid_bpf_device_init() failed and the SRCU struct was never
> +	 * initialized.  Release the device directly.

That should be in the changelog, not in the code, as it can, and will,
change over time.

thanks,

greg k-h