line6_alloc_sysex_buffer() is never called in atomic context.
The call chain ending up at line6_alloc_sysex_buffer() are: [1] line6_alloc_sysex_buffer() <- pod_alloc_sysex_buffer() <- pod_set_system_param_int() <- snd_pod_control_monitor_put() snd_pod_control_monitor_put() is only set as ".put" in snd_kcontrol_new structure "pod_control_monitor". And ->put() is called in snd_ctl_elem_write() (sound/core/control.c) snd_ctl_elem_write() are called in ctl_elem_write_user() and snd_ctl_elem_write_user(). These functions are not called in atomic context.
Despite never getting called from atomic context, line6_alloc_sysex_buffer() calls kmalloc with GFP_ATOMIC, which waits busily for allocation. GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL to avoid busy waiting.
This is found by a static analysis tool named DCNS written by myself.
Signed-off-by: Jia-Ju Bai baijiaju1990@gmail.com --- sound/usb/line6/driver.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/usb/line6/driver.c b/sound/usb/line6/driver.c index 0ff5a7d..70a509e 100644 --- a/sound/usb/line6/driver.c +++ b/sound/usb/line6/driver.c @@ -265,7 +265,7 @@ int line6_send_sysex_message(struct usb_line6 *line6, const char *buffer, char *line6_alloc_sysex_buffer(struct usb_line6 *line6, int code1, int code2, int size) { - char *buffer = kmalloc(size + SYSEX_EXTRA_SIZE, GFP_ATOMIC); + char *buffer = kmalloc(size + SYSEX_EXTRA_SIZE, GFP_KERNEL);
if (!buffer) return NULL;