After checking all possible call chains to fore200e_open here, my tool finds that fore200e_open is never called in atomic context. And fore200e_open calls mutex_lock which can sleep later, thus it is supposed fore200e_open can call functions which may sleep. Thus GFP_ATOMIC is not necessary, and it can be replaced with GFP_KERNEL.
This is found by a static analysis tool named DCNS written by myself. Signed-off-by: Jia-Ju Bai <baijiaju1...@gmail.com> --- drivers/atm/fore200e.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/atm/fore200e.c b/drivers/atm/fore200e.c index 6ebc4e4..bdffb85 100644 --- a/drivers/atm/fore200e.c +++ b/drivers/atm/fore200e.c @@ -1422,7 +1422,7 @@ int bsq_audit(int where, struct host_bsq* bsq, int scheme, int magn) spin_unlock_irqrestore(&fore200e->q_lock, flags); - fore200e_vcc = kzalloc(sizeof(struct fore200e_vcc), GFP_ATOMIC); + fore200e_vcc = kzalloc(sizeof(struct fore200e_vcc), GFP_KERNEL); if (fore200e_vcc == NULL) { vc_map->vcc = NULL; return -ENOMEM; -- 1.7.9.5