new_pcichild_device() is not called in atomic context. The call chain ending up at new_pcichild_device() is: [1] new_pcichild_device() <- pci_devices_present_work() pci_devices_present_work() is only set in INIT_WORK(). Despite never getting called from atomic context, new_pcichild_device() calls kzalloc 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@xxxxxxxxx> --- drivers/pci/host/pci-hyperv.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.c index 0fe3ea1..289e31d 100644 --- a/drivers/pci/host/pci-hyperv.c +++ b/drivers/pci/host/pci-hyperv.c @@ -1500,7 +1500,7 @@ static struct hv_pci_dev *new_pcichild_device(struct hv_pcibus_device *hbus, unsigned long flags; int ret; - hpdev = kzalloc(sizeof(*hpdev), GFP_ATOMIC); + hpdev = kzalloc(sizeof(*hpdev), GFP_KERNEL); if (!hpdev) return NULL; -- 1.9.1
