hv_pci_onchannelcallback() is not called in atomic context. The call chain ending up at hv_pci_onchannelcallback() is: [1] hv_pci_onchannelcallback() <- hv_pci_probe() hv_pci_probe() is only set as ".probe" in hv_driver structure "hv_pci_drv". Despite never getting called from atomic context, hv_pci_onchannelcallback() 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@xxxxxxxxx> --- drivers/pci/host/pci-hyperv.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/pci/host/pci-hyperv.c b/drivers/pci/host/pci-hyperv.c index 0fe3ea1..c5c8a99 100644 --- a/drivers/pci/host/pci-hyperv.c +++ b/drivers/pci/host/pci-hyperv.c @@ -1887,7 +1887,7 @@ static void hv_pci_onchannelcallback(void *context) struct pci_dev_incoming *dev_message; struct hv_pci_dev *hpdev; - buffer = kmalloc(bufferlen, GFP_ATOMIC); + buffer = kmalloc(bufferlen, GFP_KERNEL); if (!buffer) return; @@ -1899,7 +1899,7 @@ static void hv_pci_onchannelcallback(void *context) kfree(buffer); /* Handle large packet */ bufferlen = bytes_recvd; - buffer = kmalloc(bytes_recvd, GFP_ATOMIC); + buffer = kmalloc(bytes_recvd, GFP_KERNEL); if (!buffer) return; continue; -- 1.9.1 _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
