gpio-ml-ioh: Use NUMA_NO_NODE not GFP_KERNEL

Currently, GFP_KERNEL is used as parameter of irq_alloc_descs like below.

          irq_base = irq_alloc_descs(-1, IOH_IRQ_BASE, num_ports[j],
                                     GFP_KERNEL);

This is not true.
So, this patch uses NUMA_NO_NODE not GFP_KERNEL.

Reported-by: Dan Carpenter <[email protected]>
Reported-by: David Rientjes <[email protected]>
Signed-off-by: Tomoya MORINAGA <[email protected]>
Acked-by: David Rientjes <[email protected]>
Signed-off-by: Grant Likely <[email protected]>

diff --git a/drivers/gpio/gpio-ml-ioh.c b/drivers/gpio/gpio-ml-ioh.c
index 274fd4d0792fc0b8434825130fdb6fc801253370..3aa6beec8c1e71b1e08527b43c6dbbd9e3f8f037 100644 (file)
--- a/drivers/gpio/gpio-ml-ioh.c
+++ b/drivers/gpio/gpio-ml-ioh.c
@@ -428,7 +428,7 @@ static int __devinit ioh_gpio_probe(struct pci_dev *pdev,
        chip = chip_save;
        for (j = 0; j < 8; j++, chip++) {
                irq_base = irq_alloc_descs(-1, IOH_IRQ_BASE, num_ports[j],
-                                          GFP_KERNEL);
+                                          NUMA_NO_NODE);
                if (irq_base < 0) {
                        dev_warn(&pdev->dev,
                                "ml_ioh_gpio: Failed to get IRQ base num\n");