This patch fixes a long standing bug in the PCI sysfs code, where if we were
exposing a shadow copy of the ROM (at 0xc0000-0xc0000+128k) we wouldn't
properly set the size. Any comments? Another approach would be to mess with
pdev->resource[PCI_ROM_RESOURCE] when we do the shadow fixup, but I haven't
looked to see if that would have bad implications elsewhere.
Thanks,
--
Jesse Barnes, Intel Open Source Technology Center
diff --git a/drivers/pci/pci-sysfs.c b/drivers/pci/pci-sysfs.c
index 9c71858..c23ae27 100644
--- a/drivers/pci/pci-sysfs.c
+++ b/drivers/pci/pci-sysfs.c
@@ -720,6 +720,7 @@ int __must_check pci_create_sysfs_dev_files (struct
pci_dev *pdev)
{
struct bin_attribute *attr = NULL;
int retval;
+ int rom_size = 0;
if (!sysfs_initialized)
return -EACCES;
@@ -755,12 +756,16 @@ int __must_check pci_create_sysfs_dev_files (struct
pci_dev *pdev)
goto err_vpd_file;
/* If the device has a ROM, try to expose it in sysfs. */
- if (pci_resource_len(pdev, PCI_ROM_RESOURCE) ||
- (pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW)) {
+ if (pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW)
+ rom_size = 0x20000;
+ else if (pci_resource_len(pdev, PCI_ROM_RESOURCE))
+ rom_size = pci_resource_len(pdev, PCI_ROM_RESOURCE);
+
+ if (rom_size) {
attr = kzalloc(sizeof(*attr), GFP_ATOMIC);
if (attr) {
pdev->rom_attr = attr;
- attr->size = pci_resource_len(pdev, PCI_ROM_RESOURCE);
+ attr->size = rom_size;
attr->attr.name = "rom";
attr->attr.mode = S_IRUSR;
attr->read = pci_read_rom;
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
