Re: PCI bug in 2.6.13

[Andrew Morton <akpm@osdl.org>]

Miguel <frankpoole@terra.es> wrote:
>
> After switching from 2.6.13-rc7 to 2.6.13 I've started to have corrupted
>  data written on my hard disks, there was appearing blocks of 0x00 bytes
>  in between the data. I've tracked when this started to happen in the git
>  patches and I've found that this bug is due to the changes done in the
>  file drivers/pci/setup-res.c of 2.6.13-rc7-git2 patch. Now I run a 2.6.13
>  kernel with that file unmodified and everything is ok.

Ugly.   I assume you're referring to this?

http://www.kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=blobdiff;h=50d6685dcbcce801682c9600a81be2a98f90f8a1;hp=5598b4714f77ac2efaf0f545e404b4c9163c4fcf;hb=755528c860b05fcecda1c88a2bdaffcb50760a7f;f=drivers/pci/setup-res.c


Ignore disabled ROM resources at setup

Writing even a disabled value seems to mess up some matrox graphics
cards.  It may be a card-related issue, but we may also be writing
reserved low bits in the result.

This was a fall-out of switching x86 over to the generic PCI resource
allocation code, and needs more debugging.  In particular, the old x86
code defaulted to not doing any resource allocations at all for ROM
resources.

In the meantime, this has been reported to make X happier by Helge
Hafting <helgehaf@aitel.hist.no>.


diff --git a/drivers/pci/setup-res.c b/drivers/pci/setup-res.c
--- a/drivers/pci/setup-res.c
+++ b/drivers/pci/setup-res.c
@@ -53,7 +53,9 @@ pci_update_resource(struct pci_dev *dev,
 	if (resno < 6) {
 		reg = PCI_BASE_ADDRESS_0 + 4 * resno;
 	} else if (resno == PCI_ROM_RESOURCE) {
-		new |= res->flags & IORESOURCE_ROM_ENABLE;
+		if (!(res->flags & IORESOURCE_ROM_ENABLE))
+			return;
+		new |= PCI_ROM_ADDRESS_ENABLE;
 		reg = dev->rom_base_reg;
 	} else {
 		/* Hmm, non-standard resource. */