agp: ensure GART has an address before enabling it
authorStephen Kitt <steve@sk2.org>
Mon, 31 Jan 2011 22:25:43 +0000 (14:25 -0800)
committerGreg Kroah-Hartman <gregkh@suse.de>
Thu, 17 Feb 2011 22:47:31 +0000 (14:47 -0800)
commit9487d9dcae378886b3d06452b630e211982a079a
treeb74cc057b1c93eb13843c4a355f158b67d9d8280
parent6d621e78c9e6efb6f4f2666434aa986bf6441195
agp: ensure GART has an address before enabling it

commit a70b95c017e8b518e1e069853355e4e497453dbb upstream.

Some BIOSs (eg.  the AMI BIOS on the Asus P4P800 motherboard) don't
initialise the GART address, and pcibios_assign_resources() can ignore it
because it can be marked as a host bridge (see
https://bugzilla.kernel.org/show_bug.cgi?id=24392#c5 for details).  This
was handled correctly up to 2.6.35, but the pci_enable_device() cleanup in
2.6.36 96576a9e1a0cdb8 ("agp: intel-agp: do not use PCI resources before
pci_enable_device()") means that the kernel tries to enable the GART
before assigning it an address; in such cases the GART overlaps with other
device assignments and ends up being disabled.

This patch fixes https://bugzilla.kernel.org/show_bug.cgi?id=24392

Note that I imagine efficeon-agp.c probably has the same problem, but
I can't test that and I'd like to make sure this patch is suitable for
-stable (since 2.6.36 and 2.6.37 are affected).

Signed-off-by: Stephen Kitt <steve@sk2.org>
Cc: Bjorn Helgaas <bjorn.helgaas@hp.com>
Cc: Maciej Rutecki <maciej.rutecki@gmail.com>
Cc: "Rafael J. Wysocki" <rjw@sisk.pl>
Cc: Kulikov Vasiliy <segooon@gmail.com>
Cc: Florian Mickler <florian@mickler.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Dave Airlie <airlied@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
drivers/char/agp/intel-agp.c