[Regression] [2.6.37-rc2] WARNING: at drivers/pci/pci-sysfs.c:758 pci_mmap_resource+0x133/0x190()

From: Maciej Rutecki
Date: Wed Nov 17 2010 - 15:53:27 EST

Last known good: 2.6.37-rc1
Bad kernel: 2.6.37-rc2
Hardware: VirtualBox

Description: cannot run Xwindow, booting hangs on this kernel message:
[ 143.502215] ------------[ cut here ]------------
[ 143.503050] WARNING: at drivers/pci/pci-sysfs.c:758
pci_mmap_resource+0x133/0x190()
[ 143.523361] Hardware name: VirtualBox
[ 143.523387] process "Xorg" tried to map 0x02000000 bytes at page 0x00000000
on 0000:00:02.0 BAR 0 (start 0x
[ 143.523408] Modules linked in: fuse loop sha256_generic aes_i586
aes_generic cbc dm_crypt dm_mod ohci_hcd ehci
[ 143.523823] Pid: 1665, comm: Xorg Not tainted 2.6.37-rc2 #1
[ 143.523843] Call Trace:
[ 143.523878] [<c0137f9d>] warn_slowpath_common+0x6d/0xa0
[ 143.523898] [<c0281c83>] ? pci_mmap_resource+0x133/0x190
[ 143.523916] [<c0281c83>] ? pci_mmap_resource+0x133/0x190
[ 143.523935] [<c013804e>] warn_slowpath_fmt+0x2e/0x30
[ 143.523953] [<c0281c83>] pci_mmap_resource+0x133/0x190
[ 143.523972] [<c0281ce0>] ? pci_mmap_resource_uc+0x0/0x20
[ 143.524000] [<c0281cf9>] pci_mmap_resource_uc+0x19/0x20
[ 143.526472] [<c020eba2>] mmap+0x62/0x100
[ 143.526501] [<c01ac52f>] mmap_region+0x2bf/0x410
[ 143.526522] [<c01ac8dc>] do_mmap_pgoff+0x25c/0x300
[ 143.526541] [<c01aca17>] sys_mmap_pgoff+0x97/0x130
[ 143.526569] [<c03dbd11>] syscall_call+0x7/0xb
[ 143.526584] ---[ end trace 57cafb32c03e0f5f ]---

Bisection shows:
3b519e4ea618b6943a82931630872907f9ac2c2b is the first bad commit
commit 3b519e4ea618b6943a82931630872907f9ac2c2b
Author: Martin Wilck <martin.wilck@xxxxxxxxxxxxxx>
Date: Wed Nov 10 11:03:21 2010 +0100

PCI: fix size checks for mmap() on /proc/bus/pci files

The checks for valid mmaps of PCI resources made through /proc/bus/pci
files
that were introduced in 9eff02e2042f96fb2aedd02e032eca1c5333d767 have
several
problems:

1. mmap() calls on /proc/bus/pci files are made with real file offsets >
0,
whereas under /sys/bus/pci/devices, the start of the resource corresponds
to offset 0. This may lead to false negatives in pci_mmap_fits(), which
implicitly assumes the /sys/bus/pci/devices layout.

2. The loop in proc_bus_pci_mmap doesn't skip empty resouces. This leads
to false positives, because pci_mmap_fits() doesn't treat empty resources
correctly (the calculated size is 1 << (8*sizeof(resource_size_t)-
PAGE_SHIFT)
in this case!).

3. If a user maps resources with BAR > 0, pci_mmap_fits will emit bogus
WARNINGS for the first resources that don't fit until the correct one is
found.

On many controllers the first 2-4 BARs are used, and the others are empty.
In this case, an mmap attempt will first fail on the non-empty BARs
(including the "right" BAR because of 1.) and emit bogus WARNINGS because
of 3., and finally succeed on the first empty BAR because of 2.
This is certainly not the intended behaviour.

This patch addresses all 3 issues.
Updated with an enum type for the additional parameter for pci_mmap_fits().

Cc: stable@xxxxxxxxxx
Signed-off-by: Martin Wilck <martin.wilck@xxxxxxxxxxxxxx>
Signed-off-by: Jesse Barnes <jbarnes@xxxxxxxxxxxxxxxx>

040000 040000 73eda4a589e4b0da691202a36ec4b7e38f635c64
2dd7820cdcff8b2d1e23d990eee7a5832a356c38 M drivers

Full dmesg:
http://unixy.pl/maciek/download/kernel/2.6.37-rc2/vbox/dmesg-2.6.37-rc2.txt

Config:
http://unixy.pl/maciek/download/kernel/2.6.37-rc2/vbox/config-2.6.37-rc2

lspci:
http://unixy.pl/maciek/download/kernel/2.6.37-rc2/vbox/lspci.txt

Regards
--
Maciej Rutecki
http://www.maciek.unixy.pl
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/