linux/arch/x86/kernel/apic
Brandon Phiilps ced5b697a7 x86: Avoid race condition in pci_enable_msix()
Keep chip_data in create_irq_nr and destroy_irq.

When two drivers are setting up MSI-X at the same time via
pci_enable_msix() there is a race.  See this dmesg excerpt:

[   85.170610] ixgbe 0000:02:00.1: irq 97 for MSI/MSI-X
[   85.170611]   alloc irq_desc for 99 on node -1
[   85.170613] igb 0000:08:00.1: irq 98 for MSI/MSI-X
[   85.170614]   alloc kstat_irqs on node -1
[   85.170616] alloc irq_2_iommu on node -1
[   85.170617]   alloc irq_desc for 100 on node -1
[   85.170619]   alloc kstat_irqs on node -1
[   85.170621] alloc irq_2_iommu on node -1
[   85.170625] ixgbe 0000:02:00.1: irq 99 for MSI/MSI-X
[   85.170626]   alloc irq_desc for 101 on node -1
[   85.170628] igb 0000:08:00.1: irq 100 for MSI/MSI-X
[   85.170630]   alloc kstat_irqs on node -1
[   85.170631] alloc irq_2_iommu on node -1
[   85.170635]   alloc irq_desc for 102 on node -1
[   85.170636]   alloc kstat_irqs on node -1
[   85.170639] alloc irq_2_iommu on node -1
[   85.170646] BUG: unable to handle kernel NULL pointer dereference
at 0000000000000088

As you can see igb and ixgbe are both alternating on create_irq_nr()
via pci_enable_msix() in their probe function.

ixgbe: While looping through irq_desc_ptrs[] via create_irq_nr() ixgbe
choses irq_desc_ptrs[102] and exits the loop, drops vector_lock and
calls dynamic_irq_init. Then it sets irq_desc_ptrs[102]->chip_data =
NULL via dynamic_irq_init().

igb: Grabs the vector_lock now and starts looping over irq_desc_ptrs[]
via create_irq_nr(). It gets to irq_desc_ptrs[102] and does this:

	cfg_new = irq_desc_ptrs[102]->chip_data;
	if (cfg_new->vector != 0)
		continue;

This hits the NULL deref.

Another possible race exists via pci_disable_msix() in a driver or in
the number of error paths that call free_msi_irqs():

destroy_irq()
dynamic_irq_cleanup() which sets desc->chip_data = NULL
...race window...
desc->chip_data = cfg;

Remove the save and restore code for cfg in create_irq_nr() and
destroy_irq() and take the desc->lock when checking the irq_cfg.

Reported-and-analyzed-by: Brandon Philips <bphilips@suse.de>
Signed-off-by: Yinghai Lu <yinghai@kernel.org>
LKML-Reference: <1265793639-15071-3-git-send-email-yinghai@kernel.org>
Signed-off-by: Brandon Phililps <bphilips@suse.de>
Cc: stable@kernel.org
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
2010-02-10 14:27:28 -08:00
..
apic_flat_64.c x86, apic: use physical mode for IBM summit platforms 2010-01-18 14:15:27 -08:00
apic_noop.c x86: Fix bogus warning in apic_noop.apic_write() 2009-12-07 13:16:37 +01:00
apic.c x86, apic: use logical flat for systems with <= 8 logical cpus 2010-01-18 14:15:28 -08:00
bigsmp_32.c x86, irq: Allow 0xff for /proc/irq/[n]/smp_affinity on an 8-cpu system 2009-12-17 22:03:06 -08:00
es7000_32.c x86: es7000_32.c: Use pr_<level> and add pr_fmt(fmt) 2009-12-10 08:57:49 +01:00
io_apic.c x86: Avoid race condition in pci_enable_msix() 2010-02-10 14:27:28 -08:00
ipi.c Merge commit 'v2.6.31-rc7' into x86/cleanups 2009-08-24 12:25:54 +02:00
Makefile x86, apic: Introduce the NOOP apic driver 2009-10-14 09:17:00 +02:00
nmi.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/percpu 2009-12-14 09:58:24 -08:00
numaq_32.c Merge branch 'x86-mm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2009-12-08 13:27:33 -08:00
probe_32.c x86: apic: Do not use stacked physid_mask_t 2009-11-10 05:52:07 +01:00
probe_64.c x86, apic: use logical flat for systems with <= 8 logical cpus 2010-01-18 14:15:28 -08:00
summit_32.c x86: apic: Do not use stacked physid_mask_t 2009-11-10 05:52:07 +01:00
x2apic_cluster.c x86, irq: Allow 0xff for /proc/irq/[n]/smp_affinity on an 8-cpu system 2009-12-17 22:03:06 -08:00
x2apic_phys.c x86, irq: Allow 0xff for /proc/irq/[n]/smp_affinity on an 8-cpu system 2009-12-17 22:03:06 -08:00
x2apic_uv_x.c x86, uv: Ensure hub revision set for all ACPI modes. 2010-01-15 11:09:04 -08:00