x86, ioapic: Fix potential resume deadlock
authorDaniel J Blueman <daniel.blueman@gmail.com>
Wed, 18 May 2011 23:31:31 +0000 (16:31 -0700)
committerIngo Molnar <mingo@elte.hu>
Fri, 20 May 2011 11:40:41 +0000 (13:40 +0200)
Fix a potential deadlock when resuming; here the calling
function has disabled interrupts, so we cannot sleep.

Change the memory allocation flag from GFP_KERNEL to GFP_ATOMIC.

TODO: We can do away with this memory allocation during resume
      by reusing the ioapic suspend/resume code that uses boot time
      allocated buffers, but we want to keep this -stable patch
      simple.

Signed-off-by: Daniel J Blueman <daniel.blueman@gmail.com>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Cc: <stable@kernel.org> # v2.6.38/39
Link: http://lkml.kernel.org/r/20110518233157.385970138@sbsiddha-MOBL3.sc.intel.com
Signed-off-by: Ingo Molnar <mingo@elte.hu>
arch/x86/kernel/apic/io_apic.c

index 45fd33d..df63620 100644 (file)
@@ -621,14 +621,14 @@ struct IO_APIC_route_entry **alloc_ioapic_entries(void)
        struct IO_APIC_route_entry **ioapic_entries;
 
        ioapic_entries = kzalloc(sizeof(*ioapic_entries) * nr_ioapics,
-                               GFP_KERNEL);
+                               GFP_ATOMIC);
        if (!ioapic_entries)
                return 0;
 
        for (apic = 0; apic < nr_ioapics; apic++) {
                ioapic_entries[apic] =
                        kzalloc(sizeof(struct IO_APIC_route_entry) *
-                               nr_ioapic_registers[apic], GFP_KERNEL);
+                               nr_ioapic_registers[apic], GFP_ATOMIC);
                if (!ioapic_entries[apic])
                        goto nomem;
        }