[PATCH] x86_64: ioapic virtual wire mode fix
authorVivek Goyal <vgoyal@in.ibm.com>
Wed, 11 Jan 2006 21:46:21 +0000 (22:46 +0100)
committerLinus Torvalds <torvalds@g5.osdl.org>
Thu, 12 Jan 2006 03:05:01 +0000 (19:05 -0800)
o Currently, during kexec reboot, IOAPIC is re-programmed back to virtual
  wire mode if there was an i8259 connected to it. This enables getting
  timer interrupts in second kernel in legacy mode.

o After putting into virtual wire mode, IOAPIC delivers the i8259 interrupts
  to CPU0. This works well for kexec but not for kdump as we might crash
  on a different CPU and second kernel will not see timer interrupts.

o This patch modifies the redirection table entry to deliver the timer
  interrupts to the cpu we are rebooting (instead of hardcoding to zero).
  This ensures that second kernel receives timer interrupts even on a
  non-boot cpu.

Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
arch/x86_64/kernel/io_apic.c

index 6c938c60fa3b25acaf46616f2e153543fc402954..fd1e7c36dc23f3a86625dfbb1e984accdbf16f3b 100644 (file)
@@ -1307,7 +1307,8 @@ void disable_IO_APIC(void)
                entry.dest_mode       = 0; /* Physical */
                entry.delivery_mode   = dest_ExtINT; /* ExtInt */
                entry.vector          = 0;
-               entry.dest.physical.physical_dest = 0;
+               entry.dest.physical.physical_dest =
+                                       GET_APIC_ID(apic_read(APIC_ID));
 
                /*
                 * Add it to the IO-APIC irq-routing table: