[PATCH] Do not enforce unique IO_APIC_ID check for xAPIC systems (i386)
This patch is per Andi's request to remove NO_IOAPIC_CHECK from genapic and
use heuristics to prevent unique I/O APIC ID check for systems that don't
need it. The patch disables unique I/O APIC ID check for Xeon-based and
other platforms that don't use serial APIC bus for interrupt delivery.
Andi stated that AMD systems don't need unique IO_APIC_IDs either.
Signed-off-by: Natalie Protasevich <Natalie.Protasevich@unisys.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
diff --git a/arch/i386/kernel/io_apic.c b/arch/i386/kernel/io_apic.c
index 7a324e8..51f7a5d 100644
--- a/arch/i386/kernel/io_apic.c
+++ b/arch/i386/kernel/io_apic.c
@@ -1659,6 +1659,12 @@
unsigned long flags;
/*
+ * Don't check I/O APIC IDs for xAPIC systems. They have
+ * no meaning without the serial APIC bus.
+ */
+ if (!(boot_cpu_data.x86_vendor == X86_VENDOR_INTEL && boot_cpu_data.x86 < 15))
+ return;
+ /*
* This is broken; anything with a real cpu count has to
* circumvent this idiocy regardless.
*/
@@ -1684,10 +1690,6 @@
mp_ioapics[apic].mpc_apicid = reg_00.bits.ID;
}
- /* Don't check I/O APIC IDs for some xAPIC systems. They have
- * no meaning without the serial APIC bus. */
- if (NO_IOAPIC_CHECK)
- continue;
/*
* Sanity check, is the ID really free? Every APIC in a
* system must have a unique ID or we get lots of nice