A customer of ours noticed when their machine crashed, kdump did not
work but hung instead. Using their firmware dumping solution they
grabbed a vmcore and decoded the stacks on the cpus. What they
noticed seemed to be a rare deadlock with the ioapic_lock.
CPU4:
machine_crash_shutdown
-> machine_ops.crash_shutdown
-> native_machine_crash_shutdown
-> kdump_nmi_shootdown_cpus ------> Send NMI to other CPUs
-> disable_IO_APIC
-> clear_IO_APIC
-> clear_IO_APIC_pin
-> ioapic_read_entry
-> spin_lock_irqsave(&ioapic_lock, flags)
---Infinite loop here---
CPU0:
do_IRQ
-> handle_irq
-> handle_edge_irq
-> ack_apic_edge
-> move_native_irq
-> mask_IO_APIC_irq
-> mask_IO_APIC_irq_desc
-> spin_lock_irqsave(&ioapic_lock, flags)
---Receive NMI here after getting spinlock---
-> nmi
-> do_nmi
-> crash_nmi_callback
---Infinite loop here---
The problem is that although kdump tries to shutdown minimal hardware,
it still needs to disable the IO APIC. This requires spinlocks which
may be held by another cpu. This other cpu is being held infinitely in
an NMI context by kdump in order to serialize the crashing path. Instant
deadlock.
Eric, brought up a point that because the boot code was restructured we may
not need to disable the io apic any more in the crash path. The original
concern that led to the development of disable_IO_APIC, was that the jiffies
calibration on boot up relied on the PIT timer for reference. Access
to the PIT required 8259 interrupts to be working. This wouldn't work
if the ioapic needed to be configured. So on panic path, the ioapic was
reconfigured to use virtual wire mode to allow the 8259 to passthrough.
Those concerns don't hold true now, thanks to the jiffies calibration code
not needing the PIT. As a result, we can remove this call and simplify the
locking needed in the panic path.
The same work allowed us to remove the need to disable the local apic on
shutdown too. This should allow us to jump to the second a little faster.
I tested kdump on an Ivy Bridge platform, a Pentium4 and an old athlon that
did not have an ioapic. All three were successful.
I also tested using lkdtm that would use jprobes to panic the system when
entering do_IRQ. The idea was to see how the system reacted with an
interrupt pending in the second kernel. My core2 quad successfully kdump'd
3 times in a row with no issues.
v2: removed the disable lapic code too
Cc: Eric W. Biederman <ebiederm at xmission.com>
Cc: Vivek Goyal <vgoyal at redhat.com>
Signed-off-by: Don Zickus <dzickus at redhat.com>
---
arch/x86/kernel/crash.c | 8 --------
1 files changed, 0 insertions(+), 8 deletions(-)
diff --git a/arch/x86/kernel/crash.c b/arch/x86/kernel/crash.c
index 13ad899..571f253 100644
--- a/arch/x86/kernel/crash.c
+++ b/arch/x86/kernel/crash.c
@@ -54,16 +54,12 @@ static void kdump_nmi_callback(int cpu, struct pt_regs *regs)
*/
cpu_emergency_vmxoff();
cpu_emergency_svm_disable();
-
- disable_local_APIC();
}
static void kdump_nmi_shootdown_cpus(void)
{
in_crash_kexec = 1;
nmi_shootdown_cpus(kdump_nmi_callback);
-
- disable_local_APIC();
}
#else
@@ -95,10 +91,6 @@ void native_machine_crash_shutdown(struct pt_regs *regs)
cpu_emergency_vmxoff();
cpu_emergency_svm_disable();
- lapic_shutdown();
-#if defined(CONFIG_X86_IO_APIC)
- disable_IO_APIC();
-#endif
#ifdef CONFIG_HPET_TIMER
hpet_disable();
#endif
--
1.7.7.6
