linux/arch
Aristeu Rozanski b3e15bdef6 x86, NMI watchdog: setup before enabling NMI watchdog
There's a small window when NMI watchdog is being set up that if any NMIs
are triggered, the NMI code will make make use of not initalized wd_ops
elements:
	void setup_apic_nmi_watchdog(void *unused)
	{
		if (__get_cpu_var(wd_enabled))
			return;

		/* cheap hack to support suspend/resume */
		/* if cpu0 is not active neither should the other cpus */
		if (smp_processor_id() != 0 && atomic_read(&nmi_active) <= 0)
			return;

		switch (nmi_watchdog) {
		case NMI_LOCAL_APIC:
			/* enable it before to avoid race with handler */
-->			__get_cpu_var(wd_enabled) = 1;
-->			if (lapic_watchdog_init(nmi_hz) < 0) {
(...)
	asmlinkage notrace __kprobes void default_do_nmi(struct pt_regs *regs)
	{
	(...)
			if (nmi_watchdog_tick(regs, reason))
				return;
(...)
	notrace __kprobes int
	nmi_watchdog_tick(struct pt_regs *regs, unsigned reason)
	{
	(...)
		if (!__get_cpu_var(wd_enabled))
			return rc;
		switch (nmi_watchdog) {
		case NMI_LOCAL_APIC:
			rc |= lapic_wd_event(nmi_hz);
(...)
int lapic_wd_event(unsigned nmi_hz)
{
	struct nmi_watchdog_ctlblk *wd = &__get_cpu_var(nmi_watchdog_ctlblk);
	u64 ctr;

-->	rdmsrl(wd->perfctr_msr, ctr);

and wd->*_msr will be initialized on each processor type specific setup, after
enabling NMIs for PMIs. Since the counter was just set, the chances of an
performance counter generated NMI is minimal, but any other unknown NMI would
trigger the problem. This patch fixes the problem by setting everything up
before enabling performance counter generated NMIs and will set wd_enabled
using a callback function.

Signed-off-by: Aristeu Rozanski <aris@redhat.com>
Acked-by: Don Zickus <dzickus@redhat.com>
Acked-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
History 2008-09-22 19:48:19 +02:00
..
alpha [PATCH] fix osf_getdirents() 2008-08-25 01:18:06 -04:00
arm [ARM] 5255/1: Update jornada ssp to remove build errors/warnings 2008-09-18 12:29:09 +01:00
avr32 avr32: nmi_enter() without nmi_exit() 2008-09-19 20:08:08 +02:00
blackfin Blackfin arch: Fix PM building on BF52x: No ROTWE on BF52x, add USBWE 2008-08-28 17:32:01 +08:00
cris removed unused #include <linux/version.h>'s 2008-08-23 12:14:12 -07:00
frv FRV: Wire up new system calls 2008-08-01 13:03:49 -07:00
h8300 [h8300] move include/asm-h8300 to arch/h8300/include/asm 2008-08-13 14:26:32 -07:00
ia64 KVM: ia64: 'struct fdesc' build fix 2008-09-19 16:49:01 -07:00
m32r m32r: use generic show_mem() 2008-07-26 12:00:11 -07:00
m68k m68k: Update defconfigs for 2.6.27-rc6 2008-09-10 14:15:29 -07:00
m68knommu m68k{,nommu}: Wire up new system calls 2008-08-11 10:37:34 -07:00
mips [MIPS] vmlinux.lds.S: handle .text.* 2008-09-21 14:52:58 +02:00
mn10300 MN10300: Change the fault handler to check in_atomic() not in_interrupt() 2008-09-11 10:36:08 -07:00
parisc lib: Correct printk %pF to work on all architectures 2008-09-09 11:51:15 -07:00
powerpc Fix compile failure with non modular builds 2008-09-17 09:14:42 -07:00
s390 [S390] CVE-2008-1514: prevent ptrace padding area read/write in 31-bit mode 2008-09-09 12:39:06 +02:00
sh sh64: resume_kernel fix for kernel oops built with CONFIG_BKL_PREEMPT=y. 2008-09-05 14:51:28 +09:00
sparc sparc: Fix user_regset 'n' field values. 2008-09-12 15:55:44 -07:00
sparc64 sparc64: Fix SMP bootup with CONFIG_STACK_DEBUG or ftrace. 2008-09-16 11:44:00 -07:00
um uml: fix tty-related build error 2008-07-30 09:41:45 -07:00
x86 x86, NMI watchdog: setup before enabling NMI watchdog 2008-09-22 19:48:19 +02:00
xtensa remove unneeded #include <linux/ide.h>'s 2008-08-05 18:17:00 +02:00
.gitignore arch: Ignore arch/i386 and arch/x86_64 2008-01-19 21:29:39 -08:00
Kconfig tracehook: CONFIG_HAVE_ARCH_TRACEHOOK 2008-07-26 12:00:09 -07:00