From: Jack F Vogel A bug against an xSeries system showed up recently noting that the check_nmi_watchdog() test was failing. I have been investigating it and discovered in both i386 and x86_64 the recent change to the routine to use the cpu_callin_map has uncovered a problem. Prior to that change, on an SMP box, the test was trivally passing because all cpu's were found to not yet be online, but now with the callin_map they are discovered, it goes on to test the counter and they have not yet begun to increment, so it announces a CPU is stuck and bails out. On all the systems I have access to test, the announcement of failure is also bougs... by the time you can login and check /proc/interrupts, the NMI count is happily incrementing on all CPUs. Its just that the test is being done too early. I have tried moving the call to the test around a bit, and it was always too early. I finally hit on this proposed solution, it delays the routine via a late_initcall(), seems like the right solution to me. Signed-off-by: Andrew Morton --- 25-akpm/arch/i386/kernel/apic.c | 2 -- 25-akpm/arch/i386/kernel/io_apic.c | 2 -- 25-akpm/arch/i386/kernel/nmi.c | 11 +++++++---- 25-akpm/arch/i386/kernel/smpboot.c | 3 --- 25-akpm/arch/x86_64/kernel/io_apic.c | 2 -- 25-akpm/arch/x86_64/kernel/nmi.c | 9 +++++++-- 25-akpm/include/asm-i386/apic.h | 1 - 7 files changed, 14 insertions(+), 16 deletions(-) diff -puN arch/i386/kernel/apic.c~rfc-check-nmi-watchdog-is-broken arch/i386/kernel/apic.c --- 25/arch/i386/kernel/apic.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/i386/kernel/apic.c 2005-03-30 18:28:21.000000000 -0800 @@ -1265,8 +1265,6 @@ int __init APIC_init_uniprocessor (void) setup_local_APIC(); - if (nmi_watchdog == NMI_LOCAL_APIC) - check_nmi_watchdog(); #ifdef CONFIG_X86_IO_APIC if (smp_found_config) if (!skip_ioapic_setup && nr_ioapics) diff -puN arch/i386/kernel/io_apic.c~rfc-check-nmi-watchdog-is-broken arch/i386/kernel/io_apic.c --- 25/arch/i386/kernel/io_apic.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/i386/kernel/io_apic.c 2005-03-30 18:28:21.000000000 -0800 @@ -2175,7 +2175,6 @@ static inline void check_timer(void) disable_8259A_irq(0); setup_nmi(); enable_8259A_irq(0); - check_nmi_watchdog(); } return; } @@ -2198,7 +2197,6 @@ static inline void check_timer(void) add_pin_to_irq(0, 0, pin2); if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); - check_nmi_watchdog(); } return; } diff -puN arch/i386/kernel/nmi.c~rfc-check-nmi-watchdog-is-broken arch/i386/kernel/nmi.c --- 25/arch/i386/kernel/nmi.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/i386/kernel/nmi.c 2005-03-30 18:28:21.000000000 -0800 @@ -102,20 +102,21 @@ int nmi_active; (P4_CCCR_OVF_PMI0|P4_CCCR_THRESHOLD(15)|P4_CCCR_COMPLEMENT| \ P4_CCCR_COMPARE|P4_CCCR_REQUIRED|P4_CCCR_ESCR_SELECT(4)|P4_CCCR_ENABLE) -int __init check_nmi_watchdog (void) +static int __init check_nmi_watchdog(void) { unsigned int prev_nmi_count[NR_CPUS]; int cpu; - printk(KERN_INFO "testing NMI watchdog ... "); + if (nmi_watchdog == NMI_NONE) + return 0; + + printk(KERN_INFO "Testing NMI watchdog ... "); for (cpu = 0; cpu < NR_CPUS; cpu++) prev_nmi_count[cpu] = per_cpu(irq_stat, cpu).__nmi_count; local_irq_enable(); mdelay((10*1000)/nmi_hz); // wait 10 ticks - /* FIXME: Only boot CPU is online at this stage. Check CPUs - as they come up. */ for (cpu = 0; cpu < NR_CPUS; cpu++) { #ifdef CONFIG_SMP /* Check cpu_callin_map here because that is set @@ -139,6 +140,8 @@ int __init check_nmi_watchdog (void) return 0; } +/* This needs to happen later in boot so counters are working */ +late_initcall(check_nmi_watchdog); static int __init setup_nmi_watchdog(char *str) { diff -puN arch/i386/kernel/smpboot.c~rfc-check-nmi-watchdog-is-broken arch/i386/kernel/smpboot.c --- 25/arch/i386/kernel/smpboot.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/i386/kernel/smpboot.c 2005-03-30 18:28:21.000000000 -0800 @@ -1062,9 +1062,6 @@ static void __init smp_boot_cpus(unsigne printk(KERN_WARNING "WARNING: %d siblings found for CPU%d, should be %d\n", siblings, cpu, smp_num_siblings); } - if (nmi_watchdog == NMI_LOCAL_APIC) - check_nmi_watchdog(); - smpboot_setup_io_apic(); setup_boot_APIC_clock(); diff -puN arch/x86_64/kernel/io_apic.c~rfc-check-nmi-watchdog-is-broken arch/x86_64/kernel/io_apic.c --- 25/arch/x86_64/kernel/io_apic.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/x86_64/kernel/io_apic.c 2005-03-30 18:28:21.000000000 -0800 @@ -1607,7 +1607,6 @@ static inline void check_timer(void) disable_8259A_irq(0); setup_nmi(); enable_8259A_irq(0); - check_nmi_watchdog(); } return; } @@ -1627,7 +1626,6 @@ static inline void check_timer(void) nmi_watchdog_default(); if (nmi_watchdog == NMI_IO_APIC) { setup_nmi(); - check_nmi_watchdog(); } return; } diff -puN arch/x86_64/kernel/nmi.c~rfc-check-nmi-watchdog-is-broken arch/x86_64/kernel/nmi.c --- 25/arch/x86_64/kernel/nmi.c~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:28:21.000000000 -0800 +++ 25-akpm/arch/x86_64/kernel/nmi.c 2005-03-30 18:28:21.000000000 -0800 @@ -112,17 +112,20 @@ static __init int cpu_has_lapic(void) } } -int __init check_nmi_watchdog (void) +static int __init check_nmi_watchdog (void) { int counts[NR_CPUS]; int cpu; + if (nmi_watchdog == NMI_NONE) + return 0; + if (nmi_watchdog == NMI_LOCAL_APIC && !cpu_has_lapic()) { nmi_watchdog = NMI_NONE; return -1; } - printk(KERN_INFO "testing NMI watchdog ... "); + printk(KERN_INFO "Testing NMI watchdog ... "); for (cpu = 0; cpu < NR_CPUS; cpu++) counts[cpu] = cpu_pda[cpu].__nmi_count; @@ -154,6 +157,8 @@ int __init check_nmi_watchdog (void) return 0; } +/* Have this called later during boot so counters are updating */ +late_initcall(check_nmi_watchdog); int __init setup_nmi_watchdog(char *str) { diff -puN include/asm-i386/apic.h~rfc-check-nmi-watchdog-is-broken include/asm-i386/apic.h --- 25/include/asm-i386/apic.h~rfc-check-nmi-watchdog-is-broken 2005-03-30 18:29:19.000000000 -0800 +++ 25-akpm/include/asm-i386/apic.h 2005-03-30 18:29:25.000000000 -0800 @@ -109,7 +109,6 @@ extern int APIC_init_uniprocessor (void) extern void disable_APIC_timer(void); extern void enable_APIC_timer(void); -extern int check_nmi_watchdog (void); extern void enable_NMI_through_LVT0 (void * dummy); extern unsigned int nmi_watchdog; _