On Tue, Aug 21, 2012 at 05:46:08PM -0700, Paul E. McKenney wrote: > On Tue, Aug 21, 2012 at 11:53:50PM +0000, Luck, Tony wrote: > > Thanks for the pointers. > > > > I turned on CONFIG_RCU_CPU_STALL_INFO=y and bumped RCU_STALL_RAT_DELAY > > from 2 to 20 > > > > This is the new console log. There is a minute of hang before the first > > pair of stack traces. Then hang for a minute and the second pair show > > up. > > > > Linux version 3.6.0-rc2-zx1-smp-next-20120821 (aegl@linux-bxb1) (gcc version 4.3.4 [gcc-4_3-branch revision 152973] (SUSE Linux) ) #2 SMP Tue Aug 21 16:44:17 PDT 2012 > > EFI v1.10 by HP: SALsystab=0x3fefa000 ACPI 2.0=0x3fd5e000 SMBIOS=0x3fefc000 HCDP=0x3fd5c000 > > Early serial console at MMIO 0xff5e0000 (options '9600') > > bootconsole [uart0] enabled > > PCDP: v0 at 0x3fd5c000 > > Explicit "console="; ignoring PCDP > > ACPI: RSDP 000000003fd5e000 00028 (v02 HP) > > ACPI: XSDT 000000003fd5e02c 00094 (v01 HP rx2620 00000000 HP 00000000) > > ACPI: FACP 000000003fd67390 000F4 (v03 HP rx2620 00000000 HP 00000000) > > ACPI BIOS Bug: Warning: 32/64X length mismatch in FADT/Gpe0Block: 32/16 (20120711/tbfadt-567) > > ACPI BIOS Bug: Warning: 32/64X length mismatch in FADT/Gpe1Block: 32/16 (20120711/tbfadt-567) > > ACPI: DSDT 000000003fd5e100 05F3C (v01 HP rx2620 00000007 INTL 02012044) > > ACPI: FACS 000000003fd67488 00040 > > ACPI: SPCR 000000003fd674c8 00050 (v01 HP rx2620 00000000 HP 00000000) > > ACPI: DBGP 000000003fd67518 00034 (v01 HP rx2620 00000000 HP 00000000) > > ACPI: APIC 000000003fd67610 000B0 (v01 HP rx2620 00000000 HP 00000000) > > ACPI: SPMI 000000003fd67550 00050 (v04 HP rx2620 00000000 HP 00000000) > > ACPI: CPEP 000000003fd675a0 00034 (v01 HP rx2620 00000000 HP 00000000) > > ACPI: SSDT 000000003fd64040 001D6 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd64220 00702 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd64930 00A16 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd65350 00A16 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd65d70 00A16 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd66790 00A16 (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd671b0 000EB (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: SSDT 000000003fd672a0 000EF (v01 HP rx2620 00000006 INTL 02012044) > > ACPI: Local APIC address c0000000fee00000 > > 2 CPUs available, 2 CPUs total > > warning: skipping physical page 0 > > Initial ramdisk at: 0xe00000407e9bb000 (6071698 bytes) > > SAL 3.1: HP version 3.15 > > SAL Platform features: None > > SAL: AP wakeup using external interrupt vector 0xff > > MCA related initialization done > > warning: skipping physical page 0 > > Zone ranges: > > DMA [mem 0x00004000-0xffffffff] > > Normal [mem 0x100000000-0x407ffc7fff] > > Movable zone start for each node > > Early memory node ranges > > node 0: [mem 0x00004000-0x3f4ebfff] > > node 0: [mem 0x3fc00000-0x3fd5bfff] > > node 0: [mem 0x4040000000-0x407fd2bfff] > > node 0: [mem 0x407fd98000-0x407fe07fff] > > node 0: [mem 0x407fe80000-0x407ffc7fff] > > Virtual mem_map starts at 0xa0007fffc7900000 > > Built 1 zonelists in Zone order, mobility grouping off. Total pages: 72586 > > Kernel command line: BOOT_IMAGE=scsi0:\efi\SuSE\l-zx1-smp.gz root=/dev/disk/by-id/scsi-200000e1100a5d5f2-part2 console=uart,mmio,0xff5e0000 > > PID hash table entries: 4096 (order: 1, 32768 bytes) > > Dentry cache hash table entries: 262144 (order: 7, 2097152 bytes) > > Inode-cache hash table entries: 131072 (order: 6, 1048576 bytes) > > Memory: 2048432k/2086064k available (13698k code, 37632k reserved, 5791k data, 816k init) > > SLUB: Genslabs=17, HWalign=128, Order=0-3, MinObjects=0, CPUs=2, Nodes=1 > > Hierarchical RCU implementation. > > Additional per-CPU info printed with stalls. > > RCU restricting CPUs from NR_CPUS=16 to nr_cpu_ids=2. > > NR_IRQS:768 > > ACPI: Local APIC address c0000000fee00000 > > GSI 36 (level, low) -> CPU 0 (0x0000) vector 48 > > Console: colour dummy device 80x25 > > Calibrating delay loop... 1945.60 BogoMIPS (lpj=3891200) > > pid_max: default: 32768 minimum: 301 > > Mount-cache hash table entries: 1024 > > ACPI: Core revision 20120711 > > Boot processor id 0x0/0x0 > > Fixed BSP b0 value from CPU 1 > > CPU 1: synchronized ITC with CPU 0 (last diff -3 cycles, maxerr 579 cycles) > > Brought up 2 CPUs > > Total of 2 processors activated (3891.20 BogoMIPS). > > SMBIOS 2.3 present. > > NET: Registered protocol family 16 > > ACPI: bus type pci registered > > bio: create slab <bio-0> at 0 > > ACPI: Added _OSI(Module Device) > > ACPI: Added _OSI(Processor Device) > > ACPI: Added _OSI(3.0 _SCP Extensions) > > ACPI: Added _OSI(Processor Aggregator Device) > > INFO: rcu_sched self-detected stall on CPU > > 1: (15000 ticks this GP) idle=001/140000000000001/0 > > OK, this is strange. The stacks below would lead me to believe that > the CPUs are idle. But the idle= value above says that RCU believes > that this CPU was executing in non-idle process context when the > interrupt occurred. > > OK, time to take a look at the IA64 idle loop. And I don't see any > calls to rcu_idle_enter()... Please see below for my best guess as > to where to place it and rcu_idle_exit() -- the rule is that there must > be no use of RCU read-side critical sections between the call to the > rcu_idle_enter() and the rcu_idle_exit(), so you probably know better > than I where to put them. > > void __attribute__((noreturn)) > cpu_idle (void) > { > void (*mark_idle)(int) = ia64_mark_idle; > int cpu = smp_processor_id(); > > /* endless idle loop with no priority at all */ > while (1) { > rcu_idle_enter(); /* HERE */ > if (can_do_pal_halt) { > current_thread_info()->status &= ~TS_POLLING; > /* > * TS_POLLING-cleared state must be visible before we > * test NEED_RESCHED: > */ > smp_mb(); > } else { > current_thread_info()->status |= TS_POLLING; > } > > if (!need_resched()) { > void (*idle)(void); > #ifdef CONFIG_SMP > min_xtp(); > #endif > rmb(); > if (mark_idle) > (*mark_idle)(1); > > idle = pm_idle; > if (!idle) > idle = default_idle; > (*idle)(); > if (mark_idle) > (*mark_idle)(0); > #ifdef CONFIG_SMP > normal_xtp(); > #endif > } > rcu_idle_exit(); /* AND HERE */ > schedule_preempt_disabled(); > check_pgt_cache(); > if (cpu_is_offline(cpu)) > play_dead(); > } > } > > Without the calls to rcu_idle_enter() and rcu_idle_exit(), RCU has no > way of knowing that the CPU is idle, so waits forever for a context > switch. > > Ah, I bet I know what happened... I don't see tick_nohz_idle_enter(), > so I would guess that there is no dyntick-idle, so the recent changes in > dyntick-idle didn't cause rcu_idle_enter() to be added. > > I wonder how many other architectures don't do dyntick-idle? > > Looks like about 12 more. Probably need fixing as well... Ouch, that's bad. Ok see below for the conversion of other architectures. While doing this, I realized that most of these archs just use the same cpu_idle() function, basically: void cpu_idle(void) { while (1) { + rcu_idle_enter(); while (!need_resched()) do_arch_thing(); + rcu_idle_exit(); schedule_preempt_disabled(); } } So I think it may be worth creating a "simple idle loop" generic function for those archs that they can call. This way there is less conversion to do. Now this is all a regression, so IMO we should first fix the things locally and do that generic idle loop later, since it's rather a feature. Hmm? I'm cooking the patches. diff --git a/arch/alpha/kernel/process.c b/arch/alpha/kernel/process.c index 153d3fc..2ebf7b5 100644 --- a/arch/alpha/kernel/process.c +++ b/arch/alpha/kernel/process.c @@ -28,6 +28,7 @@ #include <linux/tty.h> #include <linux/console.h> #include <linux/slab.h> +#include <linux/rcupdate.h> #include <asm/reg.h> #include <asm/uaccess.h> @@ -50,13 +51,16 @@ cpu_idle(void) { set_thread_flag(TIF_POLLING_NRFLAG); + preempt_disable(); while (1) { /* FIXME -- EV6 and LCA45 know how to power down the CPU. */ + rcu_idle_enter(); while (!need_resched()) cpu_relax(); - schedule(); + rcu_idle_exit(); + schedule_preempt_disabled(); } } diff --git a/arch/cris/kernel/process.c b/arch/cris/kernel/process.c index 66fd017..7f65be6 100644 --- a/arch/cris/kernel/process.c +++ b/arch/cris/kernel/process.c @@ -25,6 +25,7 @@ #include <linux/elfcore.h> #include <linux/mqueue.h> #include <linux/reboot.h> +#include <linux/rcupdate.h> //#define DEBUG @@ -74,6 +75,7 @@ void cpu_idle (void) { /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) { void (*idle)(void); /* @@ -86,6 +88,7 @@ void cpu_idle (void) idle = default_idle; idle(); } + rcu_idle_exit(); schedule_preempt_disabled(); } } diff --git a/arch/frv/kernel/process.c b/arch/frv/kernel/process.c index ff95f50..2eb7fa5 100644 --- a/arch/frv/kernel/process.c +++ b/arch/frv/kernel/process.c @@ -25,6 +25,7 @@ #include <linux/reboot.h> #include <linux/interrupt.h> #include <linux/pagemap.h> +#include <linux/rcupdate.h> #include <asm/asm-offsets.h> #include <asm/uaccess.h> @@ -69,12 +70,14 @@ void cpu_idle(void) { /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) { check_pgt_cache(); if (!frv_dma_inprogress && idle) idle(); } + rcu_idle_exit(); schedule_preempt_disabled(); } diff --git a/arch/h8300/kernel/process.c b/arch/h8300/kernel/process.c index 0e9c315..f153ed1 100644 --- a/arch/h8300/kernel/process.c +++ b/arch/h8300/kernel/process.c @@ -36,6 +36,7 @@ #include <linux/reboot.h> #include <linux/fs.h> #include <linux/slab.h> +#include <linux/rcupdate.h> #include <asm/uaccess.h> #include <asm/traps.h> @@ -78,8 +79,10 @@ void (*idle)(void) = default_idle; void cpu_idle(void) { while (1) { + rcu_idle_enter(); while (!need_resched()) idle(); + rcu_idle_exit(); schedule_preempt_disabled(); } } diff --git a/arch/m32r/kernel/process.c b/arch/m32r/kernel/process.c index 3a4a32b2..384e63f 100644 --- a/arch/m32r/kernel/process.c +++ b/arch/m32r/kernel/process.c @@ -26,6 +26,7 @@ #include <linux/ptrace.h> #include <linux/unistd.h> #include <linux/hardirq.h> +#include <linux/rcupdate.h> #include <asm/io.h> #include <asm/uaccess.h> @@ -82,6 +83,7 @@ void cpu_idle (void) { /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) { void (*idle)(void) = pm_idle; @@ -90,6 +92,7 @@ void cpu_idle (void) idle(); } + rcu_idle_exit(); schedule_preempt_disabled(); } } diff --git a/arch/m68k/kernel/process.c b/arch/m68k/kernel/process.c index c488e3c..ac2892e 100644 --- a/arch/m68k/kernel/process.c +++ b/arch/m68k/kernel/process.c @@ -25,6 +25,7 @@ #include <linux/reboot.h> #include <linux/init_task.h> #include <linux/mqueue.h> +#include <linux/rcupdate.h> #include <asm/uaccess.h> #include <asm/traps.h> @@ -75,8 +76,10 @@ void cpu_idle(void) { /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) idle(); + rcu_idle_exit(); schedule_preempt_disabled(); } } diff --git a/arch/mn10300/kernel/process.c b/arch/mn10300/kernel/process.c index 7dab0cd..e9cceba 100644 --- a/arch/mn10300/kernel/process.c +++ b/arch/mn10300/kernel/process.c @@ -25,6 +25,7 @@ #include <linux/err.h> #include <linux/fs.h> #include <linux/slab.h> +#include <linux/rcupdate.h> #include <asm/uaccess.h> #include <asm/pgtable.h> #include <asm/io.h> @@ -107,6 +108,7 @@ void cpu_idle(void) { /* endless idle loop with no priority at all */ for (;;) { + rcu_idle_enter(); while (!need_resched()) { void (*idle)(void); @@ -121,6 +123,7 @@ void cpu_idle(void) } idle(); } + rcu_idle_exit(); schedule_preempt_disabled(); } diff --git a/arch/parisc/kernel/process.c b/arch/parisc/kernel/process.c index d4b94b3..c54a4db 100644 --- a/arch/parisc/kernel/process.c +++ b/arch/parisc/kernel/process.c @@ -48,6 +48,7 @@ #include <linux/unistd.h> #include <linux/kallsyms.h> #include <linux/uaccess.h> +#include <linux/rcupdate.h> #include <asm/io.h> #include <asm/asm-offsets.h> @@ -69,8 +70,10 @@ void cpu_idle(void) /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) barrier(); + rcu_idle_exit(); schedule_preempt_disabled(); check_pgt_cache(); } diff --git a/arch/score/kernel/process.c b/arch/score/kernel/process.c index 2707023..637970c 100644 --- a/arch/score/kernel/process.c +++ b/arch/score/kernel/process.c @@ -27,6 +27,7 @@ #include <linux/reboot.h> #include <linux/elfcore.h> #include <linux/pm.h> +#include <linux/rcupdate.h> void (*pm_power_off)(void); EXPORT_SYMBOL(pm_power_off); @@ -50,9 +51,10 @@ void __noreturn cpu_idle(void) { /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) barrier(); - + rcu_idle_exit(); schedule_preempt_disabled(); } } diff --git a/arch/xtensa/kernel/process.c b/arch/xtensa/kernel/process.c index 2c8d6a3..bc44311 100644 --- a/arch/xtensa/kernel/process.c +++ b/arch/xtensa/kernel/process.c @@ -31,6 +31,7 @@ #include <linux/mqueue.h> #include <linux/fs.h> #include <linux/slab.h> +#include <linux/rcupdate.h> #include <asm/pgtable.h> #include <asm/uaccess.h> @@ -110,8 +111,10 @@ void cpu_idle(void) /* endless idle loop with no priority at all */ while (1) { + rcu_idle_enter(); while (!need_resched()) platform_idle(); + rcu_idle_exit(); schedule_preempt_disabled(); } } -- To unsubscribe from this list: send the line "unsubscribe linux-next" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html