linux/arch/x86
Paul E. McKenney bf51935f3e x86, rcu: fix strange load average and ksoftirqd behavior
Damien Wyart reported high ksoftirqd CPU usage (20%) on an
otherwise idle system.

The function-graph trace Damien provided:

>   799.521187 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.521371 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.521555 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.521738 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.521934 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.522068 |   1)  ksoftir-2324  |               |                rcu_check_callbacks() {
>   799.522208 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.522392 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.522575 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.522759 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.522956 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.523074 |   1)  ksoftir-2324  |               |                  rcu_check_callbacks() {
>   799.523214 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.523397 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.523579 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.523762 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.523960 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.524079 |   1)  ksoftir-2324  |               |                  rcu_check_callbacks() {
>   799.524220 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.524403 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.524587 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
>   799.524770 |   1)    <idle>-0    |               |  rcu_check_callbacks() {
> [ . . . ]

Shows rcu_check_callbacks() being invoked way too often. It should be called
once per jiffy, and here it is called no less than 22 times in about
3.5 milliseconds, meaning one call every 160 microseconds or so.

Why do we need to call rcu_pending() and rcu_check_callbacks() from the
idle loop of 32-bit x86, especially given that no other architecture does
this?

The following patch removes the call to rcu_pending() and
rcu_check_callbacks() from the x86 32-bit idle loop in order to
reduce the softirq load on idle systems.

Reported-by: Damien Wyart <damien.wyart@free.fr>
Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
2009-02-17 22:47:45 +01:00
..
boot x86 setup: fix asm constraints in vesa_store_edid 2009-01-30 23:55:25 +01:00
configs eeprom: More consistent symbol names 2009-01-26 21:19:57 +01:00
crypto crypto: crc32c-intel - Switch to shash 2008-12-25 11:01:37 +11:00
ia32 x86: introducing asm/sys_ia32.h 2008-12-29 13:18:40 +01:00
include/asm x86/paravirt: make arch_flush_lazy_mmu/cpu disable preemption 2009-02-12 23:11:58 +01:00
kernel x86, rcu: fix strange load average and ksoftirqd behavior 2009-02-17 22:47:45 +01:00
kvm KVM: change KVM to use IOMMU API 2009-01-03 14:11:07 +01:00
lguest lguest: typos fix 2009-01-30 11:34:10 +10:30
lib x86: use early clobbers in usercopy*.c 2009-01-21 09:43:17 +01:00
mach-default x86/Voyager: make it build and boot 2009-01-31 18:26:07 +01:00
mach-generic x86: rename all fields of mpc_table mpc_X to X 2009-01-05 14:08:34 +01:00
mach-rdc321x removed unused #include <linux/version.h>'s 2008-08-23 12:14:12 -07:00
mach-voyager x86/Voyager: make it build and boot 2009-01-31 18:26:07 +01:00
math-emu x86: fix math_emu register frame access 2009-02-10 00:39:14 +01:00
mm x86: CPA avoid repeated lazy mmu flush 2009-02-12 23:11:58 +01:00
oprofile Merge branch 'oprofile/ring_buffer' into oprofile/oprofile-for-tip 2009-01-08 14:27:34 +01:00
pci PCI: irq and pci_ids patch for Intel Tigerpoint DeviceIDs 2009-01-27 09:46:08 -08:00
power x86, hibernate: fix breakage on x86_32 with CONFIG_NUMA set 2008-11-12 23:28:51 +01:00
vdso Merge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6 2008-12-28 12:33:21 -08:00
video
xen xen: disable interrupts before saving in percpu 2009-02-03 17:22:40 -08:00
Kconfig x86: disable intel_iommu support by default 2009-02-05 16:48:38 +01:00
Kconfig.cpu x86/Kconfig.cpu: make Kconfig help readable in the console 2009-02-04 22:19:27 +01:00
Kconfig.debug Merge branch 'tracing-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2008-12-28 12:21:10 -08:00
Makefile x86, um: ... and asm-x86 move 2008-10-22 22:55:20 -07:00
Makefile_32.cpu x86: merge winchip-2 and winchip-2a cpu choices 2008-10-13 10:22:48 +02:00