Gitweb:
http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=ca74a6f84e68b44867022f4a4f3ec17c087c864e
Commit: ca74a6f84e68b44867022f4a4f3ec17c087c864e
Parent: 751752789162fde69474edfa15935d0a77c0bc17
Author: Andi Kleen <[EMAIL PROTECTED]>
AuthorDate: Wed Jan 30 13:33:17 2008 +0100
Committer: Ingo Molnar <[EMAIL PROTECTED]>
CommitDate: Wed Jan 30 13:33:17 2008 +0100
x86: optimize lock prefix switching to run less frequently
On VMs implemented using JITs that cache translated code changing the lock
prefixes is a quite costly operation that forces the JIT to throw away and
retranslate a lot of code.
Previously a SMP kernel would rewrite the locks once for each CPU which
is quite unnecessary. This patch changes the code to never switch at boot in
the normal case (SMP kernel booting with >1 CPU) or only once for SMP
kernel
on UP.
This makes a significant difference in boot up performance on AMD SimNow!
Also I expect it to be a little faster on native systems too because a smp
switch does a lot of text_poke()s which each synchronize the pipeline.
v1->v2: Rename max_cpus
v1->v2: Fix off by one in UP check (Thomas Gleixner)
Signed-off-by: Andi Kleen <[EMAIL PROTECTED]>
Signed-off-by: Ingo Molnar <[EMAIL PROTECTED]>
Signed-off-by: Thomas Gleixner <[EMAIL PROTECTED]>
---
arch/x86/kernel/alternative.c | 16 ++++++++++++++--
include/linux/smp.h | 2 ++
init/main.c | 16 ++++++++--------
3 files changed, 24 insertions(+), 10 deletions(-)
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index cdc4324..318a4f9 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -273,6 +273,7 @@ struct smp_alt_module {
};
static LIST_HEAD(smp_alt_modules);
static DEFINE_SPINLOCK(smp_alt);
+static int smp_mode = 1; /* protected by smp_alt */
void alternatives_smp_module_add(struct module *mod, char *name,
void *locks, void *locks_end,
@@ -354,7 +355,14 @@ void alternatives_smp_switch(int smp)
BUG_ON(!smp && (num_online_cpus() > 1));
spin_lock_irqsave(&smp_alt, flags);
- if (smp) {
+
+ /*
+ * Avoid unnecessary switches because it forces JIT based VMs to
+ * throw away all cached translations, which can be quite costly.
+ */
+ if (smp == smp_mode) {
+ /* nothing */
+ } else if (smp) {
printk(KERN_INFO "SMP alternatives: switching to SMP code\n");
clear_cpu_cap(&boot_cpu_data, X86_FEATURE_UP);
clear_cpu_cap(&cpu_data(0), X86_FEATURE_UP);
@@ -369,6 +377,7 @@ void alternatives_smp_switch(int smp)
alternatives_smp_unlock(mod->locks, mod->locks_end,
mod->text, mod->text_end);
}
+ smp_mode = smp;
spin_unlock_irqrestore(&smp_alt, flags);
}
@@ -441,7 +450,10 @@ void __init alternative_instructions(void)
alternatives_smp_module_add(NULL, "core kernel",
__smp_locks, __smp_locks_end,
_text, _etext);
- alternatives_smp_switch(0);
+
+ /* Only switch to UP mode if we don't immediately boot others */
+ if (num_possible_cpus() == 1 || setup_max_cpus <= 1)
+ alternatives_smp_switch(0);
}
#endif
apply_paravirt(__parainstructions, __parainstructions_end);
diff --git a/include/linux/smp.h b/include/linux/smp.h
index c25e66b..55232cc 100644
--- a/include/linux/smp.h
+++ b/include/linux/smp.h
@@ -78,6 +78,8 @@ int on_each_cpu(void (*func) (void *info), void *info, int
retry, int wait);
*/
void smp_prepare_boot_cpu(void);
+extern unsigned int setup_max_cpus;
+
#else /* !SMP */
/*
diff --git a/init/main.c b/init/main.c
index 3f8aba2..5843fe9 100644
--- a/init/main.c
+++ b/init/main.c
@@ -128,7 +128,7 @@ static char *ramdisk_execute_command;
#ifdef CONFIG_SMP
/* Setup configured maximum number of CPUs to activate */
-static unsigned int __initdata max_cpus = NR_CPUS;
+unsigned int __initdata setup_max_cpus = NR_CPUS;
/*
* Setup routine for controlling SMP activation
@@ -146,7 +146,7 @@ static inline void disable_ioapic_setup(void) {};
static int __init nosmp(char *str)
{
- max_cpus = 0;
+ setup_max_cpus = 0;
disable_ioapic_setup();
return 0;
}
@@ -155,8 +155,8 @@ early_param("nosmp", nosmp);
static int __init maxcpus(char *str)
{
- get_option(&str, &max_cpus);
- if (max_cpus == 0)
+ get_option(&str, &setup_max_cpus);
+ if (setup_max_cpus == 0)
disable_ioapic_setup();
return 0;
@@ -164,7 +164,7 @@ static int __init maxcpus(char *str)
early_param("maxcpus", maxcpus);
#else
-#define max_cpus NR_CPUS
+#define setup_max_cpus NR_CPUS
#endif
/*
@@ -393,7 +393,7 @@ static void __init smp_init(void)
/* FIXME: This should be done in userspace --RR */
for_each_present_cpu(cpu) {
- if (num_online_cpus() >= max_cpus)
+ if (num_online_cpus() >= setup_max_cpus)
break;
if (!cpu_online(cpu))
cpu_up(cpu);
@@ -401,7 +401,7 @@ static void __init smp_init(void)
/* Any cleanup work */
printk(KERN_INFO "Brought up %ld CPUs\n", (long)num_online_cpus());
- smp_cpus_done(max_cpus);
+ smp_cpus_done(setup_max_cpus);
}
#endif
@@ -824,7 +824,7 @@ static int __init kernel_init(void * unused)
__set_special_pids(1, 1);
cad_pid = task_pid(current);
- smp_prepare_cpus(max_cpus);
+ smp_prepare_cpus(setup_max_cpus);
do_pre_smp_initcalls();
-
To unsubscribe from this list: send the line "unsubscribe git-commits-head" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
