[PATCH] i386: Prevent early access to TSC to avoid crash on TSCless systems

[PATCH] i386: Prevent early access to TSC to avoid crash on TSCless systems

[Thomas Gleixner]

commit f9690982b8c2f9a2c65acdc113e758ec356676a3 removed the check for
cpu_khz from sched_clock(), which prevented early access to the TSC by
non obvious magic.

This is harmless as long as the CPU has a TSC. On TSCless systems this
results in an illegal instruction trap.

Replace tsc_disabled and tsc_unstable by tsc_enabled, which is only set
when the tsc is available and not unstable.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

diff --git a/arch/i386/kernel/tsc.c b/arch/i386/kernel/tsc.c
index 0e65f7a..6cb8f53 100644
--- a/arch/i386/kernel/tsc.c
+++ b/arch/i386/kernel/tsc.c
@@ -18,6 +18,8 @@
 
 #include "mach_timer.h"
 
+static int tsc_enabled;
+
 /*
  * On some systems the TSC frequency does not
  * change with the cpu frequency. So we need
@@ -105,7 +107,7 @@ unsigned long long sched_clock(void)
 	/*
 	 * Fall back to jiffies if there's no TSC available:
 	 */
-	if (tsc_unstable || unlikely(tsc_disable))
+	if (unlikely(!tsc_enabled))
 		/* No locking but a rare wrong value is not a big deal: */
 		return (jiffies_64 - INITIAL_JIFFIES) * (1000000000 / HZ);
 
@@ -283,6 +285,7 @@ void mark_tsc_unstable(void)
 {
 	if (!tsc_unstable) {
 		tsc_unstable = 1;
+		tsc_enabled = 0;
 		/* Can be called before registration */
 		if (clocksource_tsc.mult)
 			clocksource_change_rating(&clocksource_tsc, 0);
@@ -383,7 +386,9 @@ void __init tsc_init(void)
 	if (check_tsc_unstable()) {
 		clocksource_tsc.rating = 0;
 		clocksource_tsc.flags &= ~CLOCK_SOURCE_IS_CONTINUOUS;
-	}
+	} else
+		tsc_enabled = 1;
+
 	clocksource_register(&clocksource_tsc);
 
 	return;

Re: [PATCH] i386: Prevent early access to TSC to avoid crash on TSCless systems

[Ingo Molnar]

* Thomas Gleixner <tglx@linutronix.de> wrote:

> commit f9690982b8c2f9a2c65acdc113e758ec356676a3 removed the check for 
> cpu_khz from sched_clock(), which prevented early access to the TSC by 
> non obvious magic.
> 
> This is harmless as long as the CPU has a TSC. On TSCless systems this 
> results in an illegal instruction trap.
> 
> Replace tsc_disabled and tsc_unstable by tsc_enabled, which is only 
> set when the tsc is available and not unstable.
> 
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

oops, indeed! I think this should also resolve one of the bugs (TSC-less 
Cyrix?) that were reported against -mm some time ago.

Acked-by: Ingo Molnar <mingo@elte.hu>

	Ingo

Re: [PATCH] i386: Prevent early access to TSC to avoid crash on TSCless systems

[Guillaume Chazarain]

> +static int tsc_enabled;

So, now we have tsc_disable, tsc_enabled and tsc_unstable.
I can understand the latter, but this lacks orthogonality IMHO.

-- 
Guillaume

Re: [PATCH] i386: Prevent early access to TSC to avoid crash on TSCless systems

[Ingo Molnar]

* Guillaume Chazarain <guichaz@yahoo.fr> wrote:

> >+static int tsc_enabled;
> 
> So, now we have tsc_disable, tsc_enabled and tsc_unstable. I can 
> understand the latter, but this lacks orthogonality IMHO.

tsc_disable should be renamed to tsc_disable_override or so, to signal 
that it's only the mirror of the 'notsc' flag the user passed in over 
the boot line.

	Ingo