Re: PATCH: Race in 2.6.0-test2 timer code

[Andrew Morton <akpm@osdl.org>]

Ingo Molnar <mingo@elte.hu> wrote:
>
> 
> On Tue, 29 Jul 2003, Andrew Morton wrote:
> 
> > Andrea says that we need to take the per-timer lock in add_timer() and
> > del_timer(), but I haven't yet got around to working out why.
> 
> this makes no sense - in 2.6 (and in 2.4) there's no safe add_timer() /
> del_timer() use without using external SMP synchronization. (There's one
> special timer use variant involving del_timer_sync() that was safe in 2.4
> but is unsafe in 2.6, see below.)
> 

Well Andrea did mention a problem with the interval timers.  But I am not
aware of the exact details of the race which he found, and I don't
understand why del_timer() and add_timer() would be needing the per-timer
locks.

You need to export __mod_timer to modules btw.

--- 25/kernel/ksyms.c~timer-race-fixes	2003-07-29 23:27:05.000000000 -0700
+++ 25-akpm/kernel/ksyms.c	2003-07-29 23:27:49.000000000 -0700
@@ -405,8 +405,6 @@ EXPORT_SYMBOL(proc_doulongvec_ms_jiffies
 EXPORT_SYMBOL(proc_doulongvec_minmax);
 
 /* interrupt handling */
-EXPORT_SYMBOL(add_timer);
-EXPORT_SYMBOL(del_timer);
 EXPORT_SYMBOL(request_irq);
 EXPORT_SYMBOL(free_irq);
 
@@ -433,7 +431,10 @@ EXPORT_SYMBOL(probe_irq_off);
 #ifdef CONFIG_SMP
 EXPORT_SYMBOL(del_timer_sync);
 #endif
+EXPORT_SYMBOL(add_timer);
+EXPORT_SYMBOL(del_timer);
 EXPORT_SYMBOL(mod_timer);
+EXPORT_SYMBOL(__mod_timer);
 
 #ifdef HAVE_DISABLE_HLT
 EXPORT_SYMBOL(disable_hlt);

_