in x86_64 smp systems, a regression was introduced a few days after 2.6.23,
on its way to 2.6.23. Bisection leads me to commit
92cb7612aee39642d109b8d935ad265e602c0563.
The problem is that the output of 'cat /proc/cpuinfo' shows 0 as the id of
all cpus of a smp system. As far as I know, this only happens on x86_64
systems.
This problem was first noticed by Ravikiran. A simple patch follows
Signed-off-by: Glauber de Oliveira Costa <[email protected]>
Cc: Ravikiran Thirumalai <[email protected]>
---
arch/x86/kernel/setup_64.c | 1 -
1 files changed, 0 insertions(+), 1 deletions(-)
diff --git a/arch/x86/kernel/setup_64.c b/arch/x86/kernel/setup_64.c
index 0f0d9d4..eb41c99 100644
--- a/arch/x86/kernel/setup_64.c
+++ b/arch/x86/kernel/setup_64.c
@@ -858,7 +858,6 @@ void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c)
#ifdef CONFIG_SMP
c->phys_proc_id = (cpuid_ebx(1) >> 24) & 0xff;
- c->cpu_index = 0;
#endif
}
--
1.4.4.2
* Glauber de Oliveira Costa <[email protected]> wrote:
> in x86_64 smp systems, a regression was introduced a few days after
> 2.6.23, on its way to 2.6.23. Bisection leads me to commit
> 92cb7612aee39642d109b8d935ad265e602c0563.
>
> The problem is that the output of 'cat /proc/cpuinfo' shows 0 as the
> id of all cpus of a smp system. As far as I know, this only happens on
> x86_64 systems.
>
> This problem was first noticed by Ravikiran. A simple patch follows
oops, thanks for tracking this down!
Acked-by: Ingo Molnar <[email protected]>
Ingo