diff options
| author | Peter Zijlstra <peterz@infradead.org> | 2016-03-25 15:52:34 +0100 |
|---|---|---|
| committer | Thomas Gleixner <tglx@linutronix.de> | 2016-03-29 10:45:04 +0200 |
| commit | ee6825c80e870fff1a370c718ec77022ade0889b (patch) | |
| tree | ea5476978095b97d911d00e5a21c1c769c7f401a /arch/x86/ras | |
| parent | f55532a0c0b8bb6148f4e07853b876ef73bc69ca (diff) | |
| download | linux-ee6825c80e870fff1a370c718ec77022ade0889b.tar.gz | |
x86/topology: Fix AMD core count
It turns out AMD gets x86_max_cores wrong when there are compute
units.
The issue is that Linux assumes:
nr_logical_cpus = nr_cores * nr_siblings
But AMD reports its CU unit as 2 cores, but then sets num_smp_siblings
to 2 as well.
Boris: fixup ras/mce_amd_inj.c too, to compute the Node Base Core
properly, according to the new nomenclature.
Fixes: 1f12e32f4cd5 ("x86/topology: Create logical package id")
Reported-by: Xiong Zhou <jencce.kernel@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Cc: Andreas Herrmann <aherrmann@suse.com>
Cc: Andy Lutomirski <luto@kernel.org>
Link: http://lkml.kernel.org/r/20160317095220.GO6344@twins.programming.kicks-ass.net
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Diffstat (limited to 'arch/x86/ras')
| -rw-r--r-- | arch/x86/ras/mce_amd_inj.c | 3 |
1 files changed, 2 insertions, 1 deletions
diff --git a/arch/x86/ras/mce_amd_inj.c b/arch/x86/ras/mce_amd_inj.c index 55d38cfa46c2..9e02dcaef683 100644 --- a/arch/x86/ras/mce_amd_inj.c +++ b/arch/x86/ras/mce_amd_inj.c @@ -20,6 +20,7 @@ #include <linux/pci.h> #include <asm/mce.h> +#include <asm/smp.h> #include <asm/amd_nb.h> #include <asm/irq_vectors.h> @@ -206,7 +207,7 @@ static u32 get_nbc_for_node(int node_id) struct cpuinfo_x86 *c = &boot_cpu_data; u32 cores_per_node; - cores_per_node = c->x86_max_cores / amd_get_nodes_per_socket(); + cores_per_node = (c->x86_max_cores * smp_num_siblings) / amd_get_nodes_per_socket(); return cores_per_node * node_id; } |