2017-07-20 08:30:42

by Suthikulpanit, Suravee

[permalink] [raw]
Subject: [PATCH] x86/amd: Fixup cpu_core_id for family17h downcore configuration

For family17h, current cpu_core_id is directly taken from the value
CPUID_Fn8000001E_EBX[7:0] (CoreId), which is the physical ID of the
core within a die. However, on system with downcore configuration
(where not all physical cores within a die are available),
this could result in the case where cpu_core_id > (cores_per_node - 1).

Fix up the cpu_core_id by breaking down the bitfields of CoreId,
and calculate relative ID using available topology information.

Signed-off-by: Suravee Suthikulpanit <[email protected]>
---
arch/x86/kernel/cpu/amd.c | 61 ++++++++++++++++++++++++++++++++++-------------
1 file changed, 45 insertions(+), 16 deletions(-)

diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index bb5abe8..e7de105 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -310,38 +310,67 @@ static void amd_get_topology(struct cpuinfo_x86 *c)

/* get information required for multi-node processors */
if (boot_cpu_has(X86_FEATURE_TOPOEXT)) {
+ u16 l3_nshared = 0;
u32 eax, ebx, ecx, edx;

+ if (cpuid_edx(0x80000006)) {
+ cpuid_count(0x8000001d, 3, &eax, &ebx, &ecx, &edx);
+ l3_nshared = ((eax >> 14) & 0xfff) + 1;
+ }
+
cpuid(0x8000001e, &eax, &ebx, &ecx, &edx);

node_id = ecx & 0xff;
smp_num_siblings = ((ebx >> 8) & 0xff) + 1;

+ /* LLC is default to L3, which generally per-node */
+ if (l3_nshared > 0)
+ per_cpu(cpu_llc_id, cpu) = node_id;
+
if (c->x86 == 0x15)
c->cu_id = ebx & 0xff;

- if (c->x86 >= 0x17) {
- c->cpu_core_id = ebx & 0xff;
-
- if (smp_num_siblings > 1)
- c->x86_max_cores /= smp_num_siblings;
- }
-
/*
- * We may have multiple LLCs if L3 caches exist, so check if we
- * have an L3 cache by looking at the L3 cache CPUID leaf.
+ * In family 17h, the CPUID_Fn8000001E_EBX[7:0] (CoreId)
+ * is non-contiguous in downcore and non-SMT cases.
+ * Fixup the cpu_core_id to be contiguous for cores within
+ * the die.
*/
- if (cpuid_edx(0x80000006)) {
- if (c->x86 == 0x17) {
+ if (c->x86 >= 0x17) {
+ u32 tmp = c->cpu_core_id = ebx & 0xff;
+ u32 ccx_offset, cpu_offset;
+
+ if (smp_num_siblings == 1) {
/*
- * LLC is at the core complex level.
- * Core complex id is ApicId[3].
+ * CoreId bit-encoding for SMT-disabled
+ * [7:4] : die
+ * [3] : ccx
+ * [2:0] : core
*/
- per_cpu(cpu_llc_id, cpu) = c->apicid >> 3;
+ ccx_offset = ((tmp >> 3) & 1) * l3_nshared;
+ cpu_offset = tmp & 7;
} else {
- /* LLC is at the node level. */
- per_cpu(cpu_llc_id, cpu) = node_id;
+ /*
+ * CoreId bit-encoding for SMT-enabled
+ * [7:3] : die
+ * [2] : ccx
+ * [1:0] : core
+ */
+ c->x86_max_cores /= smp_num_siblings;
+ l3_nshared /= smp_num_siblings;
+ ccx_offset = ((tmp >> 2) & 1) * l3_nshared;
+ cpu_offset = tmp & 3;
}
+ c->cpu_core_id = ccx_offset + cpu_offset;
+ pr_debug("Fixup apicid=%#x, CoreId:%#x to cpu_core_id:%#x\n",
+ c->apicid, tmp, c->cpu_core_id);
+
+ /*
+ * Family17h L3 cache (LLC) is at Core Complex (CCX).
+ * There could be multiple CCXs in a node.
+ * CCX ID is ApicId[3].
+ */
+ per_cpu(cpu_llc_id, cpu) = c->apicid >> 3;
}
} else if (cpu_has(c, X86_FEATURE_NODEID_MSR)) {
u64 value;
--
2.7.4


2017-07-20 08:36:23

by Borislav Petkov

[permalink] [raw]
Subject: Re: [PATCH] x86/amd: Fixup cpu_core_id for family17h downcore configuration

On Thu, Jul 20, 2017 at 03:29:28AM -0500, Suravee Suthikulpanit wrote:
> For family17h, current cpu_core_id is directly taken from the value
> CPUID_Fn8000001E_EBX[7:0] (CoreId), which is the physical ID of the
> core within a die. However, on system with downcore configuration
> (where not all physical cores within a die are available),
> this could result in the case where cpu_core_id > (cores_per_node - 1).
>
> Fix up the cpu_core_id by breaking down the bitfields of CoreId,
> and calculate relative ID using available topology information.
>
> Signed-off-by: Suravee Suthikulpanit <[email protected]>
> ---
> arch/x86/kernel/cpu/amd.c | 61 ++++++++++++++++++++++++++++++++++-------------
> 1 file changed, 45 insertions(+), 16 deletions(-)
>
> diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
> index bb5abe8..e7de105 100644
> --- a/arch/x86/kernel/cpu/amd.c
> +++ b/arch/x86/kernel/cpu/amd.c
> @@ -310,38 +310,67 @@ static void amd_get_topology(struct cpuinfo_x86 *c)
>
> /* get information required for multi-node processors */
> if (boot_cpu_has(X86_FEATURE_TOPOEXT)) {

Please carve this whole TOPOEXT-specific logic out, into a separate
function, say __get_topoext() or so, for better readability/easier
review.

Thanks.

--
Regards/Gruss,
Boris.

SUSE Linux GmbH, GF: Felix Imendörffer, Jane Smithard, Graham Norton, HRB 21284 (AG Nürnberg)
--

2017-07-20 08:47:55

by Ingo Molnar

[permalink] [raw]
Subject: Re: [PATCH] x86/amd: Fixup cpu_core_id for family17h downcore configuration


* Borislav Petkov <[email protected]> wrote:

> On Thu, Jul 20, 2017 at 03:29:28AM -0500, Suravee Suthikulpanit wrote:
> > For family17h, current cpu_core_id is directly taken from the value
> > CPUID_Fn8000001E_EBX[7:0] (CoreId), which is the physical ID of the
> > core within a die. However, on system with downcore configuration
> > (where not all physical cores within a die are available),
> > this could result in the case where cpu_core_id > (cores_per_node - 1).
> >
> > Fix up the cpu_core_id by breaking down the bitfields of CoreId,
> > and calculate relative ID using available topology information.
> >
> > Signed-off-by: Suravee Suthikulpanit <[email protected]>
> > ---
> > arch/x86/kernel/cpu/amd.c | 61 ++++++++++++++++++++++++++++++++++-------------
> > 1 file changed, 45 insertions(+), 16 deletions(-)
> >
> > diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
> > index bb5abe8..e7de105 100644
> > --- a/arch/x86/kernel/cpu/amd.c
> > +++ b/arch/x86/kernel/cpu/amd.c
> > @@ -310,38 +310,67 @@ static void amd_get_topology(struct cpuinfo_x86 *c)
> >
> > /* get information required for multi-node processors */
> > if (boot_cpu_has(X86_FEATURE_TOPOEXT)) {
>
> Please carve this whole TOPOEXT-specific logic out, into a separate
> function, say __get_topoext() or so, for better readability/easier
> review.

Also please do the factoring out in a separate preparatory cleanup patch, so that
we can see the two changes isolated.

Thanks,

Ingo