Re: [Patch] x86, x86_64: Intel dual-core detection
On Wed, Mar 16, 2005 at 06:04:12PM +0100, Andi Kleen wrote:
> On Tue, Mar 15, 2005 at 05:36:25PM -0800, Siddha, Suresh B wrote:
> > It adds two new fields "core id" and "cpu cores" to x86 /proc/cpuinfo
> > and the "core id" field for x86_64("cpu cores" field is already present in
> > x86_64).
>
> Thanks. I have a similar patch for AMD CPUs, unfortuntely it uses
> different names ("shared cores" etc.)
2.6.11 already has "cpu cores" for x86_64. I would like to retain that.
> >
> > This patch also adds cpu_core_map similar to cpu_sibling_map.
> Called cpu_sharecore_map in my patch.
cpu_sharecore_map sounds like HT logical siblings sharing the core. Initially
I was calling it as cpu_coresibling_map but later thought cpu_core_map is
a smaller name.
>
> Hmm, which names should be chosen?
For /proc/cpuinfo, I want to stick with "cpu cores" and "core id"
But I am open for kernel variable names.
Depending on what we decide on kernel variable names, I will see if I have
to make any changes to the IPF patch we posted last week.
thanks,
suresh
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [Patch] x86, x86_64: Intel dual-core detection
On Tue, Mar 15, 2005 at 05:36:25PM -0800, Siddha, Suresh B wrote:
> Appended patch adds the support for Intel dual-core detection and displaying
> the core related information in /proc/cpuinfo.
>
> It adds two new fields "core id" and "cpu cores" to x86 /proc/cpuinfo
> and the "core id" field for x86_64("cpu cores" field is already present in
> x86_64).
Thanks. I have a similar patch for AMD CPUs, unfortuntely it uses
different names ("shared cores" etc.)
>
> Number of processor cores in a die is detected using cpuid(4) and this
> is documented in IA-32 Intel Architecture Software Developer's Manual (vol 2a)
> (http://developer.intel.com/design/pentium4/manuals/index_new.htm#sdm_vol2a)
>
> This patch also adds cpu_core_map similar to cpu_sibling_map.
Called cpu_sharecore_map in my patch.
Hmm, which names should be chosen?
-Andi
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [Patch] x86, x86_64: Intel dual-core detection
On Tue, Mar 15, 2005 at 05:36:25PM -0800, Siddha, Suresh B wrote: Appended patch adds the support for Intel dual-core detection and displaying the core related information in /proc/cpuinfo. It adds two new fields core id and cpu cores to x86 /proc/cpuinfo and the core id field for x86_64(cpu cores field is already present in x86_64). Thanks. I have a similar patch for AMD CPUs, unfortuntely it uses different names (shared cores etc.) Number of processor cores in a die is detected using cpuid(4) and this is documented in IA-32 Intel Architecture Software Developer's Manual (vol 2a) (http://developer.intel.com/design/pentium4/manuals/index_new.htm#sdm_vol2a) This patch also adds cpu_core_map similar to cpu_sibling_map. Called cpu_sharecore_map in my patch. Hmm, which names should be chosen? -Andi - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: [Patch] x86, x86_64: Intel dual-core detection
On Wed, Mar 16, 2005 at 06:04:12PM +0100, Andi Kleen wrote: On Tue, Mar 15, 2005 at 05:36:25PM -0800, Siddha, Suresh B wrote: It adds two new fields core id and cpu cores to x86 /proc/cpuinfo and the core id field for x86_64(cpu cores field is already present in x86_64). Thanks. I have a similar patch for AMD CPUs, unfortuntely it uses different names (shared cores etc.) 2.6.11 already has cpu cores for x86_64. I would like to retain that. This patch also adds cpu_core_map similar to cpu_sibling_map. Called cpu_sharecore_map in my patch. cpu_sharecore_map sounds like HT logical siblings sharing the core. Initially I was calling it as cpu_coresibling_map but later thought cpu_core_map is a smaller name. Hmm, which names should be chosen? For /proc/cpuinfo, I want to stick with cpu cores and core id But I am open for kernel variable names. Depending on what we decide on kernel variable names, I will see if I have to make any changes to the IPF patch we posted last week. thanks, suresh - To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
[Patch] x86, x86_64: Intel dual-core detection
Appended patch adds the support for Intel dual-core detection and displaying
the core related information in /proc/cpuinfo.
It adds two new fields "core id" and "cpu cores" to x86 /proc/cpuinfo
and the "core id" field for x86_64("cpu cores" field is already present in
x86_64).
Number of processor cores in a die is detected using cpuid(4) and this
is documented in IA-32 Intel Architecture Software Developer's Manual (vol 2a)
(http://developer.intel.com/design/pentium4/manuals/index_new.htm#sdm_vol2a)
This patch also adds cpu_core_map similar to cpu_sibling_map.
Signed-off-by: Suresh Siddha <[EMAIL PROTECTED]>
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/amd.c
linux-mc/arch/i386/kernel/cpu/amd.c
--- linux-2.6.11/arch/i386/kernel/cpu/amd.c 2005-03-01 23:38:26.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/amd.c 2004-11-01 16:13:46.141256624 -0800
@@ -188,6 +188,13 @@
}
display_cacheinfo(c);
+
+ if (cpuid_eax(0x8000) >= 0x8008) {
+ c->x86_num_cores = (cpuid_ecx(0x8008) & 0xff) + 1;
+ if (c->x86_num_cores & (c->x86_num_cores - 1))
+ c->x86_num_cores = 1;
+ }
+
detect_ht(c);
#ifdef CONFIG_X86_HT
@@ -199,12 +206,6 @@
if (cpu_has(c, X86_FEATURE_CMP_LEGACY))
smp_num_siblings = 1;
#endif
-
- if (cpuid_eax(0x8000) >= 0x8008) {
- c->x86_num_cores = (cpuid_ecx(0x8008) & 0xff) + 1;
- if (c->x86_num_cores & (c->x86_num_cores - 1))
- c->x86_num_cores = 1;
- }
}
static unsigned int amd_size_cache(struct cpuinfo_x86 * c, unsigned int size)
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/common.c
linux-mc/arch/i386/kernel/cpu/common.c
--- linux-2.6.11/arch/i386/kernel/cpu/common.c 2005-03-01 23:37:47.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/common.c 2004-11-07 11:34:10.237802664
-0800
@@ -441,7 +441,7 @@
void __init detect_ht(struct cpuinfo_x86 *c)
{
u32 eax, ebx, ecx, edx;
- int index_lsb, index_msb, tmp;
+ int index_msb, tmp;
int cpu = smp_processor_id();
if (!cpu_has(c, X86_FEATURE_HT))
@@ -453,7 +453,6 @@
if (smp_num_siblings == 1) {
printk(KERN_INFO "CPU: Hyper-Threading is disabled\n");
} else if (smp_num_siblings > 1 ) {
- index_lsb = 0;
index_msb = 31;
if (smp_num_siblings > NR_CPUS) {
@@ -462,21 +461,34 @@
return;
}
tmp = smp_num_siblings;
- while ((tmp & 1) == 0) {
- tmp >>=1 ;
- index_lsb++;
- }
- tmp = smp_num_siblings;
while ((tmp & 0x8000 ) == 0) {
tmp <<=1 ;
index_msb--;
}
- if (index_lsb != index_msb )
+ if (smp_num_siblings & (smp_num_siblings - 1))
index_msb++;
phys_proc_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb);
printk(KERN_INFO "CPU: Physical Processor ID: %d\n",
phys_proc_id[cpu]);
+
+ smp_num_siblings = smp_num_siblings / c->x86_num_cores;
+
+ tmp = smp_num_siblings;
+ index_msb = 31;
+ while ((tmp & 0x8000) == 0) {
+ tmp <<=1 ;
+ index_msb--;
+ }
+
+ if (smp_num_siblings & (smp_num_siblings - 1))
+ index_msb++;
+
+ cpu_core_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb);
+
+ if (c->x86_num_cores > 1)
+ printk(KERN_INFO "CPU: Processor Core ID: %d\n",
+ cpu_core_id[cpu]);
}
}
#endif
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/intel.c
linux-mc/arch/i386/kernel/cpu/intel.c
--- linux-2.6.11/arch/i386/kernel/cpu/intel.c 2005-03-01 23:37:52.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/intel.c 2004-11-01 16:13:46.187249632
-0800
@@ -77,6 +77,27 @@
}
+/*
+ * find out the number of processor cores on the die
+ */
+static int __init num_cpu_cores(struct cpuinfo_x86 *c)
+{
+ unsigned int eax;
+
+ if (c->cpuid_level < 4)
+ return 1;
+
+ __asm__("cpuid"
+ : "=a" (eax)
+ : "0" (4), "c" (0)
+ : "bx", "dx");
+
+ if (eax & 0x1f)
+ return ((eax >> 26) + 1);
+ else
+ return 1;
+}
+
static void __init init_intel(struct cpuinfo_x86 *c)
{
unsigned int l2 = 0;
@@ -139,6 +160,8 @@
if ( p )
strcpy(c->x86_model_id, p);
+ c->x86_num_cores = num_cpu_cores(c);
+
detect_ht(c);
/* Work around errata */
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/proc.c
[Patch] x86, x86_64: Intel dual-core detection
Appended patch adds the support for Intel dual-core detection and displaying
the core related information in /proc/cpuinfo.
It adds two new fields core id and cpu cores to x86 /proc/cpuinfo
and the core id field for x86_64(cpu cores field is already present in
x86_64).
Number of processor cores in a die is detected using cpuid(4) and this
is documented in IA-32 Intel Architecture Software Developer's Manual (vol 2a)
(http://developer.intel.com/design/pentium4/manuals/index_new.htm#sdm_vol2a)
This patch also adds cpu_core_map similar to cpu_sibling_map.
Signed-off-by: Suresh Siddha [EMAIL PROTECTED]
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/amd.c
linux-mc/arch/i386/kernel/cpu/amd.c
--- linux-2.6.11/arch/i386/kernel/cpu/amd.c 2005-03-01 23:38:26.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/amd.c 2004-11-01 16:13:46.141256624 -0800
@@ -188,6 +188,13 @@
}
display_cacheinfo(c);
+
+ if (cpuid_eax(0x8000) = 0x8008) {
+ c-x86_num_cores = (cpuid_ecx(0x8008) 0xff) + 1;
+ if (c-x86_num_cores (c-x86_num_cores - 1))
+ c-x86_num_cores = 1;
+ }
+
detect_ht(c);
#ifdef CONFIG_X86_HT
@@ -199,12 +206,6 @@
if (cpu_has(c, X86_FEATURE_CMP_LEGACY))
smp_num_siblings = 1;
#endif
-
- if (cpuid_eax(0x8000) = 0x8008) {
- c-x86_num_cores = (cpuid_ecx(0x8008) 0xff) + 1;
- if (c-x86_num_cores (c-x86_num_cores - 1))
- c-x86_num_cores = 1;
- }
}
static unsigned int amd_size_cache(struct cpuinfo_x86 * c, unsigned int size)
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/common.c
linux-mc/arch/i386/kernel/cpu/common.c
--- linux-2.6.11/arch/i386/kernel/cpu/common.c 2005-03-01 23:37:47.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/common.c 2004-11-07 11:34:10.237802664
-0800
@@ -441,7 +441,7 @@
void __init detect_ht(struct cpuinfo_x86 *c)
{
u32 eax, ebx, ecx, edx;
- int index_lsb, index_msb, tmp;
+ int index_msb, tmp;
int cpu = smp_processor_id();
if (!cpu_has(c, X86_FEATURE_HT))
@@ -453,7 +453,6 @@
if (smp_num_siblings == 1) {
printk(KERN_INFO CPU: Hyper-Threading is disabled\n);
} else if (smp_num_siblings 1 ) {
- index_lsb = 0;
index_msb = 31;
if (smp_num_siblings NR_CPUS) {
@@ -462,21 +461,34 @@
return;
}
tmp = smp_num_siblings;
- while ((tmp 1) == 0) {
- tmp =1 ;
- index_lsb++;
- }
- tmp = smp_num_siblings;
while ((tmp 0x8000 ) == 0) {
tmp =1 ;
index_msb--;
}
- if (index_lsb != index_msb )
+ if (smp_num_siblings (smp_num_siblings - 1))
index_msb++;
phys_proc_id[cpu] = phys_pkg_id((ebx 24) 0xFF, index_msb);
printk(KERN_INFO CPU: Physical Processor ID: %d\n,
phys_proc_id[cpu]);
+
+ smp_num_siblings = smp_num_siblings / c-x86_num_cores;
+
+ tmp = smp_num_siblings;
+ index_msb = 31;
+ while ((tmp 0x8000) == 0) {
+ tmp =1 ;
+ index_msb--;
+ }
+
+ if (smp_num_siblings (smp_num_siblings - 1))
+ index_msb++;
+
+ cpu_core_id[cpu] = phys_pkg_id((ebx 24) 0xFF, index_msb);
+
+ if (c-x86_num_cores 1)
+ printk(KERN_INFO CPU: Processor Core ID: %d\n,
+ cpu_core_id[cpu]);
}
}
#endif
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/intel.c
linux-mc/arch/i386/kernel/cpu/intel.c
--- linux-2.6.11/arch/i386/kernel/cpu/intel.c 2005-03-01 23:37:52.0
-0800
+++ linux-mc/arch/i386/kernel/cpu/intel.c 2004-11-01 16:13:46.187249632
-0800
@@ -77,6 +77,27 @@
}
+/*
+ * find out the number of processor cores on the die
+ */
+static int __init num_cpu_cores(struct cpuinfo_x86 *c)
+{
+ unsigned int eax;
+
+ if (c-cpuid_level 4)
+ return 1;
+
+ __asm__(cpuid
+ : =a (eax)
+ : 0 (4), c (0)
+ : bx, dx);
+
+ if (eax 0x1f)
+ return ((eax 26) + 1);
+ else
+ return 1;
+}
+
static void __init init_intel(struct cpuinfo_x86 *c)
{
unsigned int l2 = 0;
@@ -139,6 +160,8 @@
if ( p )
strcpy(c-x86_model_id, p);
+ c-x86_num_cores = num_cpu_cores(c);
+
detect_ht(c);
/* Work around errata */
diff -Nru linux-2.6.11/arch/i386/kernel/cpu/proc.c
linux-mc/arch/i386/kernel/cpu/proc.c
--- linux-2.6.11/arch/i386/kernel/cpu/proc.c
