summaryrefslogtreecommitdiffstats
path: root/arch/i386
diff options
context:
space:
mode:
authorSiddha, Suresh B <suresh.b.siddha@intel.com>2005-11-05 17:25:54 +0100
committerLinus Torvalds <torvalds@g5.osdl.org>2005-11-14 19:55:16 -0800
commit94605eff572b727aaad9b4b29bc358b919096503 (patch)
tree657a848d8ef34d2f94bbad3aa4e5458d2d3f2d2b /arch/i386
parente90f22edf432512219cc2952f5811961abbd164f (diff)
downloadop-kernel-dev-94605eff572b727aaad9b4b29bc358b919096503.zip
op-kernel-dev-94605eff572b727aaad9b4b29bc358b919096503.tar.gz
[PATCH] x86-64/i386: Intel HT, Multi core detection fixes
Fields obtained through cpuid vector 0x1(ebx[16:23]) and vector 0x4(eax[14:25], eax[26:31]) indicate the maximum values and might not always be the same as what is available and what OS sees. So make sure "siblings" and "cpu cores" values in /proc/cpuinfo reflect the values as seen by OS instead of what cpuid instruction says. This will also fix the buggy BIOS cases (for example where cpuid on a single core cpu says there are "2" siblings, even when HT is disabled in the BIOS. http://bugzilla.kernel.org/show_bug.cgi?id=4359) Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com> Signed-off-by: Andi Kleen <ak@suse.de> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Diffstat (limited to 'arch/i386')
-rw-r--r--arch/i386/kernel/cpu/amd.c12
-rw-r--r--arch/i386/kernel/cpu/common.c36
-rw-r--r--arch/i386/kernel/cpu/intel.c2
-rw-r--r--arch/i386/kernel/cpu/intel_cacheinfo.c2
-rw-r--r--arch/i386/kernel/cpu/proc.c7
-rw-r--r--arch/i386/kernel/smpboot.c73
6 files changed, 78 insertions, 54 deletions
diff --git a/arch/i386/kernel/cpu/amd.c b/arch/i386/kernel/cpu/amd.c
index 53a1681..e344ef88 100644
--- a/arch/i386/kernel/cpu/amd.c
+++ b/arch/i386/kernel/cpu/amd.c
@@ -206,9 +206,9 @@ static void __init init_amd(struct cpuinfo_x86 *c)
display_cacheinfo(c);
if (cpuid_eax(0x80000000) >= 0x80000008) {
- c->x86_num_cores = (cpuid_ecx(0x80000008) & 0xff) + 1;
- if (c->x86_num_cores & (c->x86_num_cores - 1))
- c->x86_num_cores = 1;
+ c->x86_max_cores = (cpuid_ecx(0x80000008) & 0xff) + 1;
+ if (c->x86_max_cores & (c->x86_max_cores - 1))
+ c->x86_max_cores = 1;
}
#ifdef CONFIG_X86_HT
@@ -217,15 +217,15 @@ static void __init init_amd(struct cpuinfo_x86 *c)
* distingush the cores. Assumes number of cores is a power
* of two.
*/
- if (c->x86_num_cores > 1) {
+ if (c->x86_max_cores > 1) {
int cpu = smp_processor_id();
unsigned bits = 0;
- while ((1 << bits) < c->x86_num_cores)
+ while ((1 << bits) < c->x86_max_cores)
bits++;
cpu_core_id[cpu] = phys_proc_id[cpu] & ((1<<bits)-1);
phys_proc_id[cpu] >>= bits;
printk(KERN_INFO "CPU %d(%d) -> Core %d\n",
- cpu, c->x86_num_cores, cpu_core_id[cpu]);
+ cpu, c->x86_max_cores, cpu_core_id[cpu]);
}
#endif
}
diff --git a/arch/i386/kernel/cpu/common.c b/arch/i386/kernel/cpu/common.c
index 35a67da..4e9c2e9 100644
--- a/arch/i386/kernel/cpu/common.c
+++ b/arch/i386/kernel/cpu/common.c
@@ -335,7 +335,7 @@ void __devinit identify_cpu(struct cpuinfo_x86 *c)
c->x86_model = c->x86_mask = 0; /* So far unknown... */
c->x86_vendor_id[0] = '\0'; /* Unset */
c->x86_model_id[0] = '\0'; /* Unset */
- c->x86_num_cores = 1;
+ c->x86_max_cores = 1;
memset(&c->x86_capability, 0, sizeof c->x86_capability);
if (!have_cpuid_p()) {
@@ -446,52 +446,44 @@ void __devinit identify_cpu(struct cpuinfo_x86 *c)
void __devinit detect_ht(struct cpuinfo_x86 *c)
{
u32 eax, ebx, ecx, edx;
- int index_msb, tmp;
+ int index_msb, core_bits;
int cpu = smp_processor_id();
+ cpuid(1, &eax, &ebx, &ecx, &edx);
+
+ c->apicid = phys_pkg_id((ebx >> 24) & 0xFF, 0);
+
if (!cpu_has(c, X86_FEATURE_HT) || cpu_has(c, X86_FEATURE_CMP_LEGACY))
return;
- cpuid(1, &eax, &ebx, &ecx, &edx);
smp_num_siblings = (ebx & 0xff0000) >> 16;
if (smp_num_siblings == 1) {
printk(KERN_INFO "CPU: Hyper-Threading is disabled\n");
} else if (smp_num_siblings > 1 ) {
- index_msb = 31;
if (smp_num_siblings > NR_CPUS) {
printk(KERN_WARNING "CPU: Unsupported number of the siblings %d", smp_num_siblings);
smp_num_siblings = 1;
return;
}
- tmp = smp_num_siblings;
- while ((tmp & 0x80000000 ) == 0) {
- tmp <<=1 ;
- index_msb--;
- }
- if (smp_num_siblings & (smp_num_siblings - 1))
- index_msb++;
+
+ index_msb = get_count_order(smp_num_siblings);
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;
+ smp_num_siblings = smp_num_siblings / c->x86_max_cores;
- tmp = smp_num_siblings;
- index_msb = 31;
- while ((tmp & 0x80000000) == 0) {
- tmp <<=1 ;
- index_msb--;
- }
+ index_msb = get_count_order(smp_num_siblings) ;
- if (smp_num_siblings & (smp_num_siblings - 1))
- index_msb++;
+ core_bits = get_count_order(c->x86_max_cores);
- cpu_core_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb);
+ cpu_core_id[cpu] = phys_pkg_id((ebx >> 24) & 0xFF, index_msb) &
+ ((1 << core_bits) - 1);
- if (c->x86_num_cores > 1)
+ if (c->x86_max_cores > 1)
printk(KERN_INFO "CPU: Processor Core ID: %d\n",
cpu_core_id[cpu]);
}
diff --git a/arch/i386/kernel/cpu/intel.c b/arch/i386/kernel/cpu/intel.c
index 43601de..8d603ba 100644
--- a/arch/i386/kernel/cpu/intel.c
+++ b/arch/i386/kernel/cpu/intel.c
@@ -157,7 +157,7 @@ static void __devinit init_intel(struct cpuinfo_x86 *c)
if ( p )
strcpy(c->x86_model_id, p);
- c->x86_num_cores = num_cpu_cores(c);
+ c->x86_max_cores = num_cpu_cores(c);
detect_ht(c);
diff --git a/arch/i386/kernel/cpu/intel_cacheinfo.c b/arch/i386/kernel/cpu/intel_cacheinfo.c
index 4dc42a1..e66d140 100644
--- a/arch/i386/kernel/cpu/intel_cacheinfo.c
+++ b/arch/i386/kernel/cpu/intel_cacheinfo.c
@@ -307,7 +307,7 @@ static void __cpuinit cache_shared_cpu_map_setup(unsigned int cpu, int index)
#ifdef CONFIG_X86_HT
else if (num_threads_sharing == smp_num_siblings)
this_leaf->shared_cpu_map = cpu_sibling_map[cpu];
- else if (num_threads_sharing == (c->x86_num_cores * smp_num_siblings))
+ else if (num_threads_sharing == (c->x86_max_cores * smp_num_siblings))
this_leaf->shared_cpu_map = cpu_core_map[cpu];
else
printk(KERN_DEBUG "Number of CPUs sharing cache didn't match "
diff --git a/arch/i386/kernel/cpu/proc.c b/arch/i386/kernel/cpu/proc.c
index 41b871e..e792131 100644
--- a/arch/i386/kernel/cpu/proc.c
+++ b/arch/i386/kernel/cpu/proc.c
@@ -94,12 +94,11 @@ static int show_cpuinfo(struct seq_file *m, void *v)
if (c->x86_cache_size >= 0)
seq_printf(m, "cache size\t: %d KB\n", c->x86_cache_size);
#ifdef CONFIG_X86_HT
- if (c->x86_num_cores * smp_num_siblings > 1) {
+ if (c->x86_max_cores * smp_num_siblings > 1) {
seq_printf(m, "physical id\t: %d\n", phys_proc_id[n]);
- seq_printf(m, "siblings\t: %d\n",
- c->x86_num_cores * smp_num_siblings);
+ seq_printf(m, "siblings\t: %d\n", cpus_weight(cpu_core_map[n]));
seq_printf(m, "core id\t\t: %d\n", cpu_core_id[n]);
- seq_printf(m, "cpu cores\t: %d\n", c->x86_num_cores);
+ seq_printf(m, "cpu cores\t: %d\n", c->booted_cores);
}
#endif
diff --git a/arch/i386/kernel/smpboot.c b/arch/i386/kernel/smpboot.c
index 01b618e..0a9c646 100644
--- a/arch/i386/kernel/smpboot.c
+++ b/arch/i386/kernel/smpboot.c
@@ -74,9 +74,11 @@ EXPORT_SYMBOL(phys_proc_id);
int cpu_core_id[NR_CPUS] __read_mostly = {[0 ... NR_CPUS-1] = BAD_APICID};
EXPORT_SYMBOL(cpu_core_id);
+/* representing HT siblings of each logical CPU */
cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly;
EXPORT_SYMBOL(cpu_sibling_map);
+/* representing HT and core siblings of each logical CPU */
cpumask_t cpu_core_map[NR_CPUS] __read_mostly;
EXPORT_SYMBOL(cpu_core_map);
@@ -444,35 +446,60 @@ static void __devinit smp_callin(void)
static int cpucount;
+/* representing cpus for which sibling maps can be computed */
+static cpumask_t cpu_sibling_setup_map;
+
static inline void
set_cpu_sibling_map(int cpu)
{
int i;
+ struct cpuinfo_x86 *c = cpu_data;
+
+ cpu_set(cpu, cpu_sibling_setup_map);
if (smp_num_siblings > 1) {
- for (i = 0; i < NR_CPUS; i++) {
- if (!cpu_isset(i, cpu_callout_map))
- continue;
- if (cpu_core_id[cpu] == cpu_core_id[i]) {
+ for_each_cpu_mask(i, cpu_sibling_setup_map) {
+ if (phys_proc_id[cpu] == phys_proc_id[i] &&
+ cpu_core_id[cpu] == cpu_core_id[i]) {
cpu_set(i, cpu_sibling_map[cpu]);
cpu_set(cpu, cpu_sibling_map[i]);
+ cpu_set(i, cpu_core_map[cpu]);
+ cpu_set(cpu, cpu_core_map[i]);
}
}
} else {
cpu_set(cpu, cpu_sibling_map[cpu]);
}
- if (current_cpu_data.x86_num_cores > 1) {
- for (i = 0; i < NR_CPUS; i++) {
- if (!cpu_isset(i, cpu_callout_map))
- continue;
- if (phys_proc_id[cpu] == phys_proc_id[i]) {
- cpu_set(i, cpu_core_map[cpu]);
- cpu_set(cpu, cpu_core_map[i]);
- }
- }
- } else {
+ if (current_cpu_data.x86_max_cores == 1) {
cpu_core_map[cpu] = cpu_sibling_map[cpu];
+ c[cpu].booted_cores = 1;
+ return;
+ }
+
+ for_each_cpu_mask(i, cpu_sibling_setup_map) {
+ if (phys_proc_id[cpu] == phys_proc_id[i]) {
+ cpu_set(i, cpu_core_map[cpu]);
+ cpu_set(cpu, cpu_core_map[i]);
+ /*
+ * Does this new cpu bringup a new core?
+ */
+ if (cpus_weight(cpu_sibling_map[cpu]) == 1) {
+ /*
+ * for each core in package, increment
+ * the booted_cores for this new cpu
+ */
+ if (first_cpu(cpu_sibling_map[i]) == i)
+ c[cpu].booted_cores++;
+ /*
+ * increment the core count for all
+ * the other cpus in this package
+ */
+ if (i != cpu)
+ c[i].booted_cores++;
+ } else if (i != cpu && !c[cpu].booted_cores)
+ c[cpu].booted_cores = c[i].booted_cores;
+ }
}
}
@@ -1096,11 +1123,8 @@ static void __init smp_boot_cpus(unsigned int max_cpus)
current_thread_info()->cpu = 0;
smp_tune_scheduling();
- cpus_clear(cpu_sibling_map[0]);
- cpu_set(0, cpu_sibling_map[0]);
- cpus_clear(cpu_core_map[0]);
- cpu_set(0, cpu_core_map[0]);
+ set_cpu_sibling_map(0);
/*
* If we couldn't find an SMP configuration at boot time,
@@ -1279,15 +1303,24 @@ static void
remove_siblinginfo(int cpu)
{
int sibling;
+ struct cpuinfo_x86 *c = cpu_data;
+ for_each_cpu_mask(sibling, cpu_core_map[cpu]) {
+ cpu_clear(cpu, cpu_core_map[sibling]);
+ /*
+ * last thread sibling in this cpu core going down
+ */
+ if (cpus_weight(cpu_sibling_map[cpu]) == 1)
+ c[sibling].booted_cores--;
+ }
+
for_each_cpu_mask(sibling, cpu_sibling_map[cpu])
cpu_clear(cpu, cpu_sibling_map[sibling]);
- for_each_cpu_mask(sibling, cpu_core_map[cpu])
- cpu_clear(cpu, cpu_core_map[sibling]);
cpus_clear(cpu_sibling_map[cpu]);
cpus_clear(cpu_core_map[cpu]);
phys_proc_id[cpu] = BAD_APICID;
cpu_core_id[cpu] = BAD_APICID;
+ cpu_clear(cpu, cpu_sibling_setup_map);
}
int __cpu_disable(void)
OpenPOWER on IntegriCloud