aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
Diffstat
-rw-r--r--sysdeps/i386/dl-machine.h7
-rw-r--r--sysdeps/x86/cacheinfo.c868
-rw-r--r--sysdeps/x86/cacheinfo.h413
-rw-r--r--sysdeps/x86/cpu-features.c12
-rw-r--r--sysdeps/x86/dl-cacheinfo.h478
-rw-r--r--sysdeps/x86/dl-get-cpu-features.c27
-rw-r--r--sysdeps/x86/include/cpu-features.h1
-rw-r--r--sysdeps/x86/libc-start.c1
-rw-r--r--sysdeps/x86_64/dl-machine.h7
9 files changed, 949 insertions, 865 deletions
diff --git a/sysdeps/i386/dl-machine.h b/sysdeps/i386/dl-machine.h
index 0f08079e48..bdc21d1a3c 100644
--- a/sysdeps/i386/dl-machine.h
+++ b/sysdeps/i386/dl-machine.h
@@ -25,7 +25,6 @@
#include <sysdep.h>
#include <tls.h>
#include <dl-tlsdesc.h>
-#include <cpu-features.c>
/* Return nonzero iff ELF header is compatible with the running host. */
static inline int __attribute__ ((unused))
@@ -248,9 +247,9 @@ static inline void __attribute__ ((unused))
dl_platform_init (void)
{
#if IS_IN (rtld)
- /* init_cpu_features has been called early from __libc_start_main in
- static executable. */
- init_cpu_features (&GLRO(dl_x86_cpu_features));
+ /* _dl_x86_init_cpu_features is a wrapper for init_cpu_features which
+ has been called early from __libc_start_main in static executable. */
+ _dl_x86_init_cpu_features ();
#else
if (GLRO(dl_platform) != NULL && *GLRO(dl_platform) == '\0')
/* Avoid an empty string which would disturb us. */
diff --git a/sysdeps/x86/cacheinfo.c b/sysdeps/x86/cacheinfo.c
index dadec5d58f..0d2fe3a2fa 100644
--- a/sysdeps/x86/cacheinfo.c
+++ b/sysdeps/x86/cacheinfo.c
@@ -1,4 +1,4 @@
-/* x86_64 cache info.
+/* x86 cache info.
Copyright (C) 2003-2020 Free Software Foundation, Inc.
This file is part of the GNU C Library.
@@ -19,473 +19,10 @@
#if IS_IN (libc)
#include <assert.h>
-#include <stdbool.h>
-#include <stdlib.h>
#include <unistd.h>
#include <cpuid.h>
-#include <init-arch.h>
-
-static const struct intel_02_cache_info
-{
- unsigned char idx;
- unsigned char assoc;
- unsigned char linesize;
- unsigned char rel_name;
- unsigned int size;
-} intel_02_known [] =
- {
-#define M(sc) ((sc) - _SC_LEVEL1_ICACHE_SIZE)
- { 0x06, 4, 32, M(_SC_LEVEL1_ICACHE_SIZE), 8192 },
- { 0x08, 4, 32, M(_SC_LEVEL1_ICACHE_SIZE), 16384 },
- { 0x09, 4, 32, M(_SC_LEVEL1_ICACHE_SIZE), 32768 },
- { 0x0a, 2, 32, M(_SC_LEVEL1_DCACHE_SIZE), 8192 },
- { 0x0c, 4, 32, M(_SC_LEVEL1_DCACHE_SIZE), 16384 },
- { 0x0d, 4, 64, M(_SC_LEVEL1_DCACHE_SIZE), 16384 },
- { 0x0e, 6, 64, M(_SC_LEVEL1_DCACHE_SIZE), 24576 },
- { 0x21, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x22, 4, 64, M(_SC_LEVEL3_CACHE_SIZE), 524288 },
- { 0x23, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 1048576 },
- { 0x25, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 2097152 },
- { 0x29, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 4194304 },
- { 0x2c, 8, 64, M(_SC_LEVEL1_DCACHE_SIZE), 32768 },
- { 0x30, 8, 64, M(_SC_LEVEL1_ICACHE_SIZE), 32768 },
- { 0x39, 4, 64, M(_SC_LEVEL2_CACHE_SIZE), 131072 },
- { 0x3a, 6, 64, M(_SC_LEVEL2_CACHE_SIZE), 196608 },
- { 0x3b, 2, 64, M(_SC_LEVEL2_CACHE_SIZE), 131072 },
- { 0x3c, 4, 64, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x3d, 6, 64, M(_SC_LEVEL2_CACHE_SIZE), 393216 },
- { 0x3e, 4, 64, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x3f, 2, 64, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x41, 4, 32, M(_SC_LEVEL2_CACHE_SIZE), 131072 },
- { 0x42, 4, 32, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x43, 4, 32, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x44, 4, 32, M(_SC_LEVEL2_CACHE_SIZE), 1048576 },
- { 0x45, 4, 32, M(_SC_LEVEL2_CACHE_SIZE), 2097152 },
- { 0x46, 4, 64, M(_SC_LEVEL3_CACHE_SIZE), 4194304 },
- { 0x47, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 8388608 },
- { 0x48, 12, 64, M(_SC_LEVEL2_CACHE_SIZE), 3145728 },
- { 0x49, 16, 64, M(_SC_LEVEL2_CACHE_SIZE), 4194304 },
- { 0x4a, 12, 64, M(_SC_LEVEL3_CACHE_SIZE), 6291456 },
- { 0x4b, 16, 64, M(_SC_LEVEL3_CACHE_SIZE), 8388608 },
- { 0x4c, 12, 64, M(_SC_LEVEL3_CACHE_SIZE), 12582912 },
- { 0x4d, 16, 64, M(_SC_LEVEL3_CACHE_SIZE), 16777216 },
- { 0x4e, 24, 64, M(_SC_LEVEL2_CACHE_SIZE), 6291456 },
- { 0x60, 8, 64, M(_SC_LEVEL1_DCACHE_SIZE), 16384 },
- { 0x66, 4, 64, M(_SC_LEVEL1_DCACHE_SIZE), 8192 },
- { 0x67, 4, 64, M(_SC_LEVEL1_DCACHE_SIZE), 16384 },
- { 0x68, 4, 64, M(_SC_LEVEL1_DCACHE_SIZE), 32768 },
- { 0x78, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 1048576 },
- { 0x79, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 131072 },
- { 0x7a, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x7b, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x7c, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 1048576 },
- { 0x7d, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 2097152 },
- { 0x7f, 2, 64, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x80, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x82, 8, 32, M(_SC_LEVEL2_CACHE_SIZE), 262144 },
- { 0x83, 8, 32, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x84, 8, 32, M(_SC_LEVEL2_CACHE_SIZE), 1048576 },
- { 0x85, 8, 32, M(_SC_LEVEL2_CACHE_SIZE), 2097152 },
- { 0x86, 4, 64, M(_SC_LEVEL2_CACHE_SIZE), 524288 },
- { 0x87, 8, 64, M(_SC_LEVEL2_CACHE_SIZE), 1048576 },
- { 0xd0, 4, 64, M(_SC_LEVEL3_CACHE_SIZE), 524288 },
- { 0xd1, 4, 64, M(_SC_LEVEL3_CACHE_SIZE), 1048576 },
- { 0xd2, 4, 64, M(_SC_LEVEL3_CACHE_SIZE), 2097152 },
- { 0xd6, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 1048576 },
- { 0xd7, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 2097152 },
- { 0xd8, 8, 64, M(_SC_LEVEL3_CACHE_SIZE), 4194304 },
- { 0xdc, 12, 64, M(_SC_LEVEL3_CACHE_SIZE), 2097152 },
- { 0xdd, 12, 64, M(_SC_LEVEL3_CACHE_SIZE), 4194304 },
- { 0xde, 12, 64, M(_SC_LEVEL3_CACHE_SIZE), 8388608 },
- { 0xe2, 16, 64, M(_SC_LEVEL3_CACHE_SIZE), 2097152 },
- { 0xe3, 16, 64, M(_SC_LEVEL3_CACHE_SIZE), 4194304 },
- { 0xe4, 16, 64, M(_SC_LEVEL3_CACHE_SIZE), 8388608 },
- { 0xea, 24, 64, M(_SC_LEVEL3_CACHE_SIZE), 12582912 },
- { 0xeb, 24, 64, M(_SC_LEVEL3_CACHE_SIZE), 18874368 },
- { 0xec, 24, 64, M(_SC_LEVEL3_CACHE_SIZE), 25165824 },
- };
-
-#define nintel_02_known (sizeof (intel_02_known) / sizeof (intel_02_known [0]))
-
-static int
-intel_02_known_compare (const void *p1, const void *p2)
-{
- const struct intel_02_cache_info *i1;
- const struct intel_02_cache_info *i2;
-
- i1 = (const struct intel_02_cache_info *) p1;
- i2 = (const struct intel_02_cache_info *) p2;
-
- if (i1->idx == i2->idx)
- return 0;
-
- return i1->idx < i2->idx ? -1 : 1;
-}
-
-
-static long int
-__attribute__ ((noinline))
-intel_check_word (int name, unsigned int value, bool *has_level_2,
- bool *no_level_2_or_3,
- const struct cpu_features *cpu_features)
-{
- if ((value & 0x80000000) != 0)
- /* The register value is reserved. */
- return 0;
-
- /* Fold the name. The _SC_ constants are always in the order SIZE,
- ASSOC, LINESIZE. */
- int folded_rel_name = (M(name) / 3) * 3;
-
- while (value != 0)
- {
- unsigned int byte = value & 0xff;
-
- if (byte == 0x40)
- {
- *no_level_2_or_3 = true;
-
- if (folded_rel_name == M(_SC_LEVEL3_CACHE_SIZE))
- /* No need to look further. */
- break;
- }
- else if (byte == 0xff)
- {
- /* CPUID leaf 0x4 contains all the information. We need to
- iterate over it. */
- unsigned int eax;
- unsigned int ebx;
- unsigned int ecx;
- unsigned int edx;
-
- unsigned int round = 0;
- while (1)
- {
- __cpuid_count (4, round, eax, ebx, ecx, edx);
-
- enum { null = 0, data = 1, inst = 2, uni = 3 } type = eax & 0x1f;
- if (type == null)
- /* That was the end. */
- break;
-
- unsigned int level = (eax >> 5) & 0x7;
-
- if ((level == 1 && type == data
- && folded_rel_name == M(_SC_LEVEL1_DCACHE_SIZE))
- || (level == 1 && type == inst
- && folded_rel_name == M(_SC_LEVEL1_ICACHE_SIZE))
- || (level == 2 && folded_rel_name == M(_SC_LEVEL2_CACHE_SIZE))
- || (level == 3 && folded_rel_name == M(_SC_LEVEL3_CACHE_SIZE))
- || (level == 4 && folded_rel_name == M(_SC_LEVEL4_CACHE_SIZE)))
- {
- unsigned int offset = M(name) - folded_rel_name;
-
- if (offset == 0)
- /* Cache size. */
- return (((ebx >> 22) + 1)
- * (((ebx >> 12) & 0x3ff) + 1)
- * ((ebx & 0xfff) + 1)
- * (ecx + 1));
- if (offset == 1)
- return (ebx >> 22) + 1;
-
- assert (offset == 2);
- return (ebx & 0xfff) + 1;
- }
-
- ++round;
- }
- /* There is no other cache information anywhere else. */
- break;
- }
- else
- {
- if (byte == 0x49 && folded_rel_name == M(_SC_LEVEL3_CACHE_SIZE))
- {
- /* Intel reused this value. For family 15, model 6 it
- specifies the 3rd level cache. Otherwise the 2nd
- level cache. */
- unsigned int family = cpu_features->basic.family;
- unsigned int model = cpu_features->basic.model;
-
- if (family == 15 && model == 6)
- {
- /* The level 3 cache is encoded for this model like
- the level 2 cache is for other models. Pretend
- the caller asked for the level 2 cache. */
- name = (_SC_LEVEL2_CACHE_SIZE
- + (name - _SC_LEVEL3_CACHE_SIZE));
- folded_rel_name = M(_SC_LEVEL2_CACHE_SIZE);
- }
- }
-
- struct intel_02_cache_info *found;
- struct intel_02_cache_info search;
-
- search.idx = byte;
- found = bsearch (&search, intel_02_known, nintel_02_known,
- sizeof (intel_02_known[0]), intel_02_known_compare);
- if (found != NULL)
- {
- if (found->rel_name == folded_rel_name)
- {
- unsigned int offset = M(name) - folded_rel_name;
-
- if (offset == 0)
- /* Cache size. */
- return found->size;
- if (offset == 1)
- return found->assoc;
-
- assert (offset == 2);
- return found->linesize;
- }
-
- if (found->rel_name == M(_SC_LEVEL2_CACHE_SIZE))
- *has_level_2 = true;
- }
- }
-
- /* Next byte for the next round. */
- value >>= 8;
- }
-
- /* Nothing found. */
- return 0;
-}
-
-
-static long int __attribute__ ((noinline))
-handle_intel (int name, const struct cpu_features *cpu_features)
-{
- unsigned int maxidx = cpu_features->basic.max_cpuid;
-
- /* Return -1 for older CPUs. */
- if (maxidx < 2)
- return -1;
-
- /* OK, we can use the CPUID instruction to get all info about the
- caches. */
- unsigned int cnt = 0;
- unsigned int max = 1;
- long int result = 0;
- bool no_level_2_or_3 = false;
- bool has_level_2 = false;
-
- while (cnt++ < max)
- {
- unsigned int eax;
- unsigned int ebx;
- unsigned int ecx;
- unsigned int edx;
- __cpuid (2, eax, ebx, ecx, edx);
-
- /* The low byte of EAX in the first round contain the number of
- rounds we have to make. At least one, the one we are already
- doing. */
- if (cnt == 1)
- {
- max = eax & 0xff;
- eax &= 0xffffff00;
- }
-
- /* Process the individual registers' value. */
- result = intel_check_word (name, eax, &has_level_2,
- &no_level_2_or_3, cpu_features);
- if (result != 0)
- return result;
-
- result = intel_check_word (name, ebx, &has_level_2,
- &no_level_2_or_3, cpu_features);
- if (result != 0)
- return result;
-
- result = intel_check_word (name, ecx, &has_level_2,
- &no_level_2_or_3, cpu_features);
- if (result != 0)
- return result;
-
- result = intel_check_word (name, edx, &has_level_2,
- &no_level_2_or_3, cpu_features);
- if (result != 0)
- return result;
- }
-
- if (name >= _SC_LEVEL2_CACHE_SIZE && name <= _SC_LEVEL3_CACHE_LINESIZE
- && no_level_2_or_3)
- return -1;
-
- return 0;
-}
-
-
-static long int __attribute__ ((noinline))
-handle_amd (int name)
-{
- unsigned int eax;
- unsigned int ebx;
- unsigned int ecx;
- unsigned int edx;
- __cpuid (0x80000000, eax, ebx, ecx, edx);
-
- /* No level 4 cache (yet). */
- if (name > _SC_LEVEL3_CACHE_LINESIZE)
- return 0;
-
- unsigned int fn = 0x80000005 + (name >= _SC_LEVEL2_CACHE_SIZE);
- if (eax < fn)
- return 0;
-
- __cpuid (fn, eax, ebx, ecx, edx);
-
- if (name < _SC_LEVEL1_DCACHE_SIZE)
- {
- name += _SC_LEVEL1_DCACHE_SIZE - _SC_LEVEL1_ICACHE_SIZE;
- ecx = edx;
- }
-
- switch (name)
- {
- case _SC_LEVEL1_DCACHE_SIZE:
- return (ecx >> 14) & 0x3fc00;
-
- case _SC_LEVEL1_DCACHE_ASSOC:
- ecx >>= 16;
- if ((ecx & 0xff) == 0xff)
- /* Fully associative. */
- return (ecx << 2) & 0x3fc00;
- return ecx & 0xff;
-
- case _SC_LEVEL1_DCACHE_LINESIZE:
- return ecx & 0xff;
-
- case _SC_LEVEL2_CACHE_SIZE:
- return (ecx & 0xf000) == 0 ? 0 : (ecx >> 6) & 0x3fffc00;
-
- case _SC_LEVEL2_CACHE_ASSOC:
- switch ((ecx >> 12) & 0xf)
- {
- case 0:
- case 1:
- case 2:
- case 4:
- return (ecx >> 12) & 0xf;
- case 6:
- return 8;
- case 8:
- return 16;
- case 10:
- return 32;
- case 11:
- return 48;
- case 12:
- return 64;
- case 13:
- return 96;
- case 14:
- return 128;
- case 15:
- return ((ecx >> 6) & 0x3fffc00) / (ecx & 0xff);
- default:
- return 0;
- }
- /* NOTREACHED */
-
- case _SC_LEVEL2_CACHE_LINESIZE:
- return (ecx & 0xf000) == 0 ? 0 : ecx & 0xff;
-
- case _SC_LEVEL3_CACHE_SIZE:
- return (edx & 0xf000) == 0 ? 0 : (edx & 0x3ffc0000) << 1;
-
- case _SC_LEVEL3_CACHE_ASSOC:
- switch ((edx >> 12) & 0xf)
- {
- case 0:
- case 1:
- case 2:
- case 4:
- return (edx >> 12) & 0xf;
- case 6:
- return 8;
- case 8:
- return 16;
- case 10:
- return 32;
- case 11:
- return 48;
- case 12:
- return 64;
- case 13:
- return 96;
- case 14:
- return 128;
- case 15:
- return ((edx & 0x3ffc0000) << 1) / (edx & 0xff);
- default:
- return 0;
- }
- /* NOTREACHED */
-
- case _SC_LEVEL3_CACHE_LINESIZE:
- return (edx & 0xf000) == 0 ? 0 : edx & 0xff;
-
- default:
- assert (! "cannot happen");
- }
- return -1;
-}
-
-
-static long int __attribute__ ((noinline))
-handle_zhaoxin (int name)
-{
- unsigned int eax;
- unsigned int ebx;
- unsigned int ecx;
- unsigned int edx;
-
- int folded_rel_name = (M(name) / 3) * 3;
-
- unsigned int round = 0;
- while (1)
- {
- __cpuid_count (4, round, eax, ebx, ecx, edx);
-
- enum { null = 0, data = 1, inst = 2, uni = 3 } type = eax & 0x1f;
- if (type == null)
- break;
-
- unsigned int level = (eax >> 5) & 0x7;
-
- if ((level == 1 && type == data
- && folded_rel_name == M(_SC_LEVEL1_DCACHE_SIZE))
- || (level == 1 && type == inst
- && folded_rel_name == M(_SC_LEVEL1_ICACHE_SIZE))
- || (level == 2 && folded_rel_name == M(_SC_LEVEL2_CACHE_SIZE))
- || (level == 3 && folded_rel_name == M(_SC_LEVEL3_CACHE_SIZE)))
- {
- unsigned int offset = M(name) - folded_rel_name;
-
- if (offset == 0)
- /* Cache size. */
- return (((ebx >> 22) + 1)
- * (((ebx >> 12) & 0x3ff) + 1)
- * ((ebx & 0xfff) + 1)
- * (ecx + 1));
- if (offset == 1)
- return (ebx >> 22) + 1;
-
- assert (offset == 2);
- return (ebx & 0xfff) + 1;
- }
-
- ++round;
- }
-
- /* Nothing found. */
- return 0;
-}
-
+#include <ldsodefs.h>
+#include <dl-cacheinfo.h>
/* Get the value of the system variable NAME. */
long int
@@ -509,395 +46,18 @@ __cache_sysconf (int name)
return 0;
}
+# ifdef SHARED
+/* NB: In libc.a, cacheinfo.h is included in libc-start.c. In libc.so,
+ cacheinfo.h is included here and call init_cacheinfo by initializing
+ a dummy function pointer via IFUNC relocation after CPU features in
+ ld.so have been initialized by DL_PLATFORM_INIT or IFUNC relocation. */
+# include <cacheinfo.h>
+# include <ifunc-init.h>
-/* Data cache size for use in memory and string routines, typically
- L1 size, rounded to multiple of 256 bytes. */
-long int __x86_data_cache_size_half attribute_hidden = 32 * 1024 / 2;
-long int __x86_data_cache_size attribute_hidden = 32 * 1024;
-/* Similar to __x86_data_cache_size_half, but not rounded. */
-long int __x86_raw_data_cache_size_half attribute_hidden = 32 * 1024 / 2;
-/* Similar to __x86_data_cache_size, but not rounded. */
-long int __x86_raw_data_cache_size attribute_hidden = 32 * 1024;
-/* Shared cache size for use in memory and string routines, typically
- L2 or L3 size, rounded to multiple of 256 bytes. */
-long int __x86_shared_cache_size_half attribute_hidden = 1024 * 1024 / 2;
-long int __x86_shared_cache_size attribute_hidden = 1024 * 1024;
-/* Similar to __x86_shared_cache_size_half, but not rounded. */
-long int __x86_raw_shared_cache_size_half attribute_hidden = 1024 * 1024 / 2;
-/* Similar to __x86_shared_cache_size, but not rounded. */
-long int __x86_raw_shared_cache_size attribute_hidden = 1024 * 1024;
-
-/* Threshold to use non temporal store. */
-long int __x86_shared_non_temporal_threshold attribute_hidden;
-
-/* Threshold to use Enhanced REP MOVSB. */
-long int __x86_rep_movsb_threshold attribute_hidden = 2048;
-
-/* Threshold to use Enhanced REP STOSB. */
-long int __x86_rep_stosb_threshold attribute_hidden = 2048;
-
-
-static void
-get_common_cache_info (long int *shared_ptr, unsigned int *threads_ptr,
- long int core)
-{
- unsigned int eax;
- unsigned int ebx;
- unsigned int ecx;
- unsigned int edx;
-
- /* Number of logical processors sharing L2 cache. */
- int threads_l2;
-
- /* Number of logical processors sharing L3 cache. */
- int threads_l3;
-
- const struct cpu_features *cpu_features = __get_cpu_features ();
- int max_cpuid = cpu_features->basic.max_cpuid;
- unsigned int family = cpu_features->basic.family;
- unsigned int model = cpu_features->basic.model;
- long int shared = *shared_ptr;
- unsigned int threads = *threads_ptr;
- bool inclusive_cache = true;
- bool support_count_mask = true;
-
- /* Try L3 first. */
- unsigned int level = 3;
-
- if (cpu_features->basic.kind == arch_kind_zhaoxin && family == 6)
- support_count_mask = false;
-
- if (shared <= 0)
- {
- /* Try L2 otherwise. */
- level = 2;
- shared = core;
- threads_l2 = 0;
- threads_l3 = -1;
- }
- else
- {
- threads_l2 = 0;
- threads_l3 = 0;
- }
-
- /* A value of 0 for the HTT bit indicates there is only a single
- logical processor. */
- if (CPU_FEATURE_USABLE (HTT))
- {
- /* Figure out the number of logical threads that share the
- highest cache level. */
- if (max_cpuid >= 4)
- {
- int i = 0;
-
- /* Query until cache level 2 and 3 are enumerated. */
- int check = 0x1 | (threads_l3 == 0) << 1;
- do
- {
- __cpuid_count (4, i++, eax, ebx, ecx, edx);
-
- /* There seems to be a bug in at least some Pentium Ds
- which sometimes fail to iterate all cache parameters.
- Do not loop indefinitely here, stop in this case and
- assume there is no such information. */
- if (cpu_features->basic.kind == arch_kind_intel
- && (eax & 0x1f) == 0 )
- goto intel_bug_no_cache_info;
-
- switch ((eax >> 5) & 0x7)
- {
- default:
- break;
- case 2:
- if ((check & 0x1))
- {
- /* Get maximum number of logical processors
- sharing L2 cache. */
- threads_l2 = (eax >> 14) & 0x3ff;
- check &= ~0x1;
- }
- break;
- case 3:
- if ((check & (0x1 << 1)))
- {
- /* Get maximum number of logical processors
- sharing L3 cache. */
- threads_l3 = (eax >> 14) & 0x3ff;
-
- /* Check if L2 and L3 caches are inclusive. */
- inclusive_cache = (edx & 0x2) != 0;
- check &= ~(0x1 << 1);
- }
- break;
- }
- }
- while (check);
+extern void __x86_cacheinfo (void) attribute_hidden;
+const void (*__x86_cacheinfo_p) (void) attribute_hidden
+ = __x86_cacheinfo;
- /* If max_cpuid >= 11, THREADS_L2/THREADS_L3 are the maximum
- numbers of addressable IDs for logical processors sharing
- the cache, instead of the maximum number of threads
- sharing the cache. */
- if (max_cpuid >= 11 && support_count_mask)
- {
- /* Find the number of logical processors shipped in
- one core and apply count mask. */
- i = 0;
-
- /* Count SMT only if there is L3 cache. Always count
- core if there is no L3 cache. */
- int count = ((threads_l2 > 0 && level == 3)
- | ((threads_l3 > 0
- || (threads_l2 > 0 && level == 2)) << 1));
-
- while (count)
- {
- __cpuid_count (11, i++, eax, ebx, ecx, edx);
-
- int shipped = ebx & 0xff;
- int type = ecx & 0xff00;
- if (shipped == 0 || type == 0)
- break;
- else if (type == 0x100)
- {
- /* Count SMT. */
- if ((count & 0x1))
- {
- int count_mask;
-
- /* Compute count mask. */
- asm ("bsr %1, %0"
- : "=r" (count_mask) : "g" (threads_l2));
- count_mask = ~(-1 << (count_mask + 1));
- threads_l2 = (shipped - 1) & count_mask;