[git pull] x86 fixes

From: Ingo Molnar
Date: Sat Sep 06 2008 - 15:02:16 EST

Linus,

Please pull the latest x86-fixes-for-linus git tree from:

git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip.git x86-fixes-for-linus

The NOP impact is a bit wider than what i'd like, but it addresses the
VirtualPC regression and it's what Peter tested - and also Peter's
solution of turning enhanced NOPs into a virtual CPUID flag that is
tested before use is kind of elegant IMO and reduces some of the mess
and risks.

The memory leak fixes are old commits freshly cherry-picked over from
another branch (i already ran a few test builds and test boots of it) -
they have a lot of test time already otherwise, i just missed their .27
impact.

Thanks,

Ingo

------------------>
Andreas Herrmann (2):
x86: pda_init(): fix memory leak when using CPU hotplug
x86: cpu_init(): fix memory leak when using CPU hotplug

Eduardo Habkost (1):
x86, xen: Use native_pte_flags instead of native_pte_val for .pte_flags

H. Peter Anvin (3):
x86: boot: stub out unimplemented CPU feature words
x86: add NOPL as a synthetic CPU feature bit
x86: use X86_FEATURE_NOPL in alternatives

Krzysztof Helt (1):
x86: delay early cpu initialization until cpuid is done

Yinghai Lu (1):
x86: move mtrr cpu cap setting early in early_init_xxxx


arch/x86/boot/cpucheck.c | 8 ++--
arch/x86/kernel/alternative.c | 36 ++++++-----------
arch/x86/kernel/cpu/amd.c | 9 ++--
arch/x86/kernel/cpu/centaur.c | 11 +++++
arch/x86/kernel/cpu/common.c | 34 +++++++++++++++-
arch/x86/kernel/cpu/common_64.c | 74 +++++++++++++++++++++++++++-------
arch/x86/kernel/cpu/cyrix.c | 32 +++++++++++++--
arch/x86/kernel/cpu/feature_names.c | 3 +-
arch/x86/xen/enlighten.c | 2 +-
include/asm-x86/cpufeature.h | 11 +++--
include/asm-x86/required-features.h | 8 +++-
11 files changed, 167 insertions(+), 61 deletions(-)

diff --git a/arch/x86/boot/cpucheck.c b/arch/x86/boot/cpucheck.c
index 4b9ae7c..4d3ff03 100644
--- a/arch/x86/boot/cpucheck.c
+++ b/arch/x86/boot/cpucheck.c
@@ -38,12 +38,12 @@ static const u32 req_flags[NCAPINTS] =
{
REQUIRED_MASK0,
REQUIRED_MASK1,
- REQUIRED_MASK2,
- REQUIRED_MASK3,
+ 0, /* REQUIRED_MASK2 not implemented in this file */
+ 0, /* REQUIRED_MASK3 not implemented in this file */
REQUIRED_MASK4,
- REQUIRED_MASK5,
+ 0, /* REQUIRED_MASK5 not implemented in this file */
REQUIRED_MASK6,
- REQUIRED_MASK7,
+ 0, /* REQUIRED_MASK7 not implemented in this file */
};

#define A32(a, b, c, d) (((d) << 24)+((c) << 16)+((b) << 8)+(a))
diff --git a/arch/x86/kernel/alternative.c b/arch/x86/kernel/alternative.c
index 2763cb3..65a0c1b 100644
--- a/arch/x86/kernel/alternative.c
+++ b/arch/x86/kernel/alternative.c
@@ -145,35 +145,25 @@ static const unsigned char *const p6_nops[ASM_NOP_MAX+1] = {
extern char __vsyscall_0;
const unsigned char *const *find_nop_table(void)
{
- return boot_cpu_data.x86_vendor != X86_VENDOR_INTEL ||
- boot_cpu_data.x86 < 6 ? k8_nops : p6_nops;
+ if (boot_cpu_data.x86_vendor == X86_VENDOR_INTEL &&
+ boot_cpu_has(X86_FEATURE_NOPL))
+ return p6_nops;
+ else
+ return k8_nops;
}

#else /* CONFIG_X86_64 */

-static const struct nop {
- int cpuid;
- const unsigned char *const *noptable;
-} noptypes[] = {
- { X86_FEATURE_K8, k8_nops },
- { X86_FEATURE_K7, k7_nops },
- { X86_FEATURE_P4, p6_nops },
- { X86_FEATURE_P3, p6_nops },
- { -1, NULL }
-};
-
const unsigned char *const *find_nop_table(void)
{
- const unsigned char *const *noptable = intel_nops;
- int i;
-
- for (i = 0; noptypes[i].cpuid >= 0; i++) {
- if (boot_cpu_has(noptypes[i].cpuid)) {
- noptable = noptypes[i].noptable;
- break;
- }
- }
- return noptable;
+ if (boot_cpu_has(X86_FEATURE_K8))
+ return k8_nops;
+ else if (boot_cpu_has(X86_FEATURE_K7))
+ return k7_nops;
+ else if (boot_cpu_has(X86_FEATURE_NOPL))
+ return p6_nops;
+ else
+ return intel_nops;
}

#endif /* CONFIG_X86_64 */
diff --git a/arch/x86/kernel/cpu/amd.c b/arch/x86/kernel/cpu/amd.c
index cae9cab..18514ed 100644
--- a/arch/x86/kernel/cpu/amd.c
+++ b/arch/x86/kernel/cpu/amd.c
@@ -31,6 +31,11 @@ static void __cpuinit early_init_amd(struct cpuinfo_x86 *c)
if (c->x86_power & (1<<8))
set_cpu_cap(c, X86_FEATURE_CONSTANT_TSC);
}
+
+ /* Set MTRR capability flag if appropriate */
+ if (c->x86_model == 13 || c->x86_model == 9 ||
+ (c->x86_model == 8 && c->x86_mask >= 8))
+ set_cpu_cap(c, X86_FEATURE_K6_MTRR);
}

static void __cpuinit init_amd(struct cpuinfo_x86 *c)
@@ -166,10 +171,6 @@ static void __cpuinit init_amd(struct cpuinfo_x86 *c)
mbytes);
}

- /* Set MTRR capability flag if appropriate */
- if (c->x86_model == 13 || c->x86_model == 9 ||
- (c->x86_model == 8 && c->x86_mask >= 8))
- set_cpu_cap(c, X86_FEATURE_K6_MTRR);
break;
}

diff --git a/arch/x86/kernel/cpu/centaur.c b/arch/x86/kernel/cpu/centaur.c
index e0f45ed..a0534c0 100644
--- a/arch/x86/kernel/cpu/centaur.c
+++ b/arch/x86/kernel/cpu/centaur.c
@@ -314,6 +314,16 @@ enum {
EAMD3D = 1<<20,
};

+static void __cpuinit early_init_centaur(struct cpuinfo_x86 *c)
+{
+ switch (c->x86) {
+ case 5:
+ /* Emulate MTRRs using Centaur's MCR. */
+ set_cpu_cap(c, X86_FEATURE_CENTAUR_MCR);
+ break;
+ }
+}
+
static void __cpuinit init_centaur(struct cpuinfo_x86 *c)
{

@@ -462,6 +472,7 @@ centaur_size_cache(struct cpuinfo_x86 *c, unsigned int size)
static struct cpu_dev centaur_cpu_dev __cpuinitdata = {
.c_vendor = "Centaur",
.c_ident = { "CentaurHauls" },
+ .c_early_init = early_init_centaur,
.c_init = init_centaur,
.c_size_cache = centaur_size_cache,
};
diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c
index 80ab20d..8aab851 100644
--- a/arch/x86/kernel/cpu/common.c
+++ b/arch/x86/kernel/cpu/common.c
@@ -13,6 +13,7 @@
#include <asm/mtrr.h>
#include <asm/mce.h>
#include <asm/pat.h>
+#include <asm/asm.h>
#ifdef CONFIG_X86_LOCAL_APIC
#include <asm/mpspec.h>
#include <asm/apic.h>
@@ -334,11 +335,40 @@ static void __init early_cpu_detect(void)

get_cpu_vendor(c, 1);

+ early_get_cap(c);
+
if (c->x86_vendor != X86_VENDOR_UNKNOWN &&
cpu_devs[c->x86_vendor]->c_early_init)
cpu_devs[c->x86_vendor]->c_early_init(c);
+}

- early_get_cap(c);
+/*
+ * The NOPL instruction is supposed to exist on all CPUs with
+ * family >= 6, unfortunately, that's not true in practice because
+ * of early VIA chips and (more importantly) broken virtualizers that
+ * are not easy to detect. Hence, probe for it based on first
+ * principles.
+ */
+static void __cpuinit detect_nopl(struct cpuinfo_x86 *c)
+{
+ const u32 nopl_signature = 0x888c53b1; /* Random number */
+ u32 has_nopl = nopl_signature;
+
+ clear_cpu_cap(c, X86_FEATURE_NOPL);
+ if (c->x86 >= 6) {
+ asm volatile("\n"
+ "1: .byte 0x0f,0x1f,0xc0\n" /* nopl %eax */
+ "2:\n"
+ " .section .fixup,\"ax\"\n"
+ "3: xor %0,%0\n"
+ " jmp 2b\n"
+ " .previous\n"
+ _ASM_EXTABLE(1b,3b)
+ : "+a" (has_nopl));
+
+ if (has_nopl == nopl_signature)
+ set_cpu_cap(c, X86_FEATURE_NOPL);
+ }
}

static void __cpuinit generic_identify(struct cpuinfo_x86 *c)
@@ -395,8 +425,8 @@ static void __cpuinit generic_identify(struct cpuinfo_x86 *c)
}

init_scattered_cpuid_features(c);
+ detect_nopl(c);
}
-
}

static void __cpuinit squash_the_stupid_serial_number(struct cpuinfo_x86 *c)
diff --git a/arch/x86/kernel/cpu/common_64.c b/arch/x86/kernel/cpu/common_64.c
index dd6e3f1..a11f5d4 100644
--- a/arch/x86/kernel/cpu/common_64.c
+++ b/arch/x86/kernel/cpu/common_64.c
@@ -18,6 +18,7 @@
#include <asm/mtrr.h>
#include <asm/mce.h>
#include <asm/pat.h>
+#include <asm/asm.h>
#include <asm/numa.h>
#ifdef CONFIG_X86_LOCAL_APIC
#include <asm/mpspec.h>
@@ -215,6 +216,39 @@ static void __init early_cpu_support_print(void)
}
}

+/*
+ * The NOPL instruction is supposed to exist on all CPUs with
+ * family >= 6, unfortunately, that's not true in practice because
+ * of early VIA chips and (more importantly) broken virtualizers that
+ * are not easy to detect. Hence, probe for it based on first
+ * principles.
+ *
+ * Note: no 64-bit chip is known to lack these, but put the code here
+ * for consistency with 32 bits, and to make it utterly trivial to
+ * diagnose the problem should it ever surface.
+ */
+static void __cpuinit detect_nopl(struct cpuinfo_x86 *c)
+{
+ const u32 nopl_signature = 0x888c53b1; /* Random number */
+ u32 has_nopl = nopl_signature;
+
+ clear_cpu_cap(c, X86_FEATURE_NOPL);
+ if (c->x86 >= 6) {
+ asm volatile("\n"
+ "1: .byte 0x0f,0x1f,0xc0\n" /* nopl %eax */
+ "2:\n"
+ " .section .fixup,\"ax\"\n"
+ "3: xor %0,%0\n"
+ " jmp 2b\n"
+ " .previous\n"
+ _ASM_EXTABLE(1b,3b)
+ : "+a" (has_nopl));
+
+ if (has_nopl == nopl_signature)
+ set_cpu_cap(c, X86_FEATURE_NOPL);
+ }
+}
+
static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c);

void __init early_cpu_init(void)
@@ -313,6 +347,8 @@ static void __cpuinit early_identify_cpu(struct cpuinfo_x86 *c)
c->x86_phys_bits = eax & 0xff;
}

+ detect_nopl(c);
+
if (c->x86_vendor != X86_VENDOR_UNKNOWN &&
cpu_devs[c->x86_vendor]->c_early_init)
cpu_devs[c->x86_vendor]->c_early_init(c);
@@ -493,17 +529,20 @@ void pda_init(int cpu)
/* others are initialized in smpboot.c */
pda->pcurrent = &init_task;
pda->irqstackptr = boot_cpu_stack;
+ pda->irqstackptr += IRQSTACKSIZE - 64;
} else {
- pda->irqstackptr = (char *)
- __get_free_pages(GFP_ATOMIC, IRQSTACK_ORDER);
- if (!pda->irqstackptr)
- panic("cannot allocate irqstack for cpu %d", cpu);
+ if (!pda->irqstackptr) {
+ pda->irqstackptr = (char *)
+ __get_free_pages(GFP_ATOMIC, IRQSTACK_ORDER);
+ if (!pda->irqstackptr)
+ panic("cannot allocate irqstack for cpu %d",
+ cpu);
+ pda->irqstackptr += IRQSTACKSIZE - 64;
+ }

if (pda->nodenumber == 0 && cpu_to_node(cpu) != NUMA_NO_NODE)
pda->nodenumber = cpu_to_node(cpu);
}
-
- pda->irqstackptr += IRQSTACKSIZE-64;
}

char boot_exception_stacks[(N_EXCEPTION_STACKS - 1) * EXCEPTION_STKSZ +
@@ -601,19 +640,22 @@ void __cpuinit cpu_init(void)
/*
* set up and load the per-CPU TSS
*/
- for (v = 0; v < N_EXCEPTION_STACKS; v++) {
+ if (!orig_ist->ist[0]) {
static const unsigned int order[N_EXCEPTION_STACKS] = {
- [0 ... N_EXCEPTION_STACKS - 1] = EXCEPTION_STACK_ORDER,
- [DEBUG_STACK - 1] = DEBUG_STACK_ORDER
+ [0 ... N_EXCEPTION_STACKS - 1] = EXCEPTION_STACK_ORDER,
+ [DEBUG_STACK - 1] = DEBUG_STACK_ORDER
};
- if (cpu) {
- estacks = (char *)__get_free_pages(GFP_ATOMIC, order[v]);
- if (!estacks)
- panic("Cannot allocate exception stack %ld %d\n",
- v, cpu);
+ for (v = 0; v < N_EXCEPTION_STACKS; v++) {
+ if (cpu) {
+ estacks = (char *)__get_free_pages(GFP_ATOMIC, order[v]);
+ if (!estacks)
+ panic("Cannot allocate exception "
+ "stack %ld %d\n", v, cpu);
+ }
+ estacks += PAGE_SIZE << order[v];
+ orig_ist->ist[v] = t->x86_tss.ist[v] =
+ (unsigned long)estacks;
}
- estacks += PAGE_SIZE << order[v];
- orig_ist->ist[v] = t->x86_tss.ist[v] = (unsigned long)estacks;
}

t->x86_tss.io_bitmap_base = offsetof(struct tss_struct, io_bitmap);
diff --git a/arch/x86/kernel/cpu/cyrix.c b/arch/x86/kernel/cpu/cyrix.c
index e710a21..898a5a2 100644
--- a/arch/x86/kernel/cpu/cyrix.c
+++ b/arch/x86/kernel/cpu/cyrix.c
@@ -15,13 +15,11 @@
/*
* Read NSC/Cyrix DEVID registers (DIR) to get more detailed info. about the CPU
*/
-static void __cpuinit do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
+static void __cpuinit __do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
{
unsigned char ccr2, ccr3;
- unsigned long flags;

/* we test for DEVID by checking whether CCR3 is writable */
- local_irq_save(flags);
ccr3 = getCx86(CX86_CCR3);
setCx86(CX86_CCR3, ccr3 ^ 0x80);
getCx86(0xc0); /* dummy to change bus */
@@ -44,9 +42,16 @@ static void __cpuinit do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
*dir0 = getCx86(CX86_DIR0);
*dir1 = getCx86(CX86_DIR1);
}
- local_irq_restore(flags);
}

+static void __cpuinit do_cyrix_devid(unsigned char *dir0, unsigned char *dir1)
+{
+ unsigned long flags;
+
+ local_irq_save(flags);
+ __do_cyrix_devid(dir0, dir1);
+ local_irq_restore(flags);
+}
/*
* Cx86_dir0_msb is a HACK needed by check_cx686_cpuid/slop in bugs.h in
* order to identify the Cyrix CPU model after we're out of setup.c
@@ -161,6 +166,24 @@ static void __cpuinit geode_configure(void)
local_irq_restore(flags);
}

+static void __cpuinit early_init_cyrix(struct cpuinfo_x86 *c)
+{
+ unsigned char dir0, dir0_msn, dir1 = 0;
+
+ __do_cyrix_devid(&dir0, &dir1);
+ dir0_msn = dir0 >> 4; /* identifies CPU "family" */
+
+ switch (dir0_msn) {
+ case 3: /* 6x86/6x86L */
+ /* Emulate MTRRs using Cyrix's ARRs. */
+ set_cpu_cap(c, X86_FEATURE_CYRIX_ARR);
+ break;
+ case 5: /* 6x86MX/M II */
+ /* Emulate MTRRs using Cyrix's ARRs. */
+ set_cpu_cap(c, X86_FEATURE_CYRIX_ARR);
+ break;
+ }
+}

static void __cpuinit init_cyrix(struct cpuinfo_x86 *c)
{
@@ -416,6 +439,7 @@ static void __cpuinit cyrix_identify(struct cpuinfo_x86 *c)
static struct cpu_dev cyrix_cpu_dev __cpuinitdata = {
.c_vendor = "Cyrix",
.c_ident = { "CyrixInstead" },
+ .c_early_init = early_init_cyrix,
.c_init = init_cyrix,
.c_identify = cyrix_identify,
};
diff --git a/arch/x86/kernel/cpu/feature_names.c b/arch/x86/kernel/cpu/feature_names.c
index e43ad4a..c901779 100644
--- a/arch/x86/kernel/cpu/feature_names.c
+++ b/arch/x86/kernel/cpu/feature_names.c
@@ -39,7 +39,8 @@ const char * const x86_cap_flags[NCAPINTS*32] = {
NULL, NULL, NULL, NULL,
"constant_tsc", "up", NULL, "arch_perfmon",
"pebs", "bts", NULL, NULL,
- "rep_good", NULL, NULL, NULL, NULL, NULL, NULL, NULL,
+ "rep_good", NULL, NULL, NULL,
+ "nopl", NULL, NULL, NULL,
NULL, NULL, NULL, NULL, NULL, NULL, NULL, NULL,

/* Intel-defined (#2) */
diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c
index 9ff6e3c..a4e201b 100644
--- a/arch/x86/xen/enlighten.c
+++ b/arch/x86/xen/enlighten.c
@@ -1324,7 +1324,7 @@ static const struct pv_mmu_ops xen_mmu_ops __initdata = {
.ptep_modify_prot_commit = __ptep_modify_prot_commit,

.pte_val = xen_pte_val,
- .pte_flags = native_pte_val,
+ .pte_flags = native_pte_flags,
.pgd_val = xen_pgd_val,

.make_pte = xen_make_pte,
diff --git a/include/asm-x86/cpufeature.h b/include/asm-x86/cpufeature.h
index 762f6a6..9489283 100644
--- a/include/asm-x86/cpufeature.h
+++ b/include/asm-x86/cpufeature.h
@@ -72,14 +72,15 @@
#define X86_FEATURE_UP (3*32+ 9) /* smp kernel running on up */
#define X86_FEATURE_FXSAVE_LEAK (3*32+10) /* FXSAVE leaks FOP/FIP/FOP */
#define X86_FEATURE_ARCH_PERFMON (3*32+11) /* Intel Architectural PerfMon */
-#define X86_FEATURE_PEBS (3*32+12) /* Precise-Event Based Sampling */
-#define X86_FEATURE_BTS (3*32+13) /* Branch Trace Store */
-#define X86_FEATURE_SYSCALL32 (3*32+14) /* syscall in ia32 userspace */
-#define X86_FEATURE_SYSENTER32 (3*32+15) /* sysenter in ia32 userspace */
+#define X86_FEATURE_PEBS (3*32+12) /* Precise-Event Based Sampling */
+#define X86_FEATURE_BTS (3*32+13) /* Branch Trace Store */
+#define X86_FEATURE_SYSCALL32 (3*32+14) /* syscall in ia32 userspace */
+#define X86_FEATURE_SYSENTER32 (3*32+15) /* sysenter in ia32 userspace */
#define X86_FEATURE_REP_GOOD (3*32+16) /* rep microcode works well on this CPU */
#define X86_FEATURE_MFENCE_RDTSC (3*32+17) /* Mfence synchronizes RDTSC */
#define X86_FEATURE_LFENCE_RDTSC (3*32+18) /* Lfence synchronizes RDTSC */
-#define X86_FEATURE_11AP (3*32+19) /* Bad local APIC aka 11AP */
+#define X86_FEATURE_11AP (3*32+19) /* Bad local APIC aka 11AP */
+#define X86_FEATURE_NOPL (3*32+20) /* The NOPL (0F 1F) instructions */

/* Intel-defined CPU features, CPUID level 0x00000001 (ecx), word 4 */
#define X86_FEATURE_XMM3 (4*32+ 0) /* Streaming SIMD Extensions-3 */
diff --git a/include/asm-x86/required-features.h b/include/asm-x86/required-features.h
index adec887..5c2ff4b 100644
--- a/include/asm-x86/required-features.h
+++ b/include/asm-x86/required-features.h
@@ -41,6 +41,12 @@
# define NEED_3DNOW 0
#endif

+#if defined(CONFIG_X86_P6_NOP) || defined(CONFIG_X86_64)
+# define NEED_NOPL (1<<(X86_FEATURE_NOPL & 31))
+#else
+# define NEED_NOPL 0
+#endif
+
#ifdef CONFIG_X86_64
#define NEED_PSE 0
#define NEED_MSR (1<<(X86_FEATURE_MSR & 31))
@@ -67,7 +73,7 @@
#define REQUIRED_MASK1 (NEED_LM|NEED_3DNOW)

#define REQUIRED_MASK2 0
-#define REQUIRED_MASK3 0
+#define REQUIRED_MASK3 (NEED_NOPL)
#define REQUIRED_MASK4 0
#define REQUIRED_MASK5 0
#define REQUIRED_MASK6 0
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/