mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Suresh Siddha <suresh.b.siddha@intel.com>
To: "H. Peter Anvin" <hpa@zytor.com>, Ingo Molnar <mingo@elte.hu>,
	Thomas Gleixner <tglx@linutronix.de>
Cc: LKML <linux-kernel@vger.kernel.org>,
	Suresh Siddha <suresh.b.siddha@intel.com>
Subject: [patch 3/3] x86: use xsaveopt in context-switch path when supported
Date: Mon, 19 Jul 2010 11:35:44 -0700	[thread overview]
Message-ID: <20100719183240.240236054@sbs-t61.sc.intel.com> (raw)
In-Reply-To: <20100719181941.971215286@sbs-t61.sc.intel.com>

[-- Attachment #1: xsaveopt_support.patch --]
[-- Type: text/plain, Size: 5151 bytes --]

xsaveopt is a more optimized form of xsave specifically designed
for the context switch usage. xsaveopt doesn't save the state that's not
modified from the prior xrstor. And if a specific feature state gets
modified to the init state, then xsaveopt just updates the header bit
in the xsave memory layout without updating the corresponding memory
layout.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
---
 arch/x86/include/asm/cpufeature.h          |    1 +
 arch/x86/include/asm/i387.h                |    2 +-
 arch/x86/include/asm/xsave.h               |    9 ++++++---
 arch/x86/kernel/cpu/addon_cpuid_features.c |   24 +++++++++++++-----------
 arch/x86/kernel/cpu/common.c               |    8 ++++++++
 5 files changed, 29 insertions(+), 15 deletions(-)

Index: tip/arch/x86/kernel/cpu/addon_cpuid_features.c
===================================================================
--- tip.orig/arch/x86/kernel/cpu/addon_cpuid_features.c
+++ tip/arch/x86/kernel/cpu/addon_cpuid_features.c
@@ -13,6 +13,7 @@ struct cpuid_bit {
 	u16 feature;
 	u8 reg;
 	u8 bit;
+	u32 leaf;
 	u32 level;
 };
 
@@ -30,15 +31,16 @@ void __cpuinit init_scattered_cpuid_feat
 	const struct cpuid_bit *cb;
 
 	static const struct cpuid_bit __cpuinitconst cpuid_bits[] = {
-		{ X86_FEATURE_IDA,   		CR_EAX, 1, 0x00000006 },
-		{ X86_FEATURE_ARAT,  		CR_EAX, 2, 0x00000006 },
-		{ X86_FEATURE_APERFMPERF,	CR_ECX, 0, 0x00000006 },
-		{ X86_FEATURE_EPB,		CR_ECX, 3, 0x00000006 },
-		{ X86_FEATURE_CPB,   		CR_EDX, 9, 0x80000007 },
-		{ X86_FEATURE_NPT,   		CR_EDX, 0, 0x8000000a },
-		{ X86_FEATURE_LBRV,  		CR_EDX, 1, 0x8000000a },
-		{ X86_FEATURE_SVML,  		CR_EDX, 2, 0x8000000a },
-		{ X86_FEATURE_NRIPS, 		CR_EDX, 3, 0x8000000a },
+		{ X86_FEATURE_IDA,		CR_EAX, 1, 0, 0x00000006 },
+		{ X86_FEATURE_ARAT,		CR_EAX, 2, 0, 0x00000006 },
+		{ X86_FEATURE_APERFMPERF,	CR_ECX, 0, 0, 0x00000006 },
+		{ X86_FEATURE_EPB,		CR_ECX, 3, 0, 0x00000006 },
+		{ X86_FEATURE_CPB,		CR_EDX, 9, 0, 0x80000007 },
+		{ X86_FEATURE_NPT,		CR_EDX, 0, 0, 0x8000000a },
+		{ X86_FEATURE_LBRV,		CR_EDX, 1, 0, 0x8000000a },
+		{ X86_FEATURE_SVML,		CR_EDX, 2, 0, 0x8000000a },
+		{ X86_FEATURE_NRIPS,		CR_EDX, 3, 0, 0x8000000a },
+		{ X86_FEATURE_XSAVEOPT,		CR_EAX, 0, 1, 0x0000000d },
 		{ 0, 0, 0, 0 }
 	};
 
@@ -50,8 +52,8 @@ void __cpuinit init_scattered_cpuid_feat
 		    max_level > (cb->level | 0xffff))
 			continue;
 
-		cpuid(cb->level, &regs[CR_EAX], &regs[CR_EBX],
-			&regs[CR_ECX], &regs[CR_EDX]);
+		cpuid_count(cb->level, cb->leaf, &regs[CR_EAX], &regs[CR_EBX],
+			    &regs[CR_ECX], &regs[CR_EDX]);
 
 		if (regs[cb->reg] & (1 << cb->bit))
 			set_cpu_cap(c, cb->feature);
Index: tip/arch/x86/include/asm/cpufeature.h
===================================================================
--- tip.orig/arch/x86/include/asm/cpufeature.h
+++ tip/arch/x86/include/asm/cpufeature.h
@@ -165,6 +165,7 @@
 #define X86_FEATURE_ARAT	(7*32+ 1) /* Always Running APIC Timer */
 #define X86_FEATURE_CPB		(7*32+ 2) /* AMD Core Performance Boost */
 #define X86_FEATURE_EPB		(7*32+ 3) /* IA32_ENERGY_PERF_BIAS support */
+#define X86_FEATURE_XSAVEOPT	(7*32+4) /* "xsaveopt" Optimized Xsave */
 
 /* Virtualization flags: Linux defined, word 8 */
 #define X86_FEATURE_TPR_SHADOW  (8*32+ 0) /* Intel TPR Shadow */
Index: tip/arch/x86/include/asm/xsave.h
===================================================================
--- tip.orig/arch/x86/include/asm/xsave.h
+++ tip/arch/x86/include/asm/xsave.h
@@ -135,8 +135,11 @@ static inline void fpu_xsave(struct fpu 
 {
 	/* This, however, we can work around by forcing the compiler to select
 	   an addressing mode that doesn't require extended registers. */
-	__asm__ __volatile__(".byte " REX_PREFIX "0x0f,0xae,0x27"
-			     : : "D" (&(fpu->state->xsave)),
-				 "a" (-1), "d"(-1) : "memory");
+	alternative_input(
+		".byte " REX_PREFIX "0x0f,0xae,0x27",
+		".byte " REX_PREFIX "0x0f,0xae,0x37",
+		X86_FEATURE_XSAVEOPT,
+		[fx] "D" (&fpu->state->xsave), "a" (-1), "d" (-1) :
+		"memory");
 }
 #endif
Index: tip/arch/x86/kernel/cpu/common.c
===================================================================
--- tip.orig/arch/x86/kernel/cpu/common.c
+++ tip/arch/x86/kernel/cpu/common.c
@@ -140,10 +140,18 @@ EXPORT_PER_CPU_SYMBOL_GPL(gdt_page);
 static int __init x86_xsave_setup(char *s)
 {
 	setup_clear_cpu_cap(X86_FEATURE_XSAVE);
+	setup_clear_cpu_cap(X86_FEATURE_XSAVEOPT);
 	return 1;
 }
 __setup("noxsave", x86_xsave_setup);
 
+static int __init x86_xsaveopt_setup(char *s)
+{
+	setup_clear_cpu_cap(X86_FEATURE_XSAVEOPT);
+	return 1;
+}
+__setup("noxsaveopt", x86_xsaveopt_setup);
+
 #ifdef CONFIG_X86_32
 static int cachesize_override __cpuinitdata = -1;
 static int disable_x86_serial_nr __cpuinitdata = 1;
Index: tip/arch/x86/include/asm/i387.h
===================================================================
--- tip.orig/arch/x86/include/asm/i387.h
+++ tip/arch/x86/include/asm/i387.h
@@ -60,7 +60,7 @@ extern int restore_i387_xstate_ia32(void
 
 static __always_inline __pure bool use_xsaveopt(void)
 {
-	return 0;
+	return static_cpu_has(X86_FEATURE_XSAVEOPT);
 }
 
 static __always_inline __pure bool use_xsave(void)



      parent reply	other threads:[~2010-07-19 18:37 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-07-19 18:35 [patch 0/3] x86: xsaveopt kernel enabling patches Suresh Siddha
2010-07-19 18:35 ` [patch 1/3] x86: track the offset, size of features state in the xsave layout Suresh Siddha
2010-07-19 18:35 ` [patch 2/3] x86: sync xsave memory layout with its header for user handling Suresh Siddha
2010-07-19 18:35 ` Suresh Siddha [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20100719183240.240236054@sbs-t61.sc.intel.com \
    --to=suresh.b.siddha@intel.com \
    --cc=hpa@zytor.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=tglx@linutronix.de \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

all inboxes | Powered by JetHome®