x86/entry/64: Use 'xorl' for faster register clearing

On some x86 CPU microarchitectures using 'xorq' to clear general-purpose registers is slower than 'xorl'. As 'xorl' is sufficient to clear all 64 bits of these registers due to zero-extension [*], switch the x86 64-bit entry code to use 'xorl'. No change in functionality and no change in code size. [*] According to Intel 64 and IA-32 Architecture Software Developer's Manual, section 3.4.1.1, the result of 32-bit operands are "zero- extended to a 64-bit result in the destination general-purpose register." The AMD64 Architecture Programmer’s Manual Volume 3, Appendix B.1, describes the same behaviour. Suggested-by: Denys Vlasenko <dvlasenk@redhat.com> Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net> Cc: Andy Lutomirski <luto@kernel.org> Cc: Arjan van de Ven <arjan@linux.intel.com> Cc: Borislav Petkov <bp@alien8.de> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: David Woodhouse <dwmw2@infradead.org> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Josh Poimboeuf <jpoimboe@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20180214175924.23065-3-linux@dominikbrodowski.net [ Improved on the changelog a bit. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
author: Dominik Brodowski <linux@dominikbrodowski.net> 2018-02-14 12:59:24 -0500
committer: Ingo Molnar <mingo@kernel.org> 2018-02-17 05:14:33 -0500
commit: ced5d0bf603fa0baee8ea889e1d70971fd210894 (patch)
tree: 6f9a705b89fc522ed4301049bdc58af4bcb7f21e
parent: 9e809d15d6b692fa061d74be7aaab1c79f6784b8 (diff)
2 files changed, 35 insertions, 35 deletions
diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h
index 196b6103edf6..5d10b7a85cad 100644
--- a/arch/x86/entry/calling.h
+++ b/arch/x86/entry/calling.h
@@ -117,25 +117,25 @@ For 32-bit we have the following conventions - kernel is built with
        pushq   %rcx            /* pt_regs->cx */
        pushq   \rax            /* pt_regs->ax */
        pushq   %r8             /* pt_regs->r8 */
-        xorq    %r8, %r8        /* nospec   r8 */
+        xorl    %r8d, %r8d      /* nospec   r8 */
        pushq   %r9             /* pt_regs->r9 */
-        xorq    %r9, %r9        /* nospec   r9 */
+        xorl    %r9d, %r9d      /* nospec   r9 */
        pushq   %r10            /* pt_regs->r10 */
-        xorq    %r10, %r10      /* nospec   r10 */
+        xorl    %r10d, %r10d    /* nospec   r10 */
        pushq   %r11            /* pt_regs->r11 */
-        xorq    %r11, %r11      /* nospec   r11*/
+        xorl    %r11d, %r11d    /* nospec   r11*/
        pushq   %rbx            /* pt_regs->rbx */
        xorl    %ebx, %ebx      /* nospec   rbx*/
        pushq   %rbp            /* pt_regs->rbp */
        xorl    %ebp, %ebp      /* nospec   rbp*/
        pushq   %r12            /* pt_regs->r12 */
-        xorq    %r12, %r12      /* nospec   r12*/
+        xorl    %r12d, %r12d    /* nospec   r12*/
        pushq   %r13            /* pt_regs->r13 */
-        xorq    %r13, %r13      /* nospec   r13*/
+        xorl    %r13d, %r13d    /* nospec   r13*/
        pushq   %r14            /* pt_regs->r14 */
-        xorq    %r14, %r14      /* nospec   r14*/
+        xorl    %r14d, %r14d    /* nospec   r14*/
        pushq   %r15            /* pt_regs->r15 */
-        xorq    %r15, %r15      /* nospec   r15*/
+        xorl    %r15d, %r15d    /* nospec   r15*/
        UNWIND_HINT_REGS
        .if \save_ret
        pushq   %rsi            /* return address on top of stack */
diff --git a/arch/x86/entry/entry_64_compat.S b/arch/x86/entry/entry_64_compat.S
index fd65e016e413..364ea4a207be 100644
--- a/arch/x86/entry/entry_64_compat.S
+++ b/arch/x86/entry/entry_64_compat.S
@@ -85,25 +85,25 @@ ENTRY(entry_SYSENTER_compat)
        pushq   %rcx                    /* pt_regs->cx */
        pushq   $-ENOSYS                /* pt_regs->ax */
        pushq   $0                      /* pt_regs->r8  = 0 */
-        xorq    %r8, %r8                /* nospec   r8 */
+        xorl    %r8d, %r8d              /* nospec   r8 */
        pushq   $0                      /* pt_regs->r9  = 0 */
-        xorq    %r9, %r9                /* nospec   r9 */
+        xorl    %r9d, %r9d              /* nospec   r9 */
        pushq   $0                      /* pt_regs->r10 = 0 */
-        xorq    %r10, %r10              /* nospec   r10 */
+        xorl    %r10d, %r10d            /* nospec   r10 */
        pushq   $0                      /* pt_regs->r11 = 0 */
-        xorq    %r11, %r11              /* nospec   r11 */
+        xorl    %r11d, %r11d            /* nospec   r11 */
        pushq   %rbx                    /* pt_regs->rbx */
        xorl    %ebx, %ebx              /* nospec   rbx */
        pushq   %rbp                    /* pt_regs->rbp (will be overwritten) */
        xorl    %ebp, %ebp              /* nospec   rbp */
        pushq   $0                      /* pt_regs->r12 = 0 */
-        xorq    %r12, %r12              /* nospec   r12 */
+        xorl    %r12d, %r12d            /* nospec   r12 */
        pushq   $0                      /* pt_regs->r13 = 0 */
-        xorq    %r13, %r13              /* nospec   r13 */
+        xorl    %r13d, %r13d            /* nospec   r13 */
        pushq   $0                      /* pt_regs->r14 = 0 */
-        xorq    %r14, %r14              /* nospec   r14 */
+        xorl    %r14d, %r14d            /* nospec   r14 */
        pushq   $0                      /* pt_regs->r15 = 0 */
-        xorq    %r15, %r15              /* nospec   r15 */
+        xorl    %r15d, %r15d            /* nospec   r15 */
        cld
        /*
@@ -224,25 +224,25 @@ GLOBAL(entry_SYSCALL_compat_after_hwframe)
        pushq   %rbp                    /* pt_regs->cx (stashed in bp) */
        pushq   $-ENOSYS                /* pt_regs->ax */
        pushq   $0                      /* pt_regs->r8  = 0 */
-        xorq    %r8, %r8                /* nospec   r8 */
+        xorl    %r8d, %r8d              /* nospec   r8 */
        pushq   $0                      /* pt_regs->r9  = 0 */
-        xorq    %r9, %r9                /* nospec   r9 */
+        xorl    %r9d, %r9d              /* nospec   r9 */
        pushq   $0                      /* pt_regs->r10 = 0 */
-        xorq    %r10, %r10              /* nospec   r10 */
+        xorl    %r10d, %r10d            /* nospec   r10 */
        pushq   $0                      /* pt_regs->r11 = 0 */
-        xorq    %r11, %r11              /* nospec   r11 */
+        xorl    %r11d, %r11d            /* nospec   r11 */
        pushq   %rbx                    /* pt_regs->rbx */
        xorl    %ebx, %ebx              /* nospec   rbx */
        pushq   %rbp                    /* pt_regs->rbp (will be overwritten) */
        xorl    %ebp, %ebp              /* nospec   rbp */
        pushq   $0                      /* pt_regs->r12 = 0 */
-        xorq    %r12, %r12              /* nospec   r12 */
+        xorl    %r12d, %r12d            /* nospec   r12 */
        pushq   $0                      /* pt_regs->r13 = 0 */
-        xorq    %r13, %r13              /* nospec   r13 */
+        xorl    %r13d, %r13d            /* nospec   r13 */
        pushq   $0                      /* pt_regs->r14 = 0 */
-        xorq    %r14, %r14              /* nospec   r14 */
+        xorl    %r14d, %r14d            /* nospec   r14 */
        pushq   $0                      /* pt_regs->r15 = 0 */
-        xorq    %r15, %r15              /* nospec   r15 */
+        xorl    %r15d, %r15d            /* nospec   r15 */
        /*
         * User mode is traced as though IRQs are on, and SYSENTER
@@ -298,9 +298,9 @@ sysret32_from_system_call:
         */
        SWITCH_TO_USER_CR3_NOSTACK scratch_reg=%r8 scratch_reg2=%r9
-        xorq    %r8, %r8
+        xorl    %r8d, %r8d
-        xorq    %r9, %r9
+        xorl    %r9d, %r9d
-        xorq    %r10, %r10
+        xorl    %r10d, %r10d
        swapgs
        sysretl
 END(entry_SYSCALL_compat)
@@ -358,25 +358,25 @@ ENTRY(entry_INT80_compat)
        pushq   %rcx                    /* pt_regs->cx */
        pushq   $-ENOSYS                /* pt_regs->ax */
        pushq   $0                      /* pt_regs->r8  = 0 */
-        xorq    %r8, %r8                /* nospec   r8 */
+        xorl    %r8d, %r8d              /* nospec   r8 */
        pushq   $0                      /* pt_regs->r9  = 0 */
-        xorq    %r9, %r9                /* nospec   r9 */
+        xorl    %r9d, %r9d              /* nospec   r9 */
        pushq   $0                      /* pt_regs->r10 = 0 */
-        xorq    %r10, %r10              /* nospec   r10 */
+        xorl    %r10d, %r10d            /* nospec   r10 */
        pushq   $0                      /* pt_regs->r11 = 0 */
-        xorq    %r11, %r11              /* nospec   r11 */
+        xorl    %r11d, %r11d            /* nospec   r11 */
        pushq   %rbx                    /* pt_regs->rbx */
        xorl    %ebx, %ebx              /* nospec   rbx */
        pushq   %rbp                    /* pt_regs->rbp */
        xorl    %ebp, %ebp              /* nospec   rbp */
        pushq   %r12                    /* pt_regs->r12 */
-        xorq    %r12, %r12              /* nospec   r12 */
+        xorl    %r12d, %r12d            /* nospec   r12 */
        pushq   %r13                    /* pt_regs->r13 */
-        xorq    %r13, %r13              /* nospec   r13 */
+        xorl    %r13d, %r13d            /* nospec   r13 */
        pushq   %r14                    /* pt_regs->r14 */
-        xorq    %r14, %r14              /* nospec   r14 */
+        xorl    %r14d, %r14d            /* nospec   r14 */
        pushq   %r15                    /* pt_regs->r15 */
-        xorq    %r15, %r15              /* nospec   r15 */
+        xorl    %r15d, %r15d            /* nospec   r15 */
        cld
        /*
author	Dominik Brodowski <linux@dominikbrodowski.net>	2018-02-14 12:59:24 -0500
committer	Ingo Molnar <mingo@kernel.org>	2018-02-17 05:14:33 -0500
commit	ced5d0bf603fa0baee8ea889e1d70971fd210894 (patch)
tree	6f9a705b89fc522ed4301049bdc58af4bcb7f21e
parent	9e809d15d6b692fa061d74be7aaab1c79f6784b8 (diff)

diff --git a/arch/x86/entry/calling.h b/arch/x86/entry/calling.h index 196b6103edf6..5d10b7a85cad 100644 --- a/arch/x86/entry/calling.h +++ b/arch/x86/entry/calling.h
@@ -117,25 +117,25 @@ For 32-bit we have the following conventions - kernel is built with
117	pushq %rcx /* pt_regs->cx */	117	pushq %rcx /* pt_regs->cx */
118	pushq \rax /* pt_regs->ax */	118	pushq \rax /* pt_regs->ax */
119	pushq %r8 /* pt_regs->r8 */	119	pushq %r8 /* pt_regs->r8 */
120	xorq %r8, %r8 /* nospec r8 */	120	xorl %r8d, %r8d /* nospec r8 */
121	pushq %r9 /* pt_regs->r9 */	121	pushq %r9 /* pt_regs->r9 */
122	xorq %r9, %r9 /* nospec r9 */	122	xorl %r9d, %r9d /* nospec r9 */
123	pushq %r10 /* pt_regs->r10 */	123	pushq %r10 /* pt_regs->r10 */
124	xorq %r10, %r10 /* nospec r10 */	124	xorl %r10d, %r10d /* nospec r10 */
125	pushq %r11 /* pt_regs->r11 */	125	pushq %r11 /* pt_regs->r11 */
126	xorq %r11, %r11 /* nospec r11*/	126	xorl %r11d, %r11d /* nospec r11*/
127	pushq %rbx /* pt_regs->rbx */	127	pushq %rbx /* pt_regs->rbx */
128	xorl %ebx, %ebx /* nospec rbx*/	128	xorl %ebx, %ebx /* nospec rbx*/
129	pushq %rbp /* pt_regs->rbp */	129	pushq %rbp /* pt_regs->rbp */
130	xorl %ebp, %ebp /* nospec rbp*/	130	xorl %ebp, %ebp /* nospec rbp*/
131	pushq %r12 /* pt_regs->r12 */	131	pushq %r12 /* pt_regs->r12 */
132	xorq %r12, %r12 /* nospec r12*/	132	xorl %r12d, %r12d /* nospec r12*/
133	pushq %r13 /* pt_regs->r13 */	133	pushq %r13 /* pt_regs->r13 */
134	xorq %r13, %r13 /* nospec r13*/	134	xorl %r13d, %r13d /* nospec r13*/
135	pushq %r14 /* pt_regs->r14 */	135	pushq %r14 /* pt_regs->r14 */
136	xorq %r14, %r14 /* nospec r14*/	136	xorl %r14d, %r14d /* nospec r14*/
137	pushq %r15 /* pt_regs->r15 */	137	pushq %r15 /* pt_regs->r15 */
138	xorq %r15, %r15 /* nospec r15*/	138	xorl %r15d, %r15d /* nospec r15*/
139	UNWIND_HINT_REGS	139	UNWIND_HINT_REGS
140	.if \save_ret	140	.if \save_ret
141	pushq %rsi /* return address on top of stack */	141	pushq %rsi /* return address on top of stack */


diff --git a/arch/x86/entry/entry_64_compat.S b/arch/x86/entry/entry_64_compat.S index fd65e016e413..364ea4a207be 100644 --- a/arch/x86/entry/entry_64_compat.S +++ b/arch/x86/entry/entry_64_compat.S
@@ -85,25 +85,25 @@ ENTRY(entry_SYSENTER_compat)
85	pushq %rcx /* pt_regs->cx */	85	pushq %rcx /* pt_regs->cx */
86	pushq $-ENOSYS /* pt_regs->ax */	86	pushq $-ENOSYS /* pt_regs->ax */
87	pushq $0 /* pt_regs->r8 = 0 */	87	pushq $0 /* pt_regs->r8 = 0 */
88	xorq %r8, %r8 /* nospec r8 */	88	xorl %r8d, %r8d /* nospec r8 */
89	pushq $0 /* pt_regs->r9 = 0 */	89	pushq $0 /* pt_regs->r9 = 0 */
90	xorq %r9, %r9 /* nospec r9 */	90	xorl %r9d, %r9d /* nospec r9 */
91	pushq $0 /* pt_regs->r10 = 0 */	91	pushq $0 /* pt_regs->r10 = 0 */
92	xorq %r10, %r10 /* nospec r10 */	92	xorl %r10d, %r10d /* nospec r10 */
93	pushq $0 /* pt_regs->r11 = 0 */	93	pushq $0 /* pt_regs->r11 = 0 */
94	xorq %r11, %r11 /* nospec r11 */	94	xorl %r11d, %r11d /* nospec r11 */
95	pushq %rbx /* pt_regs->rbx */	95	pushq %rbx /* pt_regs->rbx */
96	xorl %ebx, %ebx /* nospec rbx */	96	xorl %ebx, %ebx /* nospec rbx */
97	pushq %rbp /* pt_regs->rbp (will be overwritten) */	97	pushq %rbp /* pt_regs->rbp (will be overwritten) */
98	xorl %ebp, %ebp /* nospec rbp */	98	xorl %ebp, %ebp /* nospec rbp */
99	pushq $0 /* pt_regs->r12 = 0 */	99	pushq $0 /* pt_regs->r12 = 0 */
100	xorq %r12, %r12 /* nospec r12 */	100	xorl %r12d, %r12d /* nospec r12 */
101	pushq $0 /* pt_regs->r13 = 0 */	101	pushq $0 /* pt_regs->r13 = 0 */
102	xorq %r13, %r13 /* nospec r13 */	102	xorl %r13d, %r13d /* nospec r13 */
103	pushq $0 /* pt_regs->r14 = 0 */	103	pushq $0 /* pt_regs->r14 = 0 */
104	xorq %r14, %r14 /* nospec r14 */	104	xorl %r14d, %r14d /* nospec r14 */
105	pushq $0 /* pt_regs->r15 = 0 */	105	pushq $0 /* pt_regs->r15 = 0 */
106	xorq %r15, %r15 /* nospec r15 */	106	xorl %r15d, %r15d /* nospec r15 */
107	cld	107	cld
108		108
109	/*	109	/*
@@ -224,25 +224,25 @@ GLOBAL(entry_SYSCALL_compat_after_hwframe)
224	pushq %rbp /* pt_regs->cx (stashed in bp) */	224	pushq %rbp /* pt_regs->cx (stashed in bp) */
225	pushq $-ENOSYS /* pt_regs->ax */	225	pushq $-ENOSYS /* pt_regs->ax */
226	pushq $0 /* pt_regs->r8 = 0 */	226	pushq $0 /* pt_regs->r8 = 0 */
227	xorq %r8, %r8 /* nospec r8 */	227	xorl %r8d, %r8d /* nospec r8 */
228	pushq $0 /* pt_regs->r9 = 0 */	228	pushq $0 /* pt_regs->r9 = 0 */
229	xorq %r9, %r9 /* nospec r9 */	229	xorl %r9d, %r9d /* nospec r9 */
230	pushq $0 /* pt_regs->r10 = 0 */	230	pushq $0 /* pt_regs->r10 = 0 */
231	xorq %r10, %r10 /* nospec r10 */	231	xorl %r10d, %r10d /* nospec r10 */
232	pushq $0 /* pt_regs->r11 = 0 */	232	pushq $0 /* pt_regs->r11 = 0 */
233	xorq %r11, %r11 /* nospec r11 */	233	xorl %r11d, %r11d /* nospec r11 */
234	pushq %rbx /* pt_regs->rbx */	234	pushq %rbx /* pt_regs->rbx */
235	xorl %ebx, %ebx /* nospec rbx */	235	xorl %ebx, %ebx /* nospec rbx */
236	pushq %rbp /* pt_regs->rbp (will be overwritten) */	236	pushq %rbp /* pt_regs->rbp (will be overwritten) */
237	xorl %ebp, %ebp /* nospec rbp */	237	xorl %ebp, %ebp /* nospec rbp */
238	pushq $0 /* pt_regs->r12 = 0 */	238	pushq $0 /* pt_regs->r12 = 0 */
239	xorq %r12, %r12 /* nospec r12 */	239	xorl %r12d, %r12d /* nospec r12 */
240	pushq $0 /* pt_regs->r13 = 0 */	240	pushq $0 /* pt_regs->r13 = 0 */
241	xorq %r13, %r13 /* nospec r13 */	241	xorl %r13d, %r13d /* nospec r13 */
242	pushq $0 /* pt_regs->r14 = 0 */	242	pushq $0 /* pt_regs->r14 = 0 */
243	xorq %r14, %r14 /* nospec r14 */	243	xorl %r14d, %r14d /* nospec r14 */
244	pushq $0 /* pt_regs->r15 = 0 */	244	pushq $0 /* pt_regs->r15 = 0 */
245	xorq %r15, %r15 /* nospec r15 */	245	xorl %r15d, %r15d /* nospec r15 */
246		246
247	/*	247	/*
248	* User mode is traced as though IRQs are on, and SYSENTER	248	* User mode is traced as though IRQs are on, and SYSENTER
@@ -298,9 +298,9 @@ sysret32_from_system_call:
298	*/	298	*/
299	SWITCH_TO_USER_CR3_NOSTACK scratch_reg=%r8 scratch_reg2=%r9	299	SWITCH_TO_USER_CR3_NOSTACK scratch_reg=%r8 scratch_reg2=%r9
300		300
301	xorq %r8, %r8	301	xorl %r8d, %r8d
302	xorq %r9, %r9	302	xorl %r9d, %r9d
303	xorq %r10, %r10	303	xorl %r10d, %r10d
304	swapgs	304	swapgs
305	sysretl	305	sysretl
306	END(entry_SYSCALL_compat)	306	END(entry_SYSCALL_compat)
@@ -358,25 +358,25 @@ ENTRY(entry_INT80_compat)
358	pushq %rcx /* pt_regs->cx */	358	pushq %rcx /* pt_regs->cx */
359	pushq $-ENOSYS /* pt_regs->ax */	359	pushq $-ENOSYS /* pt_regs->ax */
360	pushq $0 /* pt_regs->r8 = 0 */	360	pushq $0 /* pt_regs->r8 = 0 */
361	xorq %r8, %r8 /* nospec r8 */	361	xorl %r8d, %r8d /* nospec r8 */
362	pushq $0 /* pt_regs->r9 = 0 */	362	pushq $0 /* pt_regs->r9 = 0 */
363	xorq %r9, %r9 /* nospec r9 */	363	xorl %r9d, %r9d /* nospec r9 */
364	pushq $0 /* pt_regs->r10 = 0 */	364	pushq $0 /* pt_regs->r10 = 0 */
365	xorq %r10, %r10 /* nospec r10 */	365	xorl %r10d, %r10d /* nospec r10 */
366	pushq $0 /* pt_regs->r11 = 0 */	366	pushq $0 /* pt_regs->r11 = 0 */
367	xorq %r11, %r11 /* nospec r11 */	367	xorl %r11d, %r11d /* nospec r11 */
368	pushq %rbx /* pt_regs->rbx */	368	pushq %rbx /* pt_regs->rbx */
369	xorl %ebx, %ebx /* nospec rbx */	369	xorl %ebx, %ebx /* nospec rbx */
370	pushq %rbp /* pt_regs->rbp */	370	pushq %rbp /* pt_regs->rbp */
371	xorl %ebp, %ebp /* nospec rbp */	371	xorl %ebp, %ebp /* nospec rbp */
372	pushq %r12 /* pt_regs->r12 */	372	pushq %r12 /* pt_regs->r12 */
373	xorq %r12, %r12 /* nospec r12 */	373	xorl %r12d, %r12d /* nospec r12 */
374	pushq %r13 /* pt_regs->r13 */	374	pushq %r13 /* pt_regs->r13 */
375	xorq %r13, %r13 /* nospec r13 */	375	xorl %r13d, %r13d /* nospec r13 */
376	pushq %r14 /* pt_regs->r14 */	376	pushq %r14 /* pt_regs->r14 */
377	xorq %r14, %r14 /* nospec r14 */	377	xorl %r14d, %r14d /* nospec r14 */
378	pushq %r15 /* pt_regs->r15 */	378	pushq %r15 /* pt_regs->r15 */
379	xorq %r15, %r15 /* nospec r15 */	379	xorl %r15d, %r15d /* nospec r15 */
380	cld	380	cld
381		381
382	/*	382	/*