Skip to contentSkip to conversation

x86/asm changes for v3.1

GIT PULL Ingo Molnar 1 participant 1 message Jul 22, 2011

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

Linus,

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

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

out-of-topic modifications in x86-asm-for-linus:
------------------------------------------------
arch/um/sys-i386/Makefile # a738669: x86: Unify rwsem assembly impleme
arch/um/sys-x86_64/Makefile # a738669: x86: Unify rwsem assembly impleme

Thanks,

Ingo

------------------>
Borislav Petkov (5):
x86, asm: Thin down SAVE/RESTORE_* asm macros x86, asm: Flip SAVE_ARGS arguments logic x86, asm: Flip RESTORE_ARGS arguments logic x86, asm: Cleanup thunk_64.S
x86, asm: Fix binutils 2.16 issue with __USER32_CS

Jan Beulich (3):
x86: Unify rwlock assembly implementation
x86: Unify rwsem assembly implementation
x86: Fix write lock scalability 64-bit issue

 arch/um/sys-i386/Makefile             |    2 +-
 arch/um/sys-x86_64/Makefile           |    2 +-
 arch/x86/ia32/ia32entry.S             |   10 +-
 arch/x86/include/asm/asm.h            |    5 +-
 arch/x86/include/asm/calling.h        |  130 +++++++++++++--------------------
 arch/x86/include/asm/frame.h          |   11 ++-
 arch/x86/include/asm/rwlock.h         |   43 +++++++++++-
 arch/x86/include/asm/segment.h        |    2 +-
 arch/x86/include/asm/spinlock.h       |   37 ++++++----
 arch/x86/include/asm/spinlock_types.h |    6 +-
 arch/x86/kernel/entry_64.S            |    6 +-
 arch/x86/lib/Makefile                 |    7 +-
 arch/x86/lib/rwlock.S                 |   44 +++++++++++
 arch/x86/lib/rwlock_64.S              |   38 ----------
 arch/x86/lib/{rwsem_64.S => rwsem.S}  |   75 +++++++++++++++----
 arch/x86/lib/semaphore_32.S           |  124 -------------------------------
 arch/x86/lib/thunk_64.S               |   45 +++++-------
 17 files changed, 263 insertions(+), 324 deletions(-)
 create mode 100644 arch/x86/lib/rwlock.S
 delete mode 100644 arch/x86/lib/rwlock_64.S
 rename arch/x86/lib/{rwsem_64.S => rwsem.S} (55%)
 delete mode 100644 arch/x86/lib/semaphore_32.S
arch/um/sys-i386/Makefile +1 -1
index b1da91c..15587ed 100644
@@ -8,7 +8,7 @@ obj-y = bug.o bugs.o checksum.o delay.o fault.o ksyms.o ldt.o ptrace.o \
 
 obj-$(CONFIG_BINFMT_ELF) += elfcore.o
 
-subarch-obj-y = lib/semaphore_32.o lib/string_32.o
+subarch-obj-y = lib/rwsem.o lib/string_32.o
 subarch-obj-$(CONFIG_HIGHMEM) += mm/highmem_32.o
 subarch-obj-$(CONFIG_MODULES) += kernel/module.o
 
arch/um/sys-x86_64/Makefile +1 -1
index c1ea9eb..61fc99a 100644
@@ -9,7 +9,7 @@ obj-y = bug.o bugs.o delay.o fault.o ldt.o mem.o ptrace.o ptrace_user.o \
 	sysrq.o ksyms.o tls.o
 
 subarch-obj-y = lib/csum-partial_64.o lib/memcpy_64.o lib/thunk_64.o \
-		lib/rwsem_64.o
+		lib/rwsem.o
 subarch-obj-$(CONFIG_MODULES) += kernel/module.o
 
 ldt-y = ../sys-i386/ldt.o
arch/x86/ia32/ia32entry.S +5 -5
index c1870dd..a0e866d 100644
@@ -143,7 +143,7 @@ ENTRY(ia32_sysenter_target)
 	CFI_REL_OFFSET rip,0
 	pushq_cfi %rax
 	cld
-	SAVE_ARGS 0,0,1
+	SAVE_ARGS 0,1,0
  	/* no need to do an access_ok check here because rbp has been
  	   32bit zero extended */ 
 1:	movl	(%rbp),%ebp
@@ -173,7 +173,7 @@ sysexit_from_sys_call:
 	andl  $~0x200,EFLAGS-R11(%rsp) 
 	movl	RIP-R11(%rsp),%edx		/* User %eip */
 	CFI_REGISTER rip,rdx
-	RESTORE_ARGS 1,24,1,1,1,1
+	RESTORE_ARGS 0,24,0,0,0,0
 	xorq	%r8,%r8
 	xorq	%r9,%r9
 	xorq	%r10,%r10
@@ -289,7 +289,7 @@ ENTRY(ia32_cstar_target)
 	 * disabled irqs and here we enable it straight after entry:
 	 */
 	ENABLE_INTERRUPTS(CLBR_NONE)
-	SAVE_ARGS 8,1,1
+	SAVE_ARGS 8,0,0
 	movl 	%eax,%eax	/* zero extension */
 	movq	%rax,ORIG_RAX-ARGOFFSET(%rsp)
 	movq	%rcx,RIP-ARGOFFSET(%rsp)
@@ -328,7 +328,7 @@ cstar_dispatch:
 	jnz sysretl_audit
 sysretl_from_sys_call:
 	andl $~TS_COMPAT,TI_status(%r10)
-	RESTORE_ARGS 1,-ARG_SKIP,1,1,1
+	RESTORE_ARGS 0,-ARG_SKIP,0,0,0
 	movl RIP-ARGOFFSET(%rsp),%ecx
 	CFI_REGISTER rip,rcx
 	movl EFLAGS-ARGOFFSET(%rsp),%r11d	
@@ -419,7 +419,7 @@ ENTRY(ia32_syscall)
 	cld
 	/* note the registers are not zero extended to the sf.
 	   this could be a problem. */
-	SAVE_ARGS 0,0,1
+	SAVE_ARGS 0,1,0
 	GET_THREAD_INFO(%r10)
 	orl   $TS_COMPAT,TI_status(%r10)
 	testl $_TIF_WORK_SYSCALL_ENTRY,TI_flags(%r10)
arch/x86/include/asm/asm.h +4 -1
index b3ed1e1..9412d65 100644
@@ -3,9 +3,11 @@
 
 #ifdef __ASSEMBLY__
 # define __ASM_FORM(x)	x
+# define __ASM_FORM_COMMA(x) x,
 # define __ASM_EX_SEC	.section __ex_table, "a"
 #else
 # define __ASM_FORM(x)	" " #x " "
+# define __ASM_FORM_COMMA(x) " " #x ","
 # define __ASM_EX_SEC	" .section __ex_table,\"a\"\n"
 #endif
 
@@ -15,7 +17,8 @@
 # define __ASM_SEL(a,b) __ASM_FORM(b)
 #endif
 
-#define __ASM_SIZE(inst)	__ASM_SEL(inst##l, inst##q)
+#define __ASM_SIZE(inst, ...)	__ASM_SEL(inst##l##__VA_ARGS__, \
+					  inst##q##__VA_ARGS__)
 #define __ASM_REG(reg)		__ASM_SEL(e##reg, r##reg)
 
 #define _ASM_PTR	__ASM_SEL(.long, .quad)
arch/x86/include/asm/calling.h +52 -78

170 lines not on this page · raw message

arch/x86/include/asm/frame.h +6 -5
index 2c6fc9e..3b629f4 100644
@@ -1,5 +1,6 @@
 #ifdef __ASSEMBLY__
 
+#include <asm/asm.h>
 #include <asm/dwarf2.h>
 
 /* The annotation hides the frame from the unwinder and makes it look
@@ -7,13 +8,13 @@
    frame pointer later */
 #ifdef CONFIG_FRAME_POINTER
 	.macro FRAME
-	pushl_cfi %ebp
-	CFI_REL_OFFSET ebp,0
-	movl %esp,%ebp
+	__ASM_SIZE(push,_cfi)	%__ASM_REG(bp)
+	CFI_REL_OFFSET		__ASM_REG(bp), 0
+	__ASM_SIZE(mov)		%__ASM_REG(sp), %__ASM_REG(bp)
 	.endm
 	.macro ENDFRAME
-	popl_cfi %ebp
-	CFI_RESTORE ebp
+	__ASM_SIZE(pop,_cfi)	%__ASM_REG(bp)
+	CFI_RESTORE		__ASM_REG(bp)
 	.endm
 #else
 	.macro FRAME
arch/x86/include/asm/rwlock.h +42 -1

49 lines not on this page · raw message

arch/x86/include/asm/segment.h +1 -1
index cd84f72..5e64171 100644
@@ -162,7 +162,7 @@
 #define GDT_ENTRY_DEFAULT_USER32_CS 4
 #define GDT_ENTRY_DEFAULT_USER_DS 5
 #define GDT_ENTRY_DEFAULT_USER_CS 6
-#define __USER32_CS   (GDT_ENTRY_DEFAULT_USER32_CS * 8 + 3)
+#define __USER32_CS   (GDT_ENTRY_DEFAULT_USER32_CS*8+3)
 #define __USER32_DS	__USER_DS
 
 #define GDT_ENTRY_TSS 8	/* needs two entries */
arch/x86/include/asm/spinlock.h +22 -15
index 3089f70..e9e51f7 100644
@@ -2,7 +2,6 @@
 #define _ASM_X86_SPINLOCK_H
 
 #include <asm/atomic.h>
-#include <asm/rwlock.h>
 #include <asm/page.h>
 #include <asm/processor.h>
 #include <linux/compiler.h>
@@ -234,7 +233,7 @@ static inline void arch_spin_unlock_wait(arch_spinlock_t *lock)
  */
 static inline int arch_read_can_lock(arch_rwlock_t *lock)
 {
-	return (int)(lock)->lock > 0;
+	return lock->lock > 0;
 }
 
 /**
@@ -243,12 +242,12 @@ static inline int arch_read_can_lock(arch_rwlock_t *lock)
  */
 static inline int arch_write_can_lock(arch_rwlock_t *lock)
 {
-	return (lock)->lock == RW_LOCK_BIAS;
+	return lock->write == WRITE_LOCK_CMP;
 }
 
 static inline void arch_read_lock(arch_rwlock_t *rw)
 {
-	asm volatile(LOCK_PREFIX " subl $1,(%0)\n\t"
+	asm volatile(LOCK_PREFIX READ_LOCK_SIZE(dec) " (%0)\n\t"
 		     "jns 1f\n"
 		     "call __read_lock_failed\n\t"
 		     "1:\n"
@@ -257,47 +256,55 @@ static inline void arch_read_lock(arch_rwlock_t *rw)
 
 static inline void arch_write_lock(arch_rwlock_t *rw)
 {
-	asm volatile(LOCK_PREFIX " subl %1,(%0)\n\t"
+	asm volatile(LOCK_PREFIX WRITE_LOCK_SUB(%1) "(%0)\n\t"
 		     "jz 1f\n"
 		     "call __write_lock_failed\n\t"
 		     "1:\n"
-		     ::LOCK_PTR_REG (rw), "i" (RW_LOCK_BIAS) : "memory");
+		     ::LOCK_PTR_REG (&rw->write), "i" (RW_LOCK_BIAS)
+		     : "memory");
 }
 
 static inline int arch_read_trylock(arch_rwlock_t *lock)
 {
-	atomic_t *count = (atomic_t *)lock;
+	READ_LOCK_ATOMIC(t) *count = (READ_LOCK_ATOMIC(t) *)lock;
 
-	if (atomic_dec_return(count) >= 0)
+	if (READ_LOCK_ATOMIC(dec_return)(count) >= 0)
 		return 1;
-	atomic_inc(count);
+	READ_LOCK_ATOMIC(inc)(count);
 	return 0;
 }
 
 static inline int arch_write_trylock(arch_rwlock_t *lock)
 {
-	atomic_t *count = (atomic_t *)lock;
+	atomic_t *count = (atomic_t *)&lock->write;
 
-	if (atomic_sub_and_test(RW_LOCK_BIAS, count))
+	if (atomic_sub_and_test(WRITE_LOCK_CMP, count))
 		return 1;
-	atomic_add(RW_LOCK_BIAS, count);
+	atomic_add(WRITE_LOCK_CMP, count);
 	return 0;
 }
 
 static inline void arch_read_unlock(arch_rwlock_t *rw)
 {
-	asm volatile(LOCK_PREFIX "incl %0" :"+m" (rw->lock) : : "memory");
+	asm volatile(LOCK_PREFIX READ_LOCK_SIZE(inc) " %0"
+		     :"+m" (rw->lock) : : "memory");
 }
 
 static inline void arch_write_unlock(arch_rwlock_t *rw)
 {
-	asm volatile(LOCK_PREFIX "addl %1, %0"
-		     : "+m" (rw->lock) : "i" (RW_LOCK_BIAS) : "memory");
+	asm volatile(LOCK_PREFIX WRITE_LOCK_ADD(%1) "%0"
+		     : "+m" (rw->write) : "i" (RW_LOCK_BIAS) : "memory");
 }
 
 #define arch_read_lock_flags(lock, flags) arch_read_lock(lock)
 #define arch_write_lock_flags(lock, flags) arch_write_lock(lock)
 
+#undef READ_LOCK_SIZE
+#undef READ_LOCK_ATOMIC
+#undef WRITE_LOCK_ADD
+#undef WRITE_LOCK_SUB
+#undef WRITE_LOCK_CMP
+
 #define arch_spin_relax(lock)	cpu_relax()
 #define arch_read_relax(lock)	cpu_relax()
 #define arch_write_relax(lock)	cpu_relax()
arch/x86/include/asm/spinlock_types.h +1 -5
index dcb48b2..7c7a486 100644
@@ -11,10 +11,6 @@ typedef struct arch_spinlock {
 
 #define __ARCH_SPIN_LOCK_UNLOCKED	{ 0 }
 
-typedef struct {
-	unsigned int lock;
-} arch_rwlock_t;
-
-#define __ARCH_RW_LOCK_UNLOCKED		{ RW_LOCK_BIAS }
+#include <asm/rwlock.h>
 
 #endif /* _ASM_X86_SPINLOCK_TYPES_H */
arch/x86/kernel/entry_64.S +3 -3
index 8a445a0..0412bcb 100644
@@ -473,7 +473,7 @@ ENTRY(system_call_after_swapgs)
 	 * and short:
 	 */
 	ENABLE_INTERRUPTS(CLBR_NONE)
-	SAVE_ARGS 8,1
+	SAVE_ARGS 8,0
 	movq  %rax,ORIG_RAX-ARGOFFSET(%rsp)
 	movq  %rcx,RIP-ARGOFFSET(%rsp)
 	CFI_REL_OFFSET rip,RIP-ARGOFFSET
@@ -508,7 +508,7 @@ sysret_check:
 	TRACE_IRQS_ON
 	movq RIP-ARGOFFSET(%rsp),%rcx
 	CFI_REGISTER	rip,rcx
-	RESTORE_ARGS 0,-ARG_SKIP,1
+	RESTORE_ARGS 1,-ARG_SKIP,0
 	/*CFI_REGISTER	rflags,r11*/
 	movq	PER_CPU_VAR(old_rsp), %rsp
 	USERGS_SYSRET64
@@ -858,7 +858,7 @@ retint_restore_args:	/* return to kernel space */
 	 */
 	TRACE_IRQS_IRETQ
 restore_args:
-	RESTORE_ARGS 0,8,0
+	RESTORE_ARGS 1,8,1
 
 irq_return:
 	INTERRUPT_RETURN
arch/x86/lib/Makefile +4 -3
index f2479f1..5d46ddc 100644
@@ -20,6 +20,8 @@ lib-y := delay.o
 lib-y += thunk_$(BITS).o
 lib-y += usercopy_$(BITS).o getuser.o putuser.o
 lib-y += memcpy_$(BITS).o
+lib-$(CONFIG_SMP) += rwlock.o
+lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem.o
 lib-$(CONFIG_INSTRUCTION_DECODER) += insn.o inat.o
 
 obj-y += msr.o msr-reg.o msr-reg-export.o
@@ -29,7 +31,7 @@ ifeq ($(CONFIG_X86_32),y)
         lib-y += atomic64_cx8_32.o
         lib-y += checksum_32.o
         lib-y += strstr_32.o
-        lib-y += semaphore_32.o string_32.o
+        lib-y += string_32.o
         lib-y += cmpxchg.o
 ifneq ($(CONFIG_X86_CMPXCHG64),y)
         lib-y += cmpxchg8b_emu.o atomic64_386_32.o
@@ -40,7 +42,6 @@ else
         lib-y += csum-partial_64.o csum-copy_64.o csum-wrappers_64.o
         lib-y += thunk_64.o clear_page_64.o copy_page_64.o
         lib-y += memmove_64.o memset_64.o
-        lib-y += copy_user_64.o rwlock_64.o copy_user_nocache_64.o
-	lib-$(CONFIG_RWSEM_XCHGADD_ALGORITHM) += rwsem_64.o
+        lib-y += copy_user_64.o copy_user_nocache_64.o
 	lib-y += cmpxchg16b_emu.o
 endif
arch/x86/lib/rwlock.S new file +44 -0

44 lines not on this page · raw message

arch/x86/lib/rwlock_64.S deleted +0 -38
deleted file mode 100644
index 05ea55f..0000000
@@ -1,38 +0,0 @@
-/* Slow paths of read/write spinlocks. */
-
-#include <linux/linkage.h>
-#include <asm/rwlock.h>
-#include <asm/alternative-asm.h>
-#include <asm/dwarf2.h>
-
-/* rdi:	pointer to rwlock_t */
-ENTRY(__write_lock_failed)
-	CFI_STARTPROC
-	LOCK_PREFIX
-	addl $RW_LOCK_BIAS,(%rdi)
-1:	rep
-	nop
-	cmpl $RW_LOCK_BIAS,(%rdi)
-	jne 1b
-	LOCK_PREFIX
-	subl $RW_LOCK_BIAS,(%rdi)
-	jnz  __write_lock_failed
-	ret
-	CFI_ENDPROC
-END(__write_lock_failed)
-
-/* rdi:	pointer to rwlock_t */
-ENTRY(__read_lock_failed)
-	CFI_STARTPROC
-	LOCK_PREFIX
-	incl (%rdi)
-1:	rep
-	nop
-	cmpl $1,(%rdi)
-	js 1b
-	LOCK_PREFIX
-	decl (%rdi)
-	js __read_lock_failed
-	ret
-	CFI_ENDPROC
-END(__read_lock_failed)
arch/x86/lib/rwsem.S renamed from arch/x86/lib/rwsem_64.S +59 -16

114 lines not on this page · raw message

arch/x86/lib/semaphore_32.S deleted +0 -124

124 lines not on this page · raw message

arch/x86/lib/thunk_64.S +18 -27

68 lines not on this page · raw message