[PATCH 4.14 08/39] arm64: Avoid redundant type conversions in xchg() and cmpxchg()

From: Greg Kroah-Hartman
Date: Fri Mar 05 2021 - 07:40:38 EST

Next message: Greg Kroah-Hartman: "[PATCH 4.14 09/39] arm64: cmpxchg: Use "K" instead of "L" for ll/sc immediate constraint"
Previous message: Greg Kroah-Hartman: "[PATCH 4.14 07/39] arm64 module: set plt* section addresses to 0x0"
In reply to: Greg Kroah-Hartman: "[PATCH 4.14 07/39] arm64 module: set plt* section addresses to 0x0"
Next in thread: Greg Kroah-Hartman: "[PATCH 4.14 09/39] arm64: cmpxchg: Use "K" instead of "L" for ll/sc immediate constraint"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

From: Will Deacon <will.deacon@xxxxxxx>

commit 5ef3fe4cecdf82fdd71ce78988403963d01444d4 upstream.

Our atomic instructions (either LSE atomics of LDXR/STXR sequences)
natively support byte, half-word, word and double-word memory accesses
so there is no need to mask the data register prior to being stored.

Signed-off-by: Will Deacon <will.deacon@xxxxxxx>
Signed-off-by: Ben Hutchings <ben@xxxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
arch/arm64/include/asm/atomic_ll_sc.h | 53 +++++++--------
arch/arm64/include/asm/atomic_lse.h | 46 ++++++-------
arch/arm64/include/asm/cmpxchg.h | 116 +++++++++++++++++-----------------
3 files changed, 108 insertions(+), 107 deletions(-)

--- a/arch/arm64/include/asm/atomic_ll_sc.h
+++ b/arch/arm64/include/asm/atomic_ll_sc.h
@@ -248,48 +248,49 @@ __LL_SC_PREFIX(atomic64_dec_if_positive(
}
__LL_SC_EXPORT(atomic64_dec_if_positive);

-#define __CMPXCHG_CASE(w, sz, name, mb, acq, rel, cl) \
-__LL_SC_INLINE unsigned long \
-__LL_SC_PREFIX(__cmpxchg_case_##name(volatile void *ptr, \
- unsigned long old, \
- unsigned long new)) \
+#define __CMPXCHG_CASE(w, sfx, name, sz, mb, acq, rel, cl) \
+__LL_SC_INLINE u##sz \
+__LL_SC_PREFIX(__cmpxchg_case_##name##sz(volatile void *ptr, \
+ unsigned long old, \
+ u##sz new)) \
{ \
- unsigned long tmp, oldval; \
+ unsigned long tmp; \
+ u##sz oldval; \
\
asm volatile( \
" prfm pstl1strm, %[v]\n" \
- "1: ld" #acq "xr" #sz "\t%" #w "[oldval], %[v]\n" \
+ "1: ld" #acq "xr" #sfx "\t%" #w "[oldval], %[v]\n" \
" eor %" #w "[tmp], %" #w "[oldval], %" #w "[old]\n" \
" cbnz %" #w "[tmp], 2f\n" \
- " st" #rel "xr" #sz "\t%w[tmp], %" #w "[new], %[v]\n" \
+ " st" #rel "xr" #sfx "\t%w[tmp], %" #w "[new], %[v]\n" \
" cbnz %w[tmp], 1b\n" \
" " #mb "\n" \
"2:" \
: [tmp] "=&r" (tmp), [oldval] "=&r" (oldval), \
- [v] "+Q" (*(unsigned long *)ptr) \
+ [v] "+Q" (*(u##sz *)ptr) \
: [old] "Lr" (old), [new] "r" (new) \
: cl); \
\
return oldval; \
} \
-__LL_SC_EXPORT(__cmpxchg_case_##name);
+__LL_SC_EXPORT(__cmpxchg_case_##name##sz);

-__CMPXCHG_CASE(w, b, 1, , , , )
-__CMPXCHG_CASE(w, h, 2, , , , )
-__CMPXCHG_CASE(w, , 4, , , , )
-__CMPXCHG_CASE( , , 8, , , , )
-__CMPXCHG_CASE(w, b, acq_1, , a, , "memory")
-__CMPXCHG_CASE(w, h, acq_2, , a, , "memory")
-__CMPXCHG_CASE(w, , acq_4, , a, , "memory")
-__CMPXCHG_CASE( , , acq_8, , a, , "memory")
-__CMPXCHG_CASE(w, b, rel_1, , , l, "memory")
-__CMPXCHG_CASE(w, h, rel_2, , , l, "memory")
-__CMPXCHG_CASE(w, , rel_4, , , l, "memory")
-__CMPXCHG_CASE( , , rel_8, , , l, "memory")
-__CMPXCHG_CASE(w, b, mb_1, dmb ish, , l, "memory")
-__CMPXCHG_CASE(w, h, mb_2, dmb ish, , l, "memory")
-__CMPXCHG_CASE(w, , mb_4, dmb ish, , l, "memory")
-__CMPXCHG_CASE( , , mb_8, dmb ish, , l, "memory")
+__CMPXCHG_CASE(w, b, , 8, , , , )
+__CMPXCHG_CASE(w, h, , 16, , , , )
+__CMPXCHG_CASE(w, , , 32, , , , )
+__CMPXCHG_CASE( , , , 64, , , , )
+__CMPXCHG_CASE(w, b, acq_, 8, , a, , "memory")
+__CMPXCHG_CASE(w, h, acq_, 16, , a, , "memory")
+__CMPXCHG_CASE(w, , acq_, 32, , a, , "memory")
+__CMPXCHG_CASE( , , acq_, 64, , a, , "memory")
+__CMPXCHG_CASE(w, b, rel_, 8, , , l, "memory")
+__CMPXCHG_CASE(w, h, rel_, 16, , , l, "memory")
+__CMPXCHG_CASE(w, , rel_, 32, , , l, "memory")
+__CMPXCHG_CASE( , , rel_, 64, , , l, "memory")
+__CMPXCHG_CASE(w, b, mb_, 8, dmb ish, , l, "memory")
+__CMPXCHG_CASE(w, h, mb_, 16, dmb ish, , l, "memory")
+__CMPXCHG_CASE(w, , mb_, 32, dmb ish, , l, "memory")
+__CMPXCHG_CASE( , , mb_, 64, dmb ish, , l, "memory")

#undef __CMPXCHG_CASE

--- a/arch/arm64/include/asm/atomic_lse.h
+++ b/arch/arm64/include/asm/atomic_lse.h
@@ -446,22 +446,22 @@ static inline long atomic64_dec_if_posit

#define __LL_SC_CMPXCHG(op) __LL_SC_CALL(__cmpxchg_case_##op)

-#define __CMPXCHG_CASE(w, sz, name, mb, cl...) \
-static inline unsigned long __cmpxchg_case_##name(volatile void *ptr, \
- unsigned long old, \
- unsigned long new) \
+#define __CMPXCHG_CASE(w, sfx, name, sz, mb, cl...) \
+static inline u##sz __cmpxchg_case_##name##sz(volatile void *ptr, \
+ unsigned long old, \
+ u##sz new) \
{ \
register unsigned long x0 asm ("x0") = (unsigned long)ptr; \
register unsigned long x1 asm ("x1") = old; \
- register unsigned long x2 asm ("x2") = new; \
+ register u##sz x2 asm ("x2") = new; \
\
asm volatile(ARM64_LSE_ATOMIC_INSN( \
/* LL/SC */ \
- __LL_SC_CMPXCHG(name) \
+ __LL_SC_CMPXCHG(name##sz) \
__nops(2), \
/* LSE atomics */ \
" mov " #w "30, %" #w "[old]\n" \
- " cas" #mb #sz "\t" #w "30, %" #w "[new], %[v]\n" \
+ " cas" #mb #sfx "\t" #w "30, %" #w "[new], %[v]\n" \
" mov %" #w "[ret], " #w "30") \
: [ret] "+r" (x0), [v] "+Q" (*(unsigned long *)ptr) \
: [old] "r" (x1), [new] "r" (x2) \
@@ -470,22 +470,22 @@ static inline unsigned long __cmpxchg_ca
return x0; \
}

-__CMPXCHG_CASE(w, b, 1, )
-__CMPXCHG_CASE(w, h, 2, )
-__CMPXCHG_CASE(w, , 4, )
-__CMPXCHG_CASE(x, , 8, )
-__CMPXCHG_CASE(w, b, acq_1, a, "memory")
-__CMPXCHG_CASE(w, h, acq_2, a, "memory")
-__CMPXCHG_CASE(w, , acq_4, a, "memory")
-__CMPXCHG_CASE(x, , acq_8, a, "memory")
-__CMPXCHG_CASE(w, b, rel_1, l, "memory")
-__CMPXCHG_CASE(w, h, rel_2, l, "memory")
-__CMPXCHG_CASE(w, , rel_4, l, "memory")
-__CMPXCHG_CASE(x, , rel_8, l, "memory")
-__CMPXCHG_CASE(w, b, mb_1, al, "memory")
-__CMPXCHG_CASE(w, h, mb_2, al, "memory")
-__CMPXCHG_CASE(w, , mb_4, al, "memory")
-__CMPXCHG_CASE(x, , mb_8, al, "memory")
+__CMPXCHG_CASE(w, b, , 8, )
+__CMPXCHG_CASE(w, h, , 16, )
+__CMPXCHG_CASE(w, , , 32, )
+__CMPXCHG_CASE(x, , , 64, )
+__CMPXCHG_CASE(w, b, acq_, 8, a, "memory")
+__CMPXCHG_CASE(w, h, acq_, 16, a, "memory")
+__CMPXCHG_CASE(w, , acq_, 32, a, "memory")
+__CMPXCHG_CASE(x, , acq_, 64, a, "memory")
+__CMPXCHG_CASE(w, b, rel_, 8, l, "memory")
+__CMPXCHG_CASE(w, h, rel_, 16, l, "memory")
+__CMPXCHG_CASE(w, , rel_, 32, l, "memory")
+__CMPXCHG_CASE(x, , rel_, 64, l, "memory")
+__CMPXCHG_CASE(w, b, mb_, 8, al, "memory")
+__CMPXCHG_CASE(w, h, mb_, 16, al, "memory")
+__CMPXCHG_CASE(w, , mb_, 32, al, "memory")
+__CMPXCHG_CASE(x, , mb_, 64, al, "memory")

#undef __LL_SC_CMPXCHG
#undef __CMPXCHG_CASE
--- a/arch/arm64/include/asm/cmpxchg.h
+++ b/arch/arm64/include/asm/cmpxchg.h
@@ -29,46 +29,46 @@
* barrier case is generated as release+dmb for the former and
* acquire+release for the latter.
*/
-#define __XCHG_CASE(w, sz, name, mb, nop_lse, acq, acq_lse, rel, cl) \
-static inline unsigned long __xchg_case_##name(unsigned long x, \
- volatile void *ptr) \
-{ \
- unsigned long ret, tmp; \
- \
- asm volatile(ARM64_LSE_ATOMIC_INSN( \
- /* LL/SC */ \
- " prfm pstl1strm, %2\n" \
- "1: ld" #acq "xr" #sz "\t%" #w "0, %2\n" \
- " st" #rel "xr" #sz "\t%w1, %" #w "3, %2\n" \
- " cbnz %w1, 1b\n" \
- " " #mb, \
- /* LSE atomics */ \
- " swp" #acq_lse #rel #sz "\t%" #w "3, %" #w "0, %2\n" \
- __nops(3) \
- " " #nop_lse) \
- : "=&r" (ret), "=&r" (tmp), "+Q" (*(unsigned long *)ptr) \
- : "r" (x) \
- : cl); \
- \
- return ret; \
+#define __XCHG_CASE(w, sfx, name, sz, mb, nop_lse, acq, acq_lse, rel, cl) \
+static inline u##sz __xchg_case_##name##sz(u##sz x, volatile void *ptr) \
+{ \
+ u##sz ret; \
+ unsigned long tmp; \
+ \
+ asm volatile(ARM64_LSE_ATOMIC_INSN( \
+ /* LL/SC */ \
+ " prfm pstl1strm, %2\n" \
+ "1: ld" #acq "xr" #sfx "\t%" #w "0, %2\n" \
+ " st" #rel "xr" #sfx "\t%w1, %" #w "3, %2\n" \
+ " cbnz %w1, 1b\n" \
+ " " #mb, \
+ /* LSE atomics */ \
+ " swp" #acq_lse #rel #sfx "\t%" #w "3, %" #w "0, %2\n" \
+ __nops(3) \
+ " " #nop_lse) \
+ : "=&r" (ret), "=&r" (tmp), "+Q" (*(u##sz *)ptr) \
+ : "r" (x) \
+ : cl); \
+ \
+ return ret; \
}

-__XCHG_CASE(w, b, 1, , , , , , )
-__XCHG_CASE(w, h, 2, , , , , , )
-__XCHG_CASE(w, , 4, , , , , , )
-__XCHG_CASE( , , 8, , , , , , )
-__XCHG_CASE(w, b, acq_1, , , a, a, , "memory")
-__XCHG_CASE(w, h, acq_2, , , a, a, , "memory")
-__XCHG_CASE(w, , acq_4, , , a, a, , "memory")
-__XCHG_CASE( , , acq_8, , , a, a, , "memory")
-__XCHG_CASE(w, b, rel_1, , , , , l, "memory")
-__XCHG_CASE(w, h, rel_2, , , , , l, "memory")
-__XCHG_CASE(w, , rel_4, , , , , l, "memory")
-__XCHG_CASE( , , rel_8, , , , , l, "memory")
-__XCHG_CASE(w, b, mb_1, dmb ish, nop, , a, l, "memory")
-__XCHG_CASE(w, h, mb_2, dmb ish, nop, , a, l, "memory")
-__XCHG_CASE(w, , mb_4, dmb ish, nop, , a, l, "memory")
-__XCHG_CASE( , , mb_8, dmb ish, nop, , a, l, "memory")
+__XCHG_CASE(w, b, , 8, , , , , , )
+__XCHG_CASE(w, h, , 16, , , , , , )
+__XCHG_CASE(w, , , 32, , , , , , )
+__XCHG_CASE( , , , 64, , , , , , )
+__XCHG_CASE(w, b, acq_, 8, , , a, a, , "memory")
+__XCHG_CASE(w, h, acq_, 16, , , a, a, , "memory")
+__XCHG_CASE(w, , acq_, 32, , , a, a, , "memory")
+__XCHG_CASE( , , acq_, 64, , , a, a, , "memory")
+__XCHG_CASE(w, b, rel_, 8, , , , , l, "memory")
+__XCHG_CASE(w, h, rel_, 16, , , , , l, "memory")
+__XCHG_CASE(w, , rel_, 32, , , , , l, "memory")
+__XCHG_CASE( , , rel_, 64, , , , , l, "memory")
+__XCHG_CASE(w, b, mb_, 8, dmb ish, nop, , a, l, "memory")
+__XCHG_CASE(w, h, mb_, 16, dmb ish, nop, , a, l, "memory")
+__XCHG_CASE(w, , mb_, 32, dmb ish, nop, , a, l, "memory")
+__XCHG_CASE( , , mb_, 64, dmb ish, nop, , a, l, "memory")

#undef __XCHG_CASE

@@ -79,13 +79,13 @@ static __always_inline unsigned long __
{ \
switch (size) { \
case 1: \
- return __xchg_case##sfx##_1(x, ptr); \
+ return __xchg_case##sfx##_8(x, ptr); \
case 2: \
- return __xchg_case##sfx##_2(x, ptr); \
+ return __xchg_case##sfx##_16(x, ptr); \
case 4: \
- return __xchg_case##sfx##_4(x, ptr); \
+ return __xchg_case##sfx##_32(x, ptr); \
case 8: \
- return __xchg_case##sfx##_8(x, ptr); \
+ return __xchg_case##sfx##_64(x, ptr); \
default: \
BUILD_BUG(); \
} \
@@ -122,13 +122,13 @@ static __always_inline unsigned long __c
{ \
switch (size) { \
case 1: \
- return __cmpxchg_case##sfx##_1(ptr, (u8)old, new); \
+ return __cmpxchg_case##sfx##_8(ptr, (u8)old, new); \
case 2: \
- return __cmpxchg_case##sfx##_2(ptr, (u16)old, new); \
+ return __cmpxchg_case##sfx##_16(ptr, (u16)old, new); \
case 4: \
- return __cmpxchg_case##sfx##_4(ptr, old, new); \
+ return __cmpxchg_case##sfx##_32(ptr, old, new); \
case 8: \
- return __cmpxchg_case##sfx##_8(ptr, old, new); \
+ return __cmpxchg_case##sfx##_64(ptr, old, new); \
default: \
BUILD_BUG(); \
} \
@@ -222,16 +222,16 @@ __CMPXCHG_GEN(_mb)
__ret; \
})

-#define __CMPWAIT_CASE(w, sz, name) \
-static inline void __cmpwait_case_##name(volatile void *ptr, \
- unsigned long val) \
+#define __CMPWAIT_CASE(w, sfx, sz) \
+static inline void __cmpwait_case_##sz(volatile void *ptr, \
+ unsigned long val) \
{ \
unsigned long tmp; \
\
asm volatile( \
" sevl\n" \
" wfe\n" \
- " ldxr" #sz "\t%" #w "[tmp], %[v]\n" \
+ " ldxr" #sfx "\t%" #w "[tmp], %[v]\n" \
" eor %" #w "[tmp], %" #w "[tmp], %" #w "[val]\n" \
" cbnz %" #w "[tmp], 1f\n" \
" wfe\n" \
@@ -240,10 +240,10 @@ static inline void __cmpwait_case_##name
: [val] "r" (val)); \
}

-__CMPWAIT_CASE(w, b, 1);
-__CMPWAIT_CASE(w, h, 2);
-__CMPWAIT_CASE(w, , 4);
-__CMPWAIT_CASE( , , 8);
+__CMPWAIT_CASE(w, b, 8);
+__CMPWAIT_CASE(w, h, 16);
+__CMPWAIT_CASE(w, , 32);
+__CMPWAIT_CASE( , , 64);

#undef __CMPWAIT_CASE

@@ -254,13 +254,13 @@ static __always_inline void __cmpwait##s
{ \
switch (size) { \
case 1: \
- return __cmpwait_case##sfx##_1(ptr, (u8)val); \
+ return __cmpwait_case##sfx##_8(ptr, (u8)val); \
case 2: \
- return __cmpwait_case##sfx##_2(ptr, (u16)val); \
+ return __cmpwait_case##sfx##_16(ptr, (u16)val); \
case 4: \
- return __cmpwait_case##sfx##_4(ptr, val); \
+ return __cmpwait_case##sfx##_32(ptr, val); \
case 8: \
- return __cmpwait_case##sfx##_8(ptr, val); \
+ return __cmpwait_case##sfx##_64(ptr, val); \
default: \
BUILD_BUG(); \
} \

Next message: Greg Kroah-Hartman: "[PATCH 4.14 09/39] arm64: cmpxchg: Use "K" instead of "L" for ll/sc immediate constraint"
Previous message: Greg Kroah-Hartman: "[PATCH 4.14 07/39] arm64 module: set plt* section addresses to 0x0"
In reply to: Greg Kroah-Hartman: "[PATCH 4.14 07/39] arm64 module: set plt* section addresses to 0x0"
Next in thread: Greg Kroah-Hartman: "[PATCH 4.14 09/39] arm64: cmpxchg: Use "K" instead of "L" for ll/sc immediate constraint"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]