[PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Andrzej Hajda]

Hi all,

The helper is tiny and there are advices we can live without it, so
I want to present few arguments why it would be good to have it:

1. Code readability/simplification/number of lines:
  - decreases number of lines,
  - it often eliminates local variables,
  - for real examples see patches 3+.

2. Presence of similar helpers in other somehow related languages/libs:

a) Rust[1]: 'replace' from std::mem module, there is also 'take'
    helper (__xchg(&x, 0)), which is the same as private helper in
    i915 - fetch_and_zero, see latest patch.
b) C++ [2]: 'exchange' from utility header.

If the idea is OK there are still 2 questions to answer:

1. Name of the helper, __xchg follows kernel conventions,
    but for me Rust names are also OK.
2. Where to put the helper:
a) as in this patchset include/linux/non-atomic/xchg.h,
    proposed by Andy Shevchenko,
b) include/linux/utils.h ? any better name? Some kind
    of container for simple helpers.

All __xchg conversions were performed using cocci script,
then manually adjusted if necessary.

There is lot of places it can be used in, I have just chosen
some of them. I can provide cocci script to detect others (not all),
if necessary.

Changes:
v2: squashed all __xchg -> __arch_xchg t one patch (Arnd)
v3: fixed alpha/xchg_local (lkp@intel.com)
v4: adjusted indentation (Heiko)
v5: added more __xchg conversions - patches 3-6, added tags

[1]: https://doc.rust-lang.org/std/mem/index.html
[2]: https://en.cppreference.com/w/cpp/header/utility

Regards
Andrzej

Andrzej Hajda (7):
  arch: rename all internal names __xchg to __arch_xchg
  linux/include: add non-atomic version of xchg
  arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
  llist: simplify __llist_del_all
  io_uring: use __xchg if possible
  qed: use __xchg if possible
  drm/i915/gt: use __xchg instead of internal helper

 arch/alpha/include/asm/cmpxchg.h              | 10 +++++-----
 arch/arc/include/asm/cmpxchg.h                |  4 ++--
 arch/arm/include/asm/cmpxchg.h                |  7 ++++---
 arch/arm/probes/uprobes/core.c                |  8 ++------
 arch/arm64/include/asm/cmpxchg.h              |  7 +++----
 arch/arm64/kernel/probes/uprobes.c            |  9 ++-------
 arch/csky/kernel/probes/uprobes.c             |  9 ++-------
 arch/hexagon/include/asm/cmpxchg.h            | 10 +++++-----
 arch/ia64/include/asm/cmpxchg.h               |  2 +-
 arch/ia64/include/uapi/asm/cmpxchg.h          |  4 ++--
 arch/loongarch/include/asm/cmpxchg.h          |  4 ++--
 arch/m68k/include/asm/cmpxchg.h               |  6 +++---
 arch/mips/include/asm/cmpxchg.h               |  4 ++--
 arch/mips/kernel/uprobes.c                    | 10 ++--------
 arch/openrisc/include/asm/cmpxchg.h           | 10 +++++-----
 arch/parisc/include/asm/cmpxchg.h             |  4 ++--
 arch/powerpc/include/asm/cmpxchg.h            |  4 ++--
 arch/powerpc/kernel/uprobes.c                 | 10 ++--------
 arch/riscv/include/asm/atomic.h               |  2 +-
 arch/riscv/include/asm/cmpxchg.h              |  4 ++--
 arch/riscv/kernel/probes/uprobes.c            |  9 ++-------
 arch/s390/include/asm/cmpxchg.h               |  8 ++++----
 arch/s390/kernel/uprobes.c                    |  7 ++-----
 arch/sh/include/asm/cmpxchg.h                 |  4 ++--
 arch/sparc/include/asm/cmpxchg_32.h           |  4 ++--
 arch/sparc/include/asm/cmpxchg_64.h           |  6 +++---
 arch/sparc/kernel/uprobes.c                   |  7 ++-----
 arch/xtensa/include/asm/cmpxchg.h             |  4 ++--
 drivers/gpu/drm/i915/gt/intel_engine_cs.c     |  2 +-
 .../gpu/drm/i915/gt/intel_engine_heartbeat.c  |  4 ++--
 .../drm/i915/gt/intel_execlists_submission.c  |  4 ++--
 drivers/gpu/drm/i915/gt/intel_ggtt.c          |  4 ++--
 drivers/gpu/drm/i915/gt/intel_gsc.c           |  2 +-
 drivers/gpu/drm/i915/gt/intel_gt.c            |  4 ++--
 drivers/gpu/drm/i915/gt/intel_gt_pm.c         |  2 +-
 drivers/gpu/drm/i915/gt/intel_lrc.c           |  6 +++---
 drivers/gpu/drm/i915/gt/intel_migrate.c       |  2 +-
 drivers/gpu/drm/i915/gt/intel_rc6.c           |  2 +-
 drivers/gpu/drm/i915/gt/intel_rps.c           |  2 +-
 drivers/gpu/drm/i915/gt/selftest_context.c    |  2 +-
 .../drm/i915/gt/selftest_ring_submission.c    |  2 +-
 drivers/gpu/drm/i915/gt/selftest_timeline.c   |  2 +-
 drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c     |  2 +-
 drivers/gpu/drm/i915/gt/uc/intel_uc.c         |  2 +-
 drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c      |  2 +-
 drivers/gpu/drm/i915/i915_utils.h             |  1 +
 include/linux/llist.h                         |  6 ++----
 include/linux/non-atomic/xchg.h               | 19 +++++++++++++++++++
 include/linux/qed/qed_chain.h                 | 19 +++++++------------
 io_uring/io_uring.c                           |  7 ++-----
 io_uring/slist.h                              |  6 ++----
 51 files changed, 126 insertions(+), 155 deletions(-)
 create mode 100644 include/linux/non-atomic/xchg.h

-- 
2.34.1

[PATCH v5 1/7] arch: rename all internal names __xchg to __arch_xchg

[Andrzej Hajda]

__xchg will be used for non-atomic xchg macro.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k]
Acked-by: Palmer Dabbelt <palmer@rivosinc.com> [riscv]
---
v2: squashed all arch patches into one
v3: fixed alpha/xchg_local, thx to lkp@intel.com
v4: adjusted indentation (Heiko)
---
 arch/alpha/include/asm/cmpxchg.h     | 10 +++++-----
 arch/arc/include/asm/cmpxchg.h       |  4 ++--
 arch/arm/include/asm/cmpxchg.h       |  7 ++++---
 arch/arm64/include/asm/cmpxchg.h     |  7 +++----
 arch/hexagon/include/asm/cmpxchg.h   | 10 +++++-----
 arch/ia64/include/asm/cmpxchg.h      |  2 +-
 arch/ia64/include/uapi/asm/cmpxchg.h |  4 ++--
 arch/loongarch/include/asm/cmpxchg.h |  4 ++--
 arch/m68k/include/asm/cmpxchg.h      |  6 +++---
 arch/mips/include/asm/cmpxchg.h      |  4 ++--
 arch/openrisc/include/asm/cmpxchg.h  | 10 +++++-----
 arch/parisc/include/asm/cmpxchg.h    |  4 ++--
 arch/powerpc/include/asm/cmpxchg.h   |  4 ++--
 arch/riscv/include/asm/atomic.h      |  2 +-
 arch/riscv/include/asm/cmpxchg.h     |  4 ++--
 arch/s390/include/asm/cmpxchg.h      |  8 ++++----
 arch/sh/include/asm/cmpxchg.h        |  4 ++--
 arch/sparc/include/asm/cmpxchg_32.h  |  4 ++--
 arch/sparc/include/asm/cmpxchg_64.h  |  6 +++---
 arch/xtensa/include/asm/cmpxchg.h    |  4 ++--
 20 files changed, 54 insertions(+), 54 deletions(-)

diff --git a/arch/alpha/include/asm/cmpxchg.h b/arch/alpha/include/asm/cmpxchg.h
index 6e0a850aa9d38c..91d4a4d9258cd2 100644
--- a/arch/alpha/include/asm/cmpxchg.h
+++ b/arch/alpha/include/asm/cmpxchg.h
@@ -6,15 +6,15 @@
  * Atomic exchange routines.
  */
 
-#define ____xchg(type, args...)		__xchg ## type ## _local(args)
+#define ____xchg(type, args...)		__arch_xchg ## type ## _local(args)
 #define ____cmpxchg(type, args...)	__cmpxchg ## type ## _local(args)
 #include <asm/xchg.h>
 
 #define xchg_local(ptr, x)						\
 ({									\
 	__typeof__(*(ptr)) _x_ = (x);					\
-	(__typeof__(*(ptr))) __xchg_local((ptr), (unsigned long)_x_,	\
-				       sizeof(*(ptr)));			\
+	(__typeof__(*(ptr))) __arch_xchg_local((ptr), (unsigned long)_x_,\
+					       sizeof(*(ptr)));		\
 })
 
 #define arch_cmpxchg_local(ptr, o, n)					\
@@ -34,7 +34,7 @@
 
 #undef ____xchg
 #undef ____cmpxchg
-#define ____xchg(type, args...)		__xchg ##type(args)
+#define ____xchg(type, args...)		__arch_xchg ##type(args)
 #define ____cmpxchg(type, args...)	__cmpxchg ##type(args)
 #include <asm/xchg.h>
 
@@ -48,7 +48,7 @@
 	__typeof__(*(ptr)) _x_ = (x);					\
 	smp_mb();							\
 	__ret = (__typeof__(*(ptr)))					\
-		__xchg((ptr), (unsigned long)_x_, sizeof(*(ptr)));	\
+		__arch_xchg((ptr), (unsigned long)_x_, sizeof(*(ptr)));	\
 	smp_mb();							\
 	__ret;								\
 })
diff --git a/arch/arc/include/asm/cmpxchg.h b/arch/arc/include/asm/cmpxchg.h
index c5b544a5fe8106..e138fde067dea5 100644
--- a/arch/arc/include/asm/cmpxchg.h
+++ b/arch/arc/include/asm/cmpxchg.h
@@ -85,7 +85,7 @@
  */
 #ifdef CONFIG_ARC_HAS_LLSC
 
-#define __xchg(ptr, val)						\
+#define __arch_xchg(ptr, val)						\
 ({									\
 	__asm__ __volatile__(						\
 	"	ex  %0, [%1]	\n"	/* set new value */	        \
@@ -102,7 +102,7 @@
 									\
 	switch(sizeof(*(_p_))) {					\
 	case 4:								\
-		_val_ = __xchg(_p_, _val_);				\
+		_val_ = __arch_xchg(_p_, _val_);			\
 		break;							\
 	default:							\
 		BUILD_BUG();						\
diff --git a/arch/arm/include/asm/cmpxchg.h b/arch/arm/include/asm/cmpxchg.h
index 4dfe538dfc689b..44667bdb4707a9 100644
--- a/arch/arm/include/asm/cmpxchg.h
+++ b/arch/arm/include/asm/cmpxchg.h
@@ -25,7 +25,8 @@
 #define swp_is_buggy
 #endif
 
-static inline unsigned long __xchg(unsigned long x, volatile void *ptr, int size)
+static inline unsigned long
+__arch_xchg(unsigned long x, volatile void *ptr, int size)
 {
 	extern void __bad_xchg(volatile void *, int);
 	unsigned long ret;
@@ -115,8 +116,8 @@ static inline unsigned long __xchg(unsigned long x, volatile void *ptr, int size
 }
 
 #define arch_xchg_relaxed(ptr, x) ({					\
-	(__typeof__(*(ptr)))__xchg((unsigned long)(x), (ptr),		\
-				   sizeof(*(ptr)));			\
+	(__typeof__(*(ptr)))__arch_xchg((unsigned long)(x), (ptr),	\
+					sizeof(*(ptr)));		\
 })
 
 #include <asm-generic/cmpxchg-local.h>
diff --git a/arch/arm64/include/asm/cmpxchg.h b/arch/arm64/include/asm/cmpxchg.h
index 497acf134d9923..c6bc5d8ec3ca41 100644
--- a/arch/arm64/include/asm/cmpxchg.h
+++ b/arch/arm64/include/asm/cmpxchg.h
@@ -62,9 +62,8 @@ __XCHG_CASE( ,  ,  mb_, 64, dmb ish, nop,  , a, l, "memory")
 #undef __XCHG_CASE
 
 #define __XCHG_GEN(sfx)							\
-static __always_inline  unsigned long __xchg##sfx(unsigned long x,	\
-					volatile void *ptr,		\
-					int size)			\
+static __always_inline unsigned long					\
+__arch_xchg##sfx(unsigned long x, volatile void *ptr, int size)		\
 {									\
 	switch (size) {							\
 	case 1:								\
@@ -93,7 +92,7 @@ __XCHG_GEN(_mb)
 ({									\
 	__typeof__(*(ptr)) __ret;					\
 	__ret = (__typeof__(*(ptr)))					\
-		__xchg##sfx((unsigned long)(x), (ptr), sizeof(*(ptr))); \
+		__arch_xchg##sfx((unsigned long)(x), (ptr), sizeof(*(ptr))); \
 	__ret;								\
 })
 
diff --git a/arch/hexagon/include/asm/cmpxchg.h b/arch/hexagon/include/asm/cmpxchg.h
index cdb705e1496af8..bf6cf5579cf459 100644
--- a/arch/hexagon/include/asm/cmpxchg.h
+++ b/arch/hexagon/include/asm/cmpxchg.h
@@ -9,7 +9,7 @@
 #define _ASM_CMPXCHG_H
 
 /*
- * __xchg - atomically exchange a register and a memory location
+ * __arch_xchg - atomically exchange a register and a memory location
  * @x: value to swap
  * @ptr: pointer to memory
  * @size:  size of the value
@@ -19,8 +19,8 @@
  * Note:  there was an errata for V2 about .new's and memw_locked.
  *
  */
-static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
-				   int size)
+static inline unsigned long
+__arch_xchg(unsigned long x, volatile void *ptr, int size)
 {
 	unsigned long retval;
 
@@ -42,8 +42,8 @@ static inline unsigned long __xchg(unsigned long x, volatile void *ptr,
  * Atomically swap the contents of a register with memory.  Should be atomic
  * between multiple CPU's and within interrupts on the same CPU.
  */
-#define arch_xchg(ptr, v) ((__typeof__(*(ptr)))__xchg((unsigned long)(v), (ptr), \
-	sizeof(*(ptr))))
+#define arch_xchg(ptr, v) ((__typeof__(*(ptr)))__arch_xchg((unsigned long)(v), (ptr), \
+							   sizeof(*(ptr))))
 
 /*
  *  see rt-mutex-design.txt; cmpxchg supposedly checks if *ptr == A and swaps.
diff --git a/arch/ia64/include/asm/cmpxchg.h b/arch/ia64/include/asm/cmpxchg.h
index 94ef844298431a..8b2e644ef6a14e 100644
--- a/arch/ia64/include/asm/cmpxchg.h
+++ b/arch/ia64/include/asm/cmpxchg.h
@@ -5,7 +5,7 @@
 #include <uapi/asm/cmpxchg.h>
 
 #define arch_xchg(ptr, x)	\
-({(__typeof__(*(ptr))) __xchg((unsigned long) (x), (ptr), sizeof(*(ptr)));})
+({(__typeof__(*(ptr))) __arch_xchg((unsigned long) (x), (ptr), sizeof(*(ptr)));})
 
 #define arch_cmpxchg(ptr, o, n)		cmpxchg_acq((ptr), (o), (n))
 #define arch_cmpxchg64(ptr, o, n)	cmpxchg_acq((ptr), (o), (n))
diff --git a/arch/ia64/include/uapi/asm/cmpxchg.h b/arch/ia64/include/uapi/asm/cmpxchg.h
index ca2e0268534384..3fec9b037051bb 100644
--- a/arch/ia64/include/uapi/asm/cmpxchg.h
+++ b/arch/ia64/include/uapi/asm/cmpxchg.h
@@ -27,7 +27,7 @@
  */
 extern void ia64_xchg_called_with_bad_pointer(void);
 
-#define __xchg(x, ptr, size)						\
+#define __arch_xchg(x, ptr, size)					\
 ({									\
 	unsigned long __xchg_result;					\
 									\
@@ -55,7 +55,7 @@ extern void ia64_xchg_called_with_bad_pointer(void);
 
 #ifndef __KERNEL__
 #define xchg(ptr, x)							\
-({(__typeof__(*(ptr))) __xchg((unsigned long) (x), (ptr), sizeof(*(ptr)));})
+({(__typeof__(*(ptr))) __arch_xchg((unsigned long) (x), (ptr), sizeof(*(ptr)));})
 #endif
 
 /*
diff --git a/arch/loongarch/include/asm/cmpxchg.h b/arch/loongarch/include/asm/cmpxchg.h
index ecfa6cf79806e6..979fde61bba8a4 100644
--- a/arch/loongarch/include/asm/cmpxchg.h
+++ b/arch/loongarch/include/asm/cmpxchg.h
@@ -62,7 +62,7 @@ static inline unsigned int __xchg_small(volatile void *ptr, unsigned int val,
 }
 
 static __always_inline unsigned long
-__xchg(volatile void *ptr, unsigned long x, int size)
+__arch_xchg(volatile void *ptr, unsigned long x, int size)
 {
 	switch (size) {
 	case 1:
@@ -87,7 +87,7 @@ __xchg(volatile void *ptr, unsigned long x, int size)
 	__typeof__(*(ptr)) __res;					\
 									\
 	__res = (__typeof__(*(ptr)))					\
-		__xchg((ptr), (unsigned long)(x), sizeof(*(ptr)));	\
+		__arch_xchg((ptr), (unsigned long)(x), sizeof(*(ptr)));	\
 									\
 	__res;								\
 })
diff --git a/arch/m68k/include/asm/cmpxchg.h b/arch/m68k/include/asm/cmpxchg.h
index 6cf464cdab067e..d7f3de9c5d6f79 100644
--- a/arch/m68k/include/asm/cmpxchg.h
+++ b/arch/m68k/include/asm/cmpxchg.h
@@ -9,7 +9,7 @@
 extern unsigned long __invalid_xchg_size(unsigned long, volatile void *, int);
 
 #ifndef CONFIG_RMW_INSNS
-static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int size)
+static inline unsigned long __arch_xchg(unsigned long x, volatile void * ptr, int size)
 {
 	unsigned long flags, tmp;
 
@@ -40,7 +40,7 @@ static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int siz
 	return x;
 }
 #else
-static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int size)
+static inline unsigned long __arch_xchg(unsigned long x, volatile void * ptr, int size)
 {
 	switch (size) {
 	case 1:
@@ -75,7 +75,7 @@ static inline unsigned long __xchg(unsigned long x, volatile void * ptr, int siz
 }
 #endif
 
-#define arch_xchg(ptr,x) ({(__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr)));})
+#define arch_xchg(ptr,x) ({(__typeof__(*(ptr)))__arch_xchg((unsigned long)(x),(ptr),sizeof(*(ptr)));})
 
 #include <asm-generic/cmpxchg-local.h>
 
diff --git a/arch/mips/include/asm/cmpxchg.h b/arch/mips/include/asm/cmpxchg.h
index 7ec9493b28614f..feed343ad483a9 100644
--- a/arch/mips/include/asm/cmpxchg.h
+++ b/arch/mips/include/asm/cmpxchg.h
@@ -68,7 +68,7 @@ extern unsigned long __xchg_small(volatile void *ptr, unsigned long val,
 				  unsigned int size);
 
 static __always_inline
-unsigned long __xchg(volatile void *ptr, unsigned long x, int size)
+unsigned long __arch_xchg(volatile void *ptr, unsigned long x, int size)
 {
 	switch (size) {
 	case 1:
@@ -102,7 +102,7 @@ unsigned long __xchg(volatile void *ptr, unsigned long x, int size)
 		smp_mb__before_llsc();					\
 									\
 	__res = (__typeof__(*(ptr)))					\
-		__xchg((ptr), (unsigned long)(x), sizeof(*(ptr)));	\
+		__arch_xchg((ptr), (unsigned long)(x), sizeof(*(ptr)));	\
 									\
 	smp_llsc_mb();							\
 									\
diff --git a/arch/openrisc/include/asm/cmpxchg.h b/arch/openrisc/include/asm/cmpxchg.h
index 79fd16162ccb6d..8ee151c072e414 100644
--- a/arch/openrisc/include/asm/cmpxchg.h
+++ b/arch/openrisc/include/asm/cmpxchg.h
@@ -147,8 +147,8 @@ static inline unsigned long __cmpxchg(volatile void *ptr, unsigned long old,
 extern unsigned long __xchg_called_with_bad_pointer(void)
 	__compiletime_error("Bad argument size for xchg");
 
-static inline unsigned long __xchg(volatile void *ptr, unsigned long with,
-		int size)
+static inline unsigned long
+__arch_xchg(volatile void *ptr, unsigned long with, int size)
 {
 	switch (size) {
 	case 1:
@@ -163,9 +163,9 @@ static inline unsigned long __xchg(volatile void *ptr, unsigned long with,
 
 #define arch_xchg(ptr, with) 						\
 	({								\
-		(__typeof__(*(ptr))) __xchg((ptr),			\
-					    (unsigned long)(with),	\
-					    sizeof(*(ptr)));		\
+		(__typeof__(*(ptr))) __arch_xchg((ptr),			\
+						 (unsigned long)(with),	\
+						 sizeof(*(ptr)));	\
 	})
 
 #endif /* __ASM_OPENRISC_CMPXCHG_H */
diff --git a/arch/parisc/include/asm/cmpxchg.h b/arch/parisc/include/asm/cmpxchg.h
index 5f274be105671e..c1d776bb16b4ed 100644
--- a/arch/parisc/include/asm/cmpxchg.h
+++ b/arch/parisc/include/asm/cmpxchg.h
@@ -22,7 +22,7 @@ extern unsigned long __xchg64(unsigned long, volatile unsigned long *);
 
 /* optimizer better get rid of switch since size is a constant */
 static inline unsigned long
-__xchg(unsigned long x, volatile void *ptr, int size)
+__arch_xchg(unsigned long x, volatile void *ptr, int size)
 {
 	switch (size) {
 #ifdef CONFIG_64BIT
@@ -49,7 +49,7 @@ __xchg(unsigned long x, volatile void *ptr, int size)
 	__typeof__(*(ptr)) __ret;					\
 	__typeof__(*(ptr)) _x_ = (x);					\
 	__ret = (__typeof__(*(ptr)))					\
-		__xchg((unsigned long)_x_, (ptr), sizeof(*(ptr)));	\
+		__arch_xchg((unsigned long)_x_, (ptr), sizeof(*(ptr)));	\
 	__ret;								\
 })
 
diff --git a/arch/powerpc/include/asm/cmpxchg.h b/arch/powerpc/include/asm/cmpxchg.h
index d0ea0571e79ab5..dbb50c06f0bf4d 100644
--- a/arch/powerpc/include/asm/cmpxchg.h
+++ b/arch/powerpc/include/asm/cmpxchg.h
@@ -229,7 +229,7 @@ __xchg_local(void *ptr, unsigned long x, unsigned int size)
 		return __xchg_u64_local(ptr, x);
 #endif
 	}
-	BUILD_BUG_ON_MSG(1, "Unsupported size for __xchg");
+	BUILD_BUG_ON_MSG(1, "Unsupported size for __xchg_local");
 	return x;
 }
 
@@ -248,7 +248,7 @@ __xchg_relaxed(void *ptr, unsigned long x, unsigned int size)
 		return __xchg_u64_relaxed(ptr, x);
 #endif
 	}
-	BUILD_BUG_ON_MSG(1, "Unsupported size for __xchg_local");
+	BUILD_BUG_ON_MSG(1, "Unsupported size for __xchg_relaxed");
 	return x;
 }
 #define arch_xchg_local(ptr,x)						     \
diff --git a/arch/riscv/include/asm/atomic.h b/arch/riscv/include/asm/atomic.h
index 0dfe9d857a762b..bba472928b5393 100644
--- a/arch/riscv/include/asm/atomic.h
+++ b/arch/riscv/include/asm/atomic.h
@@ -261,7 +261,7 @@ c_t arch_atomic##prefix##_xchg_release(atomic##prefix##_t *v, c_t n)	\
 static __always_inline							\
 c_t arch_atomic##prefix##_xchg(atomic##prefix##_t *v, c_t n)		\
 {									\
-	return __xchg(&(v->counter), n, size);				\
+	return __arch_xchg(&(v->counter), n, size);			\
 }									\
 static __always_inline							\
 c_t arch_atomic##prefix##_cmpxchg_relaxed(atomic##prefix##_t *v,	\
diff --git a/arch/riscv/include/asm/cmpxchg.h b/arch/riscv/include/asm/cmpxchg.h
index 12debce235e52d..2f4726d3cfcc25 100644
--- a/arch/riscv/include/asm/cmpxchg.h
+++ b/arch/riscv/include/asm/cmpxchg.h
@@ -114,7 +114,7 @@
 					    _x_, sizeof(*(ptr)));	\
 })
 
-#define __xchg(ptr, new, size)						\
+#define __arch_xchg(ptr, new, size)					\
 ({									\
 	__typeof__(ptr) __ptr = (ptr);					\
 	__typeof__(new) __new = (new);					\
@@ -143,7 +143,7 @@
 #define arch_xchg(ptr, x)						\
 ({									\
 	__typeof__(*(ptr)) _x_ = (x);					\
-	(__typeof__(*(ptr))) __xchg((ptr), _x_, sizeof(*(ptr)));	\
+	(__typeof__(*(ptr))) __arch_xchg((ptr), _x_, sizeof(*(ptr)));	\
 })
 
 #define xchg32(ptr, x)							\
diff --git a/arch/s390/include/asm/cmpxchg.h b/arch/s390/include/asm/cmpxchg.h
index 84c3f0d576c5b1..ca32deff9880ea 100644
--- a/arch/s390/include/asm/cmpxchg.h
+++ b/arch/s390/include/asm/cmpxchg.h
@@ -14,8 +14,8 @@
 
 void __xchg_called_with_bad_pointer(void);
 
-static __always_inline unsigned long __xchg(unsigned long x,
-					    unsigned long address, int size)
+static __always_inline unsigned long
+__arch_xchg(unsigned long x, unsigned long address, int size)
 {
 	unsigned long old;
 	int shift;
@@ -77,8 +77,8 @@ static __always_inline unsigned long __xchg(unsigned long x,
 	__typeof__(*(ptr)) __ret;					\
 									\
 	__ret = (__typeof__(*(ptr)))					\
-		__xchg((unsigned long)(x), (unsigned long)(ptr),	\
-		       sizeof(*(ptr)));					\
+		__arch_xchg((unsigned long)(x), (unsigned long)(ptr),	\
+			    sizeof(*(ptr)));				\
 	__ret;								\
 })
 
diff --git a/arch/sh/include/asm/cmpxchg.h b/arch/sh/include/asm/cmpxchg.h
index 0ed9b3f4a57796..288f6f38d98fb4 100644
--- a/arch/sh/include/asm/cmpxchg.h
+++ b/arch/sh/include/asm/cmpxchg.h
@@ -22,7 +22,7 @@
 
 extern void __xchg_called_with_bad_pointer(void);
 
-#define __xchg(ptr, x, size)				\
+#define __arch_xchg(ptr, x, size)				\
 ({							\
 	unsigned long __xchg__res;			\
 	volatile void *__xchg_ptr = (ptr);		\
@@ -46,7 +46,7 @@ extern void __xchg_called_with_bad_pointer(void);
 })
 
 #define arch_xchg(ptr,x)	\
-	((__typeof__(*(ptr)))__xchg((ptr),(unsigned long)(x), sizeof(*(ptr))))
+	((__typeof__(*(ptr)))__arch_xchg((ptr),(unsigned long)(x), sizeof(*(ptr))))
 
 /* This function doesn't exist, so you'll get a linker error
  * if something tries to do an invalid cmpxchg(). */
diff --git a/arch/sparc/include/asm/cmpxchg_32.h b/arch/sparc/include/asm/cmpxchg_32.h
index 27a57a3a7597eb..7a1339533d1d7e 100644
--- a/arch/sparc/include/asm/cmpxchg_32.h
+++ b/arch/sparc/include/asm/cmpxchg_32.h
@@ -15,7 +15,7 @@
 unsigned long __xchg_u32(volatile u32 *m, u32 new);
 void __xchg_called_with_bad_pointer(void);
 
-static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr, int size)
+static inline unsigned long __arch_xchg(unsigned long x, __volatile__ void * ptr, int size)
 {
 	switch (size) {
 	case 4:
@@ -25,7 +25,7 @@ static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr, int
 	return x;
 }
 
-#define arch_xchg(ptr,x) ({(__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr)));})
+#define arch_xchg(ptr,x) ({(__typeof__(*(ptr)))__arch_xchg((unsigned long)(x),(ptr),sizeof(*(ptr)));})
 
 /* Emulate cmpxchg() the same way we emulate atomics,
  * by hashing the object address and indexing into an array
diff --git a/arch/sparc/include/asm/cmpxchg_64.h b/arch/sparc/include/asm/cmpxchg_64.h
index 12d00a42c0a3ed..66cd61dde9ec1f 100644
--- a/arch/sparc/include/asm/cmpxchg_64.h
+++ b/arch/sparc/include/asm/cmpxchg_64.h
@@ -55,7 +55,7 @@ static inline unsigned long xchg64(__volatile__ unsigned long *m, unsigned long
 #define arch_xchg(ptr,x)							\
 ({	__typeof__(*(ptr)) __ret;					\
 	__ret = (__typeof__(*(ptr)))					\
-		__xchg((unsigned long)(x), (ptr), sizeof(*(ptr)));	\
+		__arch_xchg((unsigned long)(x), (ptr), sizeof(*(ptr)));	\
 	__ret;								\
 })
 
@@ -87,8 +87,8 @@ xchg16(__volatile__ unsigned short *m, unsigned short val)
 	return (load32 & mask) >> bit_shift;
 }
 
-static inline unsigned long __xchg(unsigned long x, __volatile__ void * ptr,
-				       int size)
+static inline unsigned long
+__arch_xchg(unsigned long x, __volatile__ void * ptr, int size)
 {
 	switch (size) {
 	case 2:
diff --git a/arch/xtensa/include/asm/cmpxchg.h b/arch/xtensa/include/asm/cmpxchg.h
index eb87810357ad88..675a11ea8de76b 100644
--- a/arch/xtensa/include/asm/cmpxchg.h
+++ b/arch/xtensa/include/asm/cmpxchg.h
@@ -170,7 +170,7 @@ static inline unsigned long xchg_u32(volatile int * m, unsigned long val)
 }
 
 #define arch_xchg(ptr,x) \
-	((__typeof__(*(ptr)))__xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
+	((__typeof__(*(ptr)))__arch_xchg((unsigned long)(x),(ptr),sizeof(*(ptr))))
 
 static inline u32 xchg_small(volatile void *ptr, u32 x, int size)
 {
@@ -203,7 +203,7 @@ static inline u32 xchg_small(volatile void *ptr, u32 x, int size)
 extern void __xchg_called_with_bad_pointer(void);
 
 static __inline__ unsigned long
-__xchg(unsigned long x, volatile void * ptr, int size)
+__arch_xchg(unsigned long x, volatile void * ptr, int size)
 {
 	switch (size) {
 	case 1:
-- 
2.34.1

[PATCH v5 2/7] linux/include: add non-atomic version of xchg

[Andrzej Hajda]

The pattern of setting variable with new value and returning old
one is very common in kernel. Usually atomicity of the operation
is not required, so xchg seems to be suboptimal and confusing in
such cases.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
---
 include/linux/non-atomic/xchg.h | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
 create mode 100644 include/linux/non-atomic/xchg.h

diff --git a/include/linux/non-atomic/xchg.h b/include/linux/non-atomic/xchg.h
new file mode 100644
index 00000000000000..f7fa5dd746f37d
--- /dev/null
+++ b/include/linux/non-atomic/xchg.h
@@ -0,0 +1,19 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _LINUX_NON_ATOMIC_XCHG_H
+#define _LINUX_NON_ATOMIC_XCHG_H
+
+/**
+ * __xchg - set variable pointed by @ptr to @val, return old value
+ * @ptr: pointer to affected variable
+ * @val: value to be written
+ *
+ * This is non-atomic variant of xchg.
+ */
+#define __xchg(ptr, val) ({		\
+	__auto_type __ptr = ptr;	\
+	__auto_type __t = *__ptr;	\
+	*__ptr = (val);			\
+	__t;				\
+})
+
+#endif
-- 
2.34.1

[PATCH v5 3/7] arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr

[Andrzej Hajda]

In all architectures, except x86, arch_uretprobe_hijack_return_addr
is just __xchg.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
---
 arch/arm/probes/uprobes/core.c     |  8 ++------
 arch/arm64/kernel/probes/uprobes.c |  9 ++-------
 arch/csky/kernel/probes/uprobes.c  |  9 ++-------
 arch/mips/kernel/uprobes.c         | 10 ++--------
 arch/powerpc/kernel/uprobes.c      | 10 ++--------
 arch/riscv/kernel/probes/uprobes.c |  9 ++-------
 arch/s390/kernel/uprobes.c         |  7 ++-----
 arch/sparc/kernel/uprobes.c        |  7 ++-----
 8 files changed, 16 insertions(+), 53 deletions(-)

diff --git a/arch/arm/probes/uprobes/core.c b/arch/arm/probes/uprobes/core.c
index f5f790c6e5f896..77ce8ae431376d 100644
--- a/arch/arm/probes/uprobes/core.c
+++ b/arch/arm/probes/uprobes/core.c
@@ -9,6 +9,7 @@
 #include <linux/highmem.h>
 #include <linux/sched.h>
 #include <linux/uprobes.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/notifier.h>
 
 #include <asm/opcodes.h>
@@ -61,12 +62,7 @@ unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr,
 				  struct pt_regs *regs)
 {
-	unsigned long orig_ret_vaddr;
-
-	orig_ret_vaddr = regs->ARM_lr;
-	/* Replace the return addr with trampoline addr */
-	regs->ARM_lr = trampoline_vaddr;
-	return orig_ret_vaddr;
+	return __xchg(&regs->ARM_lr, trampoline_vaddr);
 }
 
 int arch_uprobe_analyze_insn(struct arch_uprobe *auprobe, struct mm_struct *mm,
diff --git a/arch/arm64/kernel/probes/uprobes.c b/arch/arm64/kernel/probes/uprobes.c
index d49aef2657cdf7..d7171d30ea6681 100644
--- a/arch/arm64/kernel/probes/uprobes.c
+++ b/arch/arm64/kernel/probes/uprobes.c
@@ -3,6 +3,7 @@
  * Copyright (C) 2014-2016 Pratyush Anand <panand@redhat.com>
  */
 #include <linux/highmem.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/ptrace.h>
 #include <linux/uprobes.h>
 #include <asm/cacheflush.h>
@@ -150,13 +151,7 @@ unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr,
 				  struct pt_regs *regs)
 {
-	unsigned long orig_ret_vaddr;
-
-	orig_ret_vaddr = procedure_link_pointer(regs);
-	/* Replace the return addr with trampoline addr */
-	procedure_link_pointer_set(regs, trampoline_vaddr);
-
-	return orig_ret_vaddr;
+	return __xchg(&procedure_link_pointer(regs), trampoline_vaddr);
 }
 
 int arch_uprobe_exception_notify(struct notifier_block *self,
diff --git a/arch/csky/kernel/probes/uprobes.c b/arch/csky/kernel/probes/uprobes.c
index 2d31a12e46cfee..775fe88b5f0016 100644
--- a/arch/csky/kernel/probes/uprobes.c
+++ b/arch/csky/kernel/probes/uprobes.c
@@ -3,6 +3,7 @@
  * Copyright (C) 2014-2016 Pratyush Anand <panand@redhat.com>
  */
 #include <linux/highmem.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/ptrace.h>
 #include <linux/uprobes.h>
 #include <asm/cacheflush.h>
@@ -123,13 +124,7 @@ unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr,
 				  struct pt_regs *regs)
 {
-	unsigned long ra;
-
-	ra = regs->lr;
-
-	regs->lr = trampoline_vaddr;
-
-	return ra;
+	return __xchg(&regs->lr, trampoline_vaddr);
 }
 
 int arch_uprobe_exception_notify(struct notifier_block *self,
diff --git a/arch/mips/kernel/uprobes.c b/arch/mips/kernel/uprobes.c
index 6c063aa188e626..2b1f375c407b87 100644
--- a/arch/mips/kernel/uprobes.c
+++ b/arch/mips/kernel/uprobes.c
@@ -2,6 +2,7 @@
 #include <linux/highmem.h>
 #include <linux/kdebug.h>
 #include <linux/types.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/notifier.h>
 #include <linux/sched.h>
 #include <linux/uprobes.h>
@@ -197,14 +198,7 @@ void arch_uprobe_abort_xol(struct arch_uprobe *aup,
 unsigned long arch_uretprobe_hijack_return_addr(
 	unsigned long trampoline_vaddr, struct pt_regs *regs)
 {
-	unsigned long ra;
-
-	ra = regs->regs[31];
-
-	/* Replace the return address with the trampoline address */
-	regs->regs[31] = trampoline_vaddr;
-
-	return ra;
+	return __xchg(&regs->regs[31], trampoline_vaddr);
 }
 
 /**
diff --git a/arch/powerpc/kernel/uprobes.c b/arch/powerpc/kernel/uprobes.c
index 95a41ae9dfa755..3c15c320315e6c 100644
--- a/arch/powerpc/kernel/uprobes.c
+++ b/arch/powerpc/kernel/uprobes.c
@@ -7,6 +7,7 @@
  * Adapted from the x86 port by Ananth N Mavinakayanahalli <ananth@in.ibm.com>
  */
 #include <linux/kernel.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/sched.h>
 #include <linux/ptrace.h>
 #include <linux/uprobes.h>
@@ -197,14 +198,7 @@ bool arch_uprobe_skip_sstep(struct arch_uprobe *auprobe, struct pt_regs *regs)
 unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr, struct pt_regs *regs)
 {
-	unsigned long orig_ret_vaddr;
-
-	orig_ret_vaddr = regs->link;
-
-	/* Replace the return addr with trampoline addr */
-	regs->link = trampoline_vaddr;
-
-	return orig_ret_vaddr;
+	return __xchg(&regs->link, trampoline_vaddr);
 }
 
 bool arch_uretprobe_is_alive(struct return_instance *ret, enum rp_check ctx,
diff --git a/arch/riscv/kernel/probes/uprobes.c b/arch/riscv/kernel/probes/uprobes.c
index c976a21cd4bd5b..5c8415e216536d 100644
--- a/arch/riscv/kernel/probes/uprobes.c
+++ b/arch/riscv/kernel/probes/uprobes.c
@@ -1,6 +1,7 @@
 // SPDX-License-Identifier: GPL-2.0-only
 
 #include <linux/highmem.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/ptrace.h>
 #include <linux/uprobes.h>
 
@@ -122,13 +123,7 @@ unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr,
 				  struct pt_regs *regs)
 {
-	unsigned long ra;
-
-	ra = regs->ra;
-
-	regs->ra = trampoline_vaddr;
-
-	return ra;
+	return __xchg(&regs->ra, trampoline_vaddr);
 }
 
 int arch_uprobe_exception_notify(struct notifier_block *self,
diff --git a/arch/s390/kernel/uprobes.c b/arch/s390/kernel/uprobes.c
index b88345ef8bd9ed..18591ca40ae7e5 100644
--- a/arch/s390/kernel/uprobes.c
+++ b/arch/s390/kernel/uprobes.c
@@ -11,6 +11,7 @@
 #include <linux/compat.h>
 #include <linux/kdebug.h>
 #include <linux/sched/task_stack.h>
+#include <linux/non-atomic/xchg.h>
 
 #include <asm/switch_to.h>
 #include <asm/facility.h>
@@ -144,11 +145,7 @@ void arch_uprobe_abort_xol(struct arch_uprobe *auprobe, struct pt_regs *regs)
 unsigned long arch_uretprobe_hijack_return_addr(unsigned long trampoline,
 						struct pt_regs *regs)
 {
-	unsigned long orig;
-
-	orig = regs->gprs[14];
-	regs->gprs[14] = trampoline;
-	return orig;
+	return __xchg(&regs->gprs[14], trampoline);
 }
 
 bool arch_uretprobe_is_alive(struct return_instance *ret, enum rp_check ctx,
diff --git a/arch/sparc/kernel/uprobes.c b/arch/sparc/kernel/uprobes.c
index 1a0600206bf5c2..1c02653d6de3ba 100644
--- a/arch/sparc/kernel/uprobes.c
+++ b/arch/sparc/kernel/uprobes.c
@@ -11,6 +11,7 @@
 
 #include <linux/kernel.h>
 #include <linux/highmem.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/uprobes.h>
 #include <linux/uaccess.h>
 #include <linux/sched.h> /* For struct task_struct */
@@ -310,9 +311,5 @@ unsigned long
 arch_uretprobe_hijack_return_addr(unsigned long trampoline_vaddr,
 				  struct pt_regs *regs)
 {
-	unsigned long orig_ret_vaddr = regs->u_regs[UREG_I7];
-
-	regs->u_regs[UREG_I7] = trampoline_vaddr-8;
-
-	return orig_ret_vaddr + 8;
+	return __xchg(&regs->u_regs[UREG_I7], trampoline_vaddr - 8) + 8;
 }
-- 
2.34.1

[PATCH v5 4/7] llist: simplify __llist_del_all

[Andrzej Hajda]

llist_del_all uses xchg, let's use __xchg here.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
---
 include/linux/llist.h | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/include/linux/llist.h b/include/linux/llist.h
index 85bda2d02d65be..4dc1d185ea98ab 100644
--- a/include/linux/llist.h
+++ b/include/linux/llist.h
@@ -50,6 +50,7 @@
 
 #include <linux/atomic.h>
 #include <linux/container_of.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/stddef.h>
 #include <linux/types.h>
 
@@ -241,10 +242,7 @@ static inline struct llist_node *llist_del_all(struct llist_head *head)
 
 static inline struct llist_node *__llist_del_all(struct llist_head *head)
 {
-	struct llist_node *first = head->first;
-
-	head->first = NULL;
-	return first;
+	return __xchg(&head->first, NULL);
 }
 
 extern struct llist_node *llist_del_first(struct llist_head *head);
-- 
2.34.1

[PATCH v5 5/7] io_uring: use __xchg if possible

[Andrzej Hajda]

Recently introduced helper simplifies the code.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
---
 io_uring/io_uring.c | 7 ++-----
 io_uring/slist.h    | 6 ++----
 2 files changed, 4 insertions(+), 9 deletions(-)

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 2ac1cd8d23ea62..2b46a692d69022 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -54,6 +54,7 @@
 #include <linux/fdtable.h>
 #include <linux/mm.h>
 #include <linux/mman.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/percpu.h>
 #include <linux/slab.h>
 #include <linux/bvec.h>
@@ -1095,8 +1096,6 @@ static void __io_req_find_next_prep(struct io_kiocb *req)
 
 static inline struct io_kiocb *io_req_find_next(struct io_kiocb *req)
 {
-	struct io_kiocb *nxt;
-
 	/*
 	 * If LINK is set, we have dependent requests in this chain. If we
 	 * didn't fail this request, queue the first one up, moving any other
@@ -1105,9 +1104,7 @@ static inline struct io_kiocb *io_req_find_next(struct io_kiocb *req)
 	 */
 	if (unlikely(req->flags & IO_DISARM_MASK))
 		__io_req_find_next_prep(req);
-	nxt = req->link;
-	req->link = NULL;
-	return nxt;
+	return __xchg(&req->link, NULL);
 }
 
 static void ctx_flush_and_put(struct io_ring_ctx *ctx, bool *locked)
diff --git a/io_uring/slist.h b/io_uring/slist.h
index f27601fa46607b..d3a743a1adc626 100644
--- a/io_uring/slist.h
+++ b/io_uring/slist.h
@@ -2,6 +2,7 @@
 #define INTERNAL_IO_SLIST_H
 
 #include <linux/io_uring_types.h>
+#include <linux/non-atomic/xchg.h>
 
 #define wq_list_for_each(pos, prv, head)			\
 	for (pos = (head)->first, prv = NULL; pos; prv = pos, pos = (pos)->next)
@@ -121,10 +122,7 @@ static inline void wq_list_del(struct io_wq_work_list *list,
 static inline
 struct io_wq_work_node *wq_stack_extract(struct io_wq_work_node *stack)
 {
-	struct io_wq_work_node *node = stack->next;
-
-	stack->next = node->next;
-	return node;
+	return __xchg(&stack->next, stack->next->next);
 }
 
 static inline struct io_wq_work *wq_next_work(struct io_wq_work *work)
-- 
2.34.1

[PATCH v5 6/7] qed: use __xchg if possible

[Andrzej Hajda]

Recently introduced helper simplifies the code.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
---
 include/linux/qed/qed_chain.h | 19 +++++++------------
 1 file changed, 7 insertions(+), 12 deletions(-)

diff --git a/include/linux/qed/qed_chain.h b/include/linux/qed/qed_chain.h
index a84063492c71ae..6355d558cf01b2 100644
--- a/include/linux/qed/qed_chain.h
+++ b/include/linux/qed/qed_chain.h
@@ -11,6 +11,7 @@
 #include <asm/byteorder.h>
 #include <linux/kernel.h>
 #include <linux/list.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/sizes.h>
 #include <linux/slab.h>
 #include <linux/qed/common_hsi.h>
@@ -368,7 +369,7 @@ static inline void qed_chain_return_produced(struct qed_chain *p_chain)
  */
 static inline void *qed_chain_produce(struct qed_chain *p_chain)
 {
-	void *p_ret = NULL, *p_prod_idx, *p_prod_page_idx;
+	void *p_prod_idx, *p_prod_page_idx;
 
 	if (is_chain_u16(p_chain)) {
 		if ((p_chain->u.chain16.prod_idx &
@@ -390,11 +391,8 @@ static inline void *qed_chain_produce(struct qed_chain *p_chain)
 		p_chain->u.chain32.prod_idx++;
 	}
 
-	p_ret = p_chain->p_prod_elem;
-	p_chain->p_prod_elem = (void *)(((u8 *)p_chain->p_prod_elem) +
-					p_chain->elem_size);
-
-	return p_ret;
+	return __xchg(&p_chain->p_prod_elem,
+		      (u8 *)p_chain->p_prod_elem + p_chain->elem_size);
 }
 
 /**
@@ -439,7 +437,7 @@ static inline void qed_chain_recycle_consumed(struct qed_chain *p_chain)
  */
 static inline void *qed_chain_consume(struct qed_chain *p_chain)
 {
-	void *p_ret = NULL, *p_cons_idx, *p_cons_page_idx;
+	void *p_cons_idx, *p_cons_page_idx;
 
 	if (is_chain_u16(p_chain)) {
 		if ((p_chain->u.chain16.cons_idx &
@@ -461,11 +459,8 @@ static inline void *qed_chain_consume(struct qed_chain *p_chain)
 		p_chain->u.chain32.cons_idx++;
 	}
 
-	p_ret = p_chain->p_cons_elem;
-	p_chain->p_cons_elem = (void *)(((u8 *)p_chain->p_cons_elem) +
-					p_chain->elem_size);
-
-	return p_ret;
+	return __xchg(&p_chain->p_cons_elem,
+		      (u8 *)p_chain->p_cons_elem + p_chain->elem_size);
 }
 
 /**
-- 
2.34.1

[PATCH v5 7/7] drm/i915/gt: use __xchg instead of internal helper

[Andrzej Hajda]

Prefer core helper if available.

Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
---
 drivers/gpu/drm/i915/gt/intel_engine_cs.c            | 2 +-
 drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c     | 4 ++--
 drivers/gpu/drm/i915/gt/intel_execlists_submission.c | 4 ++--
 drivers/gpu/drm/i915/gt/intel_ggtt.c                 | 4 ++--
 drivers/gpu/drm/i915/gt/intel_gsc.c                  | 2 +-
 drivers/gpu/drm/i915/gt/intel_gt.c                   | 4 ++--
 drivers/gpu/drm/i915/gt/intel_gt_pm.c                | 2 +-
 drivers/gpu/drm/i915/gt/intel_lrc.c                  | 6 +++---
 drivers/gpu/drm/i915/gt/intel_migrate.c              | 2 +-
 drivers/gpu/drm/i915/gt/intel_rc6.c                  | 2 +-
 drivers/gpu/drm/i915/gt/intel_rps.c                  | 2 +-
 drivers/gpu/drm/i915/gt/selftest_context.c           | 2 +-
 drivers/gpu/drm/i915/gt/selftest_ring_submission.c   | 2 +-
 drivers/gpu/drm/i915/gt/selftest_timeline.c          | 2 +-
 drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c            | 2 +-
 drivers/gpu/drm/i915/gt/uc/intel_uc.c                | 2 +-
 drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c             | 2 +-
 drivers/gpu/drm/i915/i915_utils.h                    | 1 +
 18 files changed, 24 insertions(+), 23 deletions(-)

diff --git a/drivers/gpu/drm/i915/gt/intel_engine_cs.c b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
index 922f1bb22dc685..9712bfc2c6523d 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_cs.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_cs.c
@@ -1042,7 +1042,7 @@ static void cleanup_status_page(struct intel_engine_cs *engine)
 	/* Prevent writes into HWSP after returning the page to the system */
 	intel_engine_set_hwsp_writemask(engine, ~0u);
 
-	vma = fetch_and_zero(&engine->status_page.vma);
+	vma = __xchg(&engine->status_page.vma, 0);
 	if (!vma)
 		return;
 
diff --git a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
index 9a527e1f5be655..09befcc6a84fa1 100644
--- a/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
+++ b/drivers/gpu/drm/i915/gt/intel_engine_heartbeat.c
@@ -229,7 +229,7 @@ static void heartbeat(struct work_struct *wrk)
 	mutex_unlock(&ce->timeline->mutex);
 out:
 	if (!engine->i915->params.enable_hangcheck || !next_heartbeat(engine))
-		i915_request_put(fetch_and_zero(&engine->heartbeat.systole));
+		i915_request_put(__xchg(&engine->heartbeat.systole, 0));
 	intel_engine_pm_put(engine);
 }
 
@@ -244,7 +244,7 @@ void intel_engine_unpark_heartbeat(struct intel_engine_cs *engine)
 void intel_engine_park_heartbeat(struct intel_engine_cs *engine)
 {
 	if (cancel_delayed_work(&engine->heartbeat.work))
-		i915_request_put(fetch_and_zero(&engine->heartbeat.systole));
+		i915_request_put(__xchg(&engine->heartbeat.systole, 0));
 }
 
 void intel_gt_unpark_heartbeats(struct intel_gt *gt)
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index 18ffe55282e594..5c985e6fa1be2f 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -3199,7 +3199,7 @@ static void execlists_reset_cancel(struct intel_engine_cs *engine)
 		RB_CLEAR_NODE(rb);
 
 		spin_lock(&ve->base.sched_engine->lock);
-		rq = fetch_and_zero(&ve->request);
+		rq = __xchg(&ve->request, NULL);
 		if (rq) {
 			if (i915_request_mark_eio(rq)) {
 				rq->engine = engine;
@@ -3604,7 +3604,7 @@ static void rcu_virtual_context_destroy(struct work_struct *wrk)
 
 		spin_lock_irq(&ve->base.sched_engine->lock);
 
-		old = fetch_and_zero(&ve->request);
+		old = __xchg(&ve->request, NULL);
 		if (old) {
 			GEM_BUG_ON(!__i915_request_is_complete(old));
 			__i915_request_submit(old);
diff --git a/drivers/gpu/drm/i915/gt/intel_ggtt.c b/drivers/gpu/drm/i915/gt/intel_ggtt.c
index fe64c13fd3b4aa..6f441c3d3d1cef 100644
--- a/drivers/gpu/drm/i915/gt/intel_ggtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_ggtt.c
@@ -684,7 +684,7 @@ static void fini_aliasing_ppgtt(struct i915_ggtt *ggtt)
 {
 	struct i915_ppgtt *ppgtt;
 
-	ppgtt = fetch_and_zero(&ggtt->alias);
+	ppgtt = __xchg(&ggtt->alias, NULL);
 	if (!ppgtt)
 		return;
 
@@ -1238,7 +1238,7 @@ bool i915_ggtt_resume_vm(struct i915_address_space *vm)
 				   was_bound);
 
 		if (obj) { /* only used during resume => exclusive access */
-			write_domain_objs |= fetch_and_zero(&obj->write_domain);
+			write_domain_objs |= __xchg(&obj->write_domain, 0);
 			obj->read_domains |= I915_GEM_DOMAIN_GTT;
 		}
 	}
diff --git a/drivers/gpu/drm/i915/gt/intel_gsc.c b/drivers/gpu/drm/i915/gt/intel_gsc.c
index bcc3605158dbde..38fbea757ba741 100644
--- a/drivers/gpu/drm/i915/gt/intel_gsc.c
+++ b/drivers/gpu/drm/i915/gt/intel_gsc.c
@@ -70,7 +70,7 @@ gsc_ext_om_alloc(struct intel_gsc *gsc, struct intel_gsc_intf *intf, size_t size
 
 static void gsc_ext_om_destroy(struct intel_gsc_intf *intf)
 {
-	struct drm_i915_gem_object *obj = fetch_and_zero(&intf->gem_obj);
+	struct drm_i915_gem_object *obj = __xchg(&intf->gem_obj, 0);
 
 	if (!obj)
 		return;
diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c
index f0dbfc434e0773..884458507969e6 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt.c
@@ -753,7 +753,7 @@ int intel_gt_init(struct intel_gt *gt)
 	intel_uc_fini(&gt->uc);
 err_engines:
 	intel_engines_release(gt);
-	i915_vm_put(fetch_and_zero(&gt->vm));
+	i915_vm_put(__xchg(&gt->vm, 0));
 err_pm:
 	intel_gt_pm_fini(gt);
 	intel_gt_fini_scratch(gt);
@@ -800,7 +800,7 @@ void intel_gt_driver_release(struct intel_gt *gt)
 {
 	struct i915_address_space *vm;
 
-	vm = fetch_and_zero(&gt->vm);
+	vm = __xchg(&gt->vm, 0);
 	if (vm) /* FIXME being called twice on error paths :( */
 		i915_vm_put(vm);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_gt_pm.c b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
index cef3d6f5c34e01..2527c5ae72e59b 100644
--- a/drivers/gpu/drm/i915/gt/intel_gt_pm.c
+++ b/drivers/gpu/drm/i915/gt/intel_gt_pm.c
@@ -124,7 +124,7 @@ static int __gt_unpark(struct intel_wakeref *wf)
 static int __gt_park(struct intel_wakeref *wf)
 {
 	struct intel_gt *gt = container_of(wf, typeof(*gt), wakeref);
-	intel_wakeref_t wakeref = fetch_and_zero(&gt->awake);
+	intel_wakeref_t wakeref = __xchg(&gt->awake, 0);
 	struct drm_i915_private *i915 = gt->i915;
 
 	GT_TRACE(gt, "\n");
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c
index 7771a19008c604..9dfa3c10ddc85f 100644
--- a/drivers/gpu/drm/i915/gt/intel_lrc.c
+++ b/drivers/gpu/drm/i915/gt/intel_lrc.c
@@ -1144,7 +1144,7 @@ __lrc_alloc_state(struct intel_context *ce, struct intel_engine_cs *engine)
 static struct intel_timeline *
 pinned_timeline(struct intel_context *ce, struct intel_engine_cs *engine)
 {
-	struct intel_timeline *tl = fetch_and_zero(&ce->timeline);
+	struct intel_timeline *tl = __xchg(&ce->timeline, 0);
 
 	return intel_timeline_create_from_engine(engine, page_unmask_bits(tl));
 }
@@ -1261,8 +1261,8 @@ void lrc_fini(struct intel_context *ce)
 	if (!ce->state)
 		return;
 
-	intel_ring_put(fetch_and_zero(&ce->ring));
-	i915_vma_put(fetch_and_zero(&ce->state));
+	intel_ring_put(__xchg(&ce->ring, 0));
+	i915_vma_put(__xchg(&ce->state, 0));
 }
 
 void lrc_destroy(struct kref *kref)
diff --git a/drivers/gpu/drm/i915/gt/intel_migrate.c b/drivers/gpu/drm/i915/gt/intel_migrate.c
index 3f638f19879685..3eab1867a4abee 100644
--- a/drivers/gpu/drm/i915/gt/intel_migrate.c
+++ b/drivers/gpu/drm/i915/gt/intel_migrate.c
@@ -1147,7 +1147,7 @@ void intel_migrate_fini(struct intel_migrate *m)
 {
 	struct intel_context *ce;
 
-	ce = fetch_and_zero(&m->context);
+	ce = __xchg(&m->context, 0);
 	if (!ce)
 		return;
 
diff --git a/drivers/gpu/drm/i915/gt/intel_rc6.c b/drivers/gpu/drm/i915/gt/intel_rc6.c
index 5c91622dfca420..ca6b0c905accb3 100644
--- a/drivers/gpu/drm/i915/gt/intel_rc6.c
+++ b/drivers/gpu/drm/i915/gt/intel_rc6.c
@@ -702,7 +702,7 @@ void intel_rc6_fini(struct intel_rc6 *rc6)
 
 	intel_rc6_disable(rc6);
 
-	pctx = fetch_and_zero(&rc6->pctx);
+	pctx = __xchg(&rc6->pctx, 0);
 	if (pctx)
 		i915_gem_object_put(pctx);
 
diff --git a/drivers/gpu/drm/i915/gt/intel_rps.c b/drivers/gpu/drm/i915/gt/intel_rps.c
index 9ad3bc7201cbaa..e34ca33b09d2e7 100644
--- a/drivers/gpu/drm/i915/gt/intel_rps.c
+++ b/drivers/gpu/drm/i915/gt/intel_rps.c
@@ -1831,7 +1831,7 @@ static void rps_work(struct work_struct *work)
 	u32 pm_iir = 0;
 
 	spin_lock_irq(gt->irq_lock);
-	pm_iir = fetch_and_zero(&rps->pm_iir) & rps->pm_events;
+	pm_iir = __xchg(&rps->pm_iir, 0) & rps->pm_events;
 	client_boost = atomic_read(&rps->num_waiters);
 	spin_unlock_irq(gt->irq_lock);
 
diff --git a/drivers/gpu/drm/i915/gt/selftest_context.c b/drivers/gpu/drm/i915/gt/selftest_context.c
index 76fbae358072df..3f49ca1debc6ce 100644
--- a/drivers/gpu/drm/i915/gt/selftest_context.c
+++ b/drivers/gpu/drm/i915/gt/selftest_context.c
@@ -171,7 +171,7 @@ static int live_context_size(void *arg)
 		 * active state is sufficient, we are only checking that we
 		 * don't use more than we planned.
 		 */
-		saved = fetch_and_zero(&engine->default_state);
+		saved = __xchg(&engine->default_state, 0);
 
 		/* Overlaps with the execlists redzone */
 		engine->context_size += I915_GTT_PAGE_SIZE;
diff --git a/drivers/gpu/drm/i915/gt/selftest_ring_submission.c b/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
index 87ceb0f374b673..a01aaca4fbf5ff 100644
--- a/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/selftest_ring_submission.c
@@ -269,7 +269,7 @@ static int live_ctx_switch_wa(void *arg)
 		if (IS_GRAPHICS_VER(gt->i915, 4, 5))
 			continue; /* MI_STORE_DWORD is privileged! */
 
-		saved_wa = fetch_and_zero(&engine->wa_ctx.vma);
+		saved_wa = __xchg(&engine->wa_ctx.vma, 0);
 
 		intel_engine_pm_get(engine);
 		err = __live_ctx_switch_wa(engine);
diff --git a/drivers/gpu/drm/i915/gt/selftest_timeline.c b/drivers/gpu/drm/i915/gt/selftest_timeline.c
index 522d0190509ccc..d14d5159024ec7 100644
--- a/drivers/gpu/drm/i915/gt/selftest_timeline.c
+++ b/drivers/gpu/drm/i915/gt/selftest_timeline.c
@@ -892,7 +892,7 @@ static int create_watcher(struct hwsp_watcher *w,
 static int check_watcher(struct hwsp_watcher *w, const char *name,
 			 bool (*op)(u32 hwsp, u32 seqno))
 {
-	struct i915_request *rq = fetch_and_zero(&w->rq);
+	struct i915_request *rq = __xchg(&w->rq, NULL);
 	u32 offset, end;
 	int err;
 
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c
index fd21dbd2663bec..3f85d3f6fc6e92 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c
@@ -110,7 +110,7 @@ void intel_gsc_uc_fini(struct intel_gsc_uc *gsc)
 	flush_work(&gsc->work);
 
 	if (gsc->ce)
-		intel_engine_destroy_pinned_context(fetch_and_zero(&gsc->ce));
+		intel_engine_destroy_pinned_context(__xchg(&gsc->ce, NULL));
 
 	i915_vma_unpin_and_release(&gsc->local, 0);
 
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
index 9a8a1abf71d7fe..0292212cffbcb9 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c
@@ -169,7 +169,7 @@ static void __uc_capture_load_err_log(struct intel_uc *uc)
 
 static void __uc_free_load_err_log(struct intel_uc *uc)
 {
-	struct drm_i915_gem_object *log = fetch_and_zero(&uc->load_err_log);
+	struct drm_i915_gem_object *log = __xchg(&uc->load_err_log, NULL);
 
 	if (log)
 		i915_gem_object_put(log);
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
index 65672ff8260540..3f684f34469581 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c
@@ -1119,7 +1119,7 @@ void intel_uc_fw_cleanup_fetch(struct intel_uc_fw *uc_fw)
 	if (!intel_uc_fw_is_available(uc_fw))
 		return;
 
-	i915_gem_object_put(fetch_and_zero(&uc_fw->obj));
+	i915_gem_object_put(__xchg(&uc_fw->obj, NULL));
 
 	intel_uc_fw_change_status(uc_fw, INTEL_UC_FIRMWARE_SELECTED);
 }
diff --git a/drivers/gpu/drm/i915/i915_utils.h b/drivers/gpu/drm/i915/i915_utils.h
index 2c430c0c3badde..be7df2c384c832 100644
--- a/drivers/gpu/drm/i915/i915_utils.h
+++ b/drivers/gpu/drm/i915/i915_utils.h
@@ -26,6 +26,7 @@
 #define __I915_UTILS_H
 
 #include <linux/list.h>
+#include <linux/non-atomic/xchg.h>
 #include <linux/overflow.h>
 #include <linux/sched.h>
 #include <linux/string_helpers.h>
-- 
2.34.1

Re: [Intel-gfx] [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Andrzej Hajda]

Hi,

Ping on the series.
Arnd, Andrew is there anything more I can do to push the process forward?

Regards
Andrzej


On 18.01.2023 16:35, Andrzej Hajda wrote:
> Hi all,
> 
> The helper is tiny and there are advices we can live without it, so
> I want to present few arguments why it would be good to have it:
> 
> 1. Code readability/simplification/number of lines:
>    - decreases number of lines,
>    - it often eliminates local variables,
>    - for real examples see patches 3+.
> 
> 2. Presence of similar helpers in other somehow related languages/libs:
> 
> a) Rust[1]: 'replace' from std::mem module, there is also 'take'
>      helper (__xchg(&x, 0)), which is the same as private helper in
>      i915 - fetch_and_zero, see latest patch.
> b) C++ [2]: 'exchange' from utility header.
> 
> If the idea is OK there are still 2 questions to answer:
> 
> 1. Name of the helper, __xchg follows kernel conventions,
>      but for me Rust names are also OK.
> 2. Where to put the helper:
> a) as in this patchset include/linux/non-atomic/xchg.h,
>      proposed by Andy Shevchenko,
> b) include/linux/utils.h ? any better name? Some kind
>      of container for simple helpers.
> 
> All __xchg conversions were performed using cocci script,
> then manually adjusted if necessary.
> 
> There is lot of places it can be used in, I have just chosen
> some of them. I can provide cocci script to detect others (not all),
> if necessary.
> 
> Changes:
> v2: squashed all __xchg -> __arch_xchg t one patch (Arnd)
> v3: fixed alpha/xchg_local (lkp@intel.com)
> v4: adjusted indentation (Heiko)
> v5: added more __xchg conversions - patches 3-6, added tags
> 
> [1]: https://doc.rust-lang.org/std/mem/index.html
> [2]: https://en.cppreference.com/w/cpp/header/utility
> 
> Regards
> Andrzej
> 
> Andrzej Hajda (7):
>    arch: rename all internal names __xchg to __arch_xchg
>    linux/include: add non-atomic version of xchg
>    arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
>    llist: simplify __llist_del_all
>    io_uring: use __xchg if possible
>    qed: use __xchg if possible
>    drm/i915/gt: use __xchg instead of internal helper
> 
>   arch/alpha/include/asm/cmpxchg.h              | 10 +++++-----
>   arch/arc/include/asm/cmpxchg.h                |  4 ++--
>   arch/arm/include/asm/cmpxchg.h                |  7 ++++---
>   arch/arm/probes/uprobes/core.c                |  8 ++------
>   arch/arm64/include/asm/cmpxchg.h              |  7 +++----
>   arch/arm64/kernel/probes/uprobes.c            |  9 ++-------
>   arch/csky/kernel/probes/uprobes.c             |  9 ++-------
>   arch/hexagon/include/asm/cmpxchg.h            | 10 +++++-----
>   arch/ia64/include/asm/cmpxchg.h               |  2 +-
>   arch/ia64/include/uapi/asm/cmpxchg.h          |  4 ++--
>   arch/loongarch/include/asm/cmpxchg.h          |  4 ++--
>   arch/m68k/include/asm/cmpxchg.h               |  6 +++---
>   arch/mips/include/asm/cmpxchg.h               |  4 ++--
>   arch/mips/kernel/uprobes.c                    | 10 ++--------
>   arch/openrisc/include/asm/cmpxchg.h           | 10 +++++-----
>   arch/parisc/include/asm/cmpxchg.h             |  4 ++--
>   arch/powerpc/include/asm/cmpxchg.h            |  4 ++--
>   arch/powerpc/kernel/uprobes.c                 | 10 ++--------
>   arch/riscv/include/asm/atomic.h               |  2 +-
>   arch/riscv/include/asm/cmpxchg.h              |  4 ++--
>   arch/riscv/kernel/probes/uprobes.c            |  9 ++-------
>   arch/s390/include/asm/cmpxchg.h               |  8 ++++----
>   arch/s390/kernel/uprobes.c                    |  7 ++-----
>   arch/sh/include/asm/cmpxchg.h                 |  4 ++--
>   arch/sparc/include/asm/cmpxchg_32.h           |  4 ++--
>   arch/sparc/include/asm/cmpxchg_64.h           |  6 +++---
>   arch/sparc/kernel/uprobes.c                   |  7 ++-----
>   arch/xtensa/include/asm/cmpxchg.h             |  4 ++--
>   drivers/gpu/drm/i915/gt/intel_engine_cs.c     |  2 +-
>   .../gpu/drm/i915/gt/intel_engine_heartbeat.c  |  4 ++--
>   .../drm/i915/gt/intel_execlists_submission.c  |  4 ++--
>   drivers/gpu/drm/i915/gt/intel_ggtt.c          |  4 ++--
>   drivers/gpu/drm/i915/gt/intel_gsc.c           |  2 +-
>   drivers/gpu/drm/i915/gt/intel_gt.c            |  4 ++--
>   drivers/gpu/drm/i915/gt/intel_gt_pm.c         |  2 +-
>   drivers/gpu/drm/i915/gt/intel_lrc.c           |  6 +++---
>   drivers/gpu/drm/i915/gt/intel_migrate.c       |  2 +-
>   drivers/gpu/drm/i915/gt/intel_rc6.c           |  2 +-
>   drivers/gpu/drm/i915/gt/intel_rps.c           |  2 +-
>   drivers/gpu/drm/i915/gt/selftest_context.c    |  2 +-
>   .../drm/i915/gt/selftest_ring_submission.c    |  2 +-
>   drivers/gpu/drm/i915/gt/selftest_timeline.c   |  2 +-
>   drivers/gpu/drm/i915/gt/uc/intel_gsc_uc.c     |  2 +-
>   drivers/gpu/drm/i915/gt/uc/intel_uc.c         |  2 +-
>   drivers/gpu/drm/i915/gt/uc/intel_uc_fw.c      |  2 +-
>   drivers/gpu/drm/i915/i915_utils.h             |  1 +
>   include/linux/llist.h                         |  6 ++----
>   include/linux/non-atomic/xchg.h               | 19 +++++++++++++++++++
>   include/linux/qed/qed_chain.h                 | 19 +++++++------------
>   io_uring/io_uring.c                           |  7 ++-----
>   io_uring/slist.h                              |  6 ++----
>   51 files changed, 126 insertions(+), 155 deletions(-)
>   create mode 100644 include/linux/non-atomic/xchg.h
> 

Re: [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Peter Zijlstra]

On Wed, Jan 18, 2023 at 04:35:22PM +0100, Andrzej Hajda wrote:

> Andrzej Hajda (7):
>   arch: rename all internal names __xchg to __arch_xchg
>   linux/include: add non-atomic version of xchg
>   arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
>   llist: simplify __llist_del_all
>   io_uring: use __xchg if possible
>   qed: use __xchg if possible
>   drm/i915/gt: use __xchg instead of internal helper

Nothing crazy in here I suppose, I somewhat wonder why you went through
the trouble, but meh.

You want me to take this through te locking tree (for the next cycle,
not this one) where I normally take atomic things or does someone else
want this?

Re: [Intel-gfx] [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Andrzej Hajda]

On 22.02.2023 18:04, Peter Zijlstra wrote:
> On Wed, Jan 18, 2023 at 04:35:22PM +0100, Andrzej Hajda wrote:
> 
>> Andrzej Hajda (7):
>>    arch: rename all internal names __xchg to __arch_xchg
>>    linux/include: add non-atomic version of xchg
>>    arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
>>    llist: simplify __llist_del_all
>>    io_uring: use __xchg if possible
>>    qed: use __xchg if possible
>>    drm/i915/gt: use __xchg instead of internal helper
> 
> Nothing crazy in here I suppose, I somewhat wonder why you went through
> the trouble, but meh.

If you are asking why I have proposed this patchset, then the answer is 
simple, 1st I've tried to find a way to move internal i915 helper to 
core (see patch 7).
Then I was looking for possible other users of this helper. And 
apparently there are many of them, patches 3-7 shows some.


> 
> You want me to take this through te locking tree (for the next cycle,
> not this one) where I normally take atomic things or does someone else
> want this?

If you could take it I will be happy.

Regards
Andrzej

Re: [Intel-gfx] [PATCH v5 1/7] arch: rename all internal names __xchg to __arch_xchg

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:44PM +0100, Andrzej Hajda wrote:
> __xchg will be used for non-atomic xchg macro.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
> Reviewed-by: Arnd Bergmann <arnd@arndb.de>
> Acked-by: Geert Uytterhoeven <geert@linux-m68k.org> [m68k]
> Acked-by: Palmer Dabbelt <palmer@rivosinc.com> [riscv]

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 2/7] linux/include: add non-atomic version of xchg

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:45PM +0100, Andrzej Hajda wrote:
> The pattern of setting variable with new value and returning old
> one is very common in kernel. Usually atomicity of the operation
> is not required, so xchg seems to be suboptimal and confusing in
> such cases.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>
> Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 3/7] arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:46PM +0100, Andrzej Hajda wrote:
> In all architectures, except x86, arch_uretprobe_hijack_return_addr
> is just __xchg.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 4/7] llist: simplify __llist_del_all

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:47PM +0100, Andrzej Hajda wrote:
> llist_del_all uses xchg, let's use __xchg here.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 6/7] qed: use __xchg if possible

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:49PM +0100, Andrzej Hajda wrote:
> Recently introduced helper simplifies the code.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 7/7] drm/i915/gt: use __xchg instead of internal helper

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:50PM +0100, Andrzej Hajda wrote:
> Prefer core helper if available.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 5/7] io_uring: use __xchg if possible

[Andi Shyti]

Hi Andrzej,

On Wed, Jan 18, 2023 at 04:44:48PM +0100, Andrzej Hajda wrote:
> Recently introduced helper simplifies the code.
> 
> Signed-off-by: Andrzej Hajda <andrzej.hajda@intel.com>

Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>

Thanks,
Andi

Re: [Intel-gfx] [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Peter Zijlstra]

On Thu, Feb 23, 2023 at 10:24:19PM +0100, Andrzej Hajda wrote:
> On 22.02.2023 18:04, Peter Zijlstra wrote:
> > On Wed, Jan 18, 2023 at 04:35:22PM +0100, Andrzej Hajda wrote:
> > 
> > > Andrzej Hajda (7):
> > >    arch: rename all internal names __xchg to __arch_xchg
> > >    linux/include: add non-atomic version of xchg
> > >    arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
> > >    llist: simplify __llist_del_all
> > >    io_uring: use __xchg if possible
> > >    qed: use __xchg if possible
> > >    drm/i915/gt: use __xchg instead of internal helper
> > 
> > Nothing crazy in here I suppose, I somewhat wonder why you went through
> > the trouble, but meh.
> 
> If you are asking why I have proposed this patchset, then the answer is
> simple, 1st I've tried to find a way to move internal i915 helper to core
> (see patch 7).
> Then I was looking for possible other users of this helper. And apparently
> there are many of them, patches 3-7 shows some.
> 
> 
> > 
> > You want me to take this through te locking tree (for the next cycle,
> > not this one) where I normally take atomic things or does someone else
> > want this?
> 
> If you could take it I will be happy.

OK, I'll go queue it in tip/locking/core after -rc1. Thanks!

Re: [Intel-gfx] [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Jani Nikula]

On Mon, 27 Feb 2023, Peter Zijlstra <peterz@infradead.org> wrote:
> On Thu, Feb 23, 2023 at 10:24:19PM +0100, Andrzej Hajda wrote:
>> On 22.02.2023 18:04, Peter Zijlstra wrote:
>> > On Wed, Jan 18, 2023 at 04:35:22PM +0100, Andrzej Hajda wrote:
>> > 
>> > > Andrzej Hajda (7):
>> > >    arch: rename all internal names __xchg to __arch_xchg
>> > >    linux/include: add non-atomic version of xchg
>> > >    arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
>> > >    llist: simplify __llist_del_all
>> > >    io_uring: use __xchg if possible
>> > >    qed: use __xchg if possible
>> > >    drm/i915/gt: use __xchg instead of internal helper
>> > 
>> > Nothing crazy in here I suppose, I somewhat wonder why you went through
>> > the trouble, but meh.
>> 
>> If you are asking why I have proposed this patchset, then the answer is
>> simple, 1st I've tried to find a way to move internal i915 helper to core
>> (see patch 7).
>> Then I was looking for possible other users of this helper. And apparently
>> there are many of them, patches 3-7 shows some.
>> 
>> 
>> > 
>> > You want me to take this through te locking tree (for the next cycle,
>> > not this one) where I normally take atomic things or does someone else
>> > want this?
>> 
>> If you could take it I will be happy.
>
> OK, I'll go queue it in tip/locking/core after -rc1. Thanks!

Is this where the series fell between the cracks, or was there some
follow-up that I missed?

I think this would still be useful. Andrzej, would you mind rebasing and
resending if there are no objections?

BR,
Jani.


-- 
Jani Nikula, Intel

Re: [Intel-gfx] [PATCH v5 0/7] Introduce __xchg, non-atomic xchg

[Andrzej Hajda]

On 05.04.2024 16:47, Jani Nikula wrote:
> On Mon, 27 Feb 2023, Peter Zijlstra <peterz@infradead.org> wrote:
>> On Thu, Feb 23, 2023 at 10:24:19PM +0100, Andrzej Hajda wrote:
>>> On 22.02.2023 18:04, Peter Zijlstra wrote:
>>>> On Wed, Jan 18, 2023 at 04:35:22PM +0100, Andrzej Hajda wrote:
>>>>
>>>>> Andrzej Hajda (7):
>>>>>     arch: rename all internal names __xchg to __arch_xchg
>>>>>     linux/include: add non-atomic version of xchg
>>>>>     arch/*/uprobes: simplify arch_uretprobe_hijack_return_addr
>>>>>     llist: simplify __llist_del_all
>>>>>     io_uring: use __xchg if possible
>>>>>     qed: use __xchg if possible
>>>>>     drm/i915/gt: use __xchg instead of internal helper
>>>> Nothing crazy in here I suppose, I somewhat wonder why you went through
>>>> the trouble, but meh.
>>> If you are asking why I have proposed this patchset, then the answer is
>>> simple, 1st I've tried to find a way to move internal i915 helper to core
>>> (see patch 7).
>>> Then I was looking for possible other users of this helper. And apparently
>>> there are many of them, patches 3-7 shows some.
>>>
>>>
>>>> You want me to take this through te locking tree (for the next cycle,
>>>> not this one) where I normally take atomic things or does someone else
>>>> want this?
>>> If you could take it I will be happy.
>> OK, I'll go queue it in tip/locking/core after -rc1. Thanks!
> Is this where the series fell between the cracks, or was there some
> follow-up that I missed?
>
> I think this would still be useful. Andrzej, would you mind rebasing and
> resending if there are no objections?

The patchset was rejected/dropped by Linus at the pull-request stage.
He didn't like many things, but the most __xchg name. However he was 
quite positive about i915 name fetch_and_zero.
I can try to revive patchset with fetch_and_zero, and maybe 
fetch_and_set, instead of __xchg.

Regards
Andrzej

>
> BR,
> Jani.
>
>