[PATCH v2 00/12] s390: More this_cpu_*() changes

[PATCH v2 00/12] s390: More this_cpu_*() changes

[Heiko Carstens]

v2:
- Yet another brown paper bug: Re-add CPU migration check to last
  patch. Unconditionally recalculating and updating the percpu variable
  address and percpu offset register can corrupt previous context
  register state in several cases (not only the single case reported by
  Sashiko).

v1:
Most of this is only about cleaning up the percpu code after preemptible
this_cpu_*() operations have been implemented. The first ten patches are
all more or less trivial cleanup patches trying to make the code shorter
and more readable.

The only non-trivial patch is the last one, which converts s390's
this_cpu_*() operations to use a similar scheme like Mark Rutland
provided it for arm64 [1]. This allows to simplify the irq entry and exit
path, however at the cost of slightly worse code for this_cpu_*()
operations.

The simplified irq entry and exit code seems to be worth it. Usable
performance numbers are not available yet, however I don't expect big
difference to before.

[1] https://lore.kernel.org/all/20260904161758.376504-1-mark.rutland@arm.com/

Thanks,
Heiko

Heiko Carstens (12):
  s390/percpu: Fix comment typo
  s390/percpu: Add sanity check to GEN_MVIY macro
  s390/lowcore: Remove _AC() from LOWCORE_ALT_ADDRESS
  s390/percpu: Let MVIY_PERCPU() calculate alternative displacement
  s390/percpu/lowcore: Add and use LC_PERCPU lowcore offset defines
  s390/percpu: Rename inline assembly symbolic names
  s390/percpu: Use __PCPU_BEGIN() and __PCPU_END() for inline assemblies
  s390/percpu: Use percpu code section for this_cpu_cmpxchg128()
  s390/percpu: Use percpu code section for this_cpu_xchg()
  s390/percpu: Use percpu code section for this_cpu_cmpxchg()
  s390: Add CC_HAS_ASM_M_FORMAT_FLAG config option
  s390/percpu: Rework to simplify percpu_entry() and percpu_exit()

 arch/s390/Kconfig                    |   6 +
 arch/s390/include/asm/entry-percpu.h |  71 ++---
 arch/s390/include/asm/lowcore.h      |   5 +-
 arch/s390/include/asm/percpu.h       | 381 +++++++++++++++++----------
 arch/s390/include/asm/ptrace.h       |   2 +-
 arch/s390/kernel/irq.c               |  10 +-
 arch/s390/kernel/nmi.c               |   4 +-
 arch/s390/kernel/traps.c             |   4 +-
 8 files changed, 280 insertions(+), 203 deletions(-)

-- 
2.53.0

[PATCH v2 01/12] s390/percpu: Fix comment typo

[Heiko Carstens]

This is all preempt_disable() / preempt_enable() pairs and not
preempt_disable() / preempt_disable() pairs.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index feb58acf87d6..31012abe8292 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -63,7 +63,7 @@
 /*
  * Macros to be used for percpu code section based on atomic instructions.
  *
- * Avoid the need to use preempt_disable() / preempt_disable() pairs and the
+ * Avoid the need to use preempt_disable() / preempt_enable() pairs and the
  * conditional preempt_schedule_notrace() function calls which come with
  * this. The idea is that this_cpu operations based on atomic instructions are
  * guarded with mviy instructions:
-- 
2.53.0

[PATCH v2 02/12] s390/percpu: Add sanity check to GEN_MVIY macro

[Heiko Carstens]

Emit an error if the GEN_MVIY macro fails to parse the passed register
string. This will make it more obvious if something similar which was
fixed with [1] happens again.

[1] commit 101782f8945a ("s390/percpu: Fix MVIY_PERCPU() with older binutils")

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 31012abe8292..9888b38dedab 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -100,12 +100,17 @@
  */
 
 #define MVIY_PERCPU(disp, dispalt, reg)						\
-	".macro GEN_MVIY disp reg\n"						\
+	".macro GEN_MVIY disp, reg\n"						\
+	".set	.Lreg,255\n"							\
 	".irp	rs,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15\n"			\
 	"	.ifc \\reg,%%r\\rs\n"						\
-	"	mviy	\\disp(%%r0),\\rs\n"					\
+	"		.set	.Lreg,\\rs\n"					\
 	"	.endif\n"							\
 	".endr\n"								\
+	".if .Lreg == 255\n"							\
+	"	.error \"Illegal register number\"\n"				\
+	".endif\n"								\
+	"mviy	\\disp(%%r0),.Lreg\n"						\
 	".endm\n"								\
 	ALTERNATIVE("GEN_MVIY " disp    ", " reg "\n",				\
 		    "GEN_MVIY " dispalt ", " reg "\n",				\
-- 
2.53.0

[PATCH v2 03/12] s390/lowcore: Remove _AC() from LOWCORE_ALT_ADDRESS

[Heiko Carstens]

Remove the _AC() conversion from LOWCORE_ALT_ADDRESS. This changes the
type of LOWCORE_ALT_ADDRESS from unsigned long to int in C code, however
looking at the current users this should not cause any changes.

This allows to use __stringify(LOWCORE_ALT_ADDRESS) for inline
assemblies, which otherwise wouldn't work because of the UL suffix.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/lowcore.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/s390/include/asm/lowcore.h b/arch/s390/include/asm/lowcore.h
index 5cef215d30e7..f0364a9f8a19 100644
--- a/arch/s390/include/asm/lowcore.h
+++ b/arch/s390/include/asm/lowcore.h
@@ -21,7 +21,7 @@
 #define LC_ORDER 1
 #define LC_PAGES 2
 
-#define LOWCORE_ALT_ADDRESS	_AC(0x70000, UL)
+#define LOWCORE_ALT_ADDRESS	0x70000
 
 #ifndef __ASSEMBLER__
 
-- 
2.53.0

[PATCH v2 04/12] s390/percpu: Let MVIY_PERCPU() calculate alternative displacement

[Heiko Carstens]

Let MVIY_PERCPU() calculate the alternative lowcore displacement instead
of passing the same value(s) again and again in all percpu inline
assemblies.

This shortens the inline assemblies and makes them a bit more readable.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 68 +++++++++++++++-------------------
 1 file changed, 29 insertions(+), 39 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 9888b38dedab..3bc448c4c2dd 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -99,7 +99,9 @@
  *   MVIY_ALT(...)    <- end of percpu code section
  */
 
-#define MVIY_PERCPU(disp, dispalt, reg)						\
+#define LC_ALT_ADDR	__stringify(LOWCORE_ALT_ADDRESS)
+
+#define MVIY_PERCPU(disp, reg)							\
 	".macro GEN_MVIY disp, reg\n"						\
 	".set	.Lreg,255\n"							\
 	".irp	rs,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15\n"			\
@@ -112,19 +114,19 @@
 	".endif\n"								\
 	"mviy	\\disp(%%r0),.Lreg\n"						\
 	".endm\n"								\
-	ALTERNATIVE("GEN_MVIY " disp    ", " reg "\n",				\
-		    "GEN_MVIY " dispalt ", " reg "\n",				\
+	ALTERNATIVE("GEN_MVIY " disp		     " , " reg "\n",		\
+		    "GEN_MVIY " disp "+" LC_ALT_ADDR " , " reg "\n",		\
 		    ALT_FEATURE(MFEATURE_LOWCORE))				\
 	".purgem GEN_MVIY\n"
 
-#define MVIY_ALT(disp, dispalt)							\
-	ALTERNATIVE("	mviy	" disp	  "(%%r0),0\n",				\
-		    "	mviy	" dispalt "(%%r0),0\n",				\
+#define MVIY_ALT(disp)								\
+	ALTERNATIVE("	mviy	" disp		       "(%%r0),0\n",		\
+		    "	mviy	" disp "+" LC_ALT_ADDR "(%%r0),0\n",		\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
-#define AG_ALT(disp, dispalt, reg)						\
-	ALTERNATIVE("	ag	" reg ", " disp	   "(%%r0)\n",			\
-		    "	ag	" reg ", " dispalt "(%%r0)\n",			\
+#define AG_ALT(disp, reg)							\
+	ALTERNATIVE("	ag	" reg ", " disp			"(%%r0)\n",	\
+		    "	ag	" reg ", " disp "+" LC_ALT_ADDR "(%%r0)\n",	\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
 #ifndef MARCH_HAS_Z196_FEATURES
@@ -153,33 +155,29 @@ do {										\
 	if (__builtin_constant_p(val__) &&					\
 	    ((szcast)val__ > -129) && ((szcast)val__ < 128)) {			\
 		asm volatile(							\
-			MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-			AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+			MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
+			AG_ALT("%[disppcpo]", "%[ptr__]")			\
 			op2 "   0(%[ptr__]),%[val__]\n"				\
-			MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+			MVIY_ALT("%[disppcpr]")					\
 			: [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "i" ((szcast)val__),				\
 			  [disppcpr] "i" (lc_pcpr),				\
 			  [disppcpo] "i" (lc_pcpo),				\
-			  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-			  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	} else {								\
 		asm volatile(							\
-			MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-			AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+			MVIY_PERCPU("%[disppcpr]", "%[ptr__]")			\
+			AG_ALT("%[disppcpo]", "%[ptr__]")			\
 			op1 "   %[old__],%[val__],0(%[ptr__])\n"		\
-			MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+			MVIY_ALT("%[disppcpr]")					\
 			: [old__] "=&d" (old__),				\
 			  [ptr__] "+&a" (ptr__),  "+m" (*ptr__),		\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "d" (val__),					\
 			  [disppcpr] "i" (lc_pcpr),				\
 			  [disppcpo] "i" (lc_pcpo),				\
-			  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-			  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	}									\
@@ -199,18 +197,16 @@ do {										\
 	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-		AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
+		AG_ALT("%[disppcpo]","%[ptr__]")			\
 		op "	%[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+		MVIY_ALT("%[disppcpr]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
 		  [disppcpr] "i" (lc_pcpr),				\
 		  [disppcpo] "i" (lc_pcpo),				\
-		  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-		  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 	old__ + val__;							\
@@ -230,18 +226,16 @@ do {									\
 	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-		AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
+		AG_ALT("%[disppcpo]","%[ptr__]")			\
 		op "    %[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+		MVIY_ALT("%[disppcpr]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
 		  [disppcpr] "i" (lc_pcpr),				\
 		  [disppcpo] "i" (lc_pcpo),				\
-		  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-		  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
@@ -263,16 +257,14 @@ do {									\
 	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-		AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
+		AG_ALT("%[disppcpo]","%[ptr__]")			\
 		op "	%[res__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+		MVIY_ALT("%[disppcpr]")					\
 		: [res__] "=&d" (res__), [ptr__] "+&a" (ptr__),		\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [disppcpr] "i" (lc_pcpr),				\
 		  [disppcpo] "i" (lc_pcpo),				\
-		  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-		  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 		  "m" (*ptr__),						\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
@@ -294,17 +286,15 @@ do {									\
 	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]", "%[dispaltpcpr]", "%[ptr__]")\
-		AG_ALT("%[disppcpo]", "%[dispaltpcpo]", "%[ptr__]")	\
+		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
+		AG_ALT("%[disppcpo]","%[ptr__]")			\
 		op "    %[val__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[disppcpr]", "%[dispaltpcpr]")		\
+		MVIY_ALT("%[disppcpr]")					\
 		: [ptr__] "+&a" (ptr__), "=m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
 		  [disppcpr] "i" (lc_pcpr),				\
 		  [disppcpo] "i" (lc_pcpo),				\
-		  [dispaltpcpr] "i" (lc_pcpr + LOWCORE_ALT_ADDRESS),	\
-		  [dispaltpcpo] "i" (lc_pcpo + LOWCORE_ALT_ADDRESS),	\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
-- 
2.53.0

[PATCH v2 05/12] s390/percpu/lowcore: Add and use LC_PERCPU lowcore offset defines

[Heiko Carstens]

Define LC_PERCPU_OFFSET and LC_PERCPU_REGISTER offset defines and use
them in percpu code. This shortens the percpu inline assemblies and makes
them a bit more readable.

Note that because of header dependencies it is not possible to use
asm-offsets. However to avoid adding more dependencies to asm-offsets
it seems to be preferable to use the current approach in any way.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/entry-percpu.h |  2 +-
 arch/s390/include/asm/lowcore.h      |  3 ++
 arch/s390/include/asm/percpu.h       | 47 ++++++++++------------------
 3 files changed, 20 insertions(+), 32 deletions(-)

diff --git a/arch/s390/include/asm/entry-percpu.h b/arch/s390/include/asm/entry-percpu.h
index 4ef47265cfce..89c9e44f6f83 100644
--- a/arch/s390/include/asm/entry-percpu.h
+++ b/arch/s390/include/asm/entry-percpu.h
@@ -49,7 +49,7 @@ static __always_inline bool percpu_code_check(struct pt_regs *regs)
 	}
 	if ((insn & 0xff0f) != 0xe300)
 		return true;
-	disp = offsetof(struct lowcore, percpu_offset);
+	disp = LC_PERCPU_OFFSET;
 	if (machine_has_relocated_lowcore())
 		disp += LOWCORE_ALT_ADDRESS;
 	insn = (disp & 0xff000) >> 4 | (disp & 0x00fff) << 16 | 0x8;
diff --git a/arch/s390/include/asm/lowcore.h b/arch/s390/include/asm/lowcore.h
index f0364a9f8a19..f4517bbe405b 100644
--- a/arch/s390/include/asm/lowcore.h
+++ b/arch/s390/include/asm/lowcore.h
@@ -222,6 +222,9 @@ struct lowcore {
 	__u8	pad_0x1900[0x2000-0x1900];	/* 0x1900 */
 } __packed __aligned(8192);
 
+#define LC_PERCPU_OFFSET	offsetof(struct lowcore, percpu_offset)
+#define LC_PERCPU_REGISTER	offsetof(struct lowcore, percpu_register)
+
 static __always_inline struct lowcore *get_lowcore(void)
 {
 	struct lowcore *lc;
diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 3bc448c4c2dd..770fe4be65dc 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -14,17 +14,16 @@
 
 #define arch_raw_cpu_ptr(_ptr)						\
 ({									\
-	unsigned long lc_percpu, tcp_ptr__;				\
+	unsigned long tcp_ptr__;					\
 									\
 	tcp_ptr__ = (__force unsigned long)(_ptr);			\
-	lc_percpu = offsetof(struct lowcore, percpu_offset);		\
 	asm_inline volatile(						\
 	ALTERNATIVE("ag		%[__ptr__],%[offzero](%%r0)\n",		\
 		    "ag		%[__ptr__],%[offalt](%%r0)\n",		\
 		    ALT_FEATURE(MFEATURE_LOWCORE))			\
 	: [__ptr__] "+d" (tcp_ptr__)					\
-	: [offzero] "i" (lc_percpu),					\
-	  [offalt] "i" (lc_percpu + LOWCORE_ALT_ADDRESS),		\
+	: [offzero] "i" (LC_PERCPU_OFFSET),				\
+	  [offalt] "i" (LC_PERCPU_OFFSET + LOWCORE_ALT_ADDRESS),	\
 	  "m" (((struct lowcore *)0)->percpu_offset)			\
 	: "cc");							\
 	(TYPEOF_UNQUAL(*(_ptr)) __force __kernel *)tcp_ptr__;		\
@@ -144,13 +143,10 @@
 
 #define arch_this_cpu_add(pcp, val, op1, op2, szcast)				\
 do {										\
-	unsigned long lc_pcpr, lc_pcpo;						\
 	typedef typeof(pcp) pcp_op_T__;						\
 	pcp_op_T__ val__ = (val);						\
 	pcp_op_T__ old__, *ptr__;						\
 										\
-	lc_pcpr = offsetof(struct lowcore, percpu_register);			\
-	lc_pcpo = offsetof(struct lowcore, percpu_offset);			\
 	ptr__ = PERCPU_PTR(&(pcp));						\
 	if (__builtin_constant_p(val__) &&					\
 	    ((szcast)val__ > -129) && ((szcast)val__ < 128)) {			\
@@ -162,8 +158,8 @@ do {										\
 			: [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "i" ((szcast)val__),				\
-			  [disppcpr] "i" (lc_pcpr),				\
-			  [disppcpo] "i" (lc_pcpo),				\
+			  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+			  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	} else {								\
@@ -176,8 +172,8 @@ do {										\
 			  [ptr__] "+&a" (ptr__),  "+m" (*ptr__),		\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "d" (val__),					\
-			  [disppcpr] "i" (lc_pcpr),				\
-			  [disppcpo] "i" (lc_pcpo),				\
+			  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+			  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	}									\
@@ -188,13 +184,10 @@ do {										\
 
 #define arch_this_cpu_add_return(pcp, val, op)				\
 ({									\
-	unsigned long lc_pcpr, lc_pcpo;					\
 	typedef typeof(pcp) pcp_op_T__; 				\
 	pcp_op_T__ val__ = (val);					\
 	pcp_op_T__ old__, *ptr__;					\
 									\
-	lc_pcpr = offsetof(struct lowcore, percpu_register);		\
-	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
 		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
@@ -205,8 +198,8 @@ do {										\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (lc_pcpr),				\
-		  [disppcpo] "i" (lc_pcpo),				\
+		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 	old__ + val__;							\
@@ -217,13 +210,10 @@ do {										\
 
 #define arch_this_cpu_to_op(pcp, val, op)				\
 do {									\
-	unsigned long lc_pcpr, lc_pcpo;					\
 	typedef typeof(pcp) pcp_op_T__; 				\
 	pcp_op_T__ val__ = (val);					\
 	pcp_op_T__ old__, *ptr__;					\
 									\
-	lc_pcpr = offsetof(struct lowcore, percpu_register);		\
-	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
 		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
@@ -234,8 +224,8 @@ do {									\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (lc_pcpr),				\
-		  [disppcpo] "i" (lc_pcpo),				\
+		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
@@ -249,12 +239,10 @@ do {									\
 
 #define arch_this_cpu_read(pcp, op)					\
 ({									\
-	unsigned long lc_pcpr, lc_pcpo, res__;				\
 	typedef typeof(pcp) pcp_op_T__;					\
+	unsigned long res__;						\
 	pcp_op_T__ *ptr__;						\
 									\
-	lc_pcpr = offsetof(struct lowcore, percpu_register);		\
-	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
 		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
@@ -263,8 +251,8 @@ do {									\
 		MVIY_ALT("%[disppcpr]")					\
 		: [res__] "=&d" (res__), [ptr__] "+&a" (ptr__),		\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [disppcpr] "i" (lc_pcpr),				\
-		  [disppcpo] "i" (lc_pcpo),				\
+		: [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (*ptr__),						\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
@@ -278,12 +266,9 @@ do {									\
 
 #define arch_this_cpu_write(pcp, val, op)				\
 do {									\
-	unsigned long lc_pcpr, lc_pcpo;					\
 	typedef typeof(pcp) pcp_op_T__;					\
 	pcp_op_T__ *ptr__, val__ = (val);				\
 									\
-	lc_pcpr = offsetof(struct lowcore, percpu_register);		\
-	lc_pcpo = offsetof(struct lowcore, percpu_offset);		\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
 		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
@@ -293,8 +278,8 @@ do {									\
 		: [ptr__] "+&a" (ptr__), "=m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (lc_pcpr),				\
-		  [disppcpo] "i" (lc_pcpo),				\
+		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
+		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
-- 
2.53.0

[PATCH v2 06/12] s390/percpu: Rename inline assembly symbolic names

[Heiko Carstens]

Rename perpcu inline assembly symbolic names so they match (more) what is
specified with input operands. A trivial change which is supposed to make
the code a bit more readable.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 60 +++++++++++++++++-----------------
 1 file changed, 30 insertions(+), 30 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 770fe4be65dc..0df47ec2e6c7 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -151,29 +151,29 @@ do {										\
 	if (__builtin_constant_p(val__) &&					\
 	    ((szcast)val__ > -129) && ((szcast)val__ < 128)) {			\
 		asm volatile(							\
-			MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
-			AG_ALT("%[disppcpo]", "%[ptr__]")			\
+			MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
+			AG_ALT("%[lcoff]", "%[ptr__]")				\
 			op2 "   0(%[ptr__]),%[val__]\n"				\
-			MVIY_ALT("%[disppcpr]")					\
+			MVIY_ALT("%[lcreg]")					\
 			: [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "i" ((szcast)val__),				\
-			  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-			  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+			  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+			  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	} else {								\
 		asm volatile(							\
-			MVIY_PERCPU("%[disppcpr]", "%[ptr__]")			\
-			AG_ALT("%[disppcpo]", "%[ptr__]")			\
+			MVIY_PERCPU("%[lcreg]", "%[ptr__]")			\
+			AG_ALT("%[lcoff]", "%[ptr__]")				\
 			op1 "   %[old__],%[val__],0(%[ptr__])\n"		\
-			MVIY_ALT("%[disppcpr]")					\
+			MVIY_ALT("%[lcreg]")					\
 			: [old__] "=&d" (old__),				\
 			  [ptr__] "+&a" (ptr__),  "+m" (*ptr__),		\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "d" (val__),					\
-			  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-			  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+			  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+			  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	}									\
@@ -190,16 +190,16 @@ do {										\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
-		AG_ALT("%[disppcpo]","%[ptr__]")			\
+		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
+		AG_ALT("%[lcoff]","%[ptr__]")				\
 		op "	%[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[disppcpr]")					\
+		MVIY_ALT("%[lcreg]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 	old__ + val__;							\
@@ -216,16 +216,16 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
-		AG_ALT("%[disppcpo]","%[ptr__]")			\
+		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
+		AG_ALT("%[lcoff]","%[ptr__]")				\
 		op "    %[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[disppcpr]")					\
+		MVIY_ALT("%[lcreg]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
@@ -245,14 +245,14 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
-		AG_ALT("%[disppcpo]","%[ptr__]")			\
+		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
+		AG_ALT("%[lcoff]","%[ptr__]")				\
 		op "	%[res__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[disppcpr]")					\
+		MVIY_ALT("%[lcreg]")					\
 		: [res__] "=&d" (res__), [ptr__] "+&a" (ptr__),		\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+		: [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (*ptr__),						\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
@@ -271,15 +271,15 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[disppcpr]","%[ptr__]")			\
-		AG_ALT("%[disppcpo]","%[ptr__]")			\
+		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
+		AG_ALT("%[lcoff]","%[ptr__]")				\
 		op "    %[val__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[disppcpr]")					\
+		MVIY_ALT("%[lcreg]")					\
 		: [ptr__] "+&a" (ptr__), "=m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-		  [disppcpr] "i" (LC_PERCPU_REGISTER),			\
-		  [disppcpo] "i" (LC_PERCPU_OFFSET),			\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
 		: "cc");						\
 } while (0)
-- 
2.53.0

[PATCH v2 07/12] s390/percpu: Use __PCPU_BEGIN() and __PCPU_END() for inline assemblies

[Heiko Carstens]

Replace the use of MVIY_PERCPU(), AG_ALT(), and MVIY_ALT() within all
percpu inline assemblies with __PCPU_BEGIN() and __PCPU_END(), since
the begin and end of each inline assembly is the same.

Also use consistent names: e.g. __PCPU_MVIY() and __PCPU_AG() instead of
the previous MVIY_PERCPU() and AG_ALT().

Furthermore extract DEFINE_GR_NUM and UNDEF_GR_NUM which are used to
convert register strings to register numbers.

All of this should make the code more readable. No functional change.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 83 +++++++++++++++++-----------------
 1 file changed, 42 insertions(+), 41 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 0df47ec2e6c7..78c3aaa138f7 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -92,42 +92,49 @@
  *
  * Inline assemblies making use of this typically have a code sequence like:
  *
- *   MVIY_PERCPU(...) <- start of percpu code section
- *   AG_ALT(...)      <- add percpu offset; must be the second instruction
- *   atomic_op	      <- atomic op
- *   MVIY_ALT(...)    <- end of percpu code section
+ *   __PCPU_BEGIN(...) <- start of percpu code section
+ *   atomic_op	       <- atomic op
+ *   __PCPU_END(...)   <- end of percpu code section
  */
 
 #define LC_ALT_ADDR	__stringify(LOWCORE_ALT_ADDRESS)
 
-#define MVIY_PERCPU(disp, reg)							\
-	".macro GEN_MVIY disp, reg\n"						\
-	".set	.Lreg,255\n"							\
-	".irp	rs,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15\n"			\
-	"	.ifc \\reg,%%r\\rs\n"						\
-	"		.set	.Lreg,\\rs\n"					\
+#define DEFINE_GR_NUM								\
+	".macro _GR_NUM opd, gr\n"						\
+	".set	\\opd,255\n"							\
+	".irp rs,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15\n"			\
+	"	.ifc \\gr,%%r\\rs\n"						\
+	"		\\opd = \\rs\n"						\
 	"	.endif\n"							\
 	".endr\n"								\
-	".if .Lreg == 255\n"							\
+	".if \\opd == 255\n"							\
 	"	.error \"Illegal register number\"\n"				\
 	".endif\n"								\
-	"mviy	\\disp(%%r0),.Lreg\n"						\
-	".endm\n"								\
-	ALTERNATIVE("GEN_MVIY " disp		     " , " reg "\n",		\
-		    "GEN_MVIY " disp "+" LC_ALT_ADDR " , " reg "\n",		\
-		    ALT_FEATURE(MFEATURE_LOWCORE))				\
-	".purgem GEN_MVIY\n"
+	".endm\n"
 
-#define MVIY_ALT(disp)								\
-	ALTERNATIVE("	mviy	" disp		       "(%%r0),0\n",		\
-		    "	mviy	" disp "+" LC_ALT_ADDR "(%%r0),0\n",		\
+#define UNDEF_GR_NUM								\
+	".purgem _GR_NUM\n"
+
+#define __PCPU_MVIY(lcreg, imm)							\
+	ALTERNATIVE("	mviy	" lcreg			"(%%r0)," imm "\n",	\
+		    "	mviy	" lcreg "+" LC_ALT_ADDR "(%%r0)," imm "\n",	\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
-#define AG_ALT(disp, reg)							\
-	ALTERNATIVE("	ag	" reg ", " disp			"(%%r0)\n",	\
-		    "	ag	" reg ", " disp "+" LC_ALT_ADDR "(%%r0)\n",	\
+#define __PCPU_AG(reg, lcoff)							\
+	ALTERNATIVE("	ag	" reg ", " lcoff		 "(%%r0)\n",	\
+		    "	ag	" reg ", " lcoff "+" LC_ALT_ADDR "(%%r0)\n",	\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
+#define __PCPU_BEGIN(lcreg, lcoff, reg)						\
+	DEFINE_GR_NUM								\
+	"_GR_NUM .Lreg, " reg "\n"						\
+	UNDEF_GR_NUM								\
+	__PCPU_MVIY(lcreg, ".Lreg")						\
+	__PCPU_AG(reg, lcoff)
+
+#define __PCPU_END(lcreg)							\
+	__PCPU_MVIY(lcreg, "0")
+
 #ifndef MARCH_HAS_Z196_FEATURES
 
 #define this_cpu_add_4(pcp, val)	arch_this_cpu_to_op_simple(pcp, val, +)
@@ -151,10 +158,9 @@ do {										\
 	if (__builtin_constant_p(val__) &&					\
 	    ((szcast)val__ > -129) && ((szcast)val__ < 128)) {			\
 		asm volatile(							\
-			MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
-			AG_ALT("%[lcoff]", "%[ptr__]")				\
+			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 			op2 "   0(%[ptr__]),%[val__]\n"				\
-			MVIY_ALT("%[lcreg]")					\
+			__PCPU_END("%[lcreg]")					\
 			: [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
 			: [val__] "i" ((szcast)val__),				\
@@ -164,10 +170,9 @@ do {										\
 			: "cc");						\
 	} else {								\
 		asm volatile(							\
-			MVIY_PERCPU("%[lcreg]", "%[ptr__]")			\
-			AG_ALT("%[lcoff]", "%[ptr__]")				\
+			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 			op1 "   %[old__],%[val__],0(%[ptr__])\n"		\
-			MVIY_ALT("%[lcreg]")					\
+			__PCPU_END("%[lcreg]")					\
 			: [old__] "=&d" (old__),				\
 			  [ptr__] "+&a" (ptr__),  "+m" (*ptr__),		\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
@@ -190,10 +195,9 @@ do {										\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
-		AG_ALT("%[lcoff]","%[ptr__]")				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 		op "	%[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[lcreg]")					\
+		__PCPU_END("%[lcreg]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
@@ -216,10 +220,9 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
-		AG_ALT("%[lcoff]","%[ptr__]")				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 		op "    %[old__],%[val__],0(%[ptr__])\n"		\
-		MVIY_ALT("%[lcreg]")					\
+		__PCPU_END("%[lcreg]")					\
 		: [old__] "=&d" (old__),				\
 		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
@@ -245,10 +248,9 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
-		AG_ALT("%[lcoff]","%[ptr__]")				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 		op "	%[res__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[lcreg]")					\
+		__PCPU_END("%[lcreg]")					\
 		: [res__] "=&d" (res__), [ptr__] "+&a" (ptr__),		\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [lcreg] "i" (LC_PERCPU_REGISTER),			\
@@ -271,10 +273,9 @@ do {									\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		MVIY_PERCPU("%[lcreg]","%[ptr__]")			\
-		AG_ALT("%[lcoff]","%[ptr__]")				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
 		op "    %[val__],0(%[ptr__])\n"				\
-		MVIY_ALT("%[lcreg]")					\
+		__PCPU_END("%[lcreg]")					\
 		: [ptr__] "+&a" (ptr__), "=m" (*ptr__),			\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
 		: [val__] "d" (val__),					\
-- 
2.53.0

[PATCH v2 08/12] s390/percpu: Use percpu code section for this_cpu_cmpxchg128()

[Heiko Carstens]

Convert this_cpu_cmpxchg128() to make use of percpu code section
functionality. This is just for the sake of completeness, since
currently there is no in-kernel user of this_cpu_cmpxchg128().

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 24 ++++++++++++++++--------
 1 file changed, 16 insertions(+), 8 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 78c3aaa138f7..0eae65937c0d 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -312,15 +312,23 @@ do {									\
 #define this_cpu_cmpxchg128(pcp, oval, nval)				\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
-	u128 old__, new__, ret__;					\
+	u128 old__ = (oval), new__ = (nval);				\
 	pcp_op_T__ *ptr__;						\
-	old__ = oval;							\
-	new__ = nval;							\
-	preempt_disable_notrace();					\
-	ptr__ = raw_cpu_ptr(&(pcp));					\
-	ret__ = cmpxchg128((void *)ptr__, old__, new__);		\
-	preempt_enable_notrace();					\
-	ret__;								\
+									\
+	ptr__ = PERCPU_PTR(&(pcp));					\
+	asm_inline volatile(						\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
+		"	cdsg	%[old__],%[new__],0(%[ptr__])\n"	\
+		__PCPU_END("%[lcreg]")					\
+		: [old__] "+&d" (old__), [ptr__] "+&a" (ptr__),		\
+		  "+m" (*ptr__),					\
+		  "=m" (((struct lowcore *)0)->percpu_register)		\
+		: [new__] "d" (new__),					\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
+		  "m" (((struct lowcore *)0)->percpu_offset)		\
+		: "memory", "cc");					\
+	old__;								\
 })
 
 #define arch_this_cpu_xchg(pcp, nval)					\
-- 
2.53.0

[PATCH v2 09/12] s390/percpu: Use percpu code section for this_cpu_xchg()

[Heiko Carstens]

Convert this_cpu_xchg() to make use of percpu code section
functionality in order to get rid of preempt_disable() and
preempt_enable() calls.

Consider only four and eight byte usages, since there are compare and
swap instructions for one and two bytes available. For one and two
byte use the previous simple implementation.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 34 +++++++++++++++++++++++++++++-----
 1 file changed, 29 insertions(+), 5 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 0eae65937c0d..237bf8c80843 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -331,7 +331,7 @@ do {									\
 	old__;								\
 })
 
-#define arch_this_cpu_xchg(pcp, nval)					\
+#define arch_this_cpu_xchg_simple(pcp, nval)				\
 ({									\
 	typeof(pcp) *ptr__;						\
 	typeof(pcp) ret__;						\
@@ -342,10 +342,34 @@ do {									\
 	ret__;								\
 })
 
-#define this_cpu_xchg_1(pcp, nval) arch_this_cpu_xchg(pcp, nval)
-#define this_cpu_xchg_2(pcp, nval) arch_this_cpu_xchg(pcp, nval)
-#define this_cpu_xchg_4(pcp, nval) arch_this_cpu_xchg(pcp, nval)
-#define this_cpu_xchg_8(pcp, nval) arch_this_cpu_xchg(pcp, nval)
+#define arch_this_cpu_xchg(pcp, nval, ldop, csop)			\
+({									\
+	typedef typeof(pcp) pcp_op_T__;					\
+	pcp_op_T__ old__, new__ = (nval);				\
+	pcp_op_T__ *ptr__;						\
+									\
+	ptr__ = PERCPU_PTR(&(pcp));					\
+	asm_inline volatile(						\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
+		"	" ldop "	%[old__],0(%[ptr__])\n"		\
+		"0:	" csop "	%[old__],%[new__],0(%[ptr__])\n"\
+		"	jnz	0b\n"					\
+		__PCPU_END("%[lcreg]")					\
+		: [old__] "=&d" (old__),				\
+		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		  "=m" (((struct lowcore *)0)->percpu_register)		\
+		: [new__] "d" (new__),					\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
+		  "m" (((struct lowcore *)0)->percpu_offset)		\
+		: "memory", "cc");					\
+	old__;								\
+})
+
+#define this_cpu_xchg_1(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+#define this_cpu_xchg_2(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+#define this_cpu_xchg_4(pcp, nval) arch_this_cpu_xchg(pcp, nval, "l",  "cs")
+#define this_cpu_xchg_8(pcp, nval) arch_this_cpu_xchg(pcp, nval, "lg", "csg")
 
 #include <asm-generic/percpu.h>
 
-- 
2.53.0

[PATCH v2 10/12] s390/percpu: Use percpu code section for this_cpu_cmpxchg()

[Heiko Carstens]

Convert this_cpu_cmpxchg() to make use of percpu code section
functionality in order to get rid of preempt_disable() and
preempt_enable() calls.

Consider only four and eight byte usages, since there are no compare
and swap instructions for one and two bytes available. For one and
two bytes use the previous simple implementation.

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/percpu.h | 32 +++++++++++++++++++++++++++-----
 1 file changed, 27 insertions(+), 5 deletions(-)

diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 237bf8c80843..46368131b932 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -290,7 +290,7 @@ do {									\
 #define this_cpu_write_4(pcp, val) arch_this_cpu_write(pcp, val, "st")
 #define this_cpu_write_8(pcp, val) arch_this_cpu_write(pcp, val, "stg")
 
-#define arch_this_cpu_cmpxchg(pcp, oval, nval)				\
+#define arch_this_cpu_cmpxchg_simple(pcp, oval, nval)			\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
 	pcp_op_T__ ret__;						\
@@ -302,10 +302,32 @@ do {									\
 	ret__;								\
 })
 
-#define this_cpu_cmpxchg_1(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval)
-#define this_cpu_cmpxchg_2(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval)
-#define this_cpu_cmpxchg_4(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval)
-#define this_cpu_cmpxchg_8(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval)
+#define arch_this_cpu_cmpxchg(pcp, oval, nval, op)			\
+({									\
+	typedef typeof(pcp) pcp_op_T__;					\
+	pcp_op_T__ old__ = (oval), new__ = (nval);			\
+	pcp_op_T__ *ptr__;						\
+									\
+	ptr__ = PERCPU_PTR(&(pcp));					\
+	asm_inline volatile(						\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
+		op "	%[old__],%[new__],0(%[ptr__])\n"		\
+		__PCPU_END("%[lcreg]")					\
+		: [old__] "+&d" (old__),				\
+		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		  "=m" (((struct lowcore *)0)->percpu_register)		\
+		: [new__] "d" (new__),					\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
+		  "m" (((struct lowcore *)0)->percpu_offset)		\
+		: "memory", "cc");					\
+	old__;								\
+})
+
+#define this_cpu_cmpxchg_1(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+#define this_cpu_cmpxchg_2(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+#define this_cpu_cmpxchg_4(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval, "cs")
+#define this_cpu_cmpxchg_8(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval, "csg")
 
 #define this_cpu_cmpxchg64(pcp, o, n)	this_cpu_cmpxchg_8(pcp, o, n)
 
-- 
2.53.0

[PATCH v2 11/12] s390: Add CC_HAS_ASM_M_FORMAT_FLAG config option

[Heiko Carstens]

Add CONFIG_CC_HAS_ASM_M_FORMAT_FLAG as an indicator if the compiler
supports the 'M' operand modifier in inline assemblies. The 'M' modifier
allows to access the odd register of a TImode operand. Clang support for
this modifier was added recently to llvm-project [1].

The config option can be used to write conditional code depending on
compiler features.

[1] 28c23411e15f ("[SystemZ] Support the 'M' code for the odd register in inline-asm (#215481)")

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/Kconfig | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index 4b51bc6e8948..f1fbbf6eb50a 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -66,6 +66,12 @@ config CC_HAS_ASM_AOR_FORMAT_FLAGS
 	  Clang versions before 19.1.0 do not support A,
 	  O, and R inline assembly format flags.
 
+config CC_HAS_ASM_M_FORMAT_FLAG
+	def_bool !(CC_IS_CLANG && CLANG_VERSION < 240100)
+	help
+	  Clang versions before 24.1.0 do not support the
+	  M inline assembly format flags.
+
 config CC_HAS_ASM_IMMEDIATE_STRINGS
 	def_bool !(CC_IS_GCC && GCC_VERSION < 90000)
 	help
-- 
2.53.0

[PATCH v2 12/12] s390/percpu: Rework to simplify percpu_entry() and percpu_exit()

[Heiko Carstens]

The percpu code section functionality uses a rather complex method to
figure out if the register, which contains the address of the current
cpu's percpu variable, needs to be adjusted.

If an interrupt happens within a percpu code section (indicated by a
lowcore field), the instruction at the interrupted location is
checked. If it is not a specific AG instruction, the register needs to
be updated. This mechanism needs to take kprobes into account, and
enforces a specific instruction ordering.

Mark Rutland provided with a different solution for arm64 [1] which
comes without such limitations, but requires to use one more
instruction, and two more registers. Given that this simplifies
percpu_entry() and percpu_exit() it seems to be worth to go that
route.

Change s390 to implement a similar approach. This requires to encode
three register numbers into the "percpu_register" field, which is used
to indicate if a percpu code section is executed.

The used mviy instruction can write only one byte, which allows to
encode only two register numbers. Use a register pair for the inline
assemblies, and only encode the even register number of the register
pair to work around this.

The generated code changes like this for e.g. a simple this_cpu_inc():

Old:

c0 20 00 00 00 00  larl    %r2,c6 <foo+0x6> <-- load address of percpu var
b9 04 00 32        lgr     %r3,%r2          <-- pointless copy of address
eb 03 03 c0 00 52  mviy    960,3            <-- start of percpu code section
                                                - gpr 3 contains percpu var address
e3 30 03 b8 00 08  ag      %r3,952          <-- add percpu offset
eb 01 30 00 00 7a  agsi    0(%r3),1         <-- atomic inc
eb 00 03 c0 00 52  mviy    960,0            <-- end of percpu code section

New:

c0 10 00 00 00 00  larl    %r1,c6 <foo+0x6> <-- load address of percpu var
eb 21 03 c0 00 52  mviy    960,33           <-- start of percpu code section
                                                33 == 0x21:
						- gpr 1 contains percpu var address
						- gpr 2 used for percpu offset
  					        - gpr 2+1 == 3 used for current cpu's percpu var address
e3 20 03 b8 00 04  lg      %r2,952          <-- load percpu offset
b9 e8 20 31        agrk    %r3,%r1,%r2      <-- generate current cpu's percpu var address
eb 01 30 00 00 7a  agsi    0(%r3),1         <-- atomic inc
eb 00 03 c0 00 52  mviy    960,0            <-- end of percpu code section

In the above "new" code example 33 (0x21) is used as indicator value to
mark that a percpu code section is executed. This value implies that
registers 2 and 3 will (only) hold the percpu offset and the current
cpu's percpu var address. Those registers will always be updated by the
percpu_exit() code to contain the percpu offset and percpu var address of
the current cpu.

Note that because of the pointless lgr instruction in the "old" code
example the two code sequences have identical size, and it looks like
only one more register is used. However this is because of suboptimal
gcc code generation.

[1] https://lore.kernel.org/all/20260908151741.394589-1-mark.rutland@arm.com/

Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/include/asm/entry-percpu.h |  71 ++-------
 arch/s390/include/asm/percpu.h       | 222 ++++++++++++++++++---------
 arch/s390/include/asm/ptrace.h       |   2 +-
 arch/s390/kernel/irq.c               |  10 +-
 arch/s390/kernel/nmi.c               |   4 +-
 arch/s390/kernel/traps.c             |   4 +-
 6 files changed, 173 insertions(+), 140 deletions(-)

diff --git a/arch/s390/include/asm/entry-percpu.h b/arch/s390/include/asm/entry-percpu.h
index 89c9e44f6f83..92fdc9142f1d 100644
--- a/arch/s390/include/asm/entry-percpu.h
+++ b/arch/s390/include/asm/entry-percpu.h
@@ -2,79 +2,42 @@
 #ifndef ARCH_S390_ENTRY_PERCPU_H
 #define ARCH_S390_ENTRY_PERCPU_H
 
-#include <linux/kprobes.h>
+#include <linux/bitfield.h>
 #include <linux/percpu.h>
 #include <asm/lowcore.h>
 #include <asm/ptrace.h>
-#include <asm/asm-offsets.h>
 
 static __always_inline void percpu_entry(struct pt_regs *regs)
 {
 	struct lowcore *lc = get_lowcore();
 
-	if (user_mode(regs))
-		return;
 	regs->cpu = lc->cpu_nr;
 	regs->percpu_register = lc->percpu_register;
 	lc->percpu_register = 0;
 }
 
-static __always_inline bool percpu_code_check(struct pt_regs *regs)
-{
-	unsigned int insn, disp;
-	struct kprobe *p;
-
-	if (likely(user_mode(regs) || !regs->percpu_register))
-		return false;
-	/*
-	 * Within a percpu code section - check if the percpu base register
-	 * needs to be updated. This is the case if the PSW does not point to
-	 * the ADD instruction within the section.
-	 * - AG %rx,percpu_offset_in_lowcore(%r0,%r0)
-	 * which adds the percpu offset to the percpu base register.
-	 */
-	lockdep_assert_preemption_disabled();
-again:
-	insn = READ_ONCE(*(u16 *)psw_bits(regs->psw).ia);
-	if (unlikely(insn == BREAKPOINT_INSTRUCTION)) {
-		p = get_kprobe((void *)psw_bits(regs->psw).ia);
-		/*
-		 * If the kprobe is concurrently removed on a different CPU
-		 * it might not be found anymore. However text must have
-		 * been restored - try again.
-		 */
-		if (!p)
-			goto again;
-		insn = p->opcode;
-	}
-	if ((insn & 0xff0f) != 0xe300)
-		return true;
-	disp = LC_PERCPU_OFFSET;
-	if (machine_has_relocated_lowcore())
-		disp += LOWCORE_ALT_ADDRESS;
-	insn = (disp & 0xff000) >> 4 | (disp & 0x00fff) << 16 | 0x8;
-	if (*(u32 *)(psw_bits(regs->psw).ia + 2) != insn)
-		return true;
-	return false;
-}
-
-static __always_inline void percpu_exit(struct pt_regs *regs, bool needs_fixup)
+static __always_inline void percpu_exit(struct pt_regs *regs)
 {
+	unsigned char regval, regpcp, regoff, regptr;
 	struct lowcore *lc = get_lowcore();
-	unsigned char reg;
 
-	if (user_mode(regs))
+	if (!regs->percpu_register)
 		return;
-	reg = regs->percpu_register;
-	lc->percpu_register = reg;
-	if (likely(!needs_fixup))
-		return;
-	/* Check if process has been migrated to a different CPU. */
+	regval = regs->percpu_register;
+	lc->percpu_register = regval;
+	/* Migrated to a different CPU? */
 	if (regs->cpu == lc->cpu_nr)
 		return;
-	/* Fixup percpu base register */
-	regs->gprs[reg] -= __per_cpu_offset[regs->cpu];
-	regs->gprs[reg] += lc->percpu_offset;
+	regpcp = FIELD_GET(PCPU_REG_PCP, regval);
+	regoff = FIELD_GET(PCPU_REG_OFF, regval);
+	regptr = regoff + 1;
+	/*
+	 * Update register 'regoff' which contains the current CPU's percpu
+	 * offset, and recalculate and update current CPU's percpu variable
+	 * address contained in register 'regptr'.
+	 */
+	regs->gprs[regoff] = lc->percpu_offset;
+	regs->gprs[regptr] = regs->gprs[regpcp] + regs->gprs[regoff];
 }
 
 #endif
diff --git a/arch/s390/include/asm/percpu.h b/arch/s390/include/asm/percpu.h
index 46368131b932..ea9448009468 100644
--- a/arch/s390/include/asm/percpu.h
+++ b/arch/s390/include/asm/percpu.h
@@ -67,28 +67,31 @@
  * this. The idea is that this_cpu operations based on atomic instructions are
  * guarded with mviy instructions:
  *
- * - The first mviy instruction writes the register number, which contains the
- *   percpu address variable to lowcore. This also indicates that a percpu
- *   code section is executed.
+ * - The first mviy instruction writes the register number of the percpu address
+ *   variable and the even register number of a register pair (which encodes two
+ *   registers: the even register for the percpu offset and the odd register for
+ *   the percpu pointer) to lowcore. This also indicates that a percpu code
+ *   section is executed.
  *
- * - The first mviy instruction following the mviy instruction must be the ag
- *   instruction which adds the percpu offset to the percpu address register.
+ * - The mviy instruction is followed by the lg instruction which loads the
+ *   percpu offset into the even register of the pair and the agrk instruction
+ *   which adds the percpu offset and the percpu address into the odd register
+ *   of the pair (the percpu pointer register).
  *
  * - Afterwards the atomic percpu operation follows.
  *
  * - Then a second mviy instruction writes a zero to lowcore, which indicates
  *   the end of the percpu code section.
  *
- * - In case of an interrupt/exception/nmi the register number which was
- *   written to lowcore is copied to the exception frame (pt_regs), and a zero
- *   is written to lowcore.
+ * - In case of an interrupt/exception/nmi the encoded register numbers which
+ *   were written to lowcore are copied to the exception frame (pt_regs), and a
+ *   zero is written to lowcore.
  *
  * - On return to the previous context it is checked if a percpu code section
- *   was executed (saved register number not zero), and if the process was
- *   migrated to a different cpu. If the percpu offset was already added to
- *   the percpu address register (instruction address does _not_ point to the
- *   ag instruction) the content of the percpu address register is adjusted so
- *   it points to percpu variable of the new cpu.
+ *   was executed (saved register value not zero). The content of the percpu
+ *   offset register (even register of the pair) is reloaded with the current
+ *   cpu's percpu offset and the percpu pointer register (odd register of the
+ *   pair) is recalculated so it points to the percpu variable of the new cpu.
  *
  * Inline assemblies making use of this typically have a code sequence like:
  *
@@ -115,27 +118,45 @@
 #define UNDEF_GR_NUM								\
 	".purgem _GR_NUM\n"
 
+#define PCPU_REG_PCP_SHIFT		0
+#define PCPU_REG_PCP			GENMASK(3, 0)
+#define PCPU_REG_OFF_SHIFT		4
+#define PCPU_REG_OFF			GENMASK(7, 4)
+
 #define __PCPU_MVIY(lcreg, imm)							\
 	ALTERNATIVE("	mviy	" lcreg			"(%%r0)," imm "\n",	\
 		    "	mviy	" lcreg "+" LC_ALT_ADDR "(%%r0)," imm "\n",	\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
-#define __PCPU_AG(reg, lcoff)							\
-	ALTERNATIVE("	ag	" reg ", " lcoff		 "(%%r0)\n",	\
-		    "	ag	" reg ", " lcoff "+" LC_ALT_ADDR "(%%r0)\n",	\
+#define __PCPU_MVIY_REGS(lcreg, regpcp, regoff)					\
+	DEFINE_GR_NUM								\
+	"_GR_NUM .Lregpcp, " regpcp "\n"					\
+	"_GR_NUM .Lregoff, " regoff "\n"					\
+	UNDEF_GR_NUM								\
+	".if .Lregpcp == .Lregoff\n"						\
+	"	.error \"Registers must not be identical\"\n"			\
+	".endif\n"								\
+	".set .Lregval, ((.Lregpcp << " __stringify(PCPU_REG_PCP_SHIFT) ") |"	\
+	"		 (.Lregoff << " __stringify(PCPU_REG_OFF_SHIFT) "))\n"	\
+	__PCPU_MVIY(lcreg, ".Lregval")
+
+#define __PCPU_LG(regoff, lcoff)						\
+	ALTERNATIVE("lg	" regoff ", " lcoff		    "(%%r0)\n",		\
+		    "lg	" regoff ", " lcoff "+" LC_ALT_ADDR "(%%r0)\n",		\
 		    ALT_FEATURE(MFEATURE_LOWCORE))
 
-#define __PCPU_BEGIN(lcreg, lcoff, reg)						\
-	DEFINE_GR_NUM								\
-	"_GR_NUM .Lreg, " reg "\n"						\
-	UNDEF_GR_NUM								\
-	__PCPU_MVIY(lcreg, ".Lreg")						\
-	__PCPU_AG(reg, lcoff)
+#define __PCPU_AGRK(regptr, regpcp, regoff)					\
+	"	agrk	" regptr "," regpcp "," regoff "\n"
+
+#define __PCPU_BEGIN(lcreg, lcoff, regpcp, regoff, regptr)			\
+	__PCPU_MVIY_REGS(lcreg, regpcp, regoff)					\
+	__PCPU_LG(regoff, lcoff)						\
+	__PCPU_AGRK(regptr, regpcp, regoff)
 
 #define __PCPU_END(lcreg)							\
 	__PCPU_MVIY(lcreg, "0")
 
-#ifndef MARCH_HAS_Z196_FEATURES
+#if !defined(MARCH_HAS_Z196_FEATURES) || !defined(CONFIG_CC_HAS_ASM_M_FORMAT_FLAG)
 
 #define this_cpu_add_4(pcp, val)	arch_this_cpu_to_op_simple(pcp, val, +)
 #define this_cpu_add_8(pcp, val)	arch_this_cpu_to_op_simple(pcp, val, +)
@@ -146,11 +167,12 @@
 #define this_cpu_or_4(pcp, val)		arch_this_cpu_to_op_simple(pcp, val, |)
 #define this_cpu_or_8(pcp, val)		arch_this_cpu_to_op_simple(pcp, val, |)
 
-#else /* MARCH_HAS_Z196_FEATURES */
+#else /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
 
 #define arch_this_cpu_add(pcp, val, op1, op2, szcast)				\
 do {										\
 	typedef typeof(pcp) pcp_op_T__;						\
+	union register_pair pair__;						\
 	pcp_op_T__ val__ = (val);						\
 	pcp_op_T__ old__, *ptr__;						\
 										\
@@ -158,25 +180,27 @@ do {										\
 	if (__builtin_constant_p(val__) &&					\
 	    ((szcast)val__ > -129) && ((szcast)val__ < 128)) {			\
 		asm volatile(							\
-			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-			op2 "   0(%[ptr__]),%[val__]\n"				\
+			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+				      "%[pair__]","%M[pair__]")			\
+			op2 "   0(%M[pair__]),%[val__]\n"			\
 			__PCPU_END("%[lcreg]")					\
-			: [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+			: [pair__] "=&d" (pair__), "+m" (*ptr__),		\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
-			: [val__] "i" ((szcast)val__),				\
+			: [val__] "i" ((szcast)val__), [ptr__] "d" (ptr__),	\
 			  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 			  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
 			: "cc");						\
 	} else {								\
 		asm volatile(							\
-			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-			op1 "   %[old__],%[val__],0(%[ptr__])\n"		\
+			__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+				     "%[pair__]","%M[pair__]")			\
+			op1 "   %[old__],%[val__],0(%M[pair__])\n"		\
 			__PCPU_END("%[lcreg]")					\
-			: [old__] "=&d" (old__),				\
-			  [ptr__] "+&a" (ptr__),  "+m" (*ptr__),		\
+			: [old__] "=&d" (old__), [pair__] "=&d" (pair__),	\
+			  "+m" (*ptr__),					\
 			  "=m" (((struct lowcore *)0)->percpu_register)		\
-			: [val__] "d" (val__),					\
+			: [val__] "d" (val__), [ptr__] "d" (ptr__),		\
 			  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 			  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 			  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -190,18 +214,20 @@ do {										\
 #define arch_this_cpu_add_return(pcp, val, op)				\
 ({									\
 	typedef typeof(pcp) pcp_op_T__; 				\
+	union register_pair pair__;					\
 	pcp_op_T__ val__ = (val);					\
 	pcp_op_T__ old__, *ptr__;					\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		op "	%[old__],%[val__],0(%[ptr__])\n"		\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			     "%[pair__]","%M[pair__]")			\
+		op "	%[old__],%[val__],0(%M[pair__])\n"		\
 		__PCPU_END("%[lcreg]")					\
-		: [old__] "=&d" (old__),				\
-		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		: [old__] "=&d" (old__), [pair__] "=&d" (pair__),	\
+		  "+m" (*ptr__),					\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [val__] "d" (val__),					\
+		: [val__] "d" (val__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -215,18 +241,20 @@ do {										\
 #define arch_this_cpu_to_op(pcp, val, op)				\
 do {									\
 	typedef typeof(pcp) pcp_op_T__; 				\
+	union register_pair pair__;					\
 	pcp_op_T__ val__ = (val);					\
 	pcp_op_T__ old__, *ptr__;					\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		op "    %[old__],%[val__],0(%[ptr__])\n"		\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			     "%[pair__]","%M[pair__]")			\
+		op "    %[old__],%[val__],0(%M[pair__])\n"		\
 		__PCPU_END("%[lcreg]")					\
-		: [old__] "=&d" (old__),				\
-		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		: [old__] "=&d" (old__), [pair__] "=&d" (pair__),	\
+		  "+m" (*ptr__),					\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [val__] "d" (val__),					\
+		: [val__] "d" (val__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -238,22 +266,23 @@ do {									\
 #define this_cpu_or_4(pcp, val)		arch_this_cpu_to_op(pcp, val, "lao")
 #define this_cpu_or_8(pcp, val)		arch_this_cpu_to_op(pcp, val, "laog")
 
-#endif /* MARCH_HAS_Z196_FEATURES */
-
 #define arch_this_cpu_read(pcp, op)					\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
+	union register_pair pair__;					\
 	unsigned long res__;						\
 	pcp_op_T__ *ptr__;						\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		op "	%[res__],0(%[ptr__])\n"				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			      "%[pair__]","%M[pair__]")			\
+		op "	%[res__],0(%M[pair__])\n"			\
 		__PCPU_END("%[lcreg]")					\
-		: [res__] "=&d" (res__), [ptr__] "+&a" (ptr__),		\
+		: [res__] "=&d" (res__), [pair__] "=&d" (pair__),	\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [lcreg] "i" (LC_PERCPU_REGISTER),			\
+		: [ptr__] "d" (ptr__),					\
+		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (*ptr__),						\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -269,16 +298,18 @@ do {									\
 #define arch_this_cpu_write(pcp, val, op)				\
 do {									\
 	typedef typeof(pcp) pcp_op_T__;					\
+	union register_pair pair__;					\
 	pcp_op_T__ *ptr__, val__ = (val);				\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		op "    %[val__],0(%[ptr__])\n"				\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			      "%[pair__]","%M[pair__]")			\
+		op "    %[val__],0(%M[pair__])\n"			\
 		__PCPU_END("%[lcreg]")					\
-		: [ptr__] "+&a" (ptr__), "=m" (*ptr__),			\
+		: [pair__] "=&d" (pair__), "=m" (*ptr__),		\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [val__] "d" (val__),					\
+		: [val__] "d" (val__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -290,6 +321,8 @@ do {									\
 #define this_cpu_write_4(pcp, val) arch_this_cpu_write(pcp, val, "st")
 #define this_cpu_write_8(pcp, val) arch_this_cpu_write(pcp, val, "stg")
 
+#endif /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #define arch_this_cpu_cmpxchg_simple(pcp, oval, nval)			\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
@@ -302,21 +335,33 @@ do {									\
 	ret__;								\
 })
 
+#define this_cpu_cmpxchg_1(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+#define this_cpu_cmpxchg_2(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+
+#if !defined(MARCH_HAS_Z196_FEATURES) || !defined(CONFIG_CC_HAS_ASM_M_FORMAT_FLAG)
+
+#define this_cpu_cmpxchg_4(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+#define this_cpu_cmpxchg_8(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
+
+#else /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #define arch_this_cpu_cmpxchg(pcp, oval, nval, op)			\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
 	pcp_op_T__ old__ = (oval), new__ = (nval);			\
+	union register_pair pair__;					\
 	pcp_op_T__ *ptr__;						\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		op "	%[old__],%[new__],0(%[ptr__])\n"		\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			     "%[pair__]","%M[pair__]")			\
+		op "	%[old__],%[new__],0(%M[pair__])\n"		\
 		__PCPU_END("%[lcreg]")					\
-		: [old__] "+&d" (old__),				\
-		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		: [old__] "+&d" (old__), [pair__] "=&d" (pair__),	\
+		  "+m" (*ptr__),					\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [new__] "d" (new__),					\
+		: [new__] "d" (new__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -324,28 +369,45 @@ do {									\
 	old__;								\
 })
 
-#define this_cpu_cmpxchg_1(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
-#define this_cpu_cmpxchg_2(pcp, oval, nval) arch_this_cpu_cmpxchg_simple(pcp, oval, nval)
 #define this_cpu_cmpxchg_4(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval, "cs")
 #define this_cpu_cmpxchg_8(pcp, oval, nval) arch_this_cpu_cmpxchg(pcp, oval, nval, "csg")
 
+#endif /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #define this_cpu_cmpxchg64(pcp, o, n)	this_cpu_cmpxchg_8(pcp, o, n)
 
+#if !defined(MARCH_HAS_Z196_FEATURES) || !defined(CONFIG_CC_HAS_ASM_M_FORMAT_FLAG)
+
+#define this_cpu_cmpxchg128(pcp, oval, nval)				\
+({									\
+	u128 *ptr__, ret__;						\
+									\
+	preempt_disable_notrace();					\
+	ptr__ = raw_cpu_ptr(&(pcp));					\
+	ret__ = arch_cmpxchg128(ptr__, oval, nval);			\
+	preempt_enable_notrace();					\
+	ret__;								\
+})
+
+#else /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #define this_cpu_cmpxchg128(pcp, oval, nval)				\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
 	u128 old__ = (oval), new__ = (nval);				\
+	union register_pair pair__;					\
 	pcp_op_T__ *ptr__;						\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		"	cdsg	%[old__],%[new__],0(%[ptr__])\n"	\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			     "%[pair__]","%M[pair__]")			\
+		"	cdsg	%[old__],%[new__],0(%M[pair__])\n"	\
 		__PCPU_END("%[lcreg]")					\
-		: [old__] "+&d" (old__), [ptr__] "+&a" (ptr__),		\
+		: [old__] "+&d" (old__), [pair__] "=&d" (pair__),	\
 		  "+m" (*ptr__),					\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [new__] "d" (new__),					\
+		: [new__] "d" (new__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -353,6 +415,8 @@ do {									\
 	old__;								\
 })
 
+#endif /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #define arch_this_cpu_xchg_simple(pcp, nval)				\
 ({									\
 	typeof(pcp) *ptr__;						\
@@ -364,23 +428,35 @@ do {									\
 	ret__;								\
 })
 
-#define arch_this_cpu_xchg(pcp, nval, ldop, csop)			\
+#define this_cpu_xchg_1(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+#define this_cpu_xchg_2(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+
+#if !defined(MARCH_HAS_Z196_FEATURES) || !defined(CONFIG_CC_HAS_ASM_M_FORMAT_FLAG)
+
+#define this_cpu_xchg_4(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+#define this_cpu_xchg_8(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
+
+#else /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
+#define arch_this_cpu_xchg(pcp, nval, lop, cop)				\
 ({									\
 	typedef typeof(pcp) pcp_op_T__;					\
 	pcp_op_T__ old__, new__ = (nval);				\
+	union register_pair pair__;					\
 	pcp_op_T__ *ptr__;						\
 									\
 	ptr__ = PERCPU_PTR(&(pcp));					\
 	asm_inline volatile(						\
-		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]")		\
-		"	" ldop "	%[old__],0(%[ptr__])\n"		\
-		"0:	" csop "	%[old__],%[new__],0(%[ptr__])\n"\
+		__PCPU_BEGIN("%[lcreg]","%[lcoff]","%[ptr__]",		\
+			     "%[pair__]","%M[pair__]")			\
+		"	" lop "	%[old__],0(%M[pair__])\n"		\
+		"0:	" cop "	%[old__],%[new__],0(%M[pair__])\n"	\
 		"	jnz	0b\n"					\
 		__PCPU_END("%[lcreg]")					\
-		: [old__] "=&d" (old__),				\
-		  [ptr__] "+&a" (ptr__), "+m" (*ptr__),			\
+		: [old__] "=&d" (old__), [pair__] "=&d" (pair__),	\
+		  "+m" (*ptr__),					\
 		  "=m" (((struct lowcore *)0)->percpu_register)		\
-		: [new__] "d" (new__),					\
+		: [new__] "d" (new__), [ptr__] "d" (ptr__),		\
 		  [lcreg] "i" (LC_PERCPU_REGISTER),			\
 		  [lcoff] "i" (LC_PERCPU_OFFSET),			\
 		  "m" (((struct lowcore *)0)->percpu_offset)		\
@@ -388,11 +464,11 @@ do {									\
 	old__;								\
 })
 
-#define this_cpu_xchg_1(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
-#define this_cpu_xchg_2(pcp, nval) arch_this_cpu_xchg_simple(pcp, nval)
 #define this_cpu_xchg_4(pcp, nval) arch_this_cpu_xchg(pcp, nval, "l",  "cs")
 #define this_cpu_xchg_8(pcp, nval) arch_this_cpu_xchg(pcp, nval, "lg", "csg")
 
+#endif /* MARCH_HAS_Z196_FEATURES || CONFIG_CC_HAS_ASM_M_FORMAT_FLAG */
+
 #include <asm-generic/percpu.h>
 
 #endif /* __ARCH_S390_PERCPU__ */
diff --git a/arch/s390/include/asm/ptrace.h b/arch/s390/include/asm/ptrace.h
index 6411e3584283..5102d851b84b 100644
--- a/arch/s390/include/asm/ptrace.h
+++ b/arch/s390/include/asm/ptrace.h
@@ -132,7 +132,7 @@ struct pt_regs {
 	};
 	unsigned long flags;
 	unsigned long last_break;
-	unsigned int cpu;
+	unsigned int cpu
 	unsigned char percpu_register;
 };
 
diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
index c923496aa7b4..8f21bea349a6 100644
--- a/arch/s390/kernel/irq.c
+++ b/arch/s390/kernel/irq.c
@@ -144,9 +144,9 @@ static int irq_pending(struct pt_regs *regs)
 
 void noinstr do_io_irq(struct pt_regs *regs)
 {
-	bool from_idle, percpu_needs_fixup;
 	struct pt_regs *old_regs;
 	irqentry_state_t state;
+	bool from_idle;
 
 	percpu_entry(regs);
 	state = irqentry_enter(regs);
@@ -174,21 +174,20 @@ void noinstr do_io_irq(struct pt_regs *regs)
 			do_irq_async(regs, IO_INTERRUPT);
 	} while (machine_is_lpar() && irq_pending(regs));
 
-	percpu_needs_fixup = percpu_code_check(regs);
 	irq_exit_rcu();
 	set_irq_regs(old_regs);
 	irqentry_exit(regs, state);
 
 	if (from_idle)
 		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
-	percpu_exit(regs, percpu_needs_fixup);
+	percpu_exit(regs);
 }
 
 void noinstr do_ext_irq(struct pt_regs *regs)
 {
-	bool from_idle, percpu_needs_fixup;
 	struct pt_regs *old_regs;
 	irqentry_state_t state;
+	bool from_idle;
 
 	percpu_entry(regs);
 	state = irqentry_enter(regs);
@@ -214,14 +213,13 @@ void noinstr do_ext_irq(struct pt_regs *regs)
 
 	do_irq_async(regs, EXT_INTERRUPT);
 
-	percpu_needs_fixup = percpu_code_check(regs);
 	irq_exit_rcu();
 	set_irq_regs(old_regs);
 	irqentry_exit(regs, state);
 
 	if (from_idle)
 		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
-	percpu_exit(regs, percpu_needs_fixup);
+	percpu_exit(regs);
 }
 
 static void show_msi_interrupt(struct seq_file *p, int irq)
diff --git a/arch/s390/kernel/nmi.c b/arch/s390/kernel/nmi.c
index 17297a8b63d9..56a15edc65d7 100644
--- a/arch/s390/kernel/nmi.c
+++ b/arch/s390/kernel/nmi.c
@@ -361,7 +361,6 @@ NOKPROBE_SYMBOL(s390_backup_mcck_info);
  */
 void notrace s390_do_machine_check(struct pt_regs *regs)
 {
-	bool percpu_needs_fixup;
 	static int ipd_count;
 	static DEFINE_SPINLOCK(ipd_lock);
 	static unsigned long long last_ipd;
@@ -495,9 +494,8 @@ void notrace s390_do_machine_check(struct pt_regs *regs)
 	if (mcck_pending)
 		schedule_mcck_handler();
 
-	percpu_needs_fixup = percpu_code_check(regs);
 	irqentry_nmi_exit(regs, irq_state);
-	percpu_exit(regs, percpu_needs_fixup);
+	percpu_exit(regs);
 }
 NOKPROBE_SYMBOL(s390_do_machine_check);
 
diff --git a/arch/s390/kernel/traps.c b/arch/s390/kernel/traps.c
index b6ba4465f59d..dd051046170b 100644
--- a/arch/s390/kernel/traps.c
+++ b/arch/s390/kernel/traps.c
@@ -338,7 +338,6 @@ static void (*pgm_check_table[128])(struct pt_regs *regs);
 void noinstr __do_pgm_check(struct pt_regs *regs, unsigned long flags)
 {
 	struct lowcore *lc = get_lowcore();
-	bool percpu_needs_fixup;
 	irqentry_state_t state;
 	struct pgm_stat *stat;
 	unsigned int trapnr;
@@ -400,9 +399,8 @@ void noinstr __do_pgm_check(struct pt_regs *regs, unsigned long flags)
 		pgm_check_table[trapnr](regs);
 out:
 	local_irq_disable();
-	percpu_needs_fixup = percpu_code_check(regs);
 	irqentry_exit(regs, state);
-	percpu_exit(regs, percpu_needs_fixup);
+	percpu_exit(regs);
 }
 
 static int pgm_check_stat_show(struct seq_file *p, void *v)
-- 
2.53.0

Re: [PATCH v2 12/12] s390/percpu: Rework to simplify percpu_entry() and percpu_exit()

[Heiko Carstens]

On Fri, Sep 18, 2026 at 10:48:42AM +0200, Heiko Carstens wrote:
> diff --git a/arch/s390/include/asm/ptrace.h b/arch/s390/include/asm/ptrace.h
> index 6411e3584283..5102d851b84b 100644
> --- a/arch/s390/include/asm/ptrace.h
> +++ b/arch/s390/include/asm/ptrace.h
> @@ -132,7 +132,7 @@ struct pt_regs {
>  	};
>  	unsigned long flags;
>  	unsigned long last_break;
> -	unsigned int cpu;
> +	unsigned int cpu

Looks like I forgot to amend a compile fix (missing semicolon) before
sending. I'm not going to send a new version for just this.

Re: [PATCH v2 00/12] s390: More this_cpu_*() changes

[Mark Rutland]

Hi Heiko,

This looks neat!

On Fri, Sep 18, 2026 at 10:48:30AM +0200, Heiko Carstens wrote:
> v2:
> - Yet another brown paper bug: Re-add CPU migration check to last
>   patch. Unconditionally recalculating and updating the percpu variable
>   address and percpu offset register can corrupt previous context
>   register state in several cases (not only the single case reported by
>   Sashiko).

Do all of those case boil down to:

* During exception entry from context A to context B, a nested exception
  is taken from context B to context C before percpu_entry() is called,
  and the nested exception handler updates B's live regs using A's
  percpu register indices.

* During exception return from context B to context A, a nested
  exception is taken from context B to context C after percpu_entry() is
  called, and the nested exception handler updates B's live regs using
  A's percpu register indices.

... or was there another case that you spotted (e.g. some case where
percpu_{entry,exit}() aren't called)?

On arm64, I think we avoid those two cases by virtue of not taking
nested (architectural) exceptions during those windows, and by not
applying fixups for SDEI events (SW NMIs that can be taken during those
windows).

If there's another case to consider, it'd be handy to know.

Mark.

Re: [PATCH v2 12/12] s390/percpu: Rework to simplify percpu_entry() and percpu_exit()

[Mark Rutland]

On Fri, Sep 18, 2026 at 10:48:42AM +0200, Heiko Carstens wrote:
> The percpu code section functionality uses a rather complex method to
> figure out if the register, which contains the address of the current
> cpu's percpu variable, needs to be adjusted.
> 
> If an interrupt happens within a percpu code section (indicated by a
> lowcore field), the instruction at the interrupted location is
> checked. If it is not a specific AG instruction, the register needs to
> be updated. This mechanism needs to take kprobes into account, and
> enforces a specific instruction ordering.
> 
> Mark Rutland provided with a different solution for arm64 [1] which
> comes without such limitations, but requires to use one more
> instruction, and two more registers. Given that this simplifies
> percpu_entry() and percpu_exit() it seems to be worth to go that
> route.
> 
> Change s390 to implement a similar approach. This requires to encode
> three register numbers into the "percpu_register" field, which is used
> to indicate if a percpu code section is executed.
> 
> The used mviy instruction can write only one byte, which allows to
> encode only two register numbers. Use a register pair for the inline
> assemblies, and only encode the even register number of the register
> pair to work around this.

Neat trick!

I have one minor comment below, but this looks good to me regardless.

> -static __always_inline void percpu_exit(struct pt_regs *regs, bool needs_fixup)
> +static __always_inline void percpu_exit(struct pt_regs *regs)
>  {
> +	unsigned char regval, regpcp, regoff, regptr;
>  	struct lowcore *lc = get_lowcore();
> -	unsigned char reg;
>  
> -	if (user_mode(regs))
> +	if (!regs->percpu_register)
>  		return;
> -	reg = regs->percpu_register;
> -	lc->percpu_register = reg;
> -	if (likely(!needs_fixup))
> -		return;
> -	/* Check if process has been migrated to a different CPU. */
> +	regval = regs->percpu_register;
> +	lc->percpu_register = regval;
> +	/* Migrated to a different CPU? */
>  	if (regs->cpu == lc->cpu_nr)
>  		return;
> -	/* Fixup percpu base register */
> -	regs->gprs[reg] -= __per_cpu_offset[regs->cpu];
> -	regs->gprs[reg] += lc->percpu_offset;
> +	regpcp = FIELD_GET(PCPU_REG_PCP, regval);
> +	regoff = FIELD_GET(PCPU_REG_OFF, regval);
> +	regptr = regoff + 1;
> +	/*
> +	 * Update register 'regoff' which contains the current CPU's percpu
> +	 * offset, and recalculate and update current CPU's percpu variable
> +	 * address contained in register 'regptr'.
> +	 */
> +	regs->gprs[regoff] = lc->percpu_offset;
> +	regs->gprs[regptr] = regs->gprs[regpcp] + regs->gprs[regoff];
>  }

> +#define PCPU_REG_PCP_SHIFT		0
> +#define PCPU_REG_PCP			GENMASK(3, 0)
> +#define PCPU_REG_OFF_SHIFT		4
> +#define PCPU_REG_OFF			GENMASK(7, 4)
> +
>  #define __PCPU_MVIY(lcreg, imm)							\
>  	ALTERNATIVE("	mviy	" lcreg			"(%%r0)," imm "\n",	\
>  		    "	mviy	" lcreg "+" LC_ALT_ADDR "(%%r0)," imm "\n",	\
>  		    ALT_FEATURE(MFEATURE_LOWCORE))
>  
> -#define __PCPU_AG(reg, lcoff)							\
> -	ALTERNATIVE("	ag	" reg ", " lcoff		 "(%%r0)\n",	\
> -		    "	ag	" reg ", " lcoff "+" LC_ALT_ADDR "(%%r0)\n",	\

> +#define __PCPU_MVIY_REGS(lcreg, regpcp, regoff)					\
> +	DEFINE_GR_NUM								\
> +	"_GR_NUM .Lregpcp, " regpcp "\n"					\
> +	"_GR_NUM .Lregoff, " regoff "\n"					\
> +	UNDEF_GR_NUM								\
> +	".if .Lregpcp == .Lregoff\n"						\
> +	"	.error \"Registers must not be identical\"\n"			\
> +	".endif\n"								\
> +	".set .Lregval, ((.Lregpcp << " __stringify(PCPU_REG_PCP_SHIFT) ") |"	\
> +	"		 (.Lregoff << " __stringify(PCPU_REG_OFF_SHIFT) "))\n"	\
> +	__PCPU_MVIY(lcreg, ".Lregval")
> +
> +#define __PCPU_LG(regoff, lcoff)						\
> +	ALTERNATIVE("lg	" regoff ", " lcoff		    "(%%r0)\n",		\
> +		    "lg	" regoff ", " lcoff "+" LC_ALT_ADDR "(%%r0)\n",		\
>  		    ALT_FEATURE(MFEATURE_LOWCORE))
>  
> -#define __PCPU_BEGIN(lcreg, lcoff, reg)						\
> -	DEFINE_GR_NUM								\
> -	"_GR_NUM .Lreg, " reg "\n"						\
> -	UNDEF_GR_NUM								\
> -	__PCPU_MVIY(lcreg, ".Lreg")						\
> -	__PCPU_AG(reg, lcoff)
> +#define __PCPU_AGRK(regptr, regpcp, regoff)					\
> +	"	agrk	" regptr "," regpcp "," regoff "\n"
> +
> +#define __PCPU_BEGIN(lcreg, lcoff, regpcp, regoff, regptr)			\
> +	__PCPU_MVIY_REGS(lcreg, regpcp, regoff)					\
> +	__PCPU_LG(regoff, lcoff)						\
> +	__PCPU_AGRK(regptr, regpcp, regoff)

Since percpu_exit() relies on regptr being regoff + 1, maybe it's worth
having a check here to verify that? Perhaps in __PCPU_MVIY_REGS() with
the uniqueness check.

Hopefully that never goes wrong, but catching a violation at build time
might save some future pain.

Mark.

Re: [PATCH v2 00/12] s390: More this_cpu_*() changes

[Heiko Carstens]

On Fri, Sep 18, 2026 at 11:31:25AM +0100, Mark Rutland wrote:
> > v2:
> > - Yet another brown paper bug: Re-add CPU migration check to last
> >   patch. Unconditionally recalculating and updating the percpu variable
> >   address and percpu offset register can corrupt previous context
> >   register state in several cases (not only the single case reported by
> >   Sashiko).
> 
> Do all of those case boil down to:
> 
> * During exception entry from context A to context B, a nested exception
>   is taken from context B to context C before percpu_entry() is called,
>   and the nested exception handler updates B's live regs using A's
>   percpu register indices.
> 
> * During exception return from context B to context A, a nested
>   exception is taken from context B to context C after percpu_entry() is
                                                         ^^^^^^^^^^^^
That should have been percpu_exit(), I guess.

>   called, and the nested exception handler updates B's live regs using
>   A's percpu register indices.
> 
> ... or was there another case that you spotted (e.g. some case where
> percpu_{entry,exit}() aren't called)?

Yes, all scenarios boil down to exactly the above.

> On arm64, I think we avoid those two cases by virtue of not taking
> nested (architectural) exceptions during those windows, and by not
> applying fixups for SDEI events (SW NMIs that can be taken during those
> windows).
> 
> If there's another case to consider, it'd be handy to know.

I'm not aware of more cases.

Re: [PATCH v2 12/12] s390/percpu: Rework to simplify percpu_entry() and percpu_exit()

[Heiko Carstens]

On Fri, Sep 18, 2026 at 11:40:21AM +0100, Mark Rutland wrote:
> On Fri, Sep 18, 2026 at 10:48:42AM +0200, Heiko Carstens wrote:
> > +#define __PCPU_AGRK(regptr, regpcp, regoff)					\
> > +	"	agrk	" regptr "," regpcp "," regoff "\n"
> > +
> > +#define __PCPU_BEGIN(lcreg, lcoff, regpcp, regoff, regptr)			\
> > +	__PCPU_MVIY_REGS(lcreg, regpcp, regoff)					\
> > +	__PCPU_LG(regoff, lcoff)						\
> > +	__PCPU_AGRK(regptr, regpcp, regoff)
> 
> Since percpu_exit() relies on regptr being regoff + 1, maybe it's worth
> having a check here to verify that? Perhaps in __PCPU_MVIY_REGS() with
> the uniqueness check.
> 
> Hopefully that never goes wrong, but catching a violation at build time
> might save some future pain.

Yes, that makes sense. Especially considering that there was already a
previous bug which could have been caught at build time.

Thanks a lot for taking a look!

Re: [PATCH v2 00/12] s390: More this_cpu_*() changes

[Mark Rutland]

On Fri, Sep 18, 2026 at 12:59:22PM +0200, Heiko Carstens wrote:
> On Fri, Sep 18, 2026 at 11:31:25AM +0100, Mark Rutland wrote:
> > > v2:
> > > - Yet another brown paper bug: Re-add CPU migration check to last
> > >   patch. Unconditionally recalculating and updating the percpu variable
> > >   address and percpu offset register can corrupt previous context
> > >   register state in several cases (not only the single case reported by
> > >   Sashiko).
> > 
> > Do all of those case boil down to:
> > 
> > * During exception entry from context A to context B, a nested exception
> >   is taken from context B to context C before percpu_entry() is called,
> >   and the nested exception handler updates B's live regs using A's
> >   percpu register indices.
> > 
> > * During exception return from context B to context A, a nested
> >   exception is taken from context B to context C after percpu_entry() is
>                                                          ^^^^^^^^^^^^
> That should have been percpu_exit(), I guess.

Whoops; yes!

> >   called, and the nested exception handler updates B's live regs using
> >   A's percpu register indices.
> > 
> > ... or was there another case that you spotted (e.g. some case where
> > percpu_{entry,exit}() aren't called)?
> 
> Yes, all scenarios boil down to exactly the above.
> 
> > On arm64, I think we avoid those two cases by virtue of not taking
> > nested (architectural) exceptions during those windows, and by not
> > applying fixups for SDEI events (SW NMIs that can be taken during those
> > windows).
> > 
> > If there's another case to consider, it'd be handy to know.
> 
> I'm not aware of more cases.

Perfect, thanks for confirming!

Mark.