[PW_SID:1120193] riscv: fix PIO helpers and add optimized percpu ops by linux-riscv-bot · Pull Request #2198 · linux-riscv/linux

linux-riscv-bot · 2026-07-02T06:22:03Z

PR for series 1120193 applied to workflow__riscv__fixes

Name: riscv: fix PIO helpers and add optimized percpu ops
URL: https://patchwork.kernel.org/project/linux-riscv/list/?series=1120193
Version: 7

When port I/O is not supported, exposing the port-string helpers is both unnecessary and can make clang diagnose null-pointer arithmetic from the PCI_IOBASE based address expression. Keep the MMIO string helpers available as before, but only provide the port I/O variants when CONFIG_HAS_IOPORT is enabled. Signed-off-by: Yunhui Cui <cuiyunhui@bytedance.com> Signed-off-by: Linux RISC-V bot <linux.riscv.bot@gmail.com>

Add RISC-V specific this_cpu helpers so common percpu operations can use short architecture sequences instead of the generic implementation. Native-width operations use AMOs, while 8/16-bit operations use Zabha when available and a local 32-bit LR/SC fallback otherwise. Signed-off-by: Yunhui Cui <cuiyunhui@bytedance.com> Signed-off-by: Linux RISC-V bot <linux.riscv.bot@gmail.com>

RISC-V percpu addressing currently derives the base offset from the CPU number and __per_cpu_offset[]. Cache the current CPU percpu offset in thread_info so percpu accesses can load it directly. Keep the cached value up to date for the boot CPU, context switches and secondary CPU bringup. Initialize secondary idle tasks before they start running so early percpu accesses use the secondary CPU offset rather than inheriting the boot CPU value. Link: https://lists.riscv.org/g/tech-privileged/topic/risc_v_tech_arch_review/113437553?page=2 Signed-off-by: Yunhui Cui <cuiyunhui@bytedance.com> Signed-off-by: Linux RISC-V bot <linux.riscv.bot@gmail.com>

linux-riscv-bot · 2026-07-02T06:27:19Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
build-rv32-defconfig
Desc: Builds riscv32 defconfig
Duration: 118.69 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:03:10Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
build-rv64-clang-allmodconfig
Desc: Builds riscv64 allmodconfig with Clang, and checks for errors and added warnings
Duration: 2149.64 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:49:03Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
build-rv64-gcc-allmodconfig
Desc: Builds riscv64 allmodconfig with GCC, and checks for errors and added warnings
Duration: 2750.81 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:49:26Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
build-rv64-nommu-k210-defconfig
Desc: Builds riscv64 defconfig with NOMMU for K210
Duration: 21.80 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:49:51Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
build-rv64-nommu-k210-virt
Desc: Builds riscv64 defconfig with NOMMU for the virt platform
Duration: 23.59 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:49:54Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
checkpatch
Desc: Runs checkpatch.pl on the patch
Duration: 0.75 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:15Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
dtb-warn-rv64
Desc: Checks for Device Tree warnings/errors
Duration: 79.40 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:17Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
header-inline
Desc: Detects static functions without inline keyword in header files
Duration: 0.27 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:19Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
kdoc
Desc: Detects for kdoc errors
Duration: 0.78 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:21Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
module-param
Desc: Detect module_param changes
Duration: 0.27 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:23Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
verify-fixes
Desc: Verifies that the Fixes: tags exist
Duration: 0.26 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:51:25Z

Patch 1: "[v7,1/3] riscv: io: avoid null-pointer arithmetic in PIO helpers"
verify-signedoff
Desc: Verifies that Signed-off-by: tags are correct
Duration: 0.30 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T07:53:42Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
build-rv32-defconfig
Desc: Builds riscv32 defconfig
Duration: 119.81 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T08:33:41Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
build-rv64-clang-allmodconfig
Desc: Builds riscv64 allmodconfig with Clang, and checks for errors and added warnings
Duration: 2397.49 seconds
Result: ERROR
Output:

Redirect to /build/tmp.HWm8Ahpk8M and /build/tmp.B4n3mw7Tmr
Tree base:
ee5e7be2d5f23 ("riscv: io: avoid null-pointer arithmetic in PIO helpers")
Building the whole tree with the patch
Building the tree before the patch
Building the tree with the patch
New errors added:
--- /build/tmp.JTc52DT0ni	2026-07-02 08:33:24.158060585 +0000
+++ /build/tmp.eYe1LtCudY	2026-07-02 08:33:24.160060567 +0000
@@ -170,0 +171 @@
+      1 /build/tmpnevle9et/kernel/fork.c:268:8: warning: variable '__old' is uninitialized when used within its own initialization [-Wuninitialized]
Per-file breakdown
error/warning file pre:
error/warning file post:
pre: 220 post: 221



real	16m24.029s
user	579m27.826s
sys	121m46.431s

real	18m28.903s
user	595m56.861s
sys	125m13.250s

real	4m48.200s
user	45m32.035s
sys	26m45.842s

linux-riscv-bot · 2026-07-02T09:24:37Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
build-rv64-gcc-allmodconfig
Desc: Builds riscv64 allmodconfig with GCC, and checks for errors and added warnings
Duration: 3055.17 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:25:01Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
build-rv64-nommu-k210-defconfig
Desc: Builds riscv64 defconfig with NOMMU for K210
Duration: 21.98 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:25:26Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
build-rv64-nommu-k210-virt
Desc: Builds riscv64 defconfig with NOMMU for the virt platform
Duration: 23.85 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:25:30Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
checkpatch
Desc: Runs checkpatch.pl on the patch
Duration: 2.03 seconds
Result: ERROR
Output:

WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?
#19: 
new file mode 100644

ERROR: Macros with complex values should be enclosed in parentheses
#68: FILE: arch/riscv/include/asm/percpu.h:45:
+#define PERCPU_OP(name, amo_insn)					\
+	__PERCPU_AMO_OP_CASE(.w, name, 32, amo_insn)			\
+	__PERCPU_AMO_OP_CASE(.d, name, 64, amo_insn)

BUT SEE:

   do {} while (0) advice is over-stated in a few situations:

   The more obvious case is macros, like MODULE_PARM_DESC, invoked at
   file-scope, where C disallows code (it must be in functions).  See
   $exceptions if you have one to add by name.

   More troublesome is declarative macros used at top of new scope,
   like DECLARE_PER_CPU.  These might just compile with a do-while-0
   wrapper, but would be incorrect.  Most of these are handled by
   detecting struct,union,etc declaration primitives in $exceptions.

   Theres also macros called inside an if (block), which "return" an
   expression.  These cannot do-while, and need a ({}) wrapper.

   Enjoy this qualification while we work to improve our heuristics.

ERROR: Macros with complex values should be enclosed in parentheses
#84: FILE: arch/riscv/include/asm/percpu.h:61:
+#define __PERCPU_AMO_RET_OP_CASE(sfx, name, sz, amo_insn)		\
+static inline u##sz							\
+__percpu_##name##_return_amo_case_##sz(void *ptr, unsigned long val)	\
+{									\
+	register u##sz ret;						\
+									\
+	asm volatile (							\
+		"amo" #amo_insn #sfx " %[ret], %[val], %[ptr]"		\
+		: [ptr] "+A" (*(u##sz *)ptr), [ret] "=r" (ret)		\
+		: [val] "r" ((u##sz)(val))				\
+		: "memory");						\
+									\
+	return ret + val;						\
+}

BUT SEE:

   do {} while (0) advice is over-stated in a few situations:

   The more obvious case is macros, like MODULE_PARM_DESC, invoked at
   file-scope, where C disallows code (it must be in functions).  See
   $exceptions if you have one to add by name.

   More troublesome is declarative macros used at top of new scope,
   like DECLARE_PER_CPU.  These might just compile with a do-while-0
   wrapper, but would be incorrect.  Most of these are handled by
   detecting struct,union,etc declaration primitives in $exceptions.

   Theres also macros called inside an if (block), which "return" an
   expression.  These cannot do-while, and need a ({}) wrapper.

   Enjoy this qualification while we work to improve our heuristics.

ERROR: Macros with complex values should be enclosed in parentheses
#100: FILE: arch/riscv/include/asm/percpu.h:77:
+#define PERCPU_RET_OP(name, amo_insn)					\
+	__PERCPU_AMO_RET_OP_CASE(.w, name, 32, amo_insn)		\
+	__PERCPU_AMO_RET_OP_CASE(.d, name, 64, amo_insn)

BUT SEE:

   do {} while (0) advice is over-stated in a few situations:

   The more obvious case is macros, like MODULE_PARM_DESC, invoked at
   file-scope, where C disallows code (it must be in functions).  See
   $exceptions if you have one to add by name.

   More troublesome is declarative macros used at top of new scope,
   like DECLARE_PER_CPU.  These might just compile with a do-while-0
   wrapper, but would be incorrect.  Most of these are handled by
   detecting struct,union,etc declaration primitives in $exceptions.

   Theres also macros called inside an if (block), which "return" an
   expression.  These cannot do-while, and need a ({}) wrapper.

   Enjoy this qualification while we work to improve our heuristics.

CHECK: Macro argument 'val_type' may be better as '(val_type)' to avoid precedence issues
#114: FILE: arch/riscv/include/asm/percpu.h:91:
+#define PERCPU_8_16_OP(name, amo_insn, sz, sfx, val_type, new_val_expr, asm_op)			\
+static inline void __percpu_##name##_amo_case_##sz(void *ptr, unsigned long val)		\
+{												\
+	if (IS_ENABLED(CONFIG_RISCV_ISA_ZABHA) &&						\
+		riscv_has_extension_unlikely(RISCV_ISA_EXT_ZABHA)) {				\
+		asm volatile ("amo" #amo_insn #sfx " zero, %[val], %[ptr]"			\
+			: [ptr] "+A"(*(val_type *)ptr)						\
+			: [val] "r"((val_type)((new_val_expr) & PERCPU_8_16_GET_MASK(sz)))	\
+			: "memory");								\
+	} else {										\
+		u32 *ptr32 = PERCPU_8_16_GET_PTR32(ptr);					\
+		const unsigned long shift = PERCPU_8_16_GET_SHIFT(ptr);				\
+		const u32 mask = PERCPU_8_16_GET_MASK(sz) << shift;				\
+		const val_type val_trunc = (val_type)((new_val_expr)				\
+					   & PERCPU_8_16_GET_MASK(sz));				\
+		u32 old, new;									\
+		val_type field;									\
+												\
+		asm volatile (									\
+			"0: lr.w %0, %2\n"							\
+			"and %3, %0, %4\n"							\
+			"srl %3, %3, %5\n"							\
+			#asm_op " %3, %3, %6\n"							\
+			"and %3, %3, %8\n"							\
+			"sll %3, %3, %5\n"							\
+			"and %1, %0, %7\n"							\
+			"or %1, %1, %3\n"							\
+			"sc.w %1, %1, %2\n"							\
+			"bnez %1, 0b\n"								\
+			: "=&r"(old), "=&r"(new), "+A"(*ptr32), "=&r"(field)			\
+			: "r"(mask), "r"(shift), "r"(val_trunc), "r"(~mask),			\
+			  "r"(PERCPU_8_16_GET_MASK(sz))						\
+			: "memory");								\
+		}										\
+}

WARNING: Non-declarative macros with multiple statements should be enclosed in a do - while loop
#150: FILE: arch/riscv/include/asm/percpu.h:127:
+#define PERCPU_OP_8_16(op_name, op, expr, final_op)			\
+	PERCPU_8_16_OP(op_name, op, 8, .b, u8, expr, final_op);		\
+	PERCPU_8_16_OP(op_name, op, 16, .h, u16, expr, final_op)

BUT SEE:

   do {} while (0) advice is over-stated in a few situations:

   The more obvious case is macros, like MODULE_PARM_DESC, invoked at
   file-scope, where C disallows code (it must be in functions).  See
   $exceptions if you have one to add by name.

   More troublesome is declarative macros used at top of new scope,
   like DECLARE_PER_CPU.  These might just compile with a do-while-0
   wrapper, but would be incorrect.  Most of these are handled by
   detecting struct,union,etc declaration primitives in $exceptions.

   Theres also macros called inside an if (block), which "return" an
   expression.  These cannot do-while, and need a ({}) wrapper.

   Enjoy this qualification while we work to improve our heuristics.

total: 3 errors, 2 warnings, 1 checks, 296 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
      mechanically convert to the typical style using --fix or --fix-inplace.

Commit 7c674a8fd072 ("riscv: introduce percpu.h into include/asm") has style problems, please review.

NOTE: Ignored message types: ALLOC_SIZEOF_STRUCT CAMELCASE COMMIT_LOG_LONG_LINE GIT_COMMIT_ID MACRO_ARG_REUSE NO_AUTHOR_SIGN_OFF

NOTE: If any of the errors are false positives, please report
      them to the maintainer, see CHECKPATCH in MAINTAINERS.
CHECK: Macro argument 'val_type' may be better as '(val_type)' to avoid precedence issues
ERROR: Macros with complex values should be enclosed in parentheses
WARNING: Non-declarative macros with multiple statements should be enclosed in a do - while loop
WARNING: added, moved or deleted file(s), does MAINTAINERS need updating?

linux-riscv-bot · 2026-07-02T09:26:50Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
dtb-warn-rv64
Desc: Checks for Device Tree warnings/errors
Duration: 78.85 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:26:52Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
header-inline
Desc: Detects static functions without inline keyword in header files
Duration: 0.21 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:26:55Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
kdoc
Desc: Detects for kdoc errors
Duration: 0.73 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:26:56Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
module-param
Desc: Detect module_param changes
Duration: 0.21 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:26:58Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
verify-fixes
Desc: Verifies that the Fixes: tags exist
Duration: 0.20 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:27:00Z

Patch 2: "[v7,2/3] riscv: introduce percpu.h into include/asm"
verify-signedoff
Desc: Verifies that Signed-off-by: tags are correct
Duration: 0.24 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T09:29:14Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
build-rv32-defconfig
Desc: Builds riscv32 defconfig
Duration: 119.71 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T10:09:04Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
build-rv64-clang-allmodconfig
Desc: Builds riscv64 allmodconfig with Clang, and checks for errors and added warnings
Duration: 2387.67 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T10:58:54Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
build-rv64-gcc-allmodconfig
Desc: Builds riscv64 allmodconfig with GCC, and checks for errors and added warnings
Duration: 2989.25 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T10:59:18Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
build-rv64-nommu-k210-defconfig
Desc: Builds riscv64 defconfig with NOMMU for K210
Duration: 22.18 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T10:59:44Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
build-rv64-nommu-k210-virt
Desc: Builds riscv64 defconfig with NOMMU for the virt platform
Duration: 23.56 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T10:59:48Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
checkpatch
Desc: Runs checkpatch.pl on the patch
Duration: 2.67 seconds
Result: WARNING
Output:

CHECK: Consider using #include <linux/thread_info.h> instead of <asm/thread_info.h>
#55: FILE: arch/riscv/include/asm/percpu.h:14:
+#include <asm/thread_info.h>

total: 0 errors, 0 warnings, 1 checks, 114 lines checked

NOTE: For some of the reported defects, checkpatch may be able to
      mechanically convert to the typical style using --fix or --fix-inplace.

Commit f07a7f9c9c8e ("riscv: store percpu offset into thread_info") has style problems, please review.

NOTE: Ignored message types: ALLOC_SIZEOF_STRUCT CAMELCASE COMMIT_LOG_LONG_LINE GIT_COMMIT_ID MACRO_ARG_REUSE NO_AUTHOR_SIGN_OFF

NOTE: If any of the errors are false positives, please report
      them to the maintainer, see CHECKPATCH in MAINTAINERS.
total: 0 errors, 0 warnings, 1 checks, 114 lines checked
CHECK: Consider using #include <linux/thread_info.h> instead of <asm/thread_info.h>

linux-riscv-bot · 2026-07-02T11:01:12Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
dtb-warn-rv64
Desc: Checks for Device Tree warnings/errors
Duration: 81.91 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T11:01:13Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
header-inline
Desc: Detects static functions without inline keyword in header files
Duration: 0.26 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T11:01:16Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
kdoc
Desc: Detects for kdoc errors
Duration: 0.77 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T11:01:18Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
module-param
Desc: Detect module_param changes
Duration: 0.26 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T11:01:20Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
verify-fixes
Desc: Verifies that the Fixes: tags exist
Duration: 0.23 seconds
Result: PASS

linux-riscv-bot · 2026-07-02T11:01:22Z

Patch 3: "[v7,3/3] riscv: store percpu offset into thread_info"
verify-signedoff
Desc: Verifies that Signed-off-by: tags are correct
Duration: 0.27 seconds
Result: PASS

cuiyunhui added 3 commits July 2, 2026 06:22

linux-riscv-bot closed this Jul 3, 2026

linux-riscv-bot deleted the pw1120193 branch July 3, 2026 13:26

Uh oh!

Conversation

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

linux-riscv-bot commented Jul 2, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects