[PATCH v19 11/14] arm64: syscall: Simplify el0_svc_common() syscall exit path

Jinjie Ruan ruanjinjie at huawei.com
Mon Sep 21 20:55:07 PDT 2026


Remove the redundant nested conditional check within the system call
exit path of el0_svc_common() to streamline the exit sequence.

The fast-path block is guarded by `!IS_ENABLED(CONFIG_DEBUG_RSEQ)`,
so CONFIG_DEBUG_RSEQ is guaranteed to be disabled and the call of
rseq_syscall() is a no-op. Under this constraint, the code logic
inside the block becomes completely identical to what
the arm64_syscall_exit_to_user_mode_work() helper already does.
Replace that nested logic with a direct invocation of the helper,
eliminating redundant code.

Before:

| if (... && !IS_ENABLED(CONFIG_DEBUG_RSEQ)) {
| 	flags = read_thread_flags();
| 	if (unlikely(flags & _TIF_SYSCALL_EXIT_WORK) || flags & _TIF_SINGLESTEP)
| 		arm64_syscall_exit_to_user_mode_work(regs);
| 	return;
| }
| trace_exit:
|	arm64_syscall_exit_to_user_mode_work(regs);

After simplify:

| if (... && !IS_ENABLED(CONFIG_DEBUG_RSEQ)) {
|	arm64_syscall_exit_to_user_mode_work(regs);
|	return;
| }
| trace_exit:
|	arm64_syscall_exit_to_user_mode_work(regs);

Furthermore, Since both the conditional fast-path and the fallback
slow-path now uniformly invoke arm64_syscall_exit_to_user_mode_work(),
this explicit conditional branch is entirely redundant regardless of
whether the evaluation is true or false. Removing it collapses
the duplicated logic into a single, unconditional path.

Also remove has_syscall_work().

No functional changes.

Cc: Mark Rutland <mark.rutland at arm.com>
Cc: Will Deacon <will at kernel.org>
Cc: Catalin Marinas <catalin.marinas at arm.com>
Cc: Ada Couprie Diaz <ada.coupriediaz at arm.com>
Cc: Linus Walleij <linusw at kernel.org>
Cc: Yeoreum Yun <yeoreum.yun at arm.com>
Cc: Can Peng <pengcan at kylinos.cn>
Cc: Li Qiang <liqiang01 at kylinos.cn>
Cc: Ryan Roberts <ryan.roberts at arm.com>
Reviewed-by: Linus Walleij <linusw at kernel.org>
Signed-off-by: Jinjie Ruan <ruanjinjie at huawei.com>
---
 arch/arm64/kernel/syscall.c | 20 +-------------------
 1 file changed, 1 insertion(+), 19 deletions(-)

diff --git a/arch/arm64/kernel/syscall.c b/arch/arm64/kernel/syscall.c
index f1f0b6f3e32b..3488afd45d20 100644
--- a/arch/arm64/kernel/syscall.c
+++ b/arch/arm64/kernel/syscall.c
@@ -54,11 +54,6 @@ static void invoke_syscall(struct pt_regs *regs, unsigned int scno,
 	syscall_set_return_value(current, regs, 0, ret);
 }
 
-static inline bool has_syscall_work(unsigned long flags)
-{
-	return unlikely(flags & _TIF_SYSCALL_WORK);
-}
-
 static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
 			   const syscall_fn_t syscall_table[])
 {
@@ -95,7 +90,7 @@ static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
 		return;
 	}
 
-	if (has_syscall_work(flags)) {
+	if (unlikely(flags & _TIF_SYSCALL_WORK)) {
 		/*
 		 * The de-facto standard way to skip a system call using ptrace
 		 * is to set the system call to -1 (NO_SYSCALL) and set x0 to a
@@ -119,19 +114,6 @@ static void el0_svc_common(struct pt_regs *regs, int scno, int sc_nr,
 	}
 
 	invoke_syscall(regs, scno, sc_nr, syscall_table);
-
-	/*
-	 * The tracing status may have changed under our feet, so we have to
-	 * check again. However, if we were tracing entry, then we always trace
-	 * exit regardless, as the old entry assembly did.
-	 */
-	if (!has_syscall_work(flags) && !IS_ENABLED(CONFIG_DEBUG_RSEQ)) {
-		flags = read_thread_flags();
-		if (unlikely(flags & _TIF_SYSCALL_EXIT_WORK) || flags & _TIF_SINGLESTEP)
-			arm64_syscall_exit_to_user_mode_work(regs);
-		return;
-	}
-
 trace_exit:
 	arm64_syscall_exit_to_user_mode_work(regs);
 }
-- 
2.34.1




More information about the linux-arm-kernel mailing list