[PATCH] arm64: entry: Clean out some indirection

[PATCH] arm64: entry: Clean out some indirection

[Linus Walleij]

The conversion to generic IRQ entry left some functions
in the EL1 (kernel) IRQ entry path very shallow, so drop
the __inner_functions() where appropriate, saving some
time and stack.

This is not a fix but an optimization.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm64/kernel/entry-common.c | 22 +++-------------------
 1 file changed, 3 insertions(+), 19 deletions(-)

diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c
index f546a914f04174e37bf3578490545edebb66afd1..e34dfd93c522ae58a5f1c914fe4a5bfc6522cdae 100644
--- a/arch/arm64/kernel/entry-common.c
+++ b/arch/arm64/kernel/entry-common.c
@@ -38,16 +38,11 @@
  * This is intended to match the logic in irqentry_enter(), handling the kernel
  * mode transitions only.
  */
-static __always_inline irqentry_state_t __enter_from_kernel_mode(struct pt_regs *regs)
-{
-	return irqentry_enter(regs);
-}
-
 static noinstr irqentry_state_t enter_from_kernel_mode(struct pt_regs *regs)
 {
 	irqentry_state_t state;
 
-	state = __enter_from_kernel_mode(regs);
+	state = irqentry_enter(regs);
 	mte_check_tfsr_entry();
 	mte_disable_tco_entry(current);
 
@@ -62,17 +57,11 @@ static noinstr irqentry_state_t enter_from_kernel_mode(struct pt_regs *regs)
  * This is intended to match the logic in irqentry_exit(), handling the kernel
  * mode transitions only, and with preemption handled elsewhere.
  */
-static __always_inline void __exit_to_kernel_mode(struct pt_regs *regs,
-						  irqentry_state_t state)
-{
-	irqentry_exit(regs, state);
-}
-
 static void noinstr exit_to_kernel_mode(struct pt_regs *regs,
 					irqentry_state_t state)
 {
 	mte_check_tfsr_exit();
-	__exit_to_kernel_mode(regs, state);
+	irqentry_exit(regs, state);
 }
 
 /*
@@ -80,17 +69,12 @@ static void noinstr exit_to_kernel_mode(struct pt_regs *regs,
  * Before this function is called it is not safe to call regular kernel code,
  * instrumentable code, or any code which may trigger an exception.
  */
-static __always_inline void __enter_from_user_mode(struct pt_regs *regs)
+static __always_inline void arm64_enter_from_user_mode(struct pt_regs *regs)
 {
 	enter_from_user_mode(regs);
 	mte_disable_tco_entry(current);
 }
 
-static __always_inline void arm64_enter_from_user_mode(struct pt_regs *regs)
-{
-	__enter_from_user_mode(regs);
-}
-
 /*
  * Handle IRQ/context state management when exiting to user mode.
  * After this function returns it is not safe to call regular kernel code,

---
base-commit: 3a8660878839faadb4f1a6dd72c3179c1df56787
change-id: 20251014-arm64-skip-indirection-4ff88c27ad02

Best regards,
-- 
Linus Walleij <linus.walleij@linaro.org>

Re: [PATCH] arm64: entry: Clean out some indirection

[Linus Walleij]

On Tue, Oct 14, 2025 at 9:08 PM Linus Walleij <linus.walleij@linaro.org> wrote:

> The conversion to generic IRQ entry left some functions
> in the EL1 (kernel) IRQ entry path very shallow, so drop
> the __inner_functions() where appropriate, saving some
> time and stack.
>
> This is not a fix but an optimization.
>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

For some reason b4 didn't pick up Jinjie on the To: line,
sorry Jinjie!

(Excellent work with the generic entry by the way.)

Yours,
Linus Walleij

Re: [PATCH] arm64: entry: Clean out some indirection

[Will Deacon]

On Tue, Oct 14, 2025 at 09:08:30PM +0200, Linus Walleij wrote:
> The conversion to generic IRQ entry left some functions
> in the EL1 (kernel) IRQ entry path very shallow, so drop
> the __inner_functions() where appropriate, saving some
> time and stack.
> 
> This is not a fix but an optimization.
> 
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
> ---
>  arch/arm64/kernel/entry-common.c | 22 +++-------------------
>  1 file changed, 3 insertions(+), 19 deletions(-)

I think the code was originally structured to follow the same flow as
the generic code so that the two could be easily compared. Now that
we're moving over to the generic code, I agree that it makes sense to
clean things up.

> diff --git a/arch/arm64/kernel/entry-common.c b/arch/arm64/kernel/entry-common.c
> index f546a914f04174e37bf3578490545edebb66afd1..e34dfd93c522ae58a5f1c914fe4a5bfc6522cdae 100644
> --- a/arch/arm64/kernel/entry-common.c
> +++ b/arch/arm64/kernel/entry-common.c
> @@ -38,16 +38,11 @@
>   * This is intended to match the logic in irqentry_enter(), handling the kernel
>   * mode transitions only.

nit: this comment ^^^ (the reference to irqentry_enter()) also seems to
be out of date now.

>   */
> -static __always_inline irqentry_state_t __enter_from_kernel_mode(struct pt_regs *regs)
> -{
> -	return irqentry_enter(regs);
> -}
> -
>  static noinstr irqentry_state_t enter_from_kernel_mode(struct pt_regs *regs)
>  {
>  	irqentry_state_t state;
>  
> -	state = __enter_from_kernel_mode(regs);
> +	state = irqentry_enter(regs);
>  	mte_check_tfsr_entry();
>  	mte_disable_tco_entry(current);
>  
> @@ -62,17 +57,11 @@ static noinstr irqentry_state_t enter_from_kernel_mode(struct pt_regs *regs)
>   * This is intended to match the logic in irqentry_exit(), handling the kernel
>   * mode transitions only, and with preemption handled elsewhere.
>   */

Similarly for this one ^^^.

Will