[PATCH v2] ARM: entry: expand comment in __switch_to

[PATCH v2] ARM: entry: expand comment in __switch_to

[Linus Walleij]

From: Linus Walleij <linus.walleij@linaro.org>

As per discussion between the developers in the mail thread
linked, expand the comment in __switch_to so that readers
of the code understand what is going on.

Link: https://lore.kernel.org/linux-arm-kernel/ZxDh9biUbf9W8gNN@J2N7QTR9R3/
Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
Changes in v2:
- Rebased on v6.19-rc1
- Link to v1: https://lore.kernel.org/r/20241028-comments-in-switch-to-v1-0-7280d09671a8@linaro.org
---
 arch/arm/kernel/entry-armv.S | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/arch/arm/kernel/entry-armv.S b/arch/arm/kernel/entry-armv.S
index ef6a657c8d13..5c016c68b5a3 100644
--- a/arch/arm/kernel/entry-armv.S
+++ b/arch/arm/kernel/entry-armv.S
@@ -557,9 +557,16 @@ ENTRY(__switch_to)
 	ldmia	r4, {r4 - sl, fp, ip, lr}	@ Load all regs saved previously
 #ifdef CONFIG_VMAP_STACK
 	@
-	@ Do a dummy read from the new stack while running from the old one so
-	@ that we can rely on do_translation_fault() to fix up any stale PMD
-	@ entries covering the vmalloc region.
+	@ For a non-lazy mm switch, check_vmalloc_seq() has ensured that
+	@ that the active mm's page tables have mappings for the prev
+	@ task's stack and the next task's stack.
+	@
+	@ For a lazy mm switch the active mm's page tables have mappings
+	@ for the prev task's stack but might not have mappings for the
+	@ new taks stack. Do a dummy read from the new stack while
+	@ running from the old stack so that we can rely on
+	@ do_translation_fault() to fix up any stale PMD entries
+	@ covering the vmalloc region.
 	@
 	ldr	r2, [ip]
 #ifdef CONFIG_KASAN_VMALLOC

---
base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8
change-id: 20241028-comments-in-switch-to-0e24480e8495

Best regards,
-- 
Linus Walleij <linusw@kernel.org>

Re: [PATCH v2] ARM: entry: expand comment in __switch_to

[Mark Rutland]

Hi Linus,

On Tue, Dec 16, 2025 at 03:27:37PM +0100, Linus Walleij wrote:
> From: Linus Walleij <linus.walleij@linaro.org>
> 
> As per discussion between the developers in the mail thread
> linked, expand the comment in __switch_to so that readers
> of the code understand what is going on.
> 
> Link: https://lore.kernel.org/linux-arm-kernel/ZxDh9biUbf9W8gNN@J2N7QTR9R3/
> Suggested-by: Mark Rutland <mark.rutland@arm.com>
> Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

This looks fine to me, with two minor nits noted below. With those
fixed, feel free to add:

  Acked-by: Mark Rutland <mark.rutland@arm.com>

... and I'd recommend dropping this into the patch system.

> ---
> Changes in v2:
> - Rebased on v6.19-rc1
> - Link to v1: https://lore.kernel.org/r/20241028-comments-in-switch-to-v1-0-7280d09671a8@linaro.org
> ---
>  arch/arm/kernel/entry-armv.S | 13 ++++++++++---
>  1 file changed, 10 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/arm/kernel/entry-armv.S b/arch/arm/kernel/entry-armv.S
> index ef6a657c8d13..5c016c68b5a3 100644
> --- a/arch/arm/kernel/entry-armv.S
> +++ b/arch/arm/kernel/entry-armv.S
> @@ -557,9 +557,16 @@ ENTRY(__switch_to)
>  	ldmia	r4, {r4 - sl, fp, ip, lr}	@ Load all regs saved previously
>  #ifdef CONFIG_VMAP_STACK
>  	@
> -	@ Do a dummy read from the new stack while running from the old one so
> -	@ that we can rely on do_translation_fault() to fix up any stale PMD
> -	@ entries covering the vmalloc region.
> +	@ For a non-lazy mm switch, check_vmalloc_seq() has ensured that
> +	@ that the active mm's page tables have mappings for the prev
> +	@ task's stack and the next task's stack.
> +	@
> +	@ For a lazy mm switch the active mm's page tables have mappings
> +	@ for the prev task's stack but might not have mappings for the
> +	@ new taks stack. Do a dummy read from the new stack while

Typo: s/task/task's/

> +	@ running from the old stack so that we can rely on
> +	@ do_translation_fault() to fix up any stale PMD entries
> +	@ covering the vmalloc region.

Ard requested in:

  https://lore.kernel.org/linux-arm-kernel/CAMj1kXFn=pfJqTrMrnqT+OLi3UqAq0PieROQhbC2M3B2uAzyEQ@mail.gmail.com/

That we change this last part from:

  to fix up any stale PMD entries covering the vmalloc region.

... to:

  to populate missing PMD entries covering the new task's stack in the
  old task's page tables.

... and I think that'd be worthwhile.

Mark.

>  	@
>  	ldr	r2, [ip]
>  #ifdef CONFIG_KASAN_VMALLOC
> 
> ---
> base-commit: 8f0b4cce4481fb22653697cced8d0d04027cb1e8
> change-id: 20241028-comments-in-switch-to-0e24480e8495
> 
> Best regards,
> -- 
> Linus Walleij <linusw@kernel.org>
>