Re: [PATCH] powerpc/32: Restore clearing of MSR[RI] in interrupt_exit_kernel_prepare()

[Christophe Leroy <christophe.leroy@csgroup.eu>]

Le 01/09/2025 à 12:37, Christophe Leroy a écrit :
> Commit 13799748b957 ("powerpc/64: use interrupt restart table to speed
> up return from interrupt") removed the inconditional clearing of MSR[RI]
> when returning from interrupt into kernel. But powerpc/32 doesn't
> implement interrupt restart table hence still need MSR[RI] to be
> cleared.

This is needed but not enough. A synchronisation instruction is needed 
after clearing MSR[RI], otherwise the Instruction Breakpoint still 
triggers on a few instructions passed clearing of RI, allthough the 
LCTRL2[BRKNOMSK] is unset, meaning "Masked mode, breakpoints are 
recognized only when MSR[RI] =1 (reset value)"

We could as a isync() after __hard_EE_RI_disable(), but to minimise the 
impact it will be better to move clearing of MSR[RI] in the low-level 
assembly part of interrupt exit, just before the mtspr to SRR0 and SRR1, 
as was done before commit 6f76a01173cc ("powerpc/syscall: implement 
system call entry/exit logic in C for PPC32")

> 
> Fixes: 13799748b957 ("powerpc/64: use interrupt restart table to speed up return from interrupt")
> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
> ---
>   arch/powerpc/kernel/interrupt.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/arch/powerpc/kernel/interrupt.c b/arch/powerpc/kernel/interrupt.c
> index e0c681d0b076..e63bfde13e03 100644
> --- a/arch/powerpc/kernel/interrupt.c
> +++ b/arch/powerpc/kernel/interrupt.c
> @@ -443,6 +443,9 @@ notrace unsigned long interrupt_exit_kernel_prepare(struct pt_regs *regs)
>   
>   		if (unlikely(stack_store))
>   			__hard_EE_RI_disable();
> +#else
> +	} else {
> +		__hard_EE_RI_disable();
>   #endif /* CONFIG_PPC64 */
>   	}
>