[PATCH] s390: Revert "s390/irq/idle: Remove psw bits early"

[PATCH] s390: Revert "s390/irq/idle: Remove psw bits early"

[Heiko Carstens]

This reverts commit d8b5cf9c63143fae54a734c41e3bb55cf3f365c7.

Mikhail Zaslonko reported that linux-next doesn't boot anymore [2]. Reason
for this is recent change [2] was supposed to slightly optimize the irq
entry/exit path by removing some psw bits early in case of an idle exit.

This however is incorrect since irqentry_exit() requires the correct old
psw state at irq entry. Otherwise the embedded regs_irqs_disabled() will
not provide the correct result.

With linux-next and HRTIMER_REARM_DEFERRED this leads to the observed boot
problems, however the commit is broken in any case.

Revert the commit which introduced this.

Thanks to Peter Zijlstra for pointing out that this is a bug in the s390
entry code.

Fixes: d8b5cf9c6314 ("s390/irq/idle: Remove psw bits early") [1]
Reported-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Reported-by: Peter Zijlstra <peterz@infradead.org>
Closes: https://lore.kernel.org/r/af549a19-db99-4b16-8511-bf315177a13e@linux.ibm.com/ [2]
Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
---
 arch/s390/kernel/irq.c | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
index 7fdf960191d3..d10a17e6531d 100644
--- a/arch/s390/kernel/irq.c
+++ b/arch/s390/kernel/irq.c
@@ -147,10 +147,8 @@ void noinstr do_io_irq(struct pt_regs *regs)
 	bool from_idle;
 
 	from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT);
-	if (from_idle) {
+	if (from_idle)
 		update_timer_idle();
-		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
-	}
 
 	irq_enter_rcu();
 
@@ -176,6 +174,9 @@ void noinstr do_io_irq(struct pt_regs *regs)
 
 	set_irq_regs(old_regs);
 	irqentry_exit(regs, state);
+
+	if (from_idle)
+		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
 }
 
 void noinstr do_ext_irq(struct pt_regs *regs)
@@ -185,10 +186,8 @@ void noinstr do_ext_irq(struct pt_regs *regs)
 	bool from_idle;
 
 	from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT);
-	if (from_idle) {
+	if (from_idle)
 		update_timer_idle();
-		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
-	}
 
 	irq_enter_rcu();
 
@@ -210,6 +209,9 @@ void noinstr do_ext_irq(struct pt_regs *regs)
 	irq_exit_rcu();
 	set_irq_regs(old_regs);
 	irqentry_exit(regs, state);
+
+	if (from_idle)
+		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
 }
 
 static void show_msi_interrupt(struct seq_file *p, int irq)
-- 
2.51.0

Re: [PATCH] s390: Revert "s390/irq/idle: Remove psw bits early"

[Vasily Gorbik]

On Fri, Mar 06, 2026 at 12:19:19PM +0100, Heiko Carstens wrote:
> This reverts commit d8b5cf9c63143fae54a734c41e3bb55cf3f365c7.
> 
> Mikhail Zaslonko reported that linux-next doesn't boot anymore [2]. Reason
> for this is recent change [2] was supposed to slightly optimize the irq
> entry/exit path by removing some psw bits early in case of an idle exit.
> 
> This however is incorrect since irqentry_exit() requires the correct old
> psw state at irq entry. Otherwise the embedded regs_irqs_disabled() will
> not provide the correct result.
> 
> With linux-next and HRTIMER_REARM_DEFERRED this leads to the observed boot
> problems, however the commit is broken in any case.
> 
> Revert the commit which introduced this.
> 
> Thanks to Peter Zijlstra for pointing out that this is a bug in the s390
> entry code.
> 
> Fixes: d8b5cf9c6314 ("s390/irq/idle: Remove psw bits early") [1]
> Reported-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
> Reported-by: Peter Zijlstra <peterz@infradead.org>
> Closes: https://lore.kernel.org/r/af549a19-db99-4b16-8511-bf315177a13e@linux.ibm.com/ [2]
> Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
> ---
>  arch/s390/kernel/irq.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)

Acked-by: Vasily Gorbik <gor@linux.ibm.com>
Applied, thank you!

Re: [PATCH] s390: Revert "s390/irq/idle: Remove psw bits early"

[Mikhail Zaslonko]

On 06-Mar-26 12:19, Heiko Carstens wrote:
> This reverts commit d8b5cf9c63143fae54a734c41e3bb55cf3f365c7.
> 
> Mikhail Zaslonko reported that linux-next doesn't boot anymore [2]. Reason
> for this is recent change [2] was supposed to slightly optimize the irq
> entry/exit path by removing some psw bits early in case of an idle exit.
> 
> This however is incorrect since irqentry_exit() requires the correct old
> psw state at irq entry. Otherwise the embedded regs_irqs_disabled() will
> not provide the correct result.
> 
> With linux-next and HRTIMER_REARM_DEFERRED this leads to the observed boot
> problems, however the commit is broken in any case.
> 
> Revert the commit which introduced this.
> 
> Thanks to Peter Zijlstra for pointing out that this is a bug in the s390
> entry code.
> 
> Fixes: d8b5cf9c6314 ("s390/irq/idle: Remove psw bits early") [1]
> Reported-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
> Reported-by: Peter Zijlstra <peterz@infradead.org>
> Closes: https://lore.kernel.org/r/af549a19-db99-4b16-8511-bf315177a13e@linux.ibm.com/ [2]
> Signed-off-by: Heiko Carstens <hca@linux.ibm.com>
> ---
>  arch/s390/kernel/irq.c | 14 ++++++++------
>  1 file changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/arch/s390/kernel/irq.c b/arch/s390/kernel/irq.c
> index 7fdf960191d3..d10a17e6531d 100644
> --- a/arch/s390/kernel/irq.c
> +++ b/arch/s390/kernel/irq.c
> @@ -147,10 +147,8 @@ void noinstr do_io_irq(struct pt_regs *regs)
>  	bool from_idle;
>  
>  	from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT);
> -	if (from_idle) {
> +	if (from_idle)
>  		update_timer_idle();
> -		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
> -	}
>  
>  	irq_enter_rcu();
>  
> @@ -176,6 +174,9 @@ void noinstr do_io_irq(struct pt_regs *regs)
>  
>  	set_irq_regs(old_regs);
>  	irqentry_exit(regs, state);
> +
> +	if (from_idle)
> +		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
>  }
>  
>  void noinstr do_ext_irq(struct pt_regs *regs)
> @@ -185,10 +186,8 @@ void noinstr do_ext_irq(struct pt_regs *regs)
>  	bool from_idle;
>  
>  	from_idle = test_and_clear_cpu_flag(CIF_ENABLED_WAIT);
> -	if (from_idle) {
> +	if (from_idle)
>  		update_timer_idle();
> -		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
> -	}
>  
>  	irq_enter_rcu();
>  
> @@ -210,6 +209,9 @@ void noinstr do_ext_irq(struct pt_regs *regs)
>  	irq_exit_rcu();
>  	set_irq_regs(old_regs);
>  	irqentry_exit(regs, state);
> +
> +	if (from_idle)
> +		regs->psw.mask &= ~(PSW_MASK_EXT | PSW_MASK_IO | PSW_MASK_WAIT);
>  }
>  
>  static void show_msi_interrupt(struct seq_file *p, int irq)

Indeed, with this revert no more kernel-next boot problems observed.

Acked-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>
Tested-by: Mikhail Zaslonko <zaslonko@linux.ibm.com>