[PATCH v5.10 RESEND 0/2] x86/irq: Plug vector setup race

Jinjie Ruan posted 2 patches 1 month, 1 week ago
There is a newer version of this series
arch/x86/kernel/irq.c | 70 ++++++++++++++++++++++++++++++++++---------
1 file changed, 56 insertions(+), 14 deletions(-)
[PATCH v5.10 RESEND 0/2] x86/irq: Plug vector setup race
Posted by Jinjie Ruan 1 month, 1 week ago
There is a vector setup race, which overwrites the interrupt
descriptor in the per CPU vector array resulting in a disfunctional device.

CPU0				CPU1
				interrupt is raised in APIC IRR
				but not handled
  free_irq()
    per_cpu(vector_irq, CPU1)[vector] = VECTOR_SHUTDOWN;

  request_irq()			common_interrupt()
  				  d = this_cpu_read(vector_irq[vector]);

    per_cpu(vector_irq, CPU1)[vector] = desc;

    				  if (d == VECTOR_SHUTDOWN)
				    this_cpu_write(vector_irq[vector], VECTOR_UNUSED);

free_irq() cannot observe the pending vector in the CPU1 APIC as there is
no way to query the remote CPUs APIC IRR.

This requires that request_irq() uses the same vector/CPU as the one which
was freed, but this also can be triggered by a spurious interrupt.

Interestingly enough this problem managed to be hidden for more than a
decade.

Prevent this by reevaluating vector_irq under the vector lock, which is
held by the interrupt activation code when vector_irq is updated.

The first patch provides context for subsequent real bugfix patch.

Fixes: 9345005f4eed ("x86/irq: Fix do_IRQ() interrupt warning for cpu hotplug retriggered irqs")
Cc: stable@vger.kernel.org#5.10.x
Cc: gregkh@linuxfoundation.org

v1 -> RESEND
- Add upstream commit ID.

Jacob Pan (1):
  x86/irq: Factor out handler invocation from common_interrupt()

Thomas Gleixner (1):
  x86/irq: Plug vector setup race

 arch/x86/kernel/irq.c | 70 ++++++++++++++++++++++++++++++++++---------
 1 file changed, 56 insertions(+), 14 deletions(-)

-- 
2.34.1
Re: [PATCH v5.10 RESEND 0/2] x86/irq: Plug vector setup race
Posted by Greg KH 1 month, 1 week ago
On Fri, Aug 22, 2025 at 03:33:02AM +0000, Jinjie Ruan wrote:
> There is a vector setup race, which overwrites the interrupt
> descriptor in the per CPU vector array resulting in a disfunctional device.
> 
> CPU0				CPU1
> 				interrupt is raised in APIC IRR
> 				but not handled
>   free_irq()
>     per_cpu(vector_irq, CPU1)[vector] = VECTOR_SHUTDOWN;
> 
>   request_irq()			common_interrupt()
>   				  d = this_cpu_read(vector_irq[vector]);
> 
>     per_cpu(vector_irq, CPU1)[vector] = desc;
> 
>     				  if (d == VECTOR_SHUTDOWN)
> 				    this_cpu_write(vector_irq[vector], VECTOR_UNUSED);
> 
> free_irq() cannot observe the pending vector in the CPU1 APIC as there is
> no way to query the remote CPUs APIC IRR.
> 
> This requires that request_irq() uses the same vector/CPU as the one which
> was freed, but this also can be triggered by a spurious interrupt.
> 
> Interestingly enough this problem managed to be hidden for more than a
> decade.
> 
> Prevent this by reevaluating vector_irq under the vector lock, which is
> held by the interrupt activation code when vector_irq is updated.
> 
> The first patch provides context for subsequent real bugfix patch.
> 
> Fixes: 9345005f4eed ("x86/irq: Fix do_IRQ() interrupt warning for cpu hotplug retriggered irqs")
> Cc: stable@vger.kernel.org#5.10.x
> Cc: gregkh@linuxfoundation.org
> 
> v1 -> RESEND
> - Add upstream commit ID.
> 
> Jacob Pan (1):
>   x86/irq: Factor out handler invocation from common_interrupt()
> 
> Thomas Gleixner (1):
>   x86/irq: Plug vector setup race
> 
>  arch/x86/kernel/irq.c | 70 ++++++++++++++++++++++++++++++++++---------
>  1 file changed, 56 insertions(+), 14 deletions(-)
> 
> -- 
> 2.34.1
> 
> 

Dropping as I didn't take the patches for later kernels either.

greg k-h