Re: Patch "irqchip/gic-v4: Make sure a VPE is locked when VMAPP is issued" has been added to the 6.6-stable tree

[Marc Zyngier <maz@kernel.org>]

On Sun, 08 Sep 2024 14:36:37 +0100,
Sasha Levin <sashal@kernel.org> wrote:
> 
> This is a note to let you know that I've just added the patch titled
> 
>     irqchip/gic-v4: Make sure a VPE is locked when VMAPP is issued
> 
> to the 6.6-stable tree which can be found at:
>     http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary
> 
> The filename of the patch is:
>      irqchip-gic-v4-make-sure-a-vpe-is-locked-when-vmapp-.patch
> and it can be found in the queue-6.6 subdirectory.
> 
> If you, or anyone else, feels it should not be added to the stable tree,
> please let <stable@vger.kernel.org> know about it.
> 
> 
> 
> commit 1a232324773145ff7ce59b6a1b52b3247223f9d4
> Author: Marc Zyngier <maz@kernel.org>
> Date:   Fri Jul 5 10:31:55 2024 +0100
> 
>     irqchip/gic-v4: Make sure a VPE is locked when VMAPP is issued
>     
>     [ Upstream commit a84a07fa3100d7ad46a3d6882af25a3df9c9e7e3 ]
>     
>     In order to make sure that vpe->col_idx is correctly sampled when a VMAPP
>     command is issued, the vpe_lock must be held for the VPE. This is now
>     possible since the introduction of the per-VM vmapp_lock, which can be
>     taken before vpe_lock in the correct locking order.
>     
>     Signed-off-by: Marc Zyngier <maz@kernel.org>
>     Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
>     Tested-by: Nianyao Tang <tangnianyao@huawei.com>
>     Link: https://lore.kernel.org/r/20240705093155.871070-4-maz@kernel.org
>     Signed-off-by: Sasha Levin <sashal@kernel.org>
> 
> diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c
> index e25dea0e50c7..1e0f0e1bf481 100644
> --- a/drivers/irqchip/irq-gic-v3-its.c
> +++ b/drivers/irqchip/irq-gic-v3-its.c
> @@ -1804,7 +1804,9 @@ static void its_map_vm(struct its_node *its, struct its_vm *vm)
>  		for (i = 0; i < vm->nr_vpes; i++) {
>  			struct its_vpe *vpe = vm->vpes[i];
>  
> -			its_send_vmapp(its, vpe, true);
> +			scoped_guard(raw_spinlock, &vpe->vpe_lock)
> +				its_send_vmapp(its, vpe, true);
> +
>  			its_send_vinvall(its, vpe);
>  		}
>  	}
> @@ -1825,8 +1827,10 @@ static void its_unmap_vm(struct its_node *its, struct its_vm *vm)
>  	if (!--vm->vlpi_count[its->list_nr]) {
>  		int i;
>  
> -		for (i = 0; i < vm->nr_vpes; i++)
> +		for (i = 0; i < vm->nr_vpes; i++) {
> +			guard(raw_spinlock)(&vm->vpes[i]->vpe_lock);
>  			its_send_vmapp(its, vm->vpes[i], false);
> +		}
>  	}
>  
>  	raw_spin_unlock_irqrestore(&vmovp_lock, flags);
>

No please.

Not only you are missing the essential part of the series (the patch
introducing the per-VM lock that this change relies on), you are also
missing the fixes that followed.

So please drop this patch from the 6.6 and 6.1 queues.

	M.

-- 
Without deviation from the norm, progress is not possible.