Re: [PATCH for-4.22 2/2] xen/arm: vgic: free eSPIs using the bitmap index

On 10-Jul-26 10:31, Mykola Kvach wrote:
> Classic vGIC stores eSPIs in allocated_irqs after the regular virtual
> interrupts. vgic_reserve_virq() therefore translates an eSPI INTID to a
> compressed bitmap index before test_and_set_bit().
> 
> vgic_free_virq() still used the raw virtual INTID. Freeing INTID 4096
> would clear bit 4096 instead of the first eSPI allocation bit, which is
> outside allocated_irqs for a domain with eSPI support. That can leave
> the eSPI reserved and may corrupt memory.
> 
> Add the inverse of idx_to_virq() and use it in both reserve and free, so
> the allocation bitmap is indexed consistently. Also reject invalid
> virtual INTIDs before clearing the bitmap.
I think we can have it in 4.22 given that the release is postponed a bit.

> 
> Fixes: bdde400c6e1b ("xen/arm: vgic: add resource management for extended SPIs")
> Signed-off-by: Mykola Kvach <[email protected]>
> ---
>  xen/arch/arm/vgic.c | 27 ++++++++++++++++-----------
>  1 file changed, 16 insertions(+), 11 deletions(-)
> 
> diff --git a/xen/arch/arm/vgic.c b/xen/arch/arm/vgic.c
> index e5aca17dcb..b85710c6a7 100644
> --- a/xen/arch/arm/vgic.c
> +++ b/xen/arch/arm/vgic.c
> @@ -33,6 +33,16 @@ static inline unsigned int idx_to_virq(struct domain *d, unsigned int idx)
>      return idx;
>  }
>  
> +static inline unsigned int virq_to_idx(struct domain *d, unsigned int virq)
> +{
> +#ifdef CONFIG_GICV3_ESPI
Same as for patch 1: is_espi() shouldn't be protected as it protects itself.

~Michal