Re: [PATCH V2 2/2] include: linux: Update gather only if it's not NULL

[Robin Murphy <robin.murphy@arm.com>]

On 01/08/2024 4:34 am, Ashish Mhetre wrote:
> Gather can be NULL when unmap is called for freeing old table while
> mapping. If it's NULL then there is no need to add page for syncing
> the TLB.

But that's only because __arm_lpae_unmap() is now choosing to 
over-invalidate the table entries for simplicity. I think it would make 
more sense to handle that at the callsite, e.g. "if (gather && 
!iommu_iotlb_gather_queued(gather))".

Also doesn't this mean that bisection is broken as-is since patch #1 on 
its own is going to blow up dereferencing NULL->pgsize when it gets here?

Thanks,
Robin.

> Signed-off-by: Ashish Mhetre <amhetre@nvidia.com>
> ---
>   include/linux/iommu.h | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/include/linux/iommu.h b/include/linux/iommu.h
> index 4d47f2c33311..2a28c1ef8517 100644
> --- a/include/linux/iommu.h
> +++ b/include/linux/iommu.h
> @@ -928,6 +928,9 @@ static inline void iommu_iotlb_gather_add_page(struct iommu_domain *domain,
>   					       struct iommu_iotlb_gather *gather,
>   					       unsigned long iova, size_t size)
>   {
> +	if (!gather)
> +		return;
> +
>   	/*
>   	 * If the new page is disjoint from the current range or is mapped at
>   	 * a different granularity, then sync the TLB so that the gather