Re: [Patch v3 3/3] iommu: avoid taking iova_rbtree_lock twice

[John Garry]

On 21/01/2020 09:56, Robin Murphy wrote:
> On 18/12/2019 4:39 am, Cong Wang wrote:
> > Both find_iova() and __free_iova() take iova_rbtree_lock,
> > there is no reason to take and release it twice inside
> > free_iova().
> >
> > Fold them into one critical section by calling the unlock
> > versions instead.
>
> Makes sense to me.
>
> Reviewed-by: Robin Murphy <robin.murphy@xxxxxxx>

Reviewed-by: John Garry <john.garry@xxxxxxxxxx>

Could we at least get this patch picked up? It should be ok to take in 
isolation, since there is some debate on the other 2 patches in this 
series. Thanks

> > Cc: Joerg Roedel <joro@xxxxxxxxxx>
> > Cc: John Garry <john.garry@xxxxxxxxxx>
> > Signed-off-by: Cong Wang <xiyou.wangcong@xxxxxxxxx>
> > ---
> > Â drivers/iommu/iova.c | 8 ++++++--
> > Â 1 file changed, 6 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/iommu/iova.c b/drivers/iommu/iova.c
> > index 184d4c0e20b5..f46f8f794678 100644
> > --- a/drivers/iommu/iova.c
> > +++ b/drivers/iommu/iova.c
> > @@ -390,10 +390,14 @@ EXPORT_SYMBOL_GPL(__free_iova);
> > Â void
> > Â free_iova(struct iova_domain *iovad, unsigned long pfn)
> > Â {
> > -ÂÂÂ struct iova *iova = find_iova(iovad, pfn);
> > +ÂÂÂ unsigned long flags;
> > +ÂÂÂ struct iova *iova;
> > +ÂÂÂ spin_lock_irqsave(&iovad->iova_rbtree_lock, flags);
> > +ÂÂÂ iova = private_find_iova(iovad, pfn);
> > ÂÂÂÂÂ if (iova)
> > -ÂÂÂÂÂÂÂ __free_iova(iovad, iova);
> > +ÂÂÂÂÂÂÂ private_free_iova(iovad, iova);
> > +ÂÂÂ spin_unlock_irqrestore(&iovad->iova_rbtree_lock, flags);
> > Â }
> > Â EXPORT_SYMBOL_GPL(free_iova);
> .