Virtualization - Sep 2019 - [PATCH 1/5] iommu: Implement iommu_put_resv_regions_simple()

On Thu, Aug 29, 2019 at 01:17:48PM +0200, Thierry Reding
wrote:
> From: Thierry Reding <treding at nvidia.com>
> 
> Implement a generic function for removing reserved regions. This can be
> used by drivers that don't do anything fancy with these regions other
> than allocating memory for them.
> 
> Signed-off-by: Thierry Reding <treding at nvidia.com>
> ---
>  drivers/iommu/iommu.c | 19 +++++++++++++++++++
>  include/linux/iommu.h |  2 ++
>  2 files changed, 21 insertions(+)
> 
> diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
> index 0f585b614657..73a2a6b13507 100644
> --- a/drivers/iommu/iommu.c
> +++ b/drivers/iommu/iommu.c
> @@ -2170,6 +2170,25 @@ void iommu_put_resv_regions(struct device *dev,
struct list_head *list)
>  		ops->put_resv_regions(dev, list);
>  }
>  
> +/**
> + * iommu_put_resv_regions_simple - Reserved region driver helper
> + * @dev: device for which to free reserved regions
> + * @list: reserved region list for device
> + *
> + * IOMMU drivers can use this to implement their .put_resv_regions()
callback
> + * for simple reservations. Memory allocated for each reserved region will
be
> + * freed. If an IOMMU driver allocates additional resources per region, it
is
> + * going to have to implement a custom callback.
> + */
> +void iommu_put_resv_regions_simple(struct device *dev, struct list_head
*list)
> +{
> +	struct iommu_resv_region *entry, *next;
> +
> +	list_for_each_entry_safe(entry, next, list, list)
> +		kfree(entry);
> +}
> +EXPORT_SYMBOL(iommu_put_resv_regions_simple);
Can you call this directly from iommu_put_resv_regions() if the function
pointer in ops is NULL? That would save having to plumb the default callback
into a bunch of drivers.

Will

On Wed, Sep 18, 2019 at 04:37:38PM +0100, Will Deacon
wrote:
> On Thu, Aug 29, 2019 at 01:17:48PM +0200, Thierry Reding wrote:
> > From: Thierry Reding <treding at nvidia.com>
> > 
> > Implement a generic function for removing reserved regions. This can
be
> > used by drivers that don't do anything fancy with these regions
other
> > than allocating memory for them.
> > 
> > Signed-off-by: Thierry Reding <treding at nvidia.com>
> > ---
> >  drivers/iommu/iommu.c | 19 +++++++++++++++++++
> >  include/linux/iommu.h |  2 ++
> >  2 files changed, 21 insertions(+)
> > 
> > diff --git a/drivers/iommu/iommu.c b/drivers/iommu/iommu.c
> > index 0f585b614657..73a2a6b13507 100644
> > --- a/drivers/iommu/iommu.c
> > +++ b/drivers/iommu/iommu.c
> > @@ -2170,6 +2170,25 @@ void iommu_put_resv_regions(struct device *dev,
struct list_head *list)
> >  		ops->put_resv_regions(dev, list);
> >  }
> >  
> > +/**
> > + * iommu_put_resv_regions_simple - Reserved region driver helper
> > + * @dev: device for which to free reserved regions
> > + * @list: reserved region list for device
> > + *
> > + * IOMMU drivers can use this to implement their .put_resv_regions()
callback
> > + * for simple reservations. Memory allocated for each reserved region
will be
> > + * freed. If an IOMMU driver allocates additional resources per
region, it is
> > + * going to have to implement a custom callback.
> > + */
> > +void iommu_put_resv_regions_simple(struct device *dev, struct
list_head *list)
> > +{
> > +	struct iommu_resv_region *entry, *next;
> > +
> > +	list_for_each_entry_safe(entry, next, list, list)
> > +		kfree(entry);
> > +}
> > +EXPORT_SYMBOL(iommu_put_resv_regions_simple);
> 
> Can you call this directly from iommu_put_resv_regions() if the function
> pointer in ops is NULL? That would save having to plumb the default
callback
> into a bunch of drivers.
I probably could, but I don't think that necessarily improves things.
The reason is that that would cause the helper to get called even if the
driver doesn't support reserved regions. That's likely harmless because
in that case the list of regions passed to it would be empty. However, I
think the current way to do this, where we have to implement both hooks
for ->get_resv_regions() and ->put_resv_regions() is nicely symmetric.

Thierry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL:
<http://lists.linuxfoundation.org/pipermail/virtualization/attachments/20191016/d1f494ab/attachment.sig>