Re: [PATCH kernel v8 26/31] powerpc/iommu: Add userspace view of TCE table

[Alex Williamson <alex.williamson@redhat.com>]

On Fri, 2015-04-10 at 16:31 +1000, Alexey Kardashevskiy wrote:
> In order to support memory pre-registration, we need a way to track
> the use of every registered memory region and only allow unregistration
> if a region is not in use anymore. So we need a way to tell from what
> region the just cleared TCE was from.
> 
> This adds a userspace view of the TCE table into iommu_table struct.
> It contains userspace address, one per TCE entry. The table is only
> allocated when the ownership over an IOMMU group is taken which means
> it is only used from outside of the powernv code (such as VFIO).
> 
> Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
> ---
> Changes:
> v8:
> * added ENOMEM on failed vzalloc()
> ---
>  arch/powerpc/include/asm/iommu.h          |  6 ++++++
>  arch/powerpc/kernel/iommu.c               |  9 +++++++++
>  arch/powerpc/platforms/powernv/pci-ioda.c | 25 ++++++++++++++++++++++++-
>  3 files changed, 39 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/powerpc/include/asm/iommu.h b/arch/powerpc/include/asm/iommu.h
> index 2c08c91..a768a4d 100644
> --- a/arch/powerpc/include/asm/iommu.h
> +++ b/arch/powerpc/include/asm/iommu.h
> @@ -106,9 +106,15 @@ struct iommu_table {
>  	unsigned long *it_map;       /* A simple allocation bitmap for now */
>  	unsigned long  it_page_shift;/* table iommu page size */
>  	struct iommu_table_group *it_group;
> +	unsigned long *it_userspace; /* userspace view of the table */
>  	struct iommu_table_ops *it_ops;
>  };
>  
> +#define IOMMU_TABLE_USERSPACE_ENTRY(tbl, entry) \
> +		((tbl)->it_userspace ? \
> +			&((tbl)->it_userspace[(entry) - (tbl)->it_offset]) : \
> +			NULL)
> +
>  /* Pure 2^n version of get_order */
>  static inline __attribute_const__
>  int get_iommu_order(unsigned long size, struct iommu_table *tbl)
> diff --git a/arch/powerpc/kernel/iommu.c b/arch/powerpc/kernel/iommu.c
> index 0bcd988..833b396 100644
> --- a/arch/powerpc/kernel/iommu.c
> +++ b/arch/powerpc/kernel/iommu.c
> @@ -38,6 +38,7 @@
>  #include <linux/pci.h>
>  #include <linux/iommu.h>
>  #include <linux/sched.h>
> +#include <linux/vmalloc.h>
>  #include <asm/io.h>
>  #include <asm/prom.h>
>  #include <asm/iommu.h>
> @@ -1069,6 +1070,11 @@ static int iommu_table_take_ownership(struct iommu_table *tbl)
>  		spin_unlock(&tbl->pools[i].lock);
>  	spin_unlock_irqrestore(&tbl->large_pool.lock, flags);
>  
> +	BUG_ON(tbl->it_userspace);
> +	tbl->it_userspace = vzalloc(sizeof(*tbl->it_userspace) * tbl->it_size);
> +	if (!tbl->it_userspace)
> +		return -ENOMEM;
> +

It would really make more sense from an error path perspective in this
function if the vzalloc where done first.  Doing it at the end, you need
to consider whether anything previous needs to be un-done.  Also note
that this -ENOMEM return clobbers the -EBUSY if you fix 15/31 to return
"ret".

>  	return 0;
>  }
>  
> @@ -1102,6 +1108,9 @@ static void iommu_table_release_ownership(struct iommu_table *tbl)
>  {
>  	unsigned long flags, i, sz = (tbl->it_size + 7) >> 3;
>  
> +	vfree(tbl->it_userspace);
> +	tbl->it_userspace = NULL;
> +
>  	spin_lock_irqsave(&tbl->large_pool.lock, flags);
>  	for (i = 0; i < tbl->nr_pools; i++)
>  		spin_lock(&tbl->pools[i].lock);
> diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c
> index 751aeab..3ac523d 100644
> --- a/arch/powerpc/platforms/powernv/pci-ioda.c
> +++ b/arch/powerpc/platforms/powernv/pci-ioda.c
> @@ -26,6 +26,7 @@
>  #include <linux/iommu.h>
>  #include <linux/mmzone.h>
>  #include <linux/sizes.h>
> +#include <linux/vmalloc.h>
>  
>  #include <asm/mmzone.h>
>  #include <asm/sections.h>
> @@ -1469,6 +1470,9 @@ static void pnv_pci_free_table(struct iommu_table *tbl)
>  	if (!tbl->it_size)
>  		return;
>  
> +	vfree(tbl->it_userspace);
> +	tbl->it_userspace = NULL;
> +
>  	pnv_free_tce_table(tbl->it_base, size, tbl->it_indirect_levels);
>  	iommu_reset_table(tbl, "ioda2");
>  }
> @@ -1656,9 +1660,28 @@ static void pnv_ioda2_set_ownership(struct iommu_table_group *table_group,
>  	pnv_pci_ioda2_set_bypass(pe, !enable);
>  }
>  
> +static long pnv_pci_ioda2_create_table_with_uas(
> +		struct iommu_table_group *table_group,
> +		int num, __u32 page_shift, __u64 window_size, __u32 levels,
> +		struct iommu_table *tbl)
> +{
> +	long ret = pnv_pci_ioda2_create_table(table_group, num,
> +			page_shift, window_size, levels, tbl);
> +
> +	if (ret)
> +		return ret;
> +
> +	BUG_ON(tbl->it_userspace);
> +	tbl->it_userspace = vzalloc(sizeof(*tbl->it_userspace) * tbl->it_size);
> +	if (!tbl->it_userspace)
> +		return -ENOMEM;


So all of the work done in pnv_pci_ioda2_create_table() can just be
ignored, we undo nothing and return -ENOMEM?  Again, doing the
allocation first might make a lot more sense than slapping on an -ENOMEM
and calling the error handling "good".

> +
> +	return 0;
> +}
> +
>  static struct iommu_table_group_ops pnv_pci_ioda2_ops = {
>  	.set_ownership = pnv_ioda2_set_ownership,
> -	.create_table = pnv_pci_ioda2_create_table,
> +	.create_table = pnv_pci_ioda2_create_table_with_uas,
>  	.set_window = pnv_pci_ioda2_set_window,
>  	.unset_window = pnv_pci_ioda2_unset_window,
>  };