Impact: see number of map_sg requests in debugfs
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: see number of unmap_single requests in debugfs
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: see number of map_single requests in debugfs
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: see number of completion wait events in debugfs
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: create a new debugfs directory
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add defines to make iommu stats collection configurable
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: adds new Kconfig entry
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: cleanup
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: cleanup
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: cleanup
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: use bool instead of int for iommu->need_sync
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: use generic dev_name instead of own function
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: also hotplug devices benefit from device isolation
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: adds a new protection domain flag
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function to lockup addresses in protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function to unmap pages into protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function to map pages into protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function to attach devices to protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function to detach devices from protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function for releasing protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a generic function for allocation protection domains
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a function to remove all devices from a domain
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: inform IOMMU about state change of a device in the driver core
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add helper functions to detach a device from a domain
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: rename set_device_domain() to attach_device()
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: know how many devices are assigned to a domain
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: detect when a driver uses a device assigned otherwise
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Imapct: add a new struct member to 'struct protection_domain'

When using protection domains for dma_ops and KVM its better to know for
which subsystem it was allocated. Add a flags member to struct
protection domain for that purpose.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add a function to flush a domain id on every IOMMU
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: save unneeded logic to add and remove domains to the list

The removal of a protection domain from the iommu_pd_list is not
necessary. Another benefit is that we save complexity because we don't
have to readd it later when the device no longer uses the domain.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: refactoring of iommu_queue_inv_iommu_pages
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: split one function into three

The separate functions are required synchronize commands across all
hardware IOMMUs in the system.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: add code to release a domain id
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: change code to free pagetables from protection domains

The dma_ops_free_pagetable function can only free pagetables from
dma_ops domains. Change that to free pagetables of pure protection
domains.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Impact: function rename

The iommu_map function maps only one page. Make this clear in the
function name.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>

Signed-off-by: NIngo Brueckl <ib@wupperonline.de>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

struct dentry is one of the most critical structures in the kernel. So it's
sad to see it going neglected.

With CONFIG_PROFILING turned on (which is probably the common case at least
for distros and kernel developers), sizeof(struct dcache) == 208 here
(64-bit). This gives 19 objects per slab.

I packed d_mounted into a hole, and took another 4 bytes off the inline
name length to take the padding out from the end of the structure. This
shinks it to 200 bytes. I could have gone the other way and increased the
length to 40, but I'm aiming for a magic number, read on...

I then got rid of the d_cookie pointer. This shrinks it to 192 bytes. Rant:
why was this ever a good idea? The cookie system should increase its hash
size or use a tree or something if lookups are a problem. Also the "fast
dcookie lookups" in oprofile should be moved into the dcookie code -- how
can oprofile possibly care about the dcookie_mutex? It gets dropped after
get_dcookie() returns so it can't be providing any sort of protection.

At 192 bytes, 21 objects fit into a 4K page, saving about 3MB on my system
with ~140 000 entries allocated. 192 is also a multiple of 64, so we get
nice cacheline alignment on 64 and 32 byte line systems -- any given dentry
will now require 3 cachelines to touch all fields wheras previously it
would require 4.

I know the inline name size was chosen quite carefully, however with the
reduction in cacheline footprint, it should actually be just about as fast
to do a name lookup for a 36 character name as it was before the patch (and
faster for other sizes). The memory footprint savings for names which are
<= 32 or > 36 bytes long should more than make up for the memory cost for
33-36 byte names.

Performance is a feature...
Signed-off-by: NAl Viro <viro@zeniv.linux.org.uk>

The invlpg and sync walkers lack knowledge of large host sptes,
descending to non-existant pagetable level.

Stop at directory level in such case.

Fixes SMP Windows XP with hugepages.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>