This patch adds support for shadow paging to the 1gb page table code in KVM.
With this code the guest can use 1gb pages even if the host does not support
them.

[ Marcelo: fix shadow page collision on pmd level if a guest 1gb page is mapped
           with 4kb ptes on host level ]
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

The page walker may be used with nested paging too when accessing mmio
areas.  Make it support the additional page-level too.

[ Marcelo: fix reserved bit check for 1gb pte ]
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

With the new name and the corresponding backend changes this function
can now support multiple hugepage sizes.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Signed-off-by: NAvi Kivity <avi@redhat.com>

[avi: fix build on non-x86]
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

We use shadow_pte and spte inconsistently, switch to the shorter spelling.

Rename set_shadow_pte() to __set_spte() to avoid a conflict with the
existing set_spte(), and to indicate its lowlevelness.
Signed-off-by: NAvi Kivity <avi@redhat.com>

Since the guest and host ptes can have wildly different format, adjust
the pte accessor names to indicate on which type of pte they operate on.

No functional changes.
Signed-off-by: NAvi Kivity <avi@redhat.com>

Instead of reloading the pdptrs on every entry and exit (vmcs writes on vmx,
guest memory access on svm) extract them on demand.
Signed-off-by: NAvi Kivity <avi@redhat.com>

Dixes compilation warning:
  CC      arch/x86/kernel/io_delay.o
 arch/x86/kvm/paging_tmpl.h: In function ‘paging64_fetch’:
 arch/x86/kvm/paging_tmpl.h:279: warning: ‘sptep’ may be used uninitialized in this function
 arch/x86/kvm/paging_tmpl.h: In function ‘paging32_fetch’:
 arch/x86/kvm/paging_tmpl.h:279: warning: ‘sptep’ may be used uninitialized in this function

warning is bogus (always have a least one level), but need to shut the compiler
up.
Signed-off-by: NJaswinder Singh Rajput <jaswinderrajput@gmail.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Complexity to fix it not worthwhile the gains, as discussed
in http://article.gmane.org/gmane.comp.emulators.kvm.devel/28649.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Detect, indicate, and propagate page faults where reserved bits are set.
Take care to handle the different paging modes, each of which has different
sets of reserved bits.

[avi: fix pte reserved bits for efer.nxe=0]
Signed-off-by: NEddie Dong <eddie.dong@intel.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

The original one is for the code before refactoring.
Signed-off-by: NYaozu (Eddie) Dong <eddie.dong@intel.com>
Signed-off-by: NSheng Yang <sheng@linux.intel.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

There is no reason to update the shadow pte here because the guest pte
is only changed to dirty state.
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>

When kvm emulates an invlpg instruction, it can drop a shadow pte, but
leaves the guest tlbs intact.  This can cause memory corruption when
swapping out.

Without this the other cpu can still write to a freed host physical page.
tlb smp flush must happen if rmap_remove is called always before mmu_lock
is released because the VM will take the mmu_lock before it can finally add
the page to the freelist after swapout. mmu notifier makes it safe to flush
the tlb after freeing the page (otherwise it would never be safe) so we can do
a single flush for multiple sptes invalidated.

Cc: stable@kernel.org
Signed-off-by: NAndrea Arcangeli <aarcange@redhat.com>
Acked-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

In the paging_fetch function rmap_remove is called after setting a large
pte to non-present. This causes rmap_remove to not drop the reference to
the large page. The result is a memory leak of that page.

Cc: stable@kernel.org
Signed-off-by: NJoerg Roedel <joerg.roedel@amd.com>
Acked-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

This actually describes what is going on, rather than alerting the reader
that something strange is going on.
Signed-off-by: NAvi Kivity <avi@redhat.com>

Signed-off-by: NAvi Kivity <avi@redhat.com>

Effectively reverting to the pre walk_shadow() version -- but now
with the reusable for_each().
Signed-off-by: NAvi Kivity <avi@redhat.com>

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>

If the guest executes invlpg, peek into the pagetable and attempt to
prepopulate the shadow entry.

Also stop dirty fault updates from interfering with the fork detector.

2% improvement on RHEL3/AIM7.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Skip syncing global pages on cr3 switch (but not on cr4/cr0). This is
important for Linux 32-bit guests with PAE, where the kmap page is
marked as global.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

It is possible for a shadow page to have a parent link
pointing to a freed page. When zapping a high level table,
kvm_mmu_page_unlink_children fails to remove the parent_pte link.
For that to happen, the child must be unreachable via the shadow
tree, which can happen in shadow_walk_entry if the guest pte was
modified in between walk() and fetch(). Remove the parent pte
reference in such case.

Possible cause for oops in bug #2217430.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Allow guest pagetables to go out of sync.  Instead of emulating write
accesses to guest pagetables, or unshadowing them, we un-write-protect
the page table and allow the guest to modify it at will.  We rely on
invlpg executions to synchronize individual ptes, and will synchronize
the entire pagetable on tlb flushes.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

With pages out of sync invlpg needs to be trapped. For now simply nuke
the entry.

Untested on AMD.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Examine guest pagetable and bring the shadow back in sync. Caller is responsible
for local TLB flush before re-entering guest mode.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

It is necessary to flush all TLB's when a large spte entry is
overwritten with a normal page directory pointer.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

Convert gfn_to_pfn to use get_user_pages_fast, which can do lockless
pagetable lookups on x86. Kernel compilation on 4-way guest is 3.7%
faster on VMX.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@redhat.com>

EPT is 4 level by default in 32pae(48 bits), but the addr parameter
of kvm_shadow_walk->entry() only accept unsigned long as virtual
address, which is 32bit in 32pae. This result in SHADOW_PT_INDEX()
overflow when try to fetch level 4 index.

Fix it by extend kvm_shadow_walk->entry() to accept 64bit addr in
parameter.
Signed-off-by: NSheng Yang <sheng.yang@intel.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Signed-off-by: NAvi Kivity <avi@qumranet.com>

It is not specific to the paging mode, so can be made global (and reusable).
Signed-off-by: NAvi Kivity <avi@qumranet.com>

The shadow code assigns a pte directly in one place, which is nonatomic on
i386 can can cause random memory references.  Fix by using an atomic setter.
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Synchronize changes to host virtual addresses which are part of
a KVM memory slot to the KVM shadow mmu.  This allows pte operations
like swapping, page migration, and madvise() to transparently work
with KVM.
Signed-off-by: NAndrea Arcangeli <andrea@qumranet.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Instead of reading each pte individually, read 256 bytes worth of ptes and
batch process them.
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Signed-off-by: NAvi Kivity <avi@qumranet.com>

This patch introduces a gfn_to_pfn() function and corresponding functions like
kvm_release_pfn_dirty().  Using these new functions, we can modify the x86
MMU to no longer assume that it can always get a struct page for any given gfn.

We don't want to eliminate gfn_to_page() entirely because a number of places
assume they can do gfn_to_page() and then kmap() the results.  When we support
IO memory, gfn_to_page() will fail for IO pages although gfn_to_pfn() will
succeed.

This does not implement support for avoiding reference counting for reserved
RAM or for IO memory.  However, it should make those things pretty straight
forward.

Since we're only introducing new common symbols, I don't think it will break
the non-x86 architectures but I haven't tested those.  I've tested Intel,
AMD, NPT, and hugetlbfs with Windows and Linux guests.

[avi: fix overflow when shifting left pfns by adding casts]
Signed-off-by: NAnthony Liguori <aliguori@us.ibm.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Unify slots_lock acquision around vcpu_run(). This is simpler and less
error-prone.

Also fix some callsites that were not grabbing the lock properly.

[avi: drop slots_lock while in guest mode to avoid holding the lock
      for indefinite periods]
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>

If we populate a shadow pte due to a fault (and not speculatively due to a
pte write) then we can set the accessed bit on it, as we know it will be
set immediately on the next guest instruction.  This saves a read-modify-write
operation.
Signed-off-by: NAvi Kivity <avi@qumranet.com>

__FUNCTION__ is gcc-specific, use __func__
Signed-off-by: NHarvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>

Create large pages mappings if the guest PTE's are marked as such and
the underlying memory is hugetlbfs backed.  If the largepage contains
write-protected pages, a large pte is not used.

Gives a consistent 2% improvement for data copies on ram mounted
filesystem, without NPT/EPT.

Anthony measures a 4% improvement on 4-way kernbench, with NPT.
Signed-off-by: NMarcelo Tosatti <mtosatti@redhat.com>
Signed-off-by: NAvi Kivity <avi@qumranet.com>