A VMM may send a non-fatal signal to its threads, including vCPU tasks,
at any time, and thus may signal vCPU tasks during KVM_RUN. If a vCPU
task receives the signal while its trying to spawn the huge page recovery
vhost task, then KVM_RUN will fail due to copy_process() returning
-ERESTARTNOINTR.
Rework call_once() to mark the call complete if and only if the called
function succeeds, and plumb the function's true error code back to the
call_once() invoker. This provides userspace with the correct, non-fatal
error code so that the VMM doesn't terminate the VM on -ENOMEM, and allows
subsequent KVM_RUN a succeed by virtue of retrying creation of the NX huge
page task.
Co-developed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
[implemented the kvm user side]
Signed-off-by: Keith Busch <kbusch@kernel.org>
Message-ID: <20250227230631.303431-3-kbusch@meta.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Lets callers distinguish why the vhost task creation failed. No one
currently cares why it failed, so no real runtime change from this
patch, but that will not be the case for long.
Signed-off-by: Keith Busch <kbusch@kernel.org>
Message-ID: <20250227230631.303431-2-kbusch@meta.com>
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
When preparing vmcb02 for nested VMRUN (or state restore), "enter" guest
mode prior to initializing the MMU for nested NPT so that guest_mode is
set in the MMU's role. KVM's model is that all L2 MMUs are tagged with
guest_mode, as the behavior of hypervisor MMUs tends to be significantly
different than kernel MMUs.
Practically speaking, the bug is relatively benign, as KVM only directly
queries role.guest_mode in kvm_mmu_free_guest_mode_roots() and
kvm_mmu_page_ad_need_write_protect(), which SVM doesn't use, and in paths
that are optimizations (mmu_page_zap_pte() and
shadow_mmu_try_split_huge_pages()).
And while the role is incorprated into shadow page usage, because nested
NPT requires KVM to be using NPT for L1, reusing shadow pages across L1
and L2 is impossible as L1 MMUs will always have direct=1, while L2 MMUs
will have direct=0.
Hoist the TLB processing and setting of HF_GUEST_MASK to the beginning
of the flow instead of forcing guest_mode in the MMU, as nothing in
nested_vmcb02_prepare_control() between the old and new locations touches
TLB flush requests or HF_GUEST_MASK, i.e. there's no reason to present
inconsistent vCPU state to the MMU.
Fixes: 69cb877487 ("KVM: nSVM: move MMU setup to nested_prepare_vmcb_control")
Cc: stable@vger.kernel.org
Reported-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Reviewed-by: Yosry Ahmed <yosry.ahmed@linux.dev>
Link: https://lore.kernel.org/r/20250130010825.220346-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
When waking a VM's NX huge page recovery thread, ensure the thread is
actually alive before trying to wake it. Now that the thread is spawned
on-demand during KVM_RUN, a VM without a recovery thread is reachable via
the related module params.
BUG: kernel NULL pointer dereference, address: 0000000000000040
#PF: supervisor read access in kernel mode
#PF: error_code(0x0000) - not-present page
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 0.0.0 02/06/2015
RIP: 0010:vhost_task_wake+0x5/0x10
Call Trace:
<TASK>
set_nx_huge_pages+0xcc/0x1e0 [kvm]
param_attr_store+0x8a/0xd0
module_attr_store+0x1a/0x30
kernfs_fop_write_iter+0x12f/0x1e0
vfs_write+0x233/0x3e0
ksys_write+0x60/0xd0
do_syscall_64+0x5b/0x160
entry_SYSCALL_64_after_hwframe+0x4b/0x53
RIP: 0033:0x7f3b52710104
</TASK>
Modules linked in: kvm_intel kvm
CR2: 0000000000000040
Fixes: 931656b9e2 ("kvm: defer huge page recovery vhost task to later")
Cc: stable@vger.kernel.org
Cc: Keith Busch <kbusch@kernel.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-ID: <20250124234623.3609069-1-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Some libraries want to ensure they are single threaded before forking,
so making the kernel's kvm huge page recovery process a vhost task of
the user process breaks those. The minijail library used by crosvm is
one such affected application.
Defer the task to after the first VM_RUN call, which occurs after the
parent process has forked all its jailed processes. This needs to happen
only once for the kvm instance, so introduce some general-purpose
infrastructure for that, too. It's similar in concept to pthread_once;
except it is actually usable, because the callback takes a parameter.
Cc: Sean Christopherson <seanjc@google.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Tested-by: Alyssa Ross <hi@alyssa.is>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Message-ID: <20250123153543.2769928-1-kbusch@meta.com>
[Move call_once API to include/linux. - Paolo]
Cc: stable@vger.kernel.org
Fixes: d96c77bd4e ("KVM: x86: switch hugepage recovery thread to vhost_task")
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
As part of enabling TDX virtual machines, support support separation of
private/shared EPT into separate roots.
Confidential computing solutions almost invariably have concepts of
private and shared memory, but they may different a lot in the details.
In SEV, for example, the bit is handled more like a permission bit as
far as the page tables are concerned: the private/shared bit is not
included in the physical address.
For TDX, instead, the bit is more like a physical address bit, with
the host mapping private memory in one half of the address space and
shared in another. Furthermore, the two halves are mapped by different
EPT roots and only the shared half is managed by KVM; the private half
(also called Secure EPT in Intel documentation) gets managed by the
privileged TDX Module via SEAMCALLs.
As a result, the operations that actually change the private half of
the EPT are limited and relatively slow compared to reading a PTE. For
this reason the design for KVM is to keep a mirror of the private EPT in
host memory. This allows KVM to quickly walk the EPT and only perform the
slower private EPT operations when it needs to actually modify mid-level
private PTEs.
There are thus three sets of EPT page tables: external, mirror and
direct. In the case of TDX (the only user of this framework) the
first two cover private memory, whereas the third manages shared
memory:
external EPT - Hidden within the TDX module, modified via TDX module
calls.
mirror EPT - Bookkeeping tree used as an optimization by KVM, not
used by the processor.
direct EPT - Normal EPT that maps unencrypted shared memory.
Managed like the EPT of a normal VM.
Modifying external EPT
----------------------
Modifications to the mirrored page tables need to also perform the
same operations to the private page tables, which will be handled via
kvm_x86_ops. Although this prep series does not interact with the TDX
module at all to actually configure the private EPT, it does lay the
ground work for doing this.
In some ways updating the private EPT is as simple as plumbing PTE
modifications through to also call into the TDX module; however, the
locking is more complicated because inserting a single PTE cannot anymore
be done atomically with a single CMPXCHG. For this reason, the existing
FROZEN_SPTE mechanism is used whenever a call to the TDX module updates the
private EPT. FROZEN_SPTE acts basically as a spinlock on a PTE. Besides
protecting operation of KVM, it limits the set of cases in which the
TDX module will encounter contention on its own PTE locks.
Zapping external EPT
--------------------
While the framework tries to be relatively generic, and to be
understandable without knowing TDX much in detail, some requirements of
TDX sometimes leak; for example the private page tables also cannot be
zapped while the range has anything mapped, so the mirrored/private page
tables need to be protected from KVM operations that zap any non-leaf
PTEs, for example kvm_mmu_reset_context() or kvm_mmu_zap_all_fast().
For normal VMs, guest memory is zapped for several reasons: user
memory getting paged out by the guest, memslots getting deleted,
passthrough of devices with non-coherent DMA. Confidential computing
adds to these the conversion of memory between shared and privates. These
operations must not zap any private memory that is in use by the guest.
This is possible because the only zapping that is out of the control
of KVM/userspace is paging out userspace memory, which cannot apply to
guestmemfd operations. Thus a TDX VM will only zap private memory from
memslot deletion and from conversion between private and shared memory
which is triggered by the guest.
To avoid zapping too much memory, enums are introduced so that operations
can choose to target only private or shared memory, and thus only
direct or mirror EPT. For example:
Memslot deletion - Private and shared
MMU notifier based zapping - Shared only
Conversion to shared - Private only
Conversion to private - Shared only
Other cases of zapping will not be supported for KVM, for example
APICv update or non-coherent DMA status update; for the latter, TDX will
simply require that the CPU supports self-snoop and honor guest PAT
unconditionally for shared memory.
- Overhaul KVM's CPUID feature infrastructure to replace "governed" features
with per-vCPU tracking of the vCPU's capabailities for all features. Along
the way, refactor the code to make it easier to add/modify features, and
add a variety of self-documenting macro types to again simplify adding new
features and to help readers understand KVM's handling of existing features.
- Rework KVM's handling of VM-Exits during event vectoring to plug holes where
KVM unintentionally puts the vCPU into infinite loops in some scenarios,
e.g. if emulation is triggered by the exit, and to bring parity between VMX
and SVM.
- Add pending request and interrupt injection information to the kvm_exit and
kvm_entry tracepoints respectively.
- Fix a relatively benign flaw where KVM would end up redoing RDPKRU when
loading guest/host PKRU due to a refactoring of the kernel helpers that
didn't account for KVM's pre-checking of the need to do WRPKRU.
-----BEGIN PGP SIGNATURE-----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=32mM
-----END PGP SIGNATURE-----
Merge tag 'kvm-x86-misc-6.14' of https://github.com/kvm-x86/linux into HEAD
KVM x86 misc changes for 6.14:
- Overhaul KVM's CPUID feature infrastructure to track all vCPU capabilities
instead of just those where KVM needs to manage state and/or explicitly
enable the feature in hardware. Along the way, refactor the code to make
it easier to add features, and to make it more self-documenting how KVM
is handling each feature.
- Rework KVM's handling of VM-Exits during event vectoring; this plugs holes
where KVM unintentionally puts the vCPU into infinite loops in some scenarios
(e.g. if emulation is triggered by the exit), and brings parity between VMX
and SVM.
- Add pending request and interrupt injection information to the kvm_exit and
kvm_entry tracepoints respectively.
- Fix a relatively benign flaw where KVM would end up redoing RDPKRU when
loading guest/host PKRU, due to a refactoring of the kernel helpers that
didn't account for KVM's pre-checking of the need to do WRPKRU.
Return RET_PF* (excluding RET_PF_EMULATE/RET_PF_CONTINUE/RET_PF_INVALID)
instead of 1 in kvm_mmu_page_fault().
The callers of kvm_mmu_page_fault() are KVM page fault handlers (i.e.,
npf_interception(), handle_ept_misconfig(), __vmx_handle_ept_violation(),
kvm_handle_page_fault()). They either check if the return value is > 0 (as
in npf_interception()) or pass it further to vcpu_run() to decide whether
to break out of the kernel loop and return to the user when r <= 0.
Therefore, returning any positive value is equivalent to returning 1.
Warn if r == RET_PF_CONTINUE (which should not be a valid value) to ensure
a positive return value.
This is a preparation to allow TDX's EPT violation handler to check the
RET_PF* value and retry internally for RET_PF_RETRY.
No functional changes are intended.
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Message-ID: <20250113021138.18875-1-yan.y.zhao@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add a few sanity checks to prevent memslot GFNs from ever having alias bits
set.
Like other Coco technologies, TDX has the concept of private and shared
memory. For TDX the private and shared mappings are managed on separate
EPT roots. The private half is managed indirectly though calls into a
protected runtime environment called the TDX module, where the shared half
is managed within KVM in normal page tables.
For TDX, the shared half will be mapped in the higher alias, with a "shared
bit" set in the GPA. However, KVM will still manage it with the same
memslots as the private half. This means memslot looks ups and zapping
operations will be provided with a GFN without the shared bit set.
If these memslot GFNs ever had the bit that selects between the two aliases
it could lead to unexpected behavior in the complicated code that directs
faulting or zapping operations between the roots that map the two aliases.
As a safety measure, prevent memslots from being set at a GFN range that
contains the alias bit.
Also, check in the kvm_faultin_pfn() for the fault path. This later check
does less today, as the alias bits are specifically stripped from the GFN
being checked, however future code could possibly call in to the fault
handler in a way that skips this stripping. Since kvm_faultin_pfn() now
has many references to vcpu->kvm, extract it to local variable.
Link: https://lore.kernel.org/kvm/ZpbKqG_ZhCWxl-Fc@google.com/
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-19-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Rename kvm_tdp_mmu_invalidate_all_roots() to
kvm_tdp_mmu_invalidate_roots(), and make it enum kvm_tdp_mmu_root_types
as an argument.
kvm_tdp_mmu_invalidate_roots() is called with different root types. For
kvm_mmu_zap_all_fast() it only operates on shared roots. But when tearing
down a VM it needs to invalidate all roots. Have the callers only
invalidate the required roots instead of all roots.
Within kvm_tdp_mmu_invalidate_roots(), respect the root type
passed by checking the root type in root iterator.
Suggested-by: Chao Gao <chao.gao@intel.com>
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-17-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add the ability for the TDP MMU to maintain a mirror of a separate
mapping.
Like other Coco technologies, TDX has the concept of private and shared
memory. For TDX the private and shared mappings are managed on separate
EPT roots. The private half is managed indirectly through calls into a
protected runtime environment called the TDX module, where the shared half
is managed within KVM in normal page tables.
In order to handle both shared and private memory, KVM needs to learn to
handle faults and other operations on the correct root for the operation.
KVM could learn the concept of private roots, and operate on them by
calling out to operations that call into the TDX module. But there are two
problems with that:
1. Calls into the TDX module are relatively slow compared to the simple
accesses required to read a PTE managed directly by KVM.
2. Other Coco technologies deal with private memory completely differently
and it will make the code confusing when being read from their
perspective. Special operations added for TDX that set private or zap
private memory will have nothing to do with these other private memory
technologies. (SEV, etc).
To handle these, instead teach the TDP MMU about a new concept "mirror
roots". Such roots maintain page tables that are not actually mapped,
and are just used to traverse quickly to determine if the mid level page
tables need to be installed. When the memory be mirrored needs to actually
be changed, calls can be made to via x86_ops.
private KVM page fault |
| |
V |
private GPA | CPU protected EPTP
| | |
V | V
mirror PT root | external PT root
| | |
V | V
mirror PT --hook to propagate-->external PT
| | |
\--------------------+------\ |
| | |
| V V
| private guest page
|
|
non-encrypted memory | encrypted memory
|
Leave calling out to actually update the private page tables that are being
mirrored for later changes. Just implement the handling of MMU operations
on to mirrored roots.
In order to direct operations to correct root, add root types
KVM_DIRECT_ROOTS and KVM_MIRROR_ROOTS. Tie the usage of mirrored/direct
roots to private/shared with conditionals. It could also be implemented by
making the kvm_tdp_mmu_root_types and kvm_gfn_range_filter enum bits line
up such that conversion could be a direct assignment with a case. Don't do
this because the mapping of private to mirrored is confusing enough. So it
is worth not hiding the logic in type casting.
Cleanup the mirror root in kvm_mmu_destroy() instead of the normal place
in kvm_mmu_free_roots(), because the private root that is being cannot be
rebuilt like a normal root. It needs to persist for the lifetime of the VM.
The TDX module will also need to be provided with page tables to use for
the actual mapping being mirrored by the mirrored page tables. Allocate
these in the mapping path using the recently added
kvm_mmu_alloc_external_spt().
Don't support 2M page for now. This is avoided by forcing 4k pages in the
fault. Add a KVM_BUG_ON() to verify.
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Co-developed-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Kai Huang <kai.huang@intel.com>
Co-developed-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Co-developed-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-13-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
The kvm_tdp_mmu_alloc_root() function currently always returns 0. This
allows for the caller, mmu_alloc_direct_roots(), to call
kvm_tdp_mmu_alloc_root() and also return 0 in one line:
return kvm_tdp_mmu_alloc_root(vcpu);
So it is useful even though the return value of kvm_tdp_mmu_alloc_root()
is always the same. However, in future changes, kvm_tdp_mmu_alloc_root()
will be called twice in mmu_alloc_direct_roots(). This will force the
first call to either awkwardly handle the return value that will always
be zero or ignore it. So change kvm_tdp_mmu_alloc_root() to return void.
Do it in a separate change so the future change will be cleaner.
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20240718211230.1492011-7-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add an external pointer to struct kvm_mmu_page for TDX's private page table
and add helper functions to allocate/initialize/free a private page table
page. TDX will only be supported with the TDP MMU. Because KVM TDP MMU
doesn't use unsync_children and write_flooding_count, pack them to have
room for a pointer and use a union to avoid memory overhead.
For private GPA, CPU refers to a private page table whose contents are
encrypted. The dedicated APIs to operate on it (e.g. updating/reading its
PTE entry) are used, and their cost is expensive.
When KVM resolves the KVM page fault, it walks the page tables. To reuse
the existing KVM MMU code and mitigate the heavy cost of directly walking
the private page table allocate two sets of page tables for the private
half of the GPA space.
For the page tables that KVM will walk, allocate them like normal and refer
to them as mirror page tables. Additionally allocate one more page for the
page tables the CPU will walk, and call them external page tables. Resolve
the KVM page fault with the existing code, and do additional operations
necessary for modifying the external page table in future patches.
The relationship of the types of page tables in this scheme is depicted
below:
KVM page fault |
| |
V |
-------------+---------- |
| | |
V V |
shared GPA private GPA |
| | |
V V |
shared PT root mirror PT root | private PT root
| | | |
V V | V
shared PT mirror PT --propagate--> external PT
| | | |
| \-----------------+------\ |
| | | |
V | V V
shared guest page | private guest page
|
non-encrypted memory | encrypted memory
|
PT - Page table
Shared PT - Visible to KVM, and the CPU uses it for shared mappings.
External PT - The CPU uses it, but it is invisible to KVM. TDX module
updates this table to map private guest pages.
Mirror PT - It is visible to KVM, but the CPU doesn't use it. KVM uses
it to propagate PT change to the actual private PT.
Add a helper kvm_has_mirrored_tdp() to trigger this behavior and wire it
to the TDX vm type.
Co-developed-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Reviewed-by: Binbin Wu <binbin.wu@linux.intel.com>
Message-ID: <20240718211230.1492011-5-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Add new members to strut kvm_gfn_range to indicate which mapping
(private-vs-shared) to operate on: enum kvm_gfn_range_filter
attr_filter. Update the core zapping operations to set them appropriately.
TDX utilizes two GPA aliases for the same memslots, one for memory that is
for private memory and one that is for shared. For private memory, KVM
cannot always perform the same operations it does on memory for default
VMs, such as zapping pages and having them be faulted back in, as this
requires guest coordination. However, some operations such as guest driven
conversion of memory between private and shared should zap private memory.
Internally to the MMU, private and shared mappings are tracked on separate
roots. Mapping and zapping operations will operate on the respective GFN
alias for each root (private or shared). So zapping operations will by
default zap both aliases. Add fields in struct kvm_gfn_range to allow
callers to specify which aliases so they can only target the aliases
appropriate for their specific operation.
There was feedback that target aliases should be specified such that the
default value (0) is to operate on both aliases. Several options were
considered. Several variations of having separate bools defined such
that the default behavior was to process both aliases. They either allowed
nonsensical configurations, or were confusing for the caller. A simple
enum was also explored and was close, but was hard to process in the
caller. Instead, use an enum with the default value (0) reserved as a
disallowed value. Catch ranges that didn't have the target aliases
specified by looking for that specific value.
Set target alias with enum appropriately for these MMU operations:
- For KVM's mmu notifier callbacks, zap shared pages only because private
pages won't have a userspace mapping
- For setting memory attributes, kvm_arch_pre_set_memory_attributes()
chooses the aliases based on the attribute.
- For guest_memfd invalidations, zap private only.
Link: https://lore.kernel.org/kvm/ZivIF9vjKcuGie3s@google.com/
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Co-developed-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-3-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Prepare for a future TDX patch which asserts that atomic zapping
(i.e. zapping with mmu_lock taken for read) don't operate on mirror roots.
When tearing down a VM, all roots have to be zapped (including mirror
roots once they're in place) so do that with the mmu_lock taken for write.
kvm_mmu_uninit_tdp_mmu() is invoked either before or after executing any
atomic operations on SPTEs by vCPU threads. Therefore, it will not impact
vCPU threads performance if kvm_tdp_mmu_zap_invalidated_roots() acquires
mmu_lock for write to zap invalid roots.
Co-developed-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Yan Zhao <yan.y.zhao@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-2-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Treat slow-path TDP MMU faults as spurious if the access is allowed given
the existing SPTE to fix a benign warning (other than the WARN itself)
due to replacing a writable SPTE with a read-only SPTE, and to avoid the
unnecessary LOCK CMPXCHG and subsequent TLB flush.
If a read fault races with a write fault, fast GUP fails for any reason
when trying to "promote" the read fault to a writable mapping, and KVM
resolves the write fault first, then KVM will end up trying to install a
read-only SPTE (for a !map_writable fault) overtop a writable SPTE.
Note, it's not entirely clear why fast GUP fails, or if that's even how
KVM ends up with a !map_writable fault with a writable SPTE. If something
else is going awry, e.g. due to a bug in mmu_notifiers, then treating read
faults as spurious in this scenario could effectively mask the underlying
problem.
However, retrying the faulting access instead of overwriting an existing
SPTE is functionally correct and desirable irrespective of the WARN, and
fast GUP _can_ legitimately fail with a writable VMA, e.g. if the Accessed
bit in primary MMU's PTE is toggled and causes a PTE value mismatch. The
WARN was also recently added, specifically to track down scenarios where
KVM is unnecessarily overwrites SPTEs, i.e. treating the fault as spurious
doesn't regress KVM's bug-finding capabilities in any way. In short,
letting the WARN linger because there's a tiny chance it's due to a bug
elsewhere would be excessively paranoid.
Fixes: 1a175082b1 ("KVM: x86/mmu: WARN and flush if resolving a TDP MMU fault clears MMU-writable")
Reported-by: Lei Yang <leiyang@redhat.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=219588
Tested-by: Lei Yang <leiyang@redhat.com>
Link: https://lore.kernel.org/r/20241218213611.3181643-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
As the first step toward replacing KVM's so-called "governed features"
framework with a more comprehensive, less poorly named implementation,
replace the "kvm_governed_feature" function prefix with "guest_cpu_cap"
and rename guest_can_use() to guest_cpu_cap_has().
The "guest_cpu_cap" naming scheme mirrors that of "kvm_cpu_cap", and
provides a more clear distinction between guest capabilities, which are
KVM controlled (heh, or one might say "governed"), and guest CPUID, which
with few exceptions is fully userspace controlled.
Opportunistically rewrite the comment about XSS passthrough for SEV-ES
guests to avoid referencing so many functions, as such comments are prone
to becoming stale (case in point...).
No functional change intended.
Reviewed-by: Maxim Levitsky <mlevitsk@redhat.com>
Reviewed-by: Binbin Wu <binbin.wu@linux.intel.com>
Link: https://lore.kernel.org/r/20241128013424.4096668-40-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
kvm_vm_create_worker_thread() is meant to be used for kthreads that
can consume significant amounts of CPU time on behalf of a VM or in
response to how the VM behaves (for example how it accesses its memory).
Therefore it wants to charge the CPU time consumed by that work to
the VM's container.
However, because of these threads, cgroups which have kvm instances
inside never complete freezing. This can be trivially reproduced:
root@test ~# mkdir /sys/fs/cgroup/test
root@test ~# echo $$ > /sys/fs/cgroup/test/cgroup.procs
root@test ~# qemu-system-x86_64 -nographic -enable-kvm
and in another terminal:
root@test ~# echo 1 > /sys/fs/cgroup/test/cgroup.freeze
root@test ~# cat /sys/fs/cgroup/test/cgroup.events
populated 1
frozen 0
The cgroup freezing happens in the signal delivery path but
kvm_nx_huge_page_recovery_worker, while joining non-root cgroups, never
calls into the signal delivery path and thus never gets frozen. Because
the cgroup freezer determines whether a given cgroup is frozen by
comparing the number of frozen threads to the total number of threads
in the cgroup, the cgroup never becomes frozen and users waiting for
the state transition may hang indefinitely.
Since the worker kthread is tied to a user process, it's better if
it behaves similarly to user tasks as much as possible, including
being able to send SIGSTOP and SIGCONT. In fact, vhost_task is all
that kvm_vm_create_worker_thread() wanted to be and more: not only it
inherits the userspace process's cgroups, it has other niceties like
being parented properly in the process tree. Use it instead of the
homegrown alternative.
Incidentally, the new code is also better behaved when you flip recovery
back and forth to disabled and back to enabled. If your recovery period
is 1 minute, it will run the next recovery after 1 minute independent
of how many times you flipped the parameter.
(Commit message based on emails from Tejun).
Reported-by: Tejun Heo <tj@kernel.org>
Reported-by: Luca Boccassi <bluca@debian.org>
Acked-by: Tejun Heo <tj@kernel.org>
Tested-by: Luca Boccassi <bluca@debian.org>
Cc: stable@vger.kernel.org
Reviewed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
- Clean up and optimize KVM's handling of writes to MSR_IA32_APICBASE.
- Quirk KVM's misguided behavior of initialized certain feature MSRs to
their maximum supported feature set, which can result in KVM creating
invalid vCPU state. E.g. initializing PERF_CAPABILITIES to a non-zero
value results in the vCPU having invalid state if userspace hides PDCM
from the guest, which can lead to save/restore failures.
- Fix KVM's handling of non-canonical checks for vCPUs that support LA57
to better follow the "architecture", in quotes because the actual
behavior is poorly documented. E.g. most MSR writes and descriptor
table loads ignore CR4.LA57 and operate purely on whether the CPU
supports LA57.
- Bypass the register cache when querying CPL from kvm_sched_out(), as
filling the cache from IRQ context is generally unsafe, and harden the
cache accessors to try to prevent similar issues from occuring in the
future.
- Advertise AMD_IBPB_RET to userspace, and fix a related bug where KVM
over-advertises SPEC_CTRL when trying to support cross-vendor VMs.
- Minor cleanups
-----BEGIN PGP SIGNATURE-----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=dESL
-----END PGP SIGNATURE-----
Merge tag 'kvm-x86-misc-6.13' of https://github.com/kvm-x86/linux into HEAD
KVM x86 misc changes for 6.13
- Clean up and optimize KVM's handling of writes to MSR_IA32_APICBASE.
- Quirk KVM's misguided behavior of initialized certain feature MSRs to
their maximum supported feature set, which can result in KVM creating
invalid vCPU state. E.g. initializing PERF_CAPABILITIES to a non-zero
value results in the vCPU having invalid state if userspace hides PDCM
from the guest, which can lead to save/restore failures.
- Fix KVM's handling of non-canonical checks for vCPUs that support LA57
to better follow the "architecture", in quotes because the actual
behavior is poorly documented. E.g. most MSR writes and descriptor
table loads ignore CR4.LA57 and operate purely on whether the CPU
supports LA57.
- Bypass the register cache when querying CPL from kvm_sched_out(), as
filling the cache from IRQ context is generally unsafe, and harden the
cache accessors to try to prevent similar issues from occuring in the
future.
- Advertise AMD_IBPB_RET to userspace, and fix a related bug where KVM
over-advertises SPEC_CTRL when trying to support cross-vendor VMs.
- Minor cleanups
Drop the per-VM zapped_obsolete_pages list now that the usage from the
defunct mmu_shrinker is gone, and instead use a local list to track pages
in kvm_zap_obsolete_pages(), the sole remaining user of
zapped_obsolete_pages.
Opportunistically add an assertion to verify and document that slots_lock
must be held, i.e. that there can only be one active instance of
kvm_zap_obsolete_pages() at any given time, and by doing so also prove
that using a local list instead of a per-VM list doesn't change any
functionality (beyond trivialities like list initialization).
Signed-off-by: Vipin Sharma <vipinsh@google.com>
Link: https://lore.kernel.org/r/20241101201437.1604321-2-vipinsh@google.com
[sean: split to separate patch, write changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Remove KVM's MMU shrinker and (almost) all of its related code, as the
current implementation is very disruptive to VMs (if it ever runs),
without providing any meaningful benefit[1].
Alternatively, KVM could repurpose its shrinker, e.g. to reclaim pages
from the per-vCPU caches[2], but given that no one has complained about
lack of TDP MMU support for the shrinker in the 3+ years since the TDP MMU
was enabled by default, it's safe to say that there is likely no real use
case for initiating reclaim of KVM's page tables from the shrinker.
And while clever/cute, reclaiming the per-vCPU caches doesn't scale the
same way that reclaiming in-use page table pages does. E.g. the amount of
memory being used by a VM doesn't always directly correlate with the
number vCPUs, and even when it does, reclaiming a few pages from per-vCPU
caches likely won't make much of a dent in the VM's total memory usage,
especially for VMs with huge amounts of memory.
Lastly, if it turns out that there is a strong use case for dropping the
per-vCPU caches, re-introducing the shrinker registration is trivial
compared to the complexity of actually reclaiming pages from the caches.
[1] https://lore.kernel.org/lkml/Y45dldZnI6OIf+a5@google.com
[2] https://lore.kernel.org/kvm/20241004195540.210396-3-vipinsh@google.com
Suggested-by: Sean Christopherson <seanjc@google.com>
Suggested-by: David Matlack <dmatlack@google.com>
Signed-off-by: Vipin Sharma <vipinsh@google.com>
Link: https://lore.kernel.org/r/20241101201437.1604321-2-vipinsh@google.com
[sean: keep zapped_obsolete_pages for now, massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Rename make_huge_page_split_spte() to make_small_spte(). This ensures
that the usage of "small_spte" and "huge_spte" are consistent between
make_huge_spte() and make_small_spte().
This should also reduce some confusion as make_huge_page_split_spte()
almost reads like it will create a huge SPTE, when in fact it is
creating a small SPTE to split the huge SPTE.
No functional change intended.
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20240823235648.3236880-6-dmatlack@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Recover TDP MMU huge page mappings in-place instead of zapping them when
dirty logging is disabled, and rename functions that recover huge page
mappings when dirty logging is disabled to move away from the "zap
collapsible spte" terminology.
Before KVM flushes TLBs, guest accesses may be translated through either
the (stale) small SPTE or the (new) huge SPTE. This is already possible
when KVM is doing eager page splitting (where TLB flushes are also
batched), and when vCPUs are faulting in huge mappings (where TLBs are
flushed after the new huge SPTE is installed).
Recovering huge pages reduces the number of page faults when dirty
logging is disabled:
$ perf stat -e kvm:kvm_page_fault -- ./dirty_log_perf_test -s anonymous_hugetlb_2mb -v 64 -e -b 4g
Before: 393,599 kvm:kvm_page_fault
After: 262,575 kvm:kvm_page_fault
vCPU throughput and the latency of disabling dirty-logging are about
equal compared to zapping, but avoiding faults can be beneficial to
remove vCPU jitter in extreme scenarios.
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20240823235648.3236880-5-dmatlack@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
As a result of a recent investigation, it was determined that x86 CPUs
which support 5-level paging, don't always respect CR4.LA57 when doing
canonical checks.
In particular:
1. MSRs which contain a linear address, allow full 57-bitcanonical address
regardless of CR4.LA57 state. For example: MSR_KERNEL_GS_BASE.
2. All hidden segment bases and GDT/IDT bases also behave like MSRs.
This means that full 57-bit canonical address can be loaded to them
regardless of CR4.LA57, both using MSRS (e.g GS_BASE) and instructions
(e.g LGDT).
3. TLB invalidation instructions also allow the user to use full 57-bit
address regardless of the CR4.LA57.
Finally, it must be noted that the CPU doesn't prevent the user from
disabling 5-level paging, even when the full 57-bit canonical address is
present in one of the registers mentioned above (e.g GDT base).
In fact, this can happen without any userspace help, when the CPU enters
SMM mode - some MSRs, for example MSR_KERNEL_GS_BASE are left to contain
a non-canonical address in regard to the new mode.
Since most of the affected MSRs and all segment bases can be read and
written freely by the guest without any KVM intervention, this patch makes
the emulator closely follow hardware behavior, which means that the
emulator doesn't take in the account the guest CPUID support for 5-level
paging, and only takes in the account the host CPU support.
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Link: https://lore.kernel.org/r/20240906221824.491834-4-mlevitsk@redhat.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Drop the @max_level parameter from kvm_mmu_max_mapping_level(). All
callers pass in PG_LEVEL_NUM, so @max_level can be replaced with
PG_LEVEL_NUM in the function body.
No functional change intended.
Signed-off-by: David Matlack <dmatlack@google.com>
Link: https://lore.kernel.org/r/20240823235648.3236880-2-dmatlack@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Now that the shadow MMU and TDP MMU have identical logic for detecting
required TLB flushes when updating SPTEs, move said logic to a helper so
that the TDP MMU code can benefit from the comments that are currently
exclusive to the shadow MMU.
No functional change intended.
Link: https://lore.kernel.org/r/20241011021051.1557902-16-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Use the Accessed bit in SPTEs even when A/D bits are disabled in hardware,
i.e. propagate accessed information to SPTE.Accessed even when KVM is
doing manual tracking by making SPTEs not-present. In addition to
eliminating a small amount of code in is_accessed_spte(), this also paves
the way for preserving Accessed information when a SPTE is zapped in
response to a mmu_notifier PROTECTION event, e.g. if a SPTE is zapped
because NUMA balancing kicks in.
Note, EPT is the only flavor of paging in which A/D bits are conditionally
enabled, and the Accessed (and Dirty) bit is software-available when A/D
bits are disabled.
Note #2, there are currently no concrete plans to preserve Accessed
information. Explorations on that front were the initial catalyst, but
the cleanup is the motivation for the actual commit.
Link: https://lore.kernel.org/r/20241011021051.1557902-13-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Add a dedicated flag to track if KVM has enabled A/D bits at the module
level, instead of inferring the state based on whether or not the MMU's
shadow_accessed_mask is non-zero. This will allow defining and using
shadow_accessed_mask even when A/D bits aren't used by hardware.
Link: https://lore.kernel.org/r/20241011021051.1557902-10-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Fold the guts of mmu_spte_update_no_track() into mmu_spte_update() now
that the latter doesn't flush when clearing A/D bits, i.e. now that there
is no need to explicitly avoid TLB flushes when aging SPTEs.
Opportunistically WARN if mmu_spte_update() requests a TLB flush when
aging SPTEs, as aging should never modify a SPTE in such a way that KVM
thinks a TLB flush is needed.
Link: https://lore.kernel.org/r/20241011021051.1557902-8-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Don't force a TLB flush when an SPTE update in the shadow MMU happens to
clear the Dirty bit, as KVM unconditionally flushes TLBs when enabling
dirty logging, and when clearing dirty logs, KVM flushes based on its
software structures, not the SPTEs. I.e. the flows that care about
accurate Dirty bit information already ensure there are no stale TLB
entries.
Opportunistically drop is_dirty_spte() as mmu_spte_update() was the sole
caller.
Link: https://lore.kernel.org/r/20241011021051.1557902-6-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Don't force a TLB flush if mmu_spte_update() clears the Accessed bit, as
access tracking tolerates false negatives, as evidenced by the
mmu_notifier hooks that explicitly test and age SPTEs without doing a TLB
flush.
In practice, this is very nearly a nop. spte_write_protect() and
spte_clear_dirty() never clear the Accessed bit. make_spte() always
sets the Accessed bit for !prefetch scenarios. FNAME(sync_spte) only sets
SPTE if the protection bits are changing, i.e. if a flush will be needed
regardless of the Accessed bits. And FNAME(pte_prefetch) sets SPTE if and
only if the old SPTE is !PRESENT.
That leaves kvm_arch_async_page_ready() as the one path that will generate
a !ACCESSED SPTE *and* overwrite a PRESENT SPTE. And that's very arguably
a bug, as clobbering a valid SPTE in that case is nonsensical.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Link: https://lore.kernel.org/r/20241011021051.1557902-5-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Don't force a remote TLB flush if KVM happens to effectively "refresh" a
read-only SPTE that is still MMU-Writable, as KVM allows MMU-Writable SPTEs
to have Writable TLB entries, even if the SPTE is !Writable. Remote TLBs
need to be flushed only when creating a read-only SPTE for write-tracking,
i.e. when installing a !MMU-Writable SPTE.
In practice, especially now that KVM doesn't overwrite existing SPTEs when
prefetching, KVM will rarely "refresh" a read-only, MMU-Writable SPTE,
i.e. this is unlikely to eliminate many, if any, TLB flushes. But, more
precisely flushing makes it easier to understand exactly when KVM does and
doesn't need to flush.
Note, x86 architecturally requires relevant TLB entries to be invalidated
on a page fault, i.e. there is no risk of putting a vCPU into an infinite
loop of read-only page faults.
Cc: Yan Zhao <yan.y.zhao@intel.com>
Link: https://lore.kernel.org/r/20241011021051.1557902-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Don't mark pages/folios as accessed in the primary MMU when zapping SPTEs,
as doing so relies on kvm_pfn_to_refcounted_page(), and generally speaking
is unnecessary and wasteful. KVM participates in page aging via
mmu_notifiers, so there's no need to push "accessed" updates to the
primary MMU.
And if KVM zaps a SPTe in response to an mmu_notifier, marking it accessed
_after_ the primary MMU has decided to zap the page is likely to go
unnoticed, i.e. odds are good that, if the page is being zapped for
reclaim, the page will be swapped out regardless of whether or not KVM
marks the page accessed.
Dropping x86's use of kvm_set_pfn_accessed() also paves the way for
removing kvm_pfn_to_refcounted_page() and all its users.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-83-seanjc@google.com>
Move KVM x86's helper that "finishes" the faultin process to common KVM
so that the logic can be shared across all architectures. Note, not all
architectures implement a fast page fault path, but the gist of the
comment applies to all architectures.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-50-seanjc@google.com>
When finishing guest page faults, don't mark pages as accessed if KVM
is resuming the guest _without_ installing a mapping, i.e. if the page
isn't being used. While it's possible that marking the page accessed
could avoid minor thrashing due to reclaiming a page that the guest is
about to access, it's far more likely that the gfn=>pfn mapping was
was invalidated, e.g. due a memslot change, or because the corresponding
VMA is being modified.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-49-seanjc@google.com>
Now that all x86 page fault paths precisely track refcounted pages, use
Use kvm_page_fault.refcounted_page to put references to struct page memory
when finishing page faults. This is a baby step towards eliminating
kvm_pfn_to_refcounted_page().
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-48-seanjc@google.com>
Provide the "struct page" associated with a guest_memfd pfn as an output
from __kvm_gmem_get_pfn() so that KVM guest page fault handlers can
directly put the page instead of having to rely on
kvm_pfn_to_refcounted_page().
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-47-seanjc@google.com>
Convert KVM x86 to use the recently introduced __kvm_faultin_pfn().
Opportunstically capture the refcounted_page grabbed by KVM for use in
future changes.
No functional change intended.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-45-seanjc@google.com>
Move the marking of folios dirty from make_spte() out to its callers,
which have access to the _struct page_, not just the underlying pfn.
Once all architectures follow suit, this will allow removing KVM's ugly
hack where KVM elevates the refcount of VM_MIXEDMAP pfns that happen to
be struct page memory.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-42-seanjc@google.com>
Add a helper to finish/complete the handling of a guest page, e.g. to
mark the pages accessed and put any held references. In the near
future, this will allow improving the logic without having to copy+paste
changes into all page fault paths. And in the less near future, will
allow sharing the "finish" API across all architectures.
No functional change intended.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-41-seanjc@google.com>
Deduplicate the prefetching code for indirect and direct MMUs. The core
logic is the same, the only difference is that indirect MMUs need to
prefetch SPTEs one-at-a-time, as contiguous guest virtual addresses aren't
guaranteed to yield contiguous guest physical addresses.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-40-seanjc@google.com>
Use kvm_release_page_clean() to put prefeteched pages instead of calling
put_page() directly. This will allow de-duplicating the prefetch code
between indirect and direct MMUs.
Note, there's a small functional change as kvm_release_page_clean() marks
the page/folio as accessed. While it's not strictly guaranteed that the
guest will access the page, KVM won't intercept guest accesses, i.e. won't
mark the page accessed if it _is_ accessed by the guest (unless A/D bits
are disabled, but running without A/D bits is effectively limited to
pre-HSW Intel CPUs).
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-39-seanjc@google.com>
Prefix x86's faultin_pfn helpers with "mmu" so that the mmu-less names can
be used by common KVM for similar APIs.
No functional change intended.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-38-seanjc@google.com>
Drop @hva from __gfn_to_pfn_memslot() now that all callers pass NULL.
No functional change intended.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-19-seanjc@google.com>
Remove kvm_page_fault.hva as it is never read, only written. This will
allow removing the @hva param from __gfn_to_pfn_memslot().
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-18-seanjc@google.com>
Add a pfn error code to communicate that hva_to_pfn() failed because I/O
was needed and disallowed, and convert @async to a constant @no_wait
boolean. This will allow eliminating the @no_wait param by having callers
pass in FOLL_NOWAIT along with other FOLL_* flags.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: David Stevens <stevensd@chromium.org>
Co-developed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-17-seanjc@google.com>
Drop @atomic from the myriad "to_pfn" APIs now that all callers pass
"false", and remove a comment blurb about KVM running only the "GUP fast"
part in atomic context.
No functional change intended.
Reviewed-by: Alex Bennée <alex.bennee@linaro.org>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-13-seanjc@google.com>
Rename gfn_to_page_many_atomic() to kvm_prefetch_pages() to try and
communicate its true purpose, as the "atomic" aspect is essentially a
side effect of the fact that x86 uses the API while holding mmu_lock.
E.g. even if mmu_lock weren't held, KVM wouldn't want to fault-in pages,
as the goal is to opportunistically grab surrounding pages that have
already been accessed and/or dirtied by the host, and to do so quickly.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-12-seanjc@google.com>
Now that KVM doesn't clobber Accessed bits of shadow-present SPTEs,
e.g. when prefetching, mark folios as accessed only when zapping leaf
SPTEs, which is a rough heuristic for "only in response to an mmu_notifier
invalidation". Page aging and LRUs are tolerant of false negatives, i.e.
KVM doesn't need to be precise for correctness, and re-marking folios as
accessed when zapping entire roots or when zapping collapsible SPTEs is
expensive and adds very little value.
E.g. when a VM is dying, all of its memory is being freed; marking folios
accessed at that time provides no known value. Similarly, because KVM
marks folios as accessed when creating SPTEs, marking all folios as
accessed when userspace happens to delete a memslot doesn't add value.
The folio was marked access when the old SPTE was created, and will be
marked accessed yet again if a vCPU accesses the pfn again after reloading
a new root. Zapping collapsible SPTEs is a similar story; marking folios
accessed just because userspace disable dirty logging is a side effect of
KVM behavior, not a deliberate goal.
As an intermediate step, a.k.a. bisection point, towards *never* marking
folios accessed when dropping SPTEs, mark folios accessed when the primary
MMU might be invalidating mappings, as such zappings are not KVM initiated,
i.e. might actually be related to page aging and LRU activity.
Note, x86 is the only KVM architecture that "double dips"; every other
arch marks pfns as accessed only when mapping into the guest, not when
mapping into the guest _and_ when removing from the guest.
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-10-seanjc@google.com>
Mark pages/folios dirty when creating SPTEs to map PFNs into the guest,
not when zapping or modifying SPTEs, as marking folios dirty when zapping
or modifying SPTEs can be extremely inefficient. E.g. when KVM is zapping
collapsible SPTEs to reconstitute a hugepage after disbling dirty logging,
KVM will mark every 4KiB pfn as dirty, even though _at least_ 512 pfns are
guaranteed to be in a single folio (the SPTE couldn't potentially be huge
if that weren't the case). The problem only becomes worse for 1GiB
HugeTLB pages, as KVM can mark a single folio dirty 512*512 times.
Marking a folio dirty when mapping is functionally safe as KVM drops all
relevant SPTEs in response to an mmu_notifier invalidation, i.e. ensures
that the guest can't dirty a folio after access has been removed.
And because KVM already marks folios dirty when zapping/modifying SPTEs
for KVM reasons, i.e. not in response to an mmu_notifier invalidation,
there is no danger of "prematurely" marking a folio dirty. E.g. if a
filesystems cleans a folio without first removing write access, then there
already exists races where KVM could mark a folio dirty before remote TLBs
are flushed, i.e. before guest writes are guaranteed to stop. Furthermore,
x86 is literally the only architecture that marks folios dirty on the
backend; every other KVM architecture marks folios dirty at map time.
x86's unique behavior likely stems from the fact that x86's MMU predates
mmu_notifiers. Long, long ago, before mmu_notifiers were added, marking
pages dirty when zapping SPTEs was logical, and perhaps even necessary, as
KVM held references to pages, i.e. kept a page's refcount elevated while
the page was mapped into the guest. At the time, KVM's rmap_remove()
simply did:
if (is_writeble_pte(*spte))
kvm_release_pfn_dirty(pfn);
else
kvm_release_pfn_clean(pfn);
i.e. dropped the refcount and marked the page dirty at the same time.
After mmu_notifiers were introduced, commit acb66dd051 ("KVM: MMU:
don't hold pagecount reference for mapped sptes pages") removed the
refcount logic, but kept the dirty logic, i.e. converted the above to:
if (is_writeble_pte(*spte))
kvm_release_pfn_dirty(pfn);
And for KVM x86, that's essentially how things have stayed over the last
~15 years, without anyone revisiting *why* KVM marks pages/folios dirty at
zap/modification time, e.g. the behavior was blindly carried forward to
the TDP MMU.
Practically speaking, the only downside to marking a folio dirty during
mapping is that KVM could trigger writeback of memory that was never
actually written. Except that can't actually happen if KVM marks folios
dirty if and only if a writable SPTE is created (as done here), because
KVM always marks writable SPTEs as dirty during make_spte(). See commit
9b51a63024 ("KVM: MMU: Explicitly set D-bit for writable spte."), circa
2015.
Note, KVM's access tracking logic for prefetched SPTEs is a bit odd. If a
guest PTE is dirty and writable, KVM will create a writable SPTE, but then
mark the SPTE for access tracking. Which isn't wrong, just a bit odd, as
it results in _more_ precise dirty tracking for MMUs _without_ A/D bits.
To keep things simple, mark the folio dirty before access tracking comes
into play, as an access-tracked SPTE can be restored in the fast page
fault path, i.e. without holding mmu_lock. While writing SPTEs and
accessing memslots outside of mmu_lock is safe, marking a folio dirty is
not. E.g. if the fast path gets interrupted _just_ after setting a SPTE,
the primary MMU could theoretically invalidate and free a folio before KVM
marks it dirty. Unlike the shadow MMU, which waits for CPUs to respond to
an IPI, the TDP MMU only guarantees the page tables themselves won't be
freed (via RCU).
Opportunistically update a few stale comments.
Cc: David Matlack <dmatlack@google.com>
Tested-by: Alex Bennée <alex.bennee@linaro.org>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Tested-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-ID: <20241010182427.1434605-9-seanjc@google.com>
Keith Busch
Sean Christopherson
Paolo Bonzini
Yan Zhao
Rick Edgecombe
Isaku Yamahata
Vipin Sharma
David Matlack
Maxim Levitsky
David Stevens