linux-2.6-microblaze.git
4 years agoKVM: VMX: Clean up vmx_flush_tlb_gva()
Sean Christopherson [Fri, 20 Mar 2020 21:28:11 +0000 (14:28 -0700)]
KVM: VMX: Clean up vmx_flush_tlb_gva()

Refactor vmx_flush_tlb_gva() to remove a superfluous local variable and
clean up its comment, which is oddly located below the code it is
commenting.

No functional change intended.

Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-16-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: make Hyper-V PV TLB flush use tlb_flush_guest()
Vitaly Kuznetsov [Wed, 25 Mar 2020 16:28:24 +0000 (12:28 -0400)]
KVM: x86: make Hyper-V PV TLB flush use tlb_flush_guest()

Hyper-V PV TLB flush mechanism does TLB flush on behalf of the guest
so doing tlb_flush_all() is an overkill, switch to using tlb_flush_guest()
(just like KVM PV TLB flush mechanism) instead. Introduce
KVM_REQ_HV_TLB_FLUSH to support the change.

Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: Move "flush guest's TLB" logic to separate kvm_x86_ops hook
Sean Christopherson [Fri, 20 Mar 2020 21:28:10 +0000 (14:28 -0700)]
KVM: x86: Move "flush guest's TLB" logic to separate kvm_x86_ops hook

Add a dedicated hook to handle flushing TLB entries on behalf of the
guest, i.e. for a paravirtualized TLB flush, and use it directly instead
of bouncing through kvm_vcpu_flush_tlb().

For VMX, change the effective implementation implementation to never do
INVEPT and flush only the current context, i.e. to always flush via
INVVPID(SINGLE_CONTEXT).  The INVEPT performed by __vmx_flush_tlb() when
@invalidate_gpa=false and enable_vpid=0 is unnecessary, as it will only
flush guest-physical mappings; linear and combined mappings are flushed
by VM-Enter when VPID is disabled, and changes in the guest pages tables
do not affect guest-physical mappings.

When EPT and VPID are enabled, doing INVVPID is not required (by Intel's
architecture) to invalidate guest-physical mappings, i.e. TLB entries
that cache guest-physical mappings can live across INVVPID as the
mappings are associated with an EPTP, not a VPID.  The intent of
@invalidate_gpa is to inform vmx_flush_tlb() that it must "invalidate
gpa mappings", i.e. do INVEPT and not simply INVVPID.  Other than nested
VPID handling, which now calls vpid_sync_context() directly, the only
scenario where KVM can safely do INVVPID instead of INVEPT (when EPT is
enabled) is if KVM is flushing TLB entries from the guest's perspective,
i.e. is only required to invalidate linear mappings.

For SVM, flushing TLB entries from the guest's perspective can be done
by flushing the current ASID, as changes to the guest's page tables are
associated only with the current ASID.

Adding a dedicated ->tlb_flush_guest() paves the way toward removing
@invalidate_gpa, which is a potentially dangerous control flag as its
meaning is not exactly crystal clear, even for those who are familiar
with the subtleties of what mappings Intel CPUs are/aren't allowed to
keep across various invalidation scenarios.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-15-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: nVMX: Use vpid_sync_vcpu_addr() to emulate INVVPID with address
Sean Christopherson [Fri, 20 Mar 2020 21:28:09 +0000 (14:28 -0700)]
KVM: nVMX: Use vpid_sync_vcpu_addr() to emulate INVVPID with address

Use vpid_sync_vcpu_addr() to emulate the "individual address" variant of
INVVPID now that said function handles the fallback case of the (host)
CPU not supporting "individual address".

Note, the "vpid == 0" checks in the vpid_sync_*() helpers aren't
actually redundant with the "!operand.vpid" check in handle_invvpid(),
as the vpid passed to vpid_sync_vcpu_addr() is a KVM (host) controlled
value, i.e. vpid02 can be zero even if operand.vpid is non-zero.

No functional change intended.

Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-14-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Drop redundant capability checks in low level INVVPID helpers
Sean Christopherson [Fri, 20 Mar 2020 21:28:08 +0000 (14:28 -0700)]
KVM: VMX: Drop redundant capability checks in low level INVVPID helpers

Remove the INVVPID capabilities checks from vpid_sync_vcpu_single() and
vpid_sync_vcpu_global() now that all callers ensure the INVVPID variant
is supported.  Note, in some cases the guarantee is provided in concert
with hardware_setup(), which enables VPID if and only if at least of
invvpid_single() or invvpid_global() is supported.

Drop the WARN_ON_ONCE() from vmx_flush_tlb() as vpid_sync_vcpu_single()
will trigger a WARN() on INVVPID failure, i.e. if SINGLE_CONTEXT isn't
supported.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-13-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Handle INVVPID fallback logic in vpid_sync_vcpu_addr()
Sean Christopherson [Fri, 20 Mar 2020 21:28:07 +0000 (14:28 -0700)]
KVM: VMX: Handle INVVPID fallback logic in vpid_sync_vcpu_addr()

Directly invoke vpid_sync_context() to do a global INVVPID when the
individual address variant is not supported instead of deferring such
behavior to the caller.  This allows for additional consolidation of
code as the logic is basically identical to the emulation of the
individual address variant in handle_invvpid().

No functional change intended.

Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-12-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Move vpid_sync_vcpu_addr() down a few lines
Sean Christopherson [Fri, 20 Mar 2020 21:28:06 +0000 (14:28 -0700)]
KVM: VMX: Move vpid_sync_vcpu_addr() down a few lines

Move vpid_sync_vcpu_addr() below vpid_sync_context() so that it can be
refactored in a future patch to call vpid_sync_context() directly when
the "individual address" INVVPID variant isn't supported.

No functional change intended.

Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-11-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Use vpid_sync_context() directly when possible
Sean Christopherson [Fri, 20 Mar 2020 21:28:05 +0000 (14:28 -0700)]
KVM: VMX: Use vpid_sync_context() directly when possible

Use vpid_sync_context() directly for flows that run if and only if
enable_vpid=1, or more specifically, nested VMX flows that are gated by
vmx->nested.msrs.secondary_ctls_high.SECONDARY_EXEC_ENABLE_VPID being
set, which is allowed if and only if enable_vpid=1.  Because these flows
call __vmx_flush_tlb() with @invalidate_gpa=false, the if-statement that
decides between INVEPT and INVVPID will always go down the INVVPID path,
i.e. call vpid_sync_context() because
"enable_ept && (invalidate_gpa || !enable_vpid)" always evaluates false.

This helps pave the way toward removing @invalidate_gpa and @vpid from
__vmx_flush_tlb() and its callers.

Opportunstically drop unnecessary brackets in handle_invvpid() around an
affected __vmx_flush_tlb()->vpid_sync_context() conversion.

No functional change intended.

Reviewed-by: Miaohe Lin <linmiaohe@huawei.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-10-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Skip global INVVPID fallback if vpid==0 in vpid_sync_context()
Sean Christopherson [Fri, 20 Mar 2020 21:28:04 +0000 (14:28 -0700)]
KVM: VMX: Skip global INVVPID fallback if vpid==0 in vpid_sync_context()

Skip the global INVVPID in the unlikely scenario that vpid==0 and the
SINGLE_CONTEXT variant of INVVPID is unsupported.  If vpid==0, there's
no need to INVVPID as it's impossible to do VM-Enter with VPID enabled
and vmcs.VPID==0, i.e. there can't be any TLB entries for the vCPU with
vpid==0.  The fact that the SINGLE_CONTEXT variant isn't supported is
irrelevant.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-9-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: Sync SPTEs when injecting page/EPT fault into L1
Junaid Shahid [Fri, 20 Mar 2020 21:28:03 +0000 (14:28 -0700)]
KVM: x86: Sync SPTEs when injecting page/EPT fault into L1

When injecting a page fault or EPT violation/misconfiguration, KVM is
not syncing any shadow PTEs associated with the faulting address,
including those in previous MMUs that are associated with L1's current
EPTP (in a nested EPT scenario), nor is it flushing any hardware TLB
entries.  All this is done by kvm_mmu_invalidate_gva.

Page faults that are either !PRESENT or RSVD are exempt from the flushing,
as the CPU is not allowed to cache such translations.

Signed-off-by: Junaid Shahid <junaids@google.com>
Co-developed-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-8-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: cleanup kvm_inject_emulated_page_fault
Paolo Bonzini [Wed, 25 Mar 2020 16:50:03 +0000 (12:50 -0400)]
KVM: x86: cleanup kvm_inject_emulated_page_fault

To reconstruct the kvm_mmu to be used for page fault injection, we
can simply use fault->nested_page_fault.  This matches how
fault->nested_page_fault is assigned in the first place by
FNAME(walk_addr_generic).

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: introduce kvm_mmu_invalidate_gva
Paolo Bonzini [Tue, 24 Mar 2020 00:42:57 +0000 (20:42 -0400)]
KVM: x86: introduce kvm_mmu_invalidate_gva

Wrap the combination of mmu->invlpg and kvm_x86_ops->tlb_flush_gva
into a new function.  This function also lets us specify the host PGD to
invalidate and also the MMU, both of which will be useful in fixing and
simplifying kvm_inject_emulated_page_fault.

A nested guest's MMU however has g_context->invlpg == NULL.  Instead of
setting it to nonpaging_invlpg, make kvm_mmu_invalidate_gva the only
entry point to mmu->invlpg and make a NULL invlpg pointer equivalent
to nonpaging_invlpg, saving a retpoline.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: Export kvm_propagate_fault() (as kvm_inject_emulated_page_fault)
Sean Christopherson [Fri, 20 Mar 2020 21:28:01 +0000 (14:28 -0700)]
KVM: x86: Export kvm_propagate_fault() (as kvm_inject_emulated_page_fault)

Export the page fault propagation helper so that VMX can use it to
correctly emulate TLB invalidation on page faults in an upcoming patch.

In the (hopefully) not-too-distant future, SGX virtualization will also
want access to the helper for injecting page faults to the correct level
(L1 vs. L2) when emulating ENCLS instructions.

Rename the function to kvm_inject_emulated_page_fault() to clarify that
it is (a) injecting a fault and (b) only for page faults.  WARN if it's
invoked with an exception other than PF_VECTOR.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-6-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: nVMX: Invalidate all roots when emulating INVVPID without EPT
Junaid Shahid [Fri, 20 Mar 2020 21:28:00 +0000 (14:28 -0700)]
KVM: nVMX: Invalidate all roots when emulating INVVPID without EPT

Free all roots when emulating INVVPID for L1 and EPT is disabled, as
outstanding changes to the page tables managed by L1 need to be
recognized.  Because L1 and L2 share an MMU when EPT is disabled, and
because VPID is not tracked by the MMU role, all roots in the current
MMU (root_mmu) need to be freed, otherwise a future nested VM-Enter or
VM-Exit could do a fast CR3 switch (without a flush/sync) and consume
stale SPTEs.

Fixes: 5c614b3583e7b ("KVM: nVMX: nested VPID emulation")
Signed-off-by: Junaid Shahid <junaids@google.com>
[sean: ported to upstream KVM, reworded the comment and changelog]
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-5-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: nVMX: Invalidate all EPTP contexts when emulating INVEPT for L1
Sean Christopherson [Fri, 20 Mar 2020 21:27:59 +0000 (14:27 -0700)]
KVM: nVMX: Invalidate all EPTP contexts when emulating INVEPT for L1

Free all L2 (guest_mmu) roots when emulating INVEPT for L1.  Outstanding
changes to the EPT tables managed by L1 need to be recognized, and
relying on KVM to always flush L2's EPTP context on nested VM-Enter is
dangerous.

Similar to handle_invpcid(), rely on kvm_mmu_free_roots() to do a remote
TLB flush if necessary, e.g. if L1 has never entered L2 then there is
nothing to be done.

Nuking all L2 roots is overkill for the single-context variant, but it's
the safe and easy bet.  A more precise zap mechanism will be added in
the future.  Add a TODO to call out that KVM only needs to invalidate
affected contexts.

Fixes: 14c07ad89f4d ("x86/kvm/mmu: introduce guest_mmu")
Reported-by: Jim Mattson <jmattson@google.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-4-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: nVMX: Validate the EPTP when emulating INVEPT(EXTENT_CONTEXT)
Sean Christopherson [Fri, 20 Mar 2020 21:27:58 +0000 (14:27 -0700)]
KVM: nVMX: Validate the EPTP when emulating INVEPT(EXTENT_CONTEXT)

Signal VM-Fail for the single-context variant of INVEPT if the specified
EPTP is invalid.  Per the INEVPT pseudocode in Intel's SDM, it's subject
to the standard EPT checks:

  If VM entry with the "enable EPT" VM execution control set to 1 would
  fail due to the EPTP value then VMfail(Invalid operand to INVEPT/INVVPID);

Fixes: bfd0a56b90005 ("nEPT: Nested INVEPT")
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-3-sean.j.christopherson@intel.com>
Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Flush all EPTP/VPID contexts on remote TLB flush
Sean Christopherson [Fri, 20 Mar 2020 21:27:57 +0000 (14:27 -0700)]
KVM: VMX: Flush all EPTP/VPID contexts on remote TLB flush

Flush all EPTP/VPID contexts if a TLB flush _may_ have been triggered by
a remote or deferred TLB flush, i.e. by KVM_REQ_TLB_FLUSH.  Remote TLB
flushes require all contexts to be invalidated, not just the active
contexts, e.g. all mappings in all contexts for a given HVA need to be
invalidated on a mmu_notifier invalidation.  Similarly, the instigator
of the deferred TLB flush may be expecting all contexts to be flushed,
e.g. vmx_vcpu_load_vmcs().

Without nested VMX, flushing only the current EPTP/VPID context isn't
problematic because KVM uses a constant VPID for each vCPU, and
mmu_alloc_direct_roots() all but guarantees KVM will use a single EPTP
for L1.  In the rare case where a different EPTP is created or reused,
KVM (currently) unconditionally flushes the new EPTP context prior to
entering the guest.

With nested VMX, KVM conditionally uses a different VPID for L2, and
unconditionally uses a different EPTP for L2.  Because KVM doesn't
_intentionally_ guarantee L2's EPTP/VPID context is flushed on nested
VM-Enter, it'd be possible for a malicious L1 to attack the host and/or
different VMs by exploiting the lack of flushing for L2.

  1) Launch nested guest from malicious L1.

  2) Nested VM-Enter to L2.

  3) Access target GPA 'g'.  CPU inserts TLB entry tagged with L2's ASID
     mapping 'g' to host PFN 'x'.

  2) Nested VM-Exit to L1.

  3) L1 triggers kernel same-page merging (ksm) by duplicating/zeroing
     the page for PFN 'x'.

  4) Host kernel merges PFN 'x' with PFN 'y', i.e. unmaps PFN 'x' and
     remaps the page to PFN 'y'.  mmu_notifier sends invalidate command,
     KVM flushes TLB only for L1's ASID.

  4) Host kernel reallocates PFN 'x' to some other task/guest.

  5) Nested VM-Enter to L2.  KVM does not invalidate L2's EPTP or VPID.

  6) L2 accesses GPA 'g' and gains read/write access to PFN 'x' via its
     stale TLB entry.

However, current KVM unconditionally flushes L1's EPTP/VPID context on
nested VM-Exit.  But, that behavior is mostly unintentional, KVM doesn't
go out of its way to flush EPTP/VPID on nested VM-Enter/VM-Exit, rather
a TLB flush is guaranteed to occur prior to re-entering L1 due to
__kvm_mmu_new_cr3() always being called with skip_tlb_flush=false.  On
nested VM-Enter, this happens via kvm_init_shadow_ept_mmu() (nested EPT
enabled) or in nested_vmx_load_cr3() (nested EPT disabled).  On nested
VM-Exit it occurs via nested_vmx_load_cr3().

This also fixes a bug where a deferred TLB flush in the context of L2,
with EPT disabled, would flush L1's VPID instead of L2's VPID, as
vmx_flush_tlb() flushes L1's VPID regardless of is_guest_mode().

Cc: Vitaly Kuznetsov <vkuznets@redhat.com>
Cc: Ben Gardon <bgardon@google.com>
Cc: Jim Mattson <jmattson@google.com>
Cc: Junaid Shahid <junaids@google.com>
Cc: Liran Alon <liran.alon@oracle.com>
Cc: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: John Haxby <john.haxby@oracle.com>
Reviewed-by: Liran Alon <liran.alon@oracle.com>
Fixes: efebf0aaec3d ("KVM: nVMX: Do not flush TLB on L1<->L2 transitions if L1 uses VPID and EPT")
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200320212833.3507-2-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoselftests: kvm: Add testcase for creating max number of memslots
Wainer dos Santos Moschetta [Fri, 10 Apr 2020 23:17:07 +0000 (16:17 -0700)]
selftests: kvm: Add testcase for creating max number of memslots

This patch introduces test_add_max_memory_regions(), which checks
that a VM can have added memory slots up to the limit defined in
KVM_CAP_NR_MEMSLOTS. Then attempt to add one more slot to
verify it fails as expected.

Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-11-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Make set_memory_region_test common to all architectures
Sean Christopherson [Fri, 10 Apr 2020 23:17:06 +0000 (16:17 -0700)]
KVM: selftests: Make set_memory_region_test common to all architectures

Make set_memory_region_test available on all architectures by wrapping
the bits that are x86-specific in ifdefs.  A future testcase
to create the maximum number of memslots will be architecture
agnostic.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-10-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Add "zero" testcase to set_memory_region_test
Sean Christopherson [Fri, 10 Apr 2020 23:17:05 +0000 (16:17 -0700)]
KVM: selftests: Add "zero" testcase to set_memory_region_test

Add a testcase for running a guest with no memslots to the memory region
test.  The expected result on x86_64 is that the guest will trigger an
internal KVM error due to the initial code fetch encountering a
non-existent memslot and resulting in an emulation failure.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-9-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoselftests: kvm: Add vm_get_fd() in kvm_util
Wainer dos Santos Moschetta [Fri, 10 Apr 2020 23:17:04 +0000 (16:17 -0700)]
selftests: kvm: Add vm_get_fd() in kvm_util

Introduces the vm_get_fd() function in kvm_util which returns
the VM file descriptor.

Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-8-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Add "delete" testcase to set_memory_region_test
Sean Christopherson [Fri, 10 Apr 2020 23:17:03 +0000 (16:17 -0700)]
KVM: selftests: Add "delete" testcase to set_memory_region_test

Add a testcase for deleting memslots while the guest is running.
Like the "move" testcase, this is x86_64-only as it relies on MMIO
happening when a non-existent memslot is encountered.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-7-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: sefltests: Add explicit synchronization to move mem region test
Sean Christopherson [Fri, 10 Apr 2020 23:17:02 +0000 (16:17 -0700)]
KVM: sefltests: Add explicit synchronization to move mem region test

Use sem_post() and sem_timedwait() to synchronize test stages between
the vCPU thread and the main thread instead of using usleep() to wait
for the vCPU thread and hoping for the best.

Opportunistically refactor the code to make it suck less in general,
and to prepare for adding more testcases.

Suggested-by: Peter Xu <peterx@redhat.com>
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-6-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Add GUEST_ASSERT variants to pass values to host
Sean Christopherson [Fri, 10 Apr 2020 23:17:01 +0000 (16:17 -0700)]
KVM: selftests: Add GUEST_ASSERT variants to pass values to host

Add variants of GUEST_ASSERT to pass values back to the host, e.g. to
help debug/understand a failure when the the cause of the assert isn't
necessarily binary.

It'd probably be possible to auto-calculate the number of arguments and
just have a single GUEST_ASSERT, but there are a limited number of
variants and silently eating arguments could lead to subtle code bugs.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200410231707.7128-5-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Add util to delete memory region
Sean Christopherson [Fri, 10 Apr 2020 23:17:00 +0000 (16:17 -0700)]
KVM: selftests: Add util to delete memory region

Add a utility to delete a memory region, it will be used by x86's
set_memory_region_test.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Message-Id: <20200410231707.7128-4-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Use kernel's list instead of homebrewed replacement
Sean Christopherson [Fri, 10 Apr 2020 23:16:59 +0000 (16:16 -0700)]
KVM: selftests: Use kernel's list instead of homebrewed replacement

Replace the KVM selftests' homebrewed linked lists for vCPUs and memory
regions with the kernel's 'struct list_head'.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Message-Id: <20200410231707.7128-3-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: selftests: Take vcpu pointer instead of id in vm_vcpu_rm()
Sean Christopherson [Fri, 10 Apr 2020 23:16:58 +0000 (16:16 -0700)]
KVM: selftests: Take vcpu pointer instead of id in vm_vcpu_rm()

The sole caller of vm_vcpu_rm() already has the vcpu pointer, take it
directly instead of doing an extra lookup.

Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Reviewed-by: Wainer dos Santos Moschetta <wainersm@redhat.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Message-Id: <20200410231707.7128-2-sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: pass through CPUID(0x80000006)
Eric Northup [Wed, 15 Apr 2020 01:23:20 +0000 (18:23 -0700)]
KVM: pass through CPUID(0x80000006)

Return the host's L2 cache and TLB information for CPUID.0x80000006
instead of zeroing out the entry as part of KVM_GET_SUPPORTED_CPUID.
This allows a userspace VMM to feed KVM_GET_SUPPORTED_CPUID's output
directly into KVM_SET_CPUID2 (without breaking the guest).

Signed-off-by: Eric Northup (Google) <digitaleric@gmail.com>
Signed-off-by: Jim Mattson <jmattson@google.com>
Signed-off-by: Jon Cargille <jcargill@google.com>
Message-Id: <20200415012320.236065-1-jcargill@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: x86: Return updated timer current count register from KVM_GET_LAPIC
Peter Shier [Wed, 10 Oct 2018 22:56:53 +0000 (15:56 -0700)]
KVM: x86: Return updated timer current count register from KVM_GET_LAPIC

kvm_vcpu_ioctl_get_lapic (implements KVM_GET_LAPIC ioctl) does a bulk copy
of the LAPIC registers but must take into account that the one-shot and
periodic timer current count register is computed upon reads and is not
present in register state. When restoring LAPIC state (e.g. after
migration), restart timers from their their current count values at time of
save.

Note: When a one-shot timer expires, the code in arch/x86/kvm/lapic.c does
not zero the value of the LAPIC initial count register (emulating HW
behavior). If no other timer is run and pending prior to a subsequent
KVM_GET_LAPIC call, the returned register set will include the expired
one-shot initial count. On a subsequent KVM_SET_LAPIC call the code will
see a non-zero initial count and start a new one-shot timer using the
expired timer's count. This is a prior existing bug and will be addressed
in a separate patch. Thanks to jmattson@google.com for this find.

Signed-off-by: Peter Shier <pshier@google.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
Reviewed-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <20181010225653.238911-1-pshier@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: remove redundant assignment to variable r
Colin Ian King [Fri, 10 Apr 2020 11:35:26 +0000 (12:35 +0100)]
KVM: remove redundant assignment to variable r

The variable r is being assigned  with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Message-Id: <20200410113526.13822-1-colin.king@canonical.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Fix __svm_vcpu_run declaration.
Uros Bizjak [Thu, 9 Apr 2020 11:49:26 +0000 (13:49 +0200)]
KVM: SVM: Fix __svm_vcpu_run declaration.

The function returns no value.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Fixes: 199cd1d7b534 ("KVM: SVM: Split svm_vcpu_run inline assembly to separate file")
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Message-Id: <20200409114926.1407442-1-ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Do not setup frame pointer in __svm_vcpu_run
Uros Bizjak [Thu, 9 Apr 2020 12:04:40 +0000 (14:04 +0200)]
KVM: SVM: Do not setup frame pointer in __svm_vcpu_run

__svm_vcpu_run is a leaf function and does not need
a frame pointer.  %rbp is also destroyed a few instructions
later when guest registers are loaded.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Message-Id: <20200409120440.1427215-1-ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Fix build error due to missing release_pages() include
Borislav Petkov [Sat, 11 Apr 2020 16:09:27 +0000 (18:09 +0200)]
KVM: SVM: Fix build error due to missing release_pages() include

Fix:

  arch/x86/kvm/svm/sev.c: In function â€˜sev_pin_memory’:
  arch/x86/kvm/svm/sev.c:360:3: error: implicit declaration of function â€˜release_pages’;\
  did you mean â€˜reclaim_pages’? [-Werror=implicit-function-declaration]
    360 |   release_pages(pages, npinned);
        |   ^~~~~~~~~~~~~
        |   reclaim_pages

because svm.c includes pagemap.h but the carved out sev.c needs it too.
Triggered by a randconfig build.

Fixes: eaf78265a4ab ("KVM: SVM: Move SEV code to separate file")
Signed-off-by: Borislav Petkov <bp@suse.de>
Message-Id: <20200411160927.27954-1-bp@alien8.de>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Do not mark svm_vcpu_run with STACK_FRAME_NON_STANDARD
Uros Bizjak [Tue, 14 Apr 2020 11:36:12 +0000 (13:36 +0200)]
KVM: SVM: Do not mark svm_vcpu_run with STACK_FRAME_NON_STANDARD

svm_vcpu_run does not change stack or frame pointer anymore.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Message-Id: <20200414113612.104501-1-ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agokvm: nVMX: match comment with return type for nested_vmx_exit_reflected
Oliver Upton [Tue, 14 Apr 2020 22:12:41 +0000 (22:12 +0000)]
kvm: nVMX: match comment with return type for nested_vmx_exit_reflected

nested_vmx_exit_reflected() returns a bool, not int. As such, refer to
the return values as true/false in the comment instead of 1/0.

Signed-off-by: Oliver Upton <oupton@google.com>
Message-Id: <20200414221241.134103-1-oupton@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agokvm: nVMX: reflect MTF VM-exits if injected by L1
Oliver Upton [Tue, 14 Apr 2020 22:47:45 +0000 (22:47 +0000)]
kvm: nVMX: reflect MTF VM-exits if injected by L1

According to SDM 26.6.2, it is possible to inject an MTF VM-exit via the
VM-entry interruption-information field regardless of the 'monitor trap
flag' VM-execution control. KVM appropriately copies the VM-entry
interruption-information field from vmcs12 to vmcs02. However, if L1
has not set the 'monitor trap flag' VM-execution control, KVM fails to
reflect the subsequent MTF VM-exit into L1.

Fix this by consulting the VM-entry interruption-information field of
vmcs12 to determine if L1 has injected the MTF VM-exit. If so, reflect
the exit, regardless of the 'monitor trap flag' VM-execution control.

Fixes: 5f3d45e7f282 ("kvm/x86: add support for MONITOR_TRAP_FLAG")
Signed-off-by: Oliver Upton <oupton@google.com>
Reviewed-by: Peter Shier <pshier@google.com>
Reviewed-by: Jim Mattson <jmattson@google.com>
Message-Id: <20200414224746.240324-1-oupton@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: s390: Return last valid slot if approx index is out-of-bounds
Sean Christopherson [Wed, 8 Apr 2020 06:40:59 +0000 (23:40 -0700)]
KVM: s390: Return last valid slot if approx index is out-of-bounds

Return the index of the last valid slot from gfn_to_memslot_approx() if
its binary search loop yielded an out-of-bounds index.  The index can
be out-of-bounds if the specified gfn is less than the base of the
lowest memslot (which is also the last valid memslot).

Note, the sole caller, kvm_s390_get_cmma(), ensures used_slots is
non-zero.

Fixes: afdad61615cc3 ("KVM: s390: Fix storage attributes migration with memory slots")
Cc: stable@vger.kernel.org # 4.19.x: 0774a964ef56: KVM: Fix out of range accesses to memslots
Cc: stable@vger.kernel.org # 4.19.x
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200408064059.8957-3-sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: Check validity of resolved slot when searching memslots
Sean Christopherson [Wed, 8 Apr 2020 06:40:58 +0000 (23:40 -0700)]
KVM: Check validity of resolved slot when searching memslots

Check that the resolved slot (somewhat confusingly named 'start') is a
valid/allocated slot before doing the final comparison to see if the
specified gfn resides in the associated slot.  The resolved slot can be
invalid if the binary search loop terminated because the search index
was incremented beyond the number of used slots.

This bug has existed since the binary search algorithm was introduced,
but went unnoticed because KVM statically allocated memory for the max
number of slots, i.e. the access would only be truly out-of-bounds if
all possible slots were allocated and the specified gfn was less than
the base of the lowest memslot.  Commit 36947254e5f98 ("KVM: Dynamically
size memslot array based on number of used slots") eliminated the "all
possible slots allocated" condition and made the bug embarrasingly easy
to hit.

Fixes: 9c1a5d38780e6 ("kvm: optimize GFN to memslot lookup with large slots amount")
Reported-by: syzbot+d889b59b2bb87d4047a2@syzkaller.appspotmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com>
Message-Id: <20200408064059.8957-2-sean.j.christopherson@intel.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Enable machine check support for 32bit targets
Uros Bizjak [Tue, 14 Apr 2020 07:14:14 +0000 (09:14 +0200)]
KVM: VMX: Enable machine check support for 32bit targets

There is no reason to limit the use of do_machine_check
to 64bit targets. MCE handling works for both target familes.

Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Sean Christopherson <sean.j.christopherson@intel.com>
Cc: stable@vger.kernel.org
Fixes: a0861c02a981 ("KVM: Add VT-x machine check support")
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Message-Id: <20200414071414.45636-1-ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: move more vmentry code to assembly
Paolo Bonzini [Mon, 13 Apr 2020 07:17:58 +0000 (03:17 -0400)]
KVM: SVM: move more vmentry code to assembly

Manipulate IF around vmload/vmsave to remove the confusing usage of
local_irq_enable where interrupts are actually disabled via GIF.
And stuff the RSB immediately without waiting for a RET to avoid
Spectre-v2 attacks.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: fix compilation with modular PSP and non-modular KVM
Paolo Bonzini [Mon, 13 Apr 2020 07:20:06 +0000 (03:20 -0400)]
KVM: SVM: fix compilation with modular PSP and non-modular KVM

Use svm_sev_enabled() in order to cull all calls to PSP code.  Otherwise,
compilation fails with undefined symbols if the PSP device driver is compiled
as a module and KVM is not.

Reported-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: fix crash cleanup when KVM wasn't used
Vitaly Kuznetsov [Wed, 1 Apr 2020 08:13:48 +0000 (10:13 +0200)]
KVM: VMX: fix crash cleanup when KVM wasn't used

If KVM wasn't used at all before we crash the cleanup procedure fails with
 BUG: unable to handle page fault for address: ffffffffffffffc8
 #PF: supervisor read access in kernel mode
 #PF: error_code(0x0000) - not-present page
 PGD 23215067 P4D 23215067 PUD 23217067 PMD 0
 Oops: 0000 [#8] SMP PTI
 CPU: 0 PID: 3542 Comm: bash Kdump: loaded Tainted: G      D           5.6.0-rc2+ #823
 RIP: 0010:crash_vmclear_local_loaded_vmcss.cold+0x19/0x51 [kvm_intel]

The root cause is that loaded_vmcss_on_cpu list is not yet initialized,
we initialize it in hardware_enable() but this only happens when we start
a VM.

Previously, we used to have a bitmap with enabled CPUs and that was
preventing [masking] the issue.

Initialized loaded_vmcss_on_cpu list earlier, right before we assign
crash_vmclear_loaded_vmcss pointer. blocked_vcpu_on_cpu list and
blocked_vcpu_on_cpu_lock are moved altogether for consistency.

Fixes: 31603d4fc2bb ("KVM: VMX: Always VMCLEAR in-use VMCSes during crash with kexec support")
Signed-off-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Message-Id: <20200401081348.1345307-1-vkuznets@redhat.com>
Reviewed-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: X86: Filter out the broadcast dest for IPI fastpath
Wanpeng Li [Thu, 2 Apr 2020 08:20:26 +0000 (16:20 +0800)]
KVM: X86: Filter out the broadcast dest for IPI fastpath

Except destination shorthand, a destination value 0xffffffff is used to
broadcast interrupts, let's also filter out this for single target IPI
fastpath.

Reviewed-by: Vitaly Kuznetsov <vkuznets@redhat.com>
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <1585815626-28370-1-git-send-email-wanpengli@tencent.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge tag 'kvm-s390-master-5.7-1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Tue, 7 Apr 2020 12:31:44 +0000 (08:31 -0400)]
Merge tag 'kvm-s390-master-5.7-1' of git://git./linux/kernel/git/kvms390/linux into HEAD

KVM: s390: Fixes for vsie (nested hypervisors)

- Several fixes for corner cases of nesting. Still relevant as it might
  crash host or first level guest or temporarily leak memory.

4 years agoKVM: s390: vsie: Fix possible race when shadowing region 3 tables
David Hildenbrand [Fri, 3 Apr 2020 15:30:48 +0000 (17:30 +0200)]
KVM: s390: vsie: Fix possible race when shadowing region 3 tables

We have to properly retry again by returning -EINVAL immediately in case
somebody else instantiated the table concurrently. We missed to add the
goto in this function only. The code now matches the other, similar
shadowing functions.

We are overwriting an existing region 2 table entry. All allocated pages
are added to the crst_list to be freed later, so they are not lost
forever. However, when unshadowing the region 2 table, we wouldn't trigger
unshadowing of the original shadowed region 3 table that we replaced. It
would get unshadowed when the original region 3 table is modified. As it's
not connected to the page table hierarchy anymore, it's not going to get
used anymore. However, for a limited time, this page table will stick
around, so it's in some sense a temporary memory leak.

Identified by manual code inspection. I don't think this classifies as
stable material.

Fixes: 998f637cc4b9 ("s390/mm: avoid races on region/segment/page table shadowing")
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20200403153050.20569-4-david@redhat.com
Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
4 years agoKVM: s390: vsie: Fix delivery of addressing exceptions
David Hildenbrand [Fri, 3 Apr 2020 15:30:47 +0000 (17:30 +0200)]
KVM: s390: vsie: Fix delivery of addressing exceptions

Whenever we get an -EFAULT, we failed to read in guest 2 physical
address space. Such addressing exceptions are reported via a program
intercept to the nested hypervisor.

We faked the intercept, we have to return to guest 2. Instead, right
now we would be returning -EFAULT from the intercept handler, eventually
crashing the VM.
the correct thing to do is to return 1 as rc == 1 is the internal
representation of "we have to go back into g2".

Addressing exceptions can only happen if the g2->g3 page tables
reference invalid g2 addresses (say, either a table or the final page is
not accessible - so something that basically never happens in sane
environments.

Identified by manual code inspection.

Fixes: a3508fbe9dc6 ("KVM: s390: vsie: initial support for nested virtualization")
Cc: <stable@vger.kernel.org> # v4.8+
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20200403153050.20569-3-david@redhat.com
Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
[borntraeger@de.ibm.com: fix patch description]
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
4 years agoKVM: s390: vsie: Fix region 1 ASCE sanity shadow address checks
David Hildenbrand [Fri, 3 Apr 2020 15:30:46 +0000 (17:30 +0200)]
KVM: s390: vsie: Fix region 1 ASCE sanity shadow address checks

In case we have a region 1 the following calculation
(31 + ((gmap->asce & _ASCE_TYPE_MASK) >> 2)*11)
results in 64. As shifts beyond the size are undefined the compiler is
free to use instructions like sllg. sllg will only use 6 bits of the
shift value (here 64) resulting in no shift at all. That means that ALL
addresses will be rejected.

The can result in endless loops, e.g. when prefix cannot get mapped.

Fixes: 4be130a08420 ("s390/mm: add shadow gmap support")
Tested-by: Janosch Frank <frankja@linux.ibm.com>
Reported-by: Janosch Frank <frankja@linux.ibm.com>
Cc: <stable@vger.kernel.org> # v4.8+
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20200403153050.20569-2-david@redhat.com
Reviewed-by: Claudio Imbrenda <imbrenda@linux.ibm.com>
Reviewed-by: Christian Borntraeger <borntraeger@de.ibm.com>
[borntraeger@de.ibm.com: fix patch description, remove WARN_ON_ONCE]
Signed-off-by: Christian Borntraeger <borntraeger@de.ibm.com>
4 years agoKVM: nVMX: don't clear mtf_pending when nested events are blocked
Oliver Upton [Mon, 6 Apr 2020 20:12:37 +0000 (20:12 +0000)]
KVM: nVMX: don't clear mtf_pending when nested events are blocked

If nested events are blocked, don't clear the mtf_pending flag to avoid
missing later delivery of the MTF VM-exit.

Fixes: 5ef8acbdd687c ("KVM: nVMX: Emulate MTF when performing instruction emulation")
Signed-off-by: Oliver Upton <oupton@google.com>
Message-Id: <20200406201237.178725-1-oupton@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: VMX: Remove unnecessary exception trampoline in vmx_vmenter
Uros Bizjak [Mon, 6 Apr 2020 20:21:08 +0000 (22:21 +0200)]
KVM: VMX: Remove unnecessary exception trampoline in vmx_vmenter

The exception trampoline in .fixup section is not needed, the exception
handling code can jump directly to the label in the .text section.

Changes since v1:
- Fix commit message.

Cc: Sean Christopherson <sean.j.christopherson@intel.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Reviewed-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Message-Id: <20200406202108.74300-1-ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Split svm_vcpu_run inline assembly to separate file
Uros Bizjak [Mon, 30 Mar 2020 13:02:13 +0000 (15:02 +0200)]
KVM: SVM: Split svm_vcpu_run inline assembly to separate file

The compiler (GCC) does not like the situation, where there is inline
assembly block that clobbers all available machine registers in the
middle of the function. This situation can be found in function
svm_vcpu_run in file kvm/svm.c and results in many register spills and
fills to/from stack frame.

This patch fixes the issue with the same approach as was done for
VMX some time ago. The big inline assembly is moved to a separate
assembly .S file, taking into account all ABI requirements.

There are two main benefits of the above approach:

* elimination of several register spills and fills to/from stack
frame, and consequently smaller function .text size. The binary size
of svm_vcpu_run is lowered from 2019 to 1626 bytes.

* more efficient access to a register save array. Currently, register
save array is accessed as:

    7b00:    48 8b 98 28 02 00 00     mov    0x228(%rax),%rbx
    7b07:    48 8b 88 18 02 00 00     mov    0x218(%rax),%rcx
    7b0e:    48 8b 90 20 02 00 00     mov    0x220(%rax),%rdx

and passing ia pointer to a register array as an argument to a function one gets:

  12:    48 8b 48 08              mov    0x8(%rax),%rcx
  16:    48 8b 50 10              mov    0x10(%rax),%rdx
  1a:    48 8b 58 18              mov    0x18(%rax),%rbx

As a result, the total size, considering that the new function size is 229
bytes, gets lowered by 164 bytes.

Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Move SEV code to separate file
Joerg Roedel [Tue, 24 Mar 2020 09:41:54 +0000 (10:41 +0100)]
KVM: SVM: Move SEV code to separate file

Move the SEV specific parts of svm.c into the new sev.c file.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
Message-Id: <20200324094154.32352-5-joro@8bytes.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Move AVIC code to separate file
Joerg Roedel [Tue, 31 Mar 2020 16:17:38 +0000 (12:17 -0400)]
KVM: SVM: Move AVIC code to separate file

Move the AVIC related functions from svm.c to the new avic.c file.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
Message-Id: <20200324094154.32352-4-joro@8bytes.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoKVM: SVM: Move Nested SVM Implementation to nested.c
Joerg Roedel [Tue, 24 Mar 2020 09:41:52 +0000 (10:41 +0100)]
KVM: SVM: Move Nested SVM Implementation to nested.c

Split out the code for the nested SVM implementation and move it to a
separate file.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
Message-Id: <20200324094154.32352-3-joro@8bytes.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agokVM SVM: Move SVM related files to own sub-directory
Joerg Roedel [Tue, 24 Mar 2020 09:41:51 +0000 (10:41 +0100)]
kVM SVM: Move SVM related files to own sub-directory

Move svm.c and pmu_amd.c into their own arch/x86/kvm/svm/
subdirectory.

Signed-off-by: Joerg Roedel <jroedel@suse.de>
Message-Id: <20200324094154.32352-2-joro@8bytes.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Thu, 2 Apr 2020 22:13:15 +0000 (15:13 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm updates from Paolo Bonzini:
 "ARM:
   - GICv4.1 support

   - 32bit host removal

  PPC:
   - secure (encrypted) using under the Protected Execution Framework
     ultravisor

  s390:
   - allow disabling GISA (hardware interrupt injection) and protected
     VMs/ultravisor support.

  x86:
   - New dirty bitmap flag that sets all bits in the bitmap when dirty
     page logging is enabled; this is faster because it doesn't require
     bulk modification of the page tables.

   - Initial work on making nested SVM event injection more similar to
     VMX, and less buggy.

   - Various cleanups to MMU code (though the big ones and related
     optimizations were delayed to 5.8). Instead of using cr3 in
     function names which occasionally means eptp, KVM too has
     standardized on "pgd".

   - A large refactoring of CPUID features, which now use an array that
     parallels the core x86_features.

   - Some removal of pointer chasing from kvm_x86_ops, which will also
     be switched to static calls as soon as they are available.

   - New Tigerlake CPUID features.

   - More bugfixes, optimizations and cleanups.

  Generic:
   - selftests: cleanups, new MMU notifier stress test, steal-time test

   - CSV output for kvm_stat"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (277 commits)
  x86/kvm: fix a missing-prototypes "vmread_error"
  KVM: x86: Fix BUILD_BUG() in __cpuid_entry_get_reg() w/ CONFIG_UBSAN=y
  KVM: VMX: Add a trampoline to fix VMREAD error handling
  KVM: SVM: Annotate svm_x86_ops as __initdata
  KVM: VMX: Annotate vmx_x86_ops as __initdata
  KVM: x86: Drop __exit from kvm_x86_ops' hardware_unsetup()
  KVM: x86: Copy kvm_x86_ops by value to eliminate layer of indirection
  KVM: x86: Set kvm_x86_ops only after ->hardware_setup() completes
  KVM: VMX: Configure runtime hooks using vmx_x86_ops
  KVM: VMX: Move hardware_setup() definition below vmx_x86_ops
  KVM: x86: Move init-only kvm_x86_ops to separate struct
  KVM: Pass kvm_init()'s opaque param to additional arch funcs
  s390/gmap: return proper error code on ksm unsharing
  KVM: selftests: Fix cosmetic copy-paste error in vm_mem_region_move()
  KVM: Fix out of range accesses to memslots
  KVM: X86: Micro-optimize IPI fastpath delay
  KVM: X86: Delay read msr data iff writes ICR MSR
  KVM: PPC: Book3S HV: Add a capability for enabling secure guests
  KVM: arm64: GICv4.1: Expose HW-based SGIs in debugfs
  KVM: arm64: GICv4.1: Allow non-trapping WFI when using HW SGIs
  ...

4 years agoMerge tag 'x86-urgent-2020-04-02' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 2 Apr 2020 21:52:12 +0000 (14:52 -0700)]
Merge tag 'x86-urgent-2020-04-02' of git://git./linux/kernel/git/tip/tip

Pull x86 fix from Ingo Molnar:
 "A single fix addressing Sparse warnings. <asm/bitops.h> is changed
  non-trivially to avoid the warnings, but generated code is not
  supposed to be affected"

* tag 'x86-urgent-2020-04-02' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86: Fix bitops.h warning with a moved cast

4 years agoMerge branch 'next-integrity' of git://git.kernel.org/pub/scm/linux/kernel/git/zohar...
Linus Torvalds [Thu, 2 Apr 2020 21:49:46 +0000 (14:49 -0700)]
Merge branch 'next-integrity' of git://git./linux/kernel/git/zohar/linux-integrity

Pull integrity updates from Mimi Zohar:
 "Just a couple of updates for linux-5.7:

   - A new Kconfig option to enable IMA architecture specific runtime
     policy rules needed for secure and/or trusted boot, as requested.

   - Some message cleanup (eg. pr_fmt, additional error messages)"

* 'next-integrity' of git://git.kernel.org/pub/scm/linux/kernel/git/zohar/linux-integrity:
  ima: add a new CONFIG for loading arch-specific policies
  integrity: Remove duplicate pr_fmt definitions
  IMA: Add log statements for failure conditions
  IMA: Update KBUILD_MODNAME for IMA files to ima

4 years agoMerge branch 'akpm' (patches from Andrew)
Linus Torvalds [Thu, 2 Apr 2020 20:55:34 +0000 (13:55 -0700)]
Merge branch 'akpm' (patches from Andrew)

Merge updates from Andrew Morton:
 "A large amount of MM, plenty more to come.

  Subsystems affected by this patch series:
   - tools
   - kthread
   - kbuild
   - scripts
   - ocfs2
   - vfs
   - mm: slub, kmemleak, pagecache, gup, swap, memcg, pagemap, mremap,
         sparsemem, kasan, pagealloc, vmscan, compaction, mempolicy,
         hugetlbfs, hugetlb"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (155 commits)
  include/linux/huge_mm.h: check PageTail in hpage_nr_pages even when !THP
  mm/hugetlb: fix build failure with HUGETLB_PAGE but not HUGEBTLBFS
  selftests/vm: fix map_hugetlb length used for testing read and write
  mm/hugetlb: remove unnecessary memory fetch in PageHeadHuge()
  mm/hugetlb.c: clean code by removing unnecessary initialization
  hugetlb_cgroup: add hugetlb_cgroup reservation docs
  hugetlb_cgroup: add hugetlb_cgroup reservation tests
  hugetlb: support file_region coalescing again
  hugetlb_cgroup: support noreserve mappings
  hugetlb_cgroup: add accounting for shared mappings
  hugetlb: disable region_add file_region coalescing
  hugetlb_cgroup: add reservation accounting for private mappings
  mm/hugetlb_cgroup: fix hugetlb_cgroup migration
  hugetlb_cgroup: add interface for charge/uncharge hugetlb reservations
  hugetlb_cgroup: add hugetlb_cgroup reservation counter
  hugetlbfs: Use i_mmap_rwsem to address page fault/truncate race
  hugetlbfs: use i_mmap_rwsem for more pmd sharing synchronization
  mm/memblock.c: remove redundant assignment to variable max_addr
  mm: mempolicy: require at least one nodeid for MPOL_PREFERRED
  mm: mempolicy: use VM_BUG_ON_VMA in queue_pages_test_walk()
  ...

4 years agoMerge tag 'xfs-5.7-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Thu, 2 Apr 2020 20:02:07 +0000 (13:02 -0700)]
Merge tag 'xfs-5.7-merge-8' of git://git./fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "There's a lot going on this cycle with cleanups in the log code, the
  btree code, and the xattr code.

  We're tightening of metadata validation and online fsck checking, and
  introducing a common btree rebuilding library so that we can refactor
  xfs_repair and introduce online repair in a future cycle.

  We also fixed a few visible bugs -- most notably there's one in
  getdents that we introduced in 5.6; and a fix for hangs when disabling
  quotas.

  This series has been running fstests & other QA in the background for
  over a week and looks good so far.

  I anticipate sending a second pull request next week. That batch will
  change how xfs interacts with memory reclaim; how the log batches and
  throttles log items; how hard writes near ENOSPC will try to squeeze
  more space out of the filesystem; and hopefully fix the last of the
  umount hangs after a catastrophic failure. That should ease a lot of
  problems when running at the limits, but for now I'm leaving that in
  for-next for another week to make sure we got all the subtleties
  right.

  Summary:

   - Fix a hard to trigger race between iclog error checking and log
     shutdown.

   - Strengthen the AGF verifier.

   - Ratelimit some of the more spammy error messages.

   - Remove the icdinode uid/gid members and just use the ones in the
     vfs inode.

   - Hold ILOCK across insert/collapse range.

   - Clean up the extended attribute interfaces.

   - Clean up the attr flags mess.

   - Restore PF_MEMALLOC after exiting xfsaild thread to avoid
     triggering warnings in the process accounting code.

   - Remove the flexibly-sized array from struct xfs_agfl to eliminate
     compiler warnings about unaligned pointers and packed structures.

   - Various macro and typedef removals.

   - Stale metadata buffers if we decide they're corrupt outside of a
     verifier.

   - Check directory data/block/free block owners.

   - Fix a UAF when aborting inactivation of a corrupt xattr fork.

   - Teach online scrub to report failed directory and attr name lookups
     as a metadata corruption instead of a runtime error.

   - Avoid potential buffer overflows in sysfs files by using scnprintf.

   - Fix a regression in getdents lookups due to a mistake in pointer
     arithmetic.

   - Refactor btree cursor private data structures to use anonymous
     unions.

   - Cleanups in the log unmounting code.

   - Fix a potential mishandling of ENOMEM errors on multi-block
     directory buffer lookups.

   - Fix an incorrect test in the block allocation code.

   - Cleanups and name prefix shortening in the scrub code.

   - Introduce btree bulk loading code for online repair and scrub.

   - Fix a quotaoff log item leak (and hang) when the fs goes down
     midway through a quotaoff operation.

   - Remove di_version from the incore inode.

   - Refactor some of the log shutdown checking code.

   - Record the forcing of the log unmount records in the log force
     counters.

   - Fix a longstanding bug where quotacheck would purge the
     administrator's default quota grace interval and warning limits.

   - Reduce memory usage when scrubbing directory and xattr trees.

   - Don't let fsfreeze race with GETFSMAP or online scrub.

   - Handle bio_add_page failures more gracefully in xlog_write_iclog"

* tag 'xfs-5.7-merge-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (108 commits)
  xfs: prohibit fs freezing when using empty transactions
  xfs: shutdown on failure to add page to log bio
  xfs: directory bestfree check should release buffers
  xfs: drop all altpath buffers at the end of the sibling check
  xfs: preserve default grace interval during quotacheck
  xfs: remove xlog_state_want_sync
  xfs: move the ioerror check out of xlog_state_clean_iclog
  xfs: refactor xlog_state_clean_iclog
  xfs: remove the aborted parameter to xlog_state_done_syncing
  xfs: simplify log shutdown checking in xfs_log_release_iclog
  xfs: simplify the xfs_log_release_iclog calling convention
  xfs: factor out a xlog_wait_on_iclog helper
  xfs: merge xlog_cil_push into xlog_cil_push_work
  xfs: remove the di_version field from struct icdinode
  xfs: simplify a check in xfs_ioctl_setattr_check_cowextsize
  xfs: simplify di_flags2 inheritance in xfs_ialloc
  xfs: only check the superblock version for dinode size calculation
  xfs: add a new xfs_sb_version_has_v3inode helper
  xfs: fix unmount hang and memory leak on shutdown during quotaoff
  xfs: factor out quotaoff intent AIL removal and memory free
  ...

4 years agoMerge tag 'vfs-5.7-merge-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Thu, 2 Apr 2020 19:59:36 +0000 (12:59 -0700)]
Merge tag 'vfs-5.7-merge-1' of git://git./fs/xfs/xfs-linux

Pull hibernation fix from Darrick Wong:
 "Fix a regression where we broke the userspace hibernation driver by
  disallowing writes to the swap device"

* tag 'vfs-5.7-merge-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  hibernate: Allow uswsusp to write to swap

4 years agoMerge tag 'iomap-5.7-merge-2' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Thu, 2 Apr 2020 19:57:18 +0000 (12:57 -0700)]
Merge tag 'iomap-5.7-merge-2' of git://git./fs/xfs/xfs-linux

Pull iomap updates from Darrick Wong:
 "We're fixing tracepoints and comments in this cycle, so there
  shouldn't be any surprises here.

  I anticipate sending a second pull request next week with a single bug
  fix for readahead, but it's still undergoing QA.

  Summary:

   - Fix a broken tracepoint

   - Fix a broken comment"

* tag 'iomap-5.7-merge-2' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  iomap: fix comments in iomap_dio_rw
  iomap: Remove pgoff from tracepoints

4 years agoMerge branch 'work.dotdot1' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Thu, 2 Apr 2020 19:30:08 +0000 (12:30 -0700)]
Merge branch 'work.dotdot1' of git://git./linux/kernel/git/viro/vfs

Pull vfs pathwalk sanitizing from Al Viro:
 "Massive pathwalk rewrite and cleanups.

  Several iterations have been posted; hopefully this thing is getting
  readable and understandable now. Pretty much all parts of pathname
  resolutions are affected...

  The branch is identical to what has sat in -next, except for commit
  message in "lift all calls of step_into() out of follow_dotdot/
  follow_dotdot_rcu", crediting Qian Cai for reporting the bug; only
  commit message changed there."

* 'work.dotdot1' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (69 commits)
  lookup_open(): don't bother with fallbacks to lookup+create
  atomic_open(): no need to pass struct open_flags anymore
  open_last_lookups(): move complete_walk() into do_open()
  open_last_lookups(): lift O_EXCL|O_CREAT handling into do_open()
  open_last_lookups(): don't abuse complete_walk() when all we want is unlazy
  open_last_lookups(): consolidate fsnotify_create() calls
  take post-lookup part of do_last() out of loop
  link_path_walk(): sample parent's i_uid and i_mode for the last component
  __nd_alloc_stack(): make it return bool
  reserve_stack(): switch to __nd_alloc_stack()
  pick_link(): take reserving space on stack into a new helper
  pick_link(): more straightforward handling of allocation failures
  fold path_to_nameidata() into its only remaining caller
  pick_link(): pass it struct path already with normal refcounting rules
  fs/namei.c: kill follow_mount()
  non-RCU analogue of the previous commit
  helper for mount rootwards traversal
  follow_dotdot(): be lazy about changing nd->path
  follow_dotdot_rcu(): be lazy about changing nd->path
  follow_dotdot{,_rcu}(): massage loops
  ...

4 years agox86/kvm: fix a missing-prototypes "vmread_error"
Qian Cai [Thu, 2 Apr 2020 15:39:55 +0000 (11:39 -0400)]
x86/kvm: fix a missing-prototypes "vmread_error"

The commit 842f4be95899 ("KVM: VMX: Add a trampoline to fix VMREAD error
handling") removed the declaration of vmread_error() causes a W=1 build
failure with KVM_WERROR=y. Fix it by adding it back.

arch/x86/kvm/vmx/vmx.c:359:17: error: no previous prototype for 'vmread_error' [-Werror=missing-prototypes]
 asmlinkage void vmread_error(unsigned long field, bool fault)
                 ^~~~~~~~~~~~

Signed-off-by: Qian Cai <cai@lca.pw>
Message-Id: <20200402153955.1695-1-cai@lca.pw>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
4 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm...
Linus Torvalds [Thu, 2 Apr 2020 18:22:17 +0000 (11:22 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/ebiederm/user-namespace

Pull exec/proc updates from Eric Biederman:
 "This contains two significant pieces of work: the work to sort out
  proc_flush_task, and the work to solve a deadlock between strace and
  exec.

  Fixing proc_flush_task so that it no longer requires a persistent
  mount makes improvements to proc possible. The removal of the
  persistent mount solves an old regression that that caused the hidepid
  mount option to only work on remount not on mount. The regression was
  found and reported by the Android folks. This further allows Alexey
  Gladkov's work making proc mount options specific to an individual
  mount of proc to move forward.

  The work on exec starts solving a long standing issue with exec that
  it takes mutexes of blocking userspace applications, which makes exec
  extremely deadlock prone. For the moment this adds a second mutex with
  a narrower scope that handles all of the easy cases. Which makes the
  tricky cases easy to spot. With a little luck the code to solve those
  deadlocks will be ready by next merge window"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace: (25 commits)
  signal: Extend exec_id to 64bits
  pidfd: Use new infrastructure to fix deadlocks in execve
  perf: Use new infrastructure to fix deadlocks in execve
  proc: io_accounting: Use new infrastructure to fix deadlocks in execve
  proc: Use new infrastructure to fix deadlocks in execve
  kernel/kcmp.c: Use new infrastructure to fix deadlocks in execve
  kernel: doc: remove outdated comment cred.c
  mm: docs: Fix a comment in process_vm_rw_core
  selftests/ptrace: add test cases for dead-locks
  exec: Fix a deadlock in strace
  exec: Add exec_update_mutex to replace cred_guard_mutex
  exec: Move exec_mmap right after de_thread in flush_old_exec
  exec: Move cleanup of posix timers on exec out of de_thread
  exec: Factor unshare_sighand out of de_thread and call it separately
  exec: Only compute current once in flush_old_exec
  pid: Improve the comment about waiting in zap_pid_ns_processes
  proc: Remove the now unnecessary internal mount of proc
  uml: Create a private mount of proc for mconsole
  uml: Don't consult current to find the proc_mnt in mconsole_proc
  proc: Use a list of inodes to flush from proc
  ...

4 years agoinclude/linux/huge_mm.h: check PageTail in hpage_nr_pages even when !THP
Matthew Wilcox (Oracle) [Thu, 2 Apr 2020 04:11:58 +0000 (21:11 -0700)]
include/linux/huge_mm.h: check PageTail in hpage_nr_pages even when !THP

It's even more important to check that we don't have a tail page when
calling hpage_nr_pages() when THP are disabled.

Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Link: http://lkml.kernel.org/r/20200318140253.6141-4-willy@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/hugetlb: fix build failure with HUGETLB_PAGE but not HUGEBTLBFS
Christophe Leroy [Thu, 2 Apr 2020 04:11:54 +0000 (21:11 -0700)]
mm/hugetlb: fix build failure with HUGETLB_PAGE but not HUGEBTLBFS

When CONFIG_HUGETLB_PAGE is set but not CONFIG_HUGETLBFS, the following
build failure is encoutered:

  In file included from arch/powerpc/mm/fault.c:33:0:
  include/linux/hugetlb.h: In function 'hstate_inode':
  include/linux/hugetlb.h:477:9: error: implicit declaration of function 'HUGETLBFS_SB' [-Werror=implicit-function-declaration]
    return HUGETLBFS_SB(i->i_sb)->hstate;
           ^
  include/linux/hugetlb.h:477:30: error: invalid type argument of '->' (have 'int')
    return HUGETLBFS_SB(i->i_sb)->hstate;
                                ^

Gate hstate_inode() with CONFIG_HUGETLBFS instead of CONFIG_HUGETLB_PAGE.

Fixes: a137e1cc6d6e ("hugetlbfs: per mount huge page sizes")
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Baoquan He <bhe@redhat.com>
Cc: Nishanth Aravamudan <nacc@us.ibm.com>
Cc: Nick Piggin <npiggin@suse.de>
Cc: Adam Litke <agl@us.ibm.com>
Cc: Andi Kleen <ak@suse.de>
Link: http://lkml.kernel.org/r/7e8c3a3c9a587b9cd8a2f146df32a421b961f3a2.1584432148.git.christophe.leroy@c-s.fr
Link: https://patchwork.ozlabs.org/patch/1255548/#2386036
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoselftests/vm: fix map_hugetlb length used for testing read and write
Christophe Leroy [Thu, 2 Apr 2020 04:11:51 +0000 (21:11 -0700)]
selftests/vm: fix map_hugetlb length used for testing read and write

Commit fa7b9a805c79 ("tools/selftest/vm: allow choosing mem size and page
size in map_hugetlb") added the possibility to change the size of memory
mapped for the test, but left the read and write test using the default
value.  This is unnoticed when mapping a length greater than the default
one, but segfaults otherwise.

Fix read_bytes() and write_bytes() by giving them the real length.

Also fix the call to munmap().

Fixes: fa7b9a805c79 ("tools/selftest/vm: allow choosing mem size and page size in map_hugetlb")
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Leonardo Bras <leonardo@linux.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Shuah Khan <shuah@kernel.org>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/9a404a13c871c4bd0ba9ede68f69a1225180dd7e.1580978385.git.christophe.leroy@c-s.fr
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/hugetlb: remove unnecessary memory fetch in PageHeadHuge()
Vlastimil Babka [Thu, 2 Apr 2020 04:11:48 +0000 (21:11 -0700)]
mm/hugetlb: remove unnecessary memory fetch in PageHeadHuge()

Commit f1e61557f023 ("mm: pack compound_dtor and compound_order into one
word in struct page") changed compound_dtor from a pointer to an array
index in order to pack it.  To check if page has the hugeltbfs
compound_dtor, we can just compare the index directly without fetching the
function pointer.  Said commit did that with PageHuge() and we can do the
same with PageHeadHuge() to make the code a bit smaller and faster.

Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Neha Agarwal <nehaagarwal@google.com>
Link: http://lkml.kernel.org/r/20200311172440.6988-1-vbabka@suse.cz
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/hugetlb.c: clean code by removing unnecessary initialization
Mateusz Nosek [Thu, 2 Apr 2020 04:11:45 +0000 (21:11 -0700)]
mm/hugetlb.c: clean code by removing unnecessary initialization

Previously variable 'check_addr' was initialized, but was not read later
before reassigning.  So the initialization can be removed.

Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Link: http://lkml.kernel.org/r/20200303212354.25226-1-mateusznosek0@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add hugetlb_cgroup reservation docs
Mina Almasry [Thu, 2 Apr 2020 04:11:41 +0000 (21:11 -0700)]
hugetlb_cgroup: add hugetlb_cgroup reservation docs

Add docs for how to use hugetlb_cgroup reservations, and their behavior.

Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-9-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add hugetlb_cgroup reservation tests
Mina Almasry [Thu, 2 Apr 2020 04:11:38 +0000 (21:11 -0700)]
hugetlb_cgroup: add hugetlb_cgroup reservation tests

The tests use both shared and private mapped hugetlb memory, and monitors
the hugetlb usage counter as well as the hugetlb reservation counter.
They test different configurations such as hugetlb memory usage via
hugetlbfs, or MAP_HUGETLB, or shmget/shmat, and with and without
MAP_POPULATE.

Also add test for hugetlb reservation reparenting, since this is a subtle
issue.

Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: Sandipan Das <sandipan@linux.ibm.com> [powerpc64]
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-8-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb: support file_region coalescing again
Mina Almasry [Thu, 2 Apr 2020 04:11:35 +0000 (21:11 -0700)]
hugetlb: support file_region coalescing again

An earlier patch in this series disabled file_region coalescing in order
to hang the hugetlb_cgroup uncharge info on the file_region entries.

This patch re-adds support for coalescing of file_region entries.
Essentially everytime we add an entry, we call a recursive function that
tries to coalesce the added region with the regions next to it.  The worst
case call depth for this function is 3: one to coalesce with the region
next to it, one to coalesce to the region prev, and one to reach the base
case.

This is an important performance optimization as private mappings add
their entries page by page, and we could incur big performance costs for
large mappings with lots of file_region entries in their resv_map.

[almasrymina@google.com: fix CONFIG_CGROUP_HUGETLB ifdefs]
Link: http://lkml.kernel.org/r/20200214204544.231482-1-almasrymina@google.com
[almasrymina@google.com: remove check_coalesce_bug debug code]
Link: http://lkml.kernel.org/r/20200219233610.13808-1-almasrymina@google.com
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-7-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: support noreserve mappings
Mina Almasry [Thu, 2 Apr 2020 04:11:31 +0000 (21:11 -0700)]
hugetlb_cgroup: support noreserve mappings

Support MAP_NORESERVE accounting as part of the new counter.

For each hugepage allocation, at allocation time we check if there is a
reservation for this allocation or not.  If there is a reservation for
this allocation, then this allocation was charged at reservation time, and
we don't re-account it.  If there is no reserevation for this allocation,
we charge the appropriate hugetlb_cgroup.

The hugetlb_cgroup to uncharge for this allocation is stored in
page[3].private.  We use new APIs added in an earlier patch to set this
pointer.

Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-6-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add accounting for shared mappings
Mina Almasry [Thu, 2 Apr 2020 04:11:28 +0000 (21:11 -0700)]
hugetlb_cgroup: add accounting for shared mappings

For shared mappings, the pointer to the hugetlb_cgroup to uncharge lives
in the resv_map entries, in file_region->reservation_counter.

After a call to region_chg, we charge the approprate hugetlb_cgroup, and
if successful, we pass on the hugetlb_cgroup info to a follow up
region_add call.  When a file_region entry is added to the resv_map via
region_add, we put the pointer to that cgroup in
file_region->reservation_counter.  If charging doesn't succeed, we report
the error to the caller, so that the kernel fails the reservation.

On region_del, which is when the hugetlb memory is unreserved, we also
uncharge the file_region->reservation_counter.

[akpm@linux-foundation.org: forward declare struct file_region]
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-5-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb: disable region_add file_region coalescing
Mina Almasry [Thu, 2 Apr 2020 04:11:25 +0000 (21:11 -0700)]
hugetlb: disable region_add file_region coalescing

A follow up patch in this series adds hugetlb cgroup uncharge info the
file_region entries in resv->regions.  The cgroup uncharge info may differ
for different regions, so they can no longer be coalesced at region_add
time.  So, disable region coalescing in region_add in this patch.

Behavior change:

Say a resv_map exists like this [0->1], [2->3], and [5->6].

Then a region_chg/add call comes in region_chg/add(f=0, t=5).

Old code would generate resv->regions: [0->5], [5->6].
New code would generate resv->regions: [0->1], [1->2], [2->3], [3->5],
[5->6].

Special care needs to be taken to handle the resv->adds_in_progress
variable correctly.  In the past, only 1 region would be added for every
region_chg and region_add call.  But now, each call may add multiple
regions, so we can no longer increment adds_in_progress by 1 in
region_chg, or decrement adds_in_progress by 1 after region_add or
region_abort.  Instead, region_chg calls add_reservation_in_range() to
count the number of regions needed and allocates those, and that info is
passed to region_add and region_abort to decrement adds_in_progress
correctly.

We've also modified the assumption that region_add after region_chg never
fails.  region_chg now pre-allocates at least 1 region for region_add.  If
region_add needs more regions than region_chg has allocated for it, then
it may fail.

[almasrymina@google.com: fix file_region entry allocations]
Link: http://lkml.kernel.org/r/20200219012736.20363-1-almasrymina@google.com
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Greg Thelen <gthelen@google.com>
Cc: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Link: http://lkml.kernel.org/r/20200211213128.73302-4-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add reservation accounting for private mappings
Mina Almasry [Thu, 2 Apr 2020 04:11:21 +0000 (21:11 -0700)]
hugetlb_cgroup: add reservation accounting for private mappings

Normally the pointer to the cgroup to uncharge hangs off the struct page,
and gets queried when it's time to free the page.  With hugetlb_cgroup
reservations, this is not possible.  Because it's possible for a page to
be reserved by one task and actually faulted in by another task.

The best place to put the hugetlb_cgroup pointer to uncharge for
reservations is in the resv_map.  But, because the resv_map has different
semantics for private and shared mappings, the code patch to
charge/uncharge shared and private mappings is different.  This patch
implements charging and uncharging for private mappings.

For private mappings, the counter to uncharge is in
resv_map->reservation_counter.  On initializing the resv_map this is set
to NULL.  On reservation of a region in private mapping, the tasks
hugetlb_cgroup is charged and the hugetlb_cgroup is placed is
resv_map->reservation_counter.

On hugetlb_vm_op_close, we uncharge resv_map->reservation_counter.

[akpm@linux-foundation.org: forward declare struct resv_map]
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-3-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/hugetlb_cgroup: fix hugetlb_cgroup migration
Mina Almasry [Thu, 2 Apr 2020 04:11:18 +0000 (21:11 -0700)]
mm/hugetlb_cgroup: fix hugetlb_cgroup migration

Commit c32300516047 ("hugetlb_cgroup: add interface for charge/uncharge
hugetlb reservations") mistakingly doesn't handle the migration of *both*
the reservation hugetlb_cgroup and the fault hugetlb_cgroup correctly.

What should happen is that both cgroups shuold be queried from the old
page, then both set to NULL on the old page, then both inserted into the
new page.

The mistake also creates the following warning:

mm/hugetlb_cgroup.c: In function 'hugetlb_cgroup_migrate':
mm/hugetlb_cgroup.c:777:25: warning: variable 'h_cg' set but not used
[-Wunused-but-set-variable]
  struct hugetlb_cgroup *h_cg;
                         ^~~~

Solution is to add the missing steps, namly setting the reservation
hugetlb_cgroup to NULL on the old page, and setting the fault
hugetlb_cgroup on the new page.

Fixes: c32300516047 ("hugetlb_cgroup: add interface for charge/uncharge hugetlb reservations")
Reported-by: Qian Cai <cai@lca.pw>
Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200218194727.46995-1-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add interface for charge/uncharge hugetlb reservations
Mina Almasry [Thu, 2 Apr 2020 04:11:15 +0000 (21:11 -0700)]
hugetlb_cgroup: add interface for charge/uncharge hugetlb reservations

Augments hugetlb_cgroup_charge_cgroup to be able to charge hugetlb usage
or hugetlb reservation counter.

Adds a new interface to uncharge a hugetlb_cgroup counter via
hugetlb_cgroup_uncharge_counter.

Integrates the counter with hugetlb_cgroup, via hugetlb_cgroup_init,
hugetlb_cgroup_have_usage, and hugetlb_cgroup_css_offline.

Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Link: http://lkml.kernel.org/r/20200211213128.73302-2-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlb_cgroup: add hugetlb_cgroup reservation counter
Mina Almasry [Thu, 2 Apr 2020 04:11:11 +0000 (21:11 -0700)]
hugetlb_cgroup: add hugetlb_cgroup reservation counter

These counters will track hugetlb reservations rather than hugetlb memory
faulted in.  This patch only adds the counter, following patches add the
charging and uncharging of the counter.

This is patch 1 of an 9 patch series.

Problem:

Currently tasks attempting to reserve more hugetlb memory than is
available get a failure at mmap/shmget time.  This is thanks to Hugetlbfs
Reservations [1].  However, if a task attempts to reserve more hugetlb
memory than its hugetlb_cgroup limit allows, the kernel will allow the
mmap/shmget call, but will SIGBUS the task when it attempts to fault in
the excess memory.

We have users hitting their hugetlb_cgroup limits and thus we've been
looking at this failure mode.  We'd like to improve this behavior such
that users violating the hugetlb_cgroup limits get an error on mmap/shmget
time, rather than getting SIGBUS'd when they try to fault the excess
memory in.  This gives the user an opportunity to fallback more gracefully
to non-hugetlbfs memory for example.

The underlying problem is that today's hugetlb_cgroup accounting happens
at hugetlb memory *fault* time, rather than at *reservation* time.  Thus,
enforcing the hugetlb_cgroup limit only happens at fault time, and the
offending task gets SIGBUS'd.

Proposed Solution:

A new page counter named
'hugetlb.xMB.rsvd.[limit|usage|max_usage]_in_bytes'. This counter has
slightly different semantics than
'hugetlb.xMB.[limit|usage|max_usage]_in_bytes':

- While usage_in_bytes tracks all *faulted* hugetlb memory,
  rsvd.usage_in_bytes tracks all *reserved* hugetlb memory and hugetlb
  memory faulted in without a prior reservation.

- If a task attempts to reserve more memory than limit_in_bytes allows,
  the kernel will allow it to do so.  But if a task attempts to reserve
  more memory than rsvd.limit_in_bytes, the kernel will fail this
  reservation.

This proposal is implemented in this patch series, with tests to verify
functionality and show the usage.

Alternatives considered:

1. A new cgroup, instead of only a new page_counter attached to the
   existing hugetlb_cgroup.  Adding a new cgroup seemed like a lot of code
   duplication with hugetlb_cgroup.  Keeping hugetlb related page counters
   under hugetlb_cgroup seemed cleaner as well.

2. Instead of adding a new counter, we considered adding a sysctl that
   modifies the behavior of hugetlb.xMB.[limit|usage]_in_bytes, to do
   accounting at reservation time rather than fault time.  Adding a new
   page_counter seems better as userspace could, if it wants, choose to
   enforce different cgroups differently: one via limit_in_bytes, and
   another via rsvd.limit_in_bytes.  This could be very useful if you're
   transitioning how hugetlb memory is partitioned on your system one
   cgroup at a time, for example.  Also, someone may find usage for both
   limit_in_bytes and rsvd.limit_in_bytes concurrently, and this approach
   gives them the option to do so.

Testing:
- Added tests passing.
- Used libhugetlbfs for regression testing.

[1]: https://www.kernel.org/doc/html/latest/vm/hugetlbfs_reserv.html

Signed-off-by: Mina Almasry <almasrymina@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Greg Thelen <gthelen@google.com>
Cc: Sandipan Das <sandipan@linux.ibm.com>
Link: http://lkml.kernel.org/r/20200211213128.73302-1-almasrymina@google.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlbfs: Use i_mmap_rwsem to address page fault/truncate race
Mike Kravetz [Thu, 2 Apr 2020 04:11:08 +0000 (21:11 -0700)]
hugetlbfs: Use i_mmap_rwsem to address page fault/truncate race

hugetlbfs page faults can race with truncate and hole punch operations.
Current code in the page fault path attempts to handle this by 'backing
out' operations if we encounter the race.  One obvious omission in the
current code is removing a page newly added to the page cache.  This is
pretty straight forward to address, but there is a more subtle and
difficult issue of backing out hugetlb reservations.  To handle this
correctly, the 'reservation state' before page allocation needs to be
noted so that it can be properly backed out.  There are four distinct
possibilities for reservation state: shared/reserved, shared/no-resv,
private/reserved and private/no-resv.  Backing out a reservation may
require memory allocation which could fail so that needs to be taken
into account as well.

Instead of writing the required complicated code for this rare
occurrence, just eliminate the race.  i_mmap_rwsem is now held in read
mode for the duration of page fault processing.  Hold i_mmap_rwsem in
write mode when modifying i_size.  In this way, truncation can not
proceed when page faults are being processed.  In addition, i_size
will not change during fault processing so a single check can be made
to ensure faults are not beyond (proposed) end of file.  Faults can
still race with hole punch, but that race is handled by existing code
and the use of hugetlb_fault_mutex.

With this modification, checks for races with truncation in the page
fault path can be simplified and removed.  remove_inode_hugepages no
longer needs to take hugetlb_fault_mutex in the case of truncation.
Comments are expanded to explain reasoning behind locking.

Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Hugh Dickins <hughd@google.com>
Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Prakash Sangappa <prakash.sangappa@oracle.com>
Link: http://lkml.kernel.org/r/20200316205756.146666-3-mike.kravetz@oracle.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agohugetlbfs: use i_mmap_rwsem for more pmd sharing synchronization
Mike Kravetz [Thu, 2 Apr 2020 04:11:05 +0000 (21:11 -0700)]
hugetlbfs: use i_mmap_rwsem for more pmd sharing synchronization

Patch series "hugetlbfs: use i_mmap_rwsem for more synchronization", v2.

While discussing the issue with huge_pte_offset [1], I remembered that
there were more outstanding hugetlb races.  These issues are:

1) For shared pmds, huge PTE pointers returned by huge_pte_alloc can become
   invalid via a call to huge_pmd_unshare by another thread.
2) hugetlbfs page faults can race with truncation causing invalid global
   reserve counts and state.

A previous attempt was made to use i_mmap_rwsem in this manner as
described at [2].  However, those patches were reverted starting with [3]
due to locking issues.

To effectively use i_mmap_rwsem to address the above issues it needs to be
held (in read mode) during page fault processing.  However, during fault
processing we need to lock the page we will be adding.  Lock ordering
requires we take page lock before i_mmap_rwsem.  Waiting until after
taking the page lock is too late in the fault process for the
synchronization we want to do.

To address this lock ordering issue, the following patches change the lock
ordering for hugetlb pages.  This is not too invasive as hugetlbfs
processing is done separate from core mm in many places.  However, I don't
really like this idea.  Much ugliness is contained in the new routine
hugetlb_page_mapping_lock_write() of patch 1.

The only other way I can think of to address these issues is by catching
all the races.  After catching a race, cleanup, backout, retry ...  etc,
as needed.  This can get really ugly, especially for huge page
reservations.  At one time, I started writing some of the reservation
backout code for page faults and it got so ugly and complicated I went
down the path of adding synchronization to avoid the races.  Any other
suggestions would be welcome.

[1] https://lore.kernel.org/linux-mm/1582342427-230392-1-git-send-email-longpeng2@huawei.com/
[2] https://lore.kernel.org/linux-mm/20181222223013.22193-1-mike.kravetz@oracle.com/
[3] https://lore.kernel.org/linux-mm/20190103235452.29335-1-mike.kravetz@oracle.com
[4] https://lore.kernel.org/linux-mm/1584028670.7365.182.camel@lca.pw/
[5] https://lore.kernel.org/lkml/20200312183142.108df9ac@canb.auug.org.au/

This patch (of 2):

While looking at BUGs associated with invalid huge page map counts, it was
discovered and observed that a huge pte pointer could become 'invalid' and
point to another task's page table.  Consider the following:

A task takes a page fault on a shared hugetlbfs file and calls
huge_pte_alloc to get a ptep.  Suppose the returned ptep points to a
shared pmd.

Now, another task truncates the hugetlbfs file.  As part of truncation, it
unmaps everyone who has the file mapped.  If the range being truncated is
covered by a shared pmd, huge_pmd_unshare will be called.  For all but the
last user of the shared pmd, huge_pmd_unshare will clear the pud pointing
to the pmd.  If the task in the middle of the page fault is not the last
user, the ptep returned by huge_pte_alloc now points to another task's
page table or worse.  This leads to bad things such as incorrect page
map/reference counts or invalid memory references.

To fix, expand the use of i_mmap_rwsem as follows:
- i_mmap_rwsem is held in read mode whenever huge_pmd_share is called.
  huge_pmd_share is only called via huge_pte_alloc, so callers of
  huge_pte_alloc take i_mmap_rwsem before calling.  In addition, callers
  of huge_pte_alloc continue to hold the semaphore until finished with
  the ptep.
- i_mmap_rwsem is held in write mode whenever huge_pmd_unshare is called.

One problem with this scheme is that it requires taking i_mmap_rwsem
before taking the page lock during page faults.  This is not the order
specified in the rest of mm code.  Handling of hugetlbfs pages is mostly
isolated today.  Therefore, we use this alternative locking order for
PageHuge() pages.

         mapping->i_mmap_rwsem
           hugetlb_fault_mutex (hugetlbfs specific page fault mutex)
             page->flags PG_locked (lock_page)

To help with lock ordering issues, hugetlb_page_mapping_lock_write() is
introduced to write lock the i_mmap_rwsem associated with a page.

In most cases it is easy to get address_space via vma->vm_file->f_mapping.
However, in the case of migration or memory errors for anon pages we do
not have an associated vma.  A new routine _get_hugetlb_page_mapping()
will use anon_vma to get address_space in these cases.

Signed-off-by: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Prakash Sangappa <prakash.sangappa@oracle.com>
Link: http://lkml.kernel.org/r/20200316205756.146666-2-mike.kravetz@oracle.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/memblock.c: remove redundant assignment to variable max_addr
Colin Ian King [Thu, 2 Apr 2020 04:11:01 +0000 (21:11 -0700)]
mm/memblock.c: remove redundant assignment to variable max_addr

The variable max_addr is being initialized with a value that is never read
and it is being updated later with a new value.  The initialization is
redundant and can be removed.

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Reviewed-by: Mike Rapoport <rppt@linux.ibm.com>
Link: http://lkml.kernel.org/r/20200228235003.112718-1-colin.king@canonical.com
Addresses-Coverity: ("Unused value")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: mempolicy: require at least one nodeid for MPOL_PREFERRED
Randy Dunlap [Thu, 2 Apr 2020 04:10:58 +0000 (21:10 -0700)]
mm: mempolicy: require at least one nodeid for MPOL_PREFERRED

Using an empty (malformed) nodelist that is not caught during mount option
parsing leads to a stack-out-of-bounds access.

The option string that was used was: "mpol=prefer:,".  However,
MPOL_PREFERRED requires a single node number, which is not being provided
here.

Add a check that 'nodes' is not empty after parsing for MPOL_PREFERRED's
nodeid.

Fixes: 095f1fc4ebf3 ("mempolicy: rework shmem mpol parsing and display")
Reported-by: Entropy Moe <3ntr0py1337@gmail.com>
Reported-by: syzbot+b055b1a6b2b958707a21@syzkaller.appspotmail.com
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Tested-by: syzbot+b055b1a6b2b958707a21@syzkaller.appspotmail.com
Cc: Lee Schermerhorn <lee.schermerhorn@hp.com>
Link: http://lkml.kernel.org/r/89526377-7eb6-b662-e1d8-4430928abde9@infradead.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: mempolicy: use VM_BUG_ON_VMA in queue_pages_test_walk()
Yang Shi [Thu, 2 Apr 2020 04:10:55 +0000 (21:10 -0700)]
mm: mempolicy: use VM_BUG_ON_VMA in queue_pages_test_walk()

The VM_BUG_ON() is already used by queue_pages_test_walk(), it sounds
better to dump more debug information by using VM_BUG_ON_VMA() to help
debugging.

Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: "Li Xinhai" <lixinhai.lxh@gmail.com>
Cc: Qian Cai <cai@lca.pw>
Link: http://lkml.kernel.org/r/1579068565-110432-1-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/mempolicy: check hugepage migration is supported by arch in vma_migratable()
Li Xinhai [Thu, 2 Apr 2020 04:10:52 +0000 (21:10 -0700)]
mm/mempolicy: check hugepage migration is supported by arch in vma_migratable()

vma_migratable() is called to check if pages in vma can be migrated before
go ahead to further actions.  Currently it is used in below code path:

- task_numa_work
- mbind
- move_pages

For hugetlb mapping, whether vma is migratable or not is determined by:
- CONFIG_ARCH_ENABLE_HUGEPAGE_MIGRATION
- arch_hugetlb_migration_supported

Issue: current code only checks for CONFIG_ARCH_ENABLE_HUGEPAGE_MIGRATION
alone, and no code should use it directly.  (note that current code in
vma_migratable don't cause failure or bug because
unmap_and_move_huge_page() will catch unsupported hugepage and handle it
properly)

This patch checks the two factors by hugepage_migration_supported for
impoving code logic and robustness.  It will enable early bail out of
hugepage migration procedure, but because currently all architecture
supporting hugepage migration is able to support all page size, we would
not see performance gain with this patch applied.

vma_migratable() is moved to mm/mempolicy.c, because of the circular
reference of mempolicy.h and hugetlb.h cause defining it as inline not
feasible.

Signed-off-by: Li Xinhai <lixinhai.lxh@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Anshuman Khandual <anshuman.khandual@arm.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Link: http://lkml.kernel.org/r/1579786179-30633-1-git-send-email-lixinhai.lxh@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/mempolicy: support MPOL_MF_STRICT for huge page mapping
Li Xinhai [Thu, 2 Apr 2020 04:10:48 +0000 (21:10 -0700)]
mm/mempolicy: support MPOL_MF_STRICT for huge page mapping

MPOL_MF_STRICT is used in mbind() for purposes:

(1) MPOL_MF_STRICT is set alone without MPOL_MF_MOVE or
    MPOL_MF_MOVE_ALL, to check if there is misplaced page and return -EIO;

(2) MPOL_MF_STRICT is set with MPOL_MF_MOVE or MPOL_MF_MOVE_ALL, to
    check if there is misplaced page which is failed to isolate, or page
    is success on isolate but failed to move, and return -EIO.

For non hugepage mapping, (1) and (2) are implemented as expectation.  For
hugepage mapping, (1) is not implemented.  And in (2), the part about
failed to isolate and report -EIO is not implemented.

This patch implements the missed parts for hugepage mapping.  Benefits
with it applied:

- User space can apply same code logic to handle mbind() on hugepage and
  non hugepage mapping;

- Reliably using MPOL_MF_STRICT alone to check whether there is
  misplaced page or not when bind policy on address range, especially for
  address range which contains both hugepage and non hugepage mapping.

Analysis of potential impact to existing users:

- If MPOL_MF_STRICT alone was previously used, hugetlb pages not
  following the memory policy would not cause an EIO error.  After this
  change, hugetlb pages are treated like all other pages.  If
  MPOL_MF_STRICT alone is used and hugetlb pages do not follow memory
  policy an EIO error will be returned.

- For users who using MPOL_MF_STRICT with MPOL_MF_MOVE or
  MPOL_MF_MOVE_ALL, the semantic about some pages could not be moved will
  not be changed by this patch, because failed to isolate and failed to
  move have same effects to users, so their existing code will not be
  impacted.

In mbind man page, the note about 'MPOL_MF_STRICT is ignored on huge page
mappings' can be removed after this patch is applied.

Mike:

: The current behavior with MPOL_MF_STRICT and hugetlb pages is inconsistent
: and does not match documentation (as described above).  The special
: behavior for hugetlb pages ideally should have been removed when hugetlb
: page migration was introduced.  It is unlikely that anyone relies on
: today's inconsistent behavior, and removing one more case of special
: handling for hugetlb pages is a good thing.

Signed-off-by: Li Xinhai <lixinhai.lxh@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Reviewed-by: Naoya Horiguchi <naoya.horiguchi@nec.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: linux-man <linux-man@vger.kernel.org>
Link: http://lkml.kernel.org/r/1581559627-6206-1-git-send-email-lixinhai.lxh@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/compaction.c: clean code by removing unnecessary assignment
Mateusz Nosek [Thu, 2 Apr 2020 04:10:45 +0000 (21:10 -0700)]
mm/compaction.c: clean code by removing unnecessary assignment

Previously 0 was assigned to variable 'last_migrated_pfn'.  But the
variable is not read after that, so the assignment can be removed.

Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Link: http://lkml.kernel.org/r/20200318174509.15021-1-mateusznosek0@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/compaction: Disable compact_unevictable_allowed on RT
Sebastian Andrzej Siewior [Thu, 2 Apr 2020 04:10:42 +0000 (21:10 -0700)]
mm/compaction: Disable compact_unevictable_allowed on RT

Since commit 5bbe3547aa3ba ("mm: allow compaction of unevictable pages")
it is allowed to examine mlocked pages and compact them by default.  On
-RT even minor pagefaults are problematic because it may take a few 100us
to resolve them and until then the task is blocked.

Make compact_unevictable_allowed = 0 default and issue a warning on RT if
it is changed.

[bigeasy@linutronix.de: v5]
Link: https://lore.kernel.org/linux-mm/20190710144138.qyn4tuttdq6h7kqx@linutronix.de/
Link: http://lkml.kernel.org/r/20200319165536.ovi75tsr2seared4@linutronix.de
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Link: https://lore.kernel.org/linux-mm/20190710144138.qyn4tuttdq6h7kqx@linutronix.de/
Link: http://lkml.kernel.org/r/20200303202225.nhqc3v5gwlb7x6et@linutronix.de
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/compaction: really limit compact_unevictable_allowed to 0 and 1
Sebastian Andrzej Siewior [Thu, 2 Apr 2020 04:10:38 +0000 (21:10 -0700)]
mm/compaction: really limit compact_unevictable_allowed to 0 and 1

The proc file `compact_unevictable_allowed' should allow 0 and 1 only, the
`extra*' attribues have been set properly but without
proc_dointvec_minmax() as the `proc_handler' the limit will not be
enforced.

Use proc_dointvec_minmax() as the `proc_handler' to enfoce the valid
specified range.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Luis Chamberlain <mcgrof@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Iurii Zaikin <yzaikin@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Link: http://lkml.kernel.org/r/20200303202054.gsosv7fsx2ma3cic@linutronix.de
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm, compaction: fully assume capture is not NULL in compact_zone_order()
Vlastimil Babka [Thu, 2 Apr 2020 04:10:35 +0000 (21:10 -0700)]
mm, compaction: fully assume capture is not NULL in compact_zone_order()

Dan reports:

The patch 5e1f0f098b46: "mm, compaction: capture a page under direct
compaction" from Mar 5, 2019, leads to the following Smatch complaint:

    mm/compaction.c:2321 compact_zone_order()
     error: we previously assumed 'capture' could be null (see line 2313)

mm/compaction.c
  2288  static enum compact_result compact_zone_order(struct zone *zone, int order,
  2289                  gfp_t gfp_mask, enum compact_priority prio,
  2290                  unsigned int alloc_flags, int classzone_idx,
  2291                  struct page **capture)
                                      ^^^^^^^

  2313 if (capture)
                    ^^^^^^^
Check for NULL

  2314 current->capture_control = &capc;
  2315
  2316 ret = compact_zone(&cc, &capc);
  2317
  2318 VM_BUG_ON(!list_empty(&cc.freepages));
  2319 VM_BUG_ON(!list_empty(&cc.migratepages));
  2320
  2321 *capture = capc.page;
                ^^^^^^^^
Unchecked dereference.

  2322 current->capture_control = NULL;
  2323

In practice this is not an issue, as the only caller path passes non-NULL
capture:

__alloc_pages_direct_compact()
  struct page *page = NULL;
  try_to_compact_pages(capture = &page);
    compact_zone_order(capture = capture);

So let's remove the unnecessary check, which should also make Smatch happy.

Fixes: 5e1f0f098b46 ("mm, compaction: capture a page under direct compaction")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Suggested-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Mel Gorman <mgorman@techsingularity.net>
Link: http://lkml.kernel.org/r/18b0df3c-0589-d96c-23fa-040798fee187@suse.cz
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm,thp,compaction,cma: allow THP migration for CMA allocations
Rik van Riel [Thu, 2 Apr 2020 04:10:31 +0000 (21:10 -0700)]
mm,thp,compaction,cma: allow THP migration for CMA allocations

The code to implement THP migrations already exists, and the code for CMA
to clear out a region of memory already exists.

Only a few small tweaks are needed to allow CMA to move THP memory when
attempting an allocation from alloc_contig_range.

With these changes, migrating THPs from a CMA area works when allocating a
1GB hugepage from CMA memory.

[riel@surriel.com: fix hugetlbfs pages per Mike, cleanup per Vlastimil]
Link: http://lkml.kernel.org/r/20200228104700.0af2f18d@imladris.surriel.com
Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Zi Yan <ziy@nvidia.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: David Rientjes <rientjes@google.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Link: http://lkml.kernel.org/r/20200227213238.1298752-2-riel@surriel.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm,compaction,cma: add alloc_contig flag to compact_control
Rik van Riel [Thu, 2 Apr 2020 04:10:28 +0000 (21:10 -0700)]
mm,compaction,cma: add alloc_contig flag to compact_control

Patch series "fix THP migration for CMA allocations", v2.

Transparent huge pages are allocated with __GFP_MOVABLE, and can end up in
CMA memory blocks.  Transparent huge pages also have most of the
infrastructure in place to allow migration.

However, a few pieces were missing, causing THP migration to fail when
attempting to use CMA to allocate 1GB hugepages.

With these patches in place, THP migration from CMA blocks seems to work,
both for anonymous THPs and for tmpfs/shmem THPs.

This patch (of 2):

Add information to struct compact_control to indicate that the allocator
would really like to clear out this specific part of memory, used by for
example CMA.

Signed-off-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Zi Yan <ziy@nvidia.com>
Cc: Joonsoo Kim <js1304@gmail.com>
Link: http://lkml.kernel.org/r/20200227213238.1298752-1-riel@surriel.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agoselftests: vm: drop dependencies on page flags from mlock2 tests
Michal Hocko [Thu, 2 Apr 2020 04:10:25 +0000 (21:10 -0700)]
selftests: vm: drop dependencies on page flags from mlock2 tests

It was noticed that mlock2 tests are failing after 9c4e6b1a7027f ("mm,
mlock, vmscan: no more skipping pagevecs") because the patch has changed
the timing on when the page is added to the unevictable LRU list and thus
gains the unevictable page flag.

The test was just too dependent on the implementation details which were
true at the time when it was introduced.  Page flags and the timing when
they are set is something no userspace should ever depend on.  The test
should be testing only for the user observable contract of the tested
syscalls.  Those are defined pretty well for the mlock and there are other
means for testing them.  In fact this is already done and testing for page
flags can be safely dropped to achieve the aimed purpose.  Present bits
can be checked by /proc/<pid>/smaps RSS field and the locking state by
VmFlags although I would argue that Locked: field would be more
appropriate.

Drop all the page flag machinery and considerably simplify the test.  This
should be more robust for future kernel changes while checking the
promised contract is still valid.

Fixes: 9c4e6b1a7027f ("mm, mlock, vmscan: no more skipping pagevecs")
Reported-by: Rafael Aquini <aquini@redhat.com>
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Rafael Aquini <aquini@redhat.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Eric B Munson <emunson@akamai.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: <stable@vger.kernel.org>
Link: http://lkml.kernel.org/r/20200324154218.GS19542@dhcp22.suse.cz
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/vmscan.c: do_try_to_free_pages(): clean code by removing unnecessary assignment
Mateusz Nosek [Thu, 2 Apr 2020 04:10:21 +0000 (21:10 -0700)]
mm/vmscan.c: do_try_to_free_pages(): clean code by removing unnecessary assignment

sc->memcg_low_skipped resets skipped_deactivate to 0 but this is not
needed as this code path is never reachable with skipped_deactivate != 0
due to previous sc->skipped_deactivate branch.

[mhocko@kernel.org: rewrite changelog]
Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Link: http://lkml.kernel.org/r/20200319165938.23354-1-mateusznosek0@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/vmscan.c: make may_enter_fs bool in shrink_page_list()
Kirill Tkhai [Thu, 2 Apr 2020 04:10:18 +0000 (21:10 -0700)]
mm/vmscan.c: make may_enter_fs bool in shrink_page_list()

This gives some size improvement:

$size mm/vmscan.o (before)
   text    data     bss     dec     hex filename
  53670   24123      12   77805   12fed mm/vmscan.o

$size mm/vmscan.o (after)
   text    data     bss     dec     hex filename
  53648   24123      12   77783   12fd7 mm/vmscan.o

Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Link: http://lkml.kernel.org/r/Message-ID:
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/vmscan.c: clean code by removing unnecessary assignment
Mateusz Nosek [Thu, 2 Apr 2020 04:10:15 +0000 (21:10 -0700)]
mm/vmscan.c: clean code by removing unnecessary assignment

Previously 0 was assigned to variable 'lruvec_size', but the variable was
never read later.  So the assignment can be removed.

Fixes: f87bccde6a7d ("mm/vmscan: remove unused lru_pages argument")
Signed-off-by: Mateusz Nosek <mateusznosek0@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Reviewed-by: Wei Yang <richard.weiyang@gmail.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Link: http://lkml.kernel.org/r/20200229214022.11853-1-mateusznosek0@gmail.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/vmscan.c: fix data races using kswapd_classzone_idx
Qian Cai [Thu, 2 Apr 2020 04:10:12 +0000 (21:10 -0700)]
mm/vmscan.c: fix data races using kswapd_classzone_idx

pgdat->kswapd_classzone_idx could be accessed concurrently in
wakeup_kswapd().  Plain writes and reads without any lock protection
result in data races.  Fix them by adding a pair of READ|WRITE_ONCE() as
well as saving a branch (compilers might well optimize the original code
in an unintentional way anyway).  While at it, also take care of
pgdat->kswapd_order and non-kswapd threads in allow_direct_reclaim().  The
data races were reported by KCSAN,

 BUG: KCSAN: data-race in wakeup_kswapd / wakeup_kswapd

 write to 0xffff9f427ffff2dc of 4 bytes by task 7454 on cpu 13:
  wakeup_kswapd+0xf1/0x400
  wakeup_kswapd at mm/vmscan.c:3967
  wake_all_kswapds+0x59/0xc0
  wake_all_kswapds at mm/page_alloc.c:4241
  __alloc_pages_slowpath+0xdcc/0x1290
  __alloc_pages_slowpath at mm/page_alloc.c:4512
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x16e/0x6f0
  __handle_mm_fault+0xcd5/0xd40
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 1 lock held by mtest01/7454:
  #0: ffff9f425afe8808 (&mm->mmap_sem#2){++++}, at:
 do_page_fault+0x143/0x6f9
 do_user_addr_fault at arch/x86/mm/fault.c:1405
 (inlined by) do_page_fault at arch/x86/mm/fault.c:1539
 irq event stamp: 6944085
 count_memcg_event_mm+0x1a6/0x270
 count_memcg_event_mm+0x119/0x270
 __do_softirq+0x34c/0x57c
 irq_exit+0xa2/0xc0

 read to 0xffff9f427ffff2dc of 4 bytes by task 7472 on cpu 38:
  wakeup_kswapd+0xc8/0x400
  wake_all_kswapds+0x59/0xc0
  __alloc_pages_slowpath+0xdcc/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x16e/0x6f0
  __handle_mm_fault+0xcd5/0xd40
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

 1 lock held by mtest01/7472:
  #0: ffff9f425a9ac148 (&mm->mmap_sem#2){++++}, at:
 do_page_fault+0x143/0x6f9
 irq event stamp: 6793561
 count_memcg_event_mm+0x1a6/0x270
 count_memcg_event_mm+0x119/0x270
 __do_softirq+0x34c/0x57c
 irq_exit+0xa2/0xc0

 BUG: KCSAN: data-race in kswapd / wakeup_kswapd

 write to 0xffff90973ffff2dc of 4 bytes by task 820 on cpu 6:
  kswapd+0x27c/0x8d0
  kthread+0x1e0/0x200
  ret_from_fork+0x27/0x50

 read to 0xffff90973ffff2dc of 4 bytes by task 6299 on cpu 0:
  wakeup_kswapd+0xf3/0x450
  wake_all_kswapds+0x59/0xc0
  __alloc_pages_slowpath+0xdcc/0x1290
  __alloc_pages_nodemask+0x3bb/0x450
  alloc_pages_vma+0x8a/0x2c0
  do_anonymous_page+0x170/0x700
  __handle_mm_fault+0xc9f/0xd00
  handle_mm_fault+0xfc/0x2f0
  do_page_fault+0x263/0x6f9
  page_fault+0x34/0x40

Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Marco Elver <elver@google.com>
Cc: Matthew Wilcox <willy@infradead.org>
Link: http://lkml.kernel.org/r/1582749472-5171-1-git-send-email-cai@lca.pw
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm/vmscan.c: remove cpu online notification for now
Wei Yang [Thu, 2 Apr 2020 04:10:09 +0000 (21:10 -0700)]
mm/vmscan.c: remove cpu online notification for now

kswapd kernel thread starts either with a CPU affinity set to the full cpu
mask of its target node or without any affinity at all if the node is
CPUless.  There is a cpu hotplug callback (kswapd_cpu_online) that
implements an elaborate way to update this mask when a cpu is onlined.

It is not really clear whether there is any actual benefit from this
scheme. Completely CPU-less NUMA nodes rarely gain a new CPU during
runtime. Drop the code for that reason. If there is a real usecase then
we can resurrect and simplify the code.

[mhocko@suse.com rewrite changelog]

Suggested-by: Michal Hocko <mhocko@suse.org>
Signed-off-by: Wei Yang <richardw.yang@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: David Rientjes <rientjes@google.com>
Link: http://lkml.kernel.org/r/20200218224422.3407-1-richardw.yang@linux.intel.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: vmscan: replace open codings to NUMA_NO_NODE
Yang Shi [Thu, 2 Apr 2020 04:10:05 +0000 (21:10 -0700)]
mm: vmscan: replace open codings to NUMA_NO_NODE

The commit 98fa15f34cb3 ("mm: replace all open encodings for
NUMA_NO_NODE") did the replacement across the kernel tree, but we got
some more in vmscan.c since then.

Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Anshuman Khandual <anshuman.khandual@arm.com>
Acked-by: Minchan Kim <minchan@kernel.org>
Acked-by: David Rientjes <rientjes@google.com>
Link: http://lkml.kernel.org/r/1581568298-45317-1-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
4 years agomm: vmpressure: use mem_cgroup_is_root API
Yang Shi [Thu, 2 Apr 2020 04:10:02 +0000 (21:10 -0700)]
mm: vmpressure: use mem_cgroup_is_root API

Use mem_cgroup_is_root() API to check if memcg is root memcg instead of
open coding.

Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Reviewed-by: David Hildenbrand <david@redhat.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: David Rientjes <rientjes@google.com>
Link: http://lkml.kernel.org/r/1581398649-125989-2-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>