linux.git/arch/x86, branch v4.19.115

ftrace/x86: Anotate text_mutex split between ftrace_arch_code_modify_post_process() and ftrace_arch_code_modify_prepare()

2020-04-02T13:28:15+00:00

commit 074376ac0e1d1fcd4fafebca86ee6158e7c20680 upstream.

ftrace_arch_code_modify_prepare() is acquiring text_mutex, while the
corresponding release is happening in ftrace_arch_code_modify_post_process().

This has already been documented in the code, but let's also make the fact
that this is intentional clear to the semantic analysis tools such as sparse.

Link: http://lkml.kernel.org/r/nycvar.YFH.7.76.1906292321170.27227@cbobk.fhfr.pm

Fixes: 39611265edc1a ("ftrace/x86: Add a comment to why we take text_mutex in ftrace_arch_code_modify_prepare()")
Fixes: d5b844a2cf507 ("ftrace/x86: Remove possible deadlock between register_kprobe() and ftrace_run_update_code()")
Signed-off-by: Jiri Kosina 
Signed-off-by: Steven Rostedt (VMware) 
Cc: Guenter Roeck 
Signed-off-by: Greg Kroah-Hartman

x86/mm: split vmalloc_sync_all()

2020-03-25T07:06:13+00:00

commit 763802b53a427ed3cbd419dbba255c414fdd9e7c upstream.

Commit 3f8fd02b1bf1 ("mm/vmalloc: Sync unmappings in
__purge_vmap_area_lazy()") introduced a call to vmalloc_sync_all() in
the vunmap() code-path.  While this change was necessary to maintain
correctness on x86-32-pae kernels, it also adds additional cycles for
architectures that don't need it.

Specifically on x86-64 with CONFIG_VMAP_STACK=y some people reported
severe performance regressions in micro-benchmarks because it now also
calls the x86-64 implementation of vmalloc_sync_all() on vunmap().  But
the vmalloc_sync_all() implementation on x86-64 is only needed for newly
created mappings.

To avoid the unnecessary work on x86-64 and to gain the performance
back, split up vmalloc_sync_all() into two functions:

	* vmalloc_sync_mappings(), and
	* vmalloc_sync_unmappings()

Most call-sites to vmalloc_sync_all() only care about new mappings being
synchronized.  The only exception is the new call-site added in the
above mentioned commit.

Shile Zhang directed us to a report of an 80% regression in reaim
throughput.

Fixes: 3f8fd02b1bf1 ("mm/vmalloc: Sync unmappings in __purge_vmap_area_lazy()")
Reported-by: kernel test robot 
Reported-by: Shile Zhang 
Signed-off-by: Joerg Roedel 
Signed-off-by: Andrew Morton 
Tested-by: Borislav Petkov 
Acked-by: Rafael J. Wysocki 	[GHES]
Cc: Dave Hansen 
Cc: Andy Lutomirski 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: Ingo Molnar 
Cc: 
Link: http://lkml.kernel.org/r/20191009124418.8286-1-joro@8bytes.org
Link: https://lists.01.org/hyperkitty/list/lkp@lists.01.org/thread/4D3JPPHBNOSPFK2KEPC6KGKS6J25AIDB/
Link: http://lkml.kernel.org/r/20191113095530.228959-1-shile.zhang@linux.alibaba.com
Signed-off-by: Linus Torvalds 
Signed-off-by: Greg Kroah-Hartman

perf/amd/uncore: Replace manual sampling check with CAP_NO_INTERRUPT flag

2020-03-20T10:55:50+00:00

[ Upstream commit f967140dfb7442e2db0868b03b961f9c59418a1b ]

Enable the sampling check in kernel/events/core.c::perf_event_open(),
which returns the more appropriate -EOPNOTSUPP.

BEFORE:

  $ sudo perf record -a -e instructions,l3_request_g1.caching_l3_cache_accesses true
  Error:
  The sys_perf_event_open() syscall returned with 22 (Invalid argument) for event (l3_request_g1.caching_l3_cache_accesses).
  /bin/dmesg | grep -i perf may provide additional information.

With nothing relevant in dmesg.

AFTER:

  $ sudo perf record -a -e instructions,l3_request_g1.caching_l3_cache_accesses true
  Error:
  l3_request_g1.caching_l3_cache_accesses: PMU Hardware doesn't support sampling/overflow-interrupts. Try 'perf stat'

Fixes: c43ca5091a37 ("perf/x86/amd: Add support for AMD NB and L2I "uncore" counters")
Signed-off-by: Kim Phillips 
Signed-off-by: Borislav Petkov 
Acked-by: Peter Zijlstra 
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20200311191323.13124-1-kim.phillips@amd.com
Signed-off-by: Sasha Levin

x86/mce: Fix logic and comments around MSR_PPIN_CTL

2020-03-18T06:14:22+00:00

commit 59b5809655bdafb0767d3fd00a3e41711aab07e6 upstream.

There are two implemented bits in the PPIN_CTL MSR:

Bit 0: LockOut (R/WO)
      Set 1 to prevent further writes to MSR_PPIN_CTL.

Bit 1: Enable_PPIN (R/W)
       If 1, enables MSR_PPIN to be accessible using RDMSR.
       If 0, an attempt to read MSR_PPIN will cause #GP.

So there are four defined values:
	0: PPIN is disabled, PPIN_CTL may be updated
	1: PPIN is disabled. PPIN_CTL is locked against updates
	2: PPIN is enabled. PPIN_CTL may be updated
	3: PPIN is enabled. PPIN_CTL is locked against updates

Code would only enable the X86_FEATURE_INTEL_PPIN feature for case "2".
When it should have done so for both case "2" and case "3".

Fix the final test to just check for the enable bit. Also fix some of
the other comments in this function.

Fixes: 3f5a7896a509 ("x86/mce: Include the PPIN in MCE records when available")
Signed-off-by: Tony Luck 
Signed-off-by: Borislav Petkov 
Cc: 
Link: https://lkml.kernel.org/r/20200226011737.9958-1-tony.luck@intel.com
Signed-off-by: Greg Kroah-Hartman

KVM: x86: clear stale x86_emulate_ctxt->intercept value

2020-03-18T06:14:21+00:00

commit 342993f96ab24d5864ab1216f46c0b199c2baf8e upstream.

After commit 07721feee46b ("KVM: nVMX: Don't emulate instructions in guest
mode") Hyper-V guests on KVM stopped booting with:

 kvm_nested_vmexit:    rip fffff802987d6169 reason EPT_VIOLATION info1 181
    info2 0 int_info 0 int_info_err 0
 kvm_page_fault:       address febd0000 error_code 181
 kvm_emulate_insn:     0:fffff802987d6169: f3 a5
 kvm_emulate_insn:     0:fffff802987d6169: f3 a5 FAIL
 kvm_inj_exception:    #UD (0x0)

"f3 a5" is a "rep movsw" instruction, which should not be intercepted
at all.  Commit c44b4c6ab80e ("KVM: emulate: clean up initializations in
init_decode_cache") reduced the number of fields cleared by
init_decode_cache() claiming that they are being cleared elsewhere,
'intercept', however, is left uncleared if the instruction does not have
any of the "slow path" flags (NotImpl, Stack, Op3264, Sse, Mmx, CheckPerm,
NearBranch, No16 and of course Intercept itself).

Fixes: c44b4c6ab80e ("KVM: emulate: clean up initializations in init_decode_cache")
Fixes: 07721feee46b ("KVM: nVMX: Don't emulate instructions in guest mode")
Cc: stable@vger.kernel.org
Suggested-by: Paolo Bonzini 
Signed-off-by: Vitaly Kuznetsov 
Reviewed-by: Sean Christopherson 
Signed-off-by: Paolo Bonzini 
Signed-off-by: Greg Kroah-Hartman

KVM: SVM: fix up incorrect backport

2020-03-16T08:52:03+00:00

When I backported 52918ed5fcf0 ("KVM: SVM: Override default MMIO mask if
memory encryption is enabled") to 4.19 (which resulted in commit
a4e761c9f63a ("KVM: SVM: Override default MMIO mask if memory encryption
is enabled")), I messed up the call to kvm_mmu_set_mmio_spte_mask()

Fix that here now.

Reported-by: Tom Lendacky 
Cc: Sean Christopherson 
Cc: Paolo Bonzini 
Signed-off-by: Greg Kroah-Hartman

efi/x86: Handle by-ref arguments covering multiple pages in mixed mode

2020-03-11T13:15:12+00:00

commit 8319e9d5ad98ffccd19f35664382c73cea216193 upstream.

The mixed mode runtime wrappers are fragile when it comes to how the
memory referred to by its pointer arguments are laid out in memory, due
to the fact that it translates these addresses to physical addresses that
the runtime services can dereference when running in 1:1 mode. Since
vmalloc'ed pages (including the vmap'ed stack) are not contiguous in the
physical address space, this scheme only works if the referenced memory
objects do not cross page boundaries.

Currently, the mixed mode runtime service wrappers require that all by-ref
arguments that live in the vmalloc space have a size that is a power of 2,
and are aligned to that same value. While this is a sensible way to
construct an object that is guaranteed not to cross a page boundary, it is
overly strict when it comes to checking whether a given object violates
this requirement, as we can simply take the physical address of the first
and the last byte, and verify that they point into the same physical page.

When this check fails, we emit a WARN(), but then simply proceed with the
call, which could cause data corruption if the next physical page belongs
to a mapping that is entirely unrelated.

Given that with vmap'ed stacks, this condition is much more likely to
trigger, let's relax the condition a bit, but fail the runtime service
call if it does trigger.

Fixes: f6697df36bdf0bf7 ("x86/efi: Prevent mixed mode boot corruption with CONFIG_VMAP_STACK=y")
Signed-off-by: Ard Biesheuvel 
Signed-off-by: Ingo Molnar 
Cc: linux-efi@vger.kernel.org
Cc: Ingo Molnar 
Cc: Thomas Gleixner 
Link: https://lore.kernel.org/r/20200221084849.26878-4-ardb@kernel.org
Signed-off-by: Greg Kroah-Hartman

efi/x86: Align GUIDs to their size in the mixed mode runtime wrapper

2020-03-11T13:15:12+00:00

commit 63056e8b5ebf41d52170e9f5ba1fc83d1855278c upstream.

Hans reports that his mixed mode systems running v5.6-rc1 kernels hit
the WARN_ON() in virt_to_phys_or_null_size(), caused by the fact that
efi_guid_t objects on the vmap'ed stack happen to be misaligned with
respect to their sizes. As a quick (i.e., backportable) fix, copy GUID
pointer arguments to the local stack into a buffer that is naturally
aligned to its size, so that it is guaranteed to cover only one
physical page.

Note that on x86, we cannot rely on the stack pointer being aligned
the way the compiler expects, so we need to allocate an 8-byte aligned
buffer of sufficient size, and copy the GUID into that buffer at an
offset that is aligned to 16 bytes.

Fixes: f6697df36bdf0bf7 ("x86/efi: Prevent mixed mode boot corruption with CONFIG_VMAP_STACK=y")
Reported-by: Hans de Goede 
Signed-off-by: Ard Biesheuvel 
Signed-off-by: Ingo Molnar 
Tested-by: Hans de Goede 
Cc: linux-efi@vger.kernel.org
Cc: Ingo Molnar 
Cc: Thomas Gleixner 
Link: https://lore.kernel.org/r/20200221084849.26878-2-ardb@kernel.org
Signed-off-by: Greg Kroah-Hartman

x86/pkeys: Manually set X86_FEATURE_OSPKE to preserve existing changes

2020-03-11T13:15:04+00:00

commit 735a6dd02222d8d070c7bb748f25895239ca8c92 upstream.

Explicitly set X86_FEATURE_OSPKE via set_cpu_cap() instead of calling
get_cpu_cap() to pull the feature bit from CPUID after enabling CR4.PKE.
Invoking get_cpu_cap() effectively wipes out any {set,clear}_cpu_cap()
changes that were made between this_cpu->c_init() and setup_pku(), as
all non-synthetic feature words are reinitialized from the CPU's CPUID
values.

Blasting away capability updates manifests most visibility when running
on a VMX capable CPU, but with VMX disabled by BIOS.  To indicate that
VMX is disabled, init_ia32_feat_ctl() clears X86_FEATURE_VMX, using
clear_cpu_cap() instead of setup_clear_cpu_cap() so that KVM can report
which CPU is misconfigured (KVM needs to probe every CPU anyways).
Restoring X86_FEATURE_VMX from CPUID causes KVM to think VMX is enabled,
ultimately leading to an unexpected #GP when KVM attempts to do VMXON.

Arguably, init_ia32_feat_ctl() should use setup_clear_cpu_cap() and let
KVM figure out a different way to report the misconfigured CPU, but VMX
is not the only feature bit that is affected, i.e. there is precedent
that tweaking feature bits via {set,clear}_cpu_cap() after ->c_init()
is expected to work.  Most notably, x86_init_rdrand()'s clearing of
X86_FEATURE_RDRAND when RDRAND malfunctions is also overwritten.

Fixes: 0697694564c8 ("x86/mm/pkeys: Actually enable Memory Protection Keys in the CPU")
Reported-by: Jacob Keller 
Signed-off-by: Sean Christopherson 
Signed-off-by: Borislav Petkov 
Acked-by: Dave Hansen 
Tested-by: Jacob Keller 
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20200226231615.13664-1-sean.j.christopherson@intel.com
Signed-off-by: Greg Kroah-Hartman

x86/xen: Distribute switch variables for initialization

2020-03-11T13:14:55+00:00

[ Upstream commit 9038ec99ceb94fb8d93ade5e236b2928f0792c7c ]

Variables declared in a switch statement before any case statements
cannot be automatically initialized with compiler instrumentation (as
they are not part of any execution flow). With GCC's proposed automatic
stack variable initialization feature, this triggers a warning (and they
don't get initialized). Clang's automatic stack variable initialization
(via CONFIG_INIT_STACK_ALL=y) doesn't throw a warning, but it also
doesn't initialize such variables[1]. Note that these warnings (or silent
skipping) happen before the dead-store elimination optimization phase,
so even when the automatic initializations are later elided in favor of
direct initializations, the warnings remain.

To avoid these problems, move such variables into the "case" where
they're used or lift them up into the main function body.

arch/x86/xen/enlighten_pv.c: In function ‘xen_write_msr_safe’:
arch/x86/xen/enlighten_pv.c:904:12: warning: statement will never be executed [-Wswitch-unreachable]
  904 |   unsigned which;
      |            ^~~~~

[1] https://bugs.llvm.org/show_bug.cgi?id=44916

Signed-off-by: Kees Cook 
Link: https://lore.kernel.org/r/20200220062318.69299-1-keescook@chromium.org
Reviewed-by: Juergen Gross 
[boris: made @which an 'unsigned int']
Signed-off-by: Boris Ostrovsky 
Signed-off-by: Sasha Levin