commit f710202b2a upstream.
After commit c104c16073 ("Kunit to check the longest symbol length"),
there is a warning when building with clang because there is now a
definition of unlikely from compiler.h in tools/include/linux, which
conflicts with the one in the instruction decoder selftest:
arch/x86/tools/insn_decoder_test.c:15:9: warning: 'unlikely' macro redefined [-Wmacro-redefined]
Remove the second unlikely() definition, as it is no longer necessary,
clearing up the warning.
Fixes: c104c16073 ("Kunit to check the longest symbol length")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Shuah Khan <skhan@linuxfoundation.org>
Link: https://lore.kernel.org/r/20250318-x86-decoder-test-fix-unlikely-redef-v1-1-74c84a7bf05b@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c104c16073 upstream.
The longest length of a symbol (KSYM_NAME_LEN) was increased to 512
in the reference [1]. This patch adds kunit test suite to check the longest
symbol length. These tests verify that the longest symbol length defined
is supported.
This test can also help other efforts for longer symbol length,
like [2].
The test suite defines one symbol with the longest possible length.
The first test verify that functions with names of the created
symbol, can be called or not.
The second test, verify that the symbols are created (or
not) in the kernel symbol table.
[1] https://lore.kernel.org/lkml/20220802015052.10452-6-ojeda@kernel.org/
[2] https://lore.kernel.org/lkml/20240605032120.3179157-1-song@kernel.org/
Link: https://lore.kernel.org/r/20250302221518.76874-1-sergio.collado@gmail.com
Tested-by: Martin Rodriguez Reboredo <yakoyoku@gmail.com>
Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>
Reviewed-by: Rae Moar <rmoar@google.com>
Signed-off-by: Sergio González Collado <sergio.collado@gmail.com>
Link: https://github.com/Rust-for-Linux/linux/issues/504
Reviewed-by: Rae Moar <rmoar@google.com>
Acked-by: David Gow <davidgow@google.com>
Signed-off-by: Shuah Khan <shuah@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 0b3bc018e8 upstream.
Two 'static inline' TDX helper functions (sc_retry() and
sc_retry_prerr()) take function pointer arguments which refer to
assembly functions. Normally, the compiler inlines the TDX helper,
realizes that the function pointer targets are completely static --
thus can be resolved at compile time -- and generates direct call
instructions.
But, other times (like when CONFIG_CC_OPTIMIZE_FOR_SIZE=y), the
compiler declines to inline the helpers and will instead generate
indirect call instructions.
Indirect calls to assembly functions require special annotation (for
various Control Flow Integrity mechanisms). But TDX assembly
functions lack the special annotations and can only be called
directly.
Annotate both the helpers as '__always_inline' to prod the compiler
into maintaining the direct calls. There is no guarantee here, but
Peter has volunteered to report the compiler bug if this assumption
ever breaks[1].
Fixes: 1e66a7e275 ("x86/virt/tdx: Handle SEAMCALL no entropy error in common code")
Fixes: df01f5ae07 ("x86/virt/tdx: Add SEAMCALL error printing for module initialization")
Signed-off-by: Kai Huang <kai.huang@intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/lkml/20250605145914.GW39944@noisy.programming.kicks-ass.net/ [1]
Link: https://lore.kernel.org/all/20250606130737.30713-1-kai.huang%40intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit ed16618c38 ]
TL;DR: SGX page reclaim touches the page to copy its contents to
secondary storage. SGX instructions do not gracefully handle machine
checks. Despite this, the existing SGX code will try to reclaim pages
that it _knows_ are poisoned. Avoid even trying to reclaim poisoned pages.
The longer story:
Pages used by an enclave only get epc_page->poison set in
arch_memory_failure() but they currently stay on sgx_active_page_list until
sgx_encl_release(), with the SGX_EPC_PAGE_RECLAIMER_TRACKED flag untouched.
epc_page->poison is not checked in the reclaimer logic meaning that, if other
conditions are met, an attempt will be made to reclaim an EPC page that was
poisoned. This is bad because 1. we don't want that page to end up added
to another enclave and 2. it is likely to cause one core to shut down
and the kernel to panic.
Specifically, reclaiming uses microcode operations including "EWB" which
accesses the EPC page contents to encrypt and write them out to non-SGX
memory. Those operations cannot handle MCEs in their accesses other than
by putting the executing core into a special shutdown state (affecting
both threads with HT.) The kernel will subsequently panic on the
remaining cores seeing the core didn't enter MCE handler(s) in time.
Call sgx_unmark_page_reclaimable() to remove the affected EPC page from
sgx_active_page_list on memory error to stop it being considered for
reclaiming.
Testing epc_page->poison in sgx_reclaim_pages() would also work but I assume
it's better to add code in the less likely paths.
The affected EPC page is not added to &node->sgx_poison_page_list until
later in sgx_encl_release()->sgx_free_epc_page() when it is EREMOVEd.
Membership on other lists doesn't change to avoid changing any of the
lists' semantics except for sgx_active_page_list. There's a "TBD" comment
in arch_memory_failure() about pre-emptive actions, the goal here is not
to address everything that it may imply.
This also doesn't completely close the time window when a memory error
notification will be fatal (for a not previously poisoned EPC page) --
the MCE can happen after sgx_reclaim_pages() has selected its candidates
or even *inside* a microcode operation (actually easy to trigger due to
the amount of time spent in them.)
The spinlock in sgx_unmark_page_reclaimable() is safe because
memory_failure() runs in process context and no spinlocks are held,
explicitly noted in a mm/memory-failure.c comment.
Signed-off-by: Andrew Zaborowski <andrew.zaborowski@intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Tony Luck <tony.luck@intel.com>
Cc: balrogg@gmail.com
Cc: linux-sgx@vger.kernel.org
Link: https://lore.kernel.org/r/20250508230429.456271-1-andrew.zaborowski@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit a0ee1d5faf upstream.
Ensure the shadow VMCS cache is evicted during an emergency reboot to
prevent potential memory corruption if the cache is evicted after reboot.
This issue was identified through code inspection, as __loaded_vmcs_clear()
flushes both the normal VMCS and the shadow VMCS.
Avoid checking the "launched" state during an emergency reboot, unlike the
behavior in __loaded_vmcs_clear(). This is important because reboot NMIs
can interfere with operations like copy_shadow_to_vmcs12(), where shadow
VMCSes are loaded directly using VMPTRLD. In such cases, if NMIs occur
right after the VMCS load, the shadow VMCSes will be active but the
"launched" state may not be set.
Fixes: 16f5b9034b ("KVM: nVMX: Copy processor-specific shadow-vmcs to VMCS12")
Cc: stable@vger.kernel.org
Signed-off-by: Chao Gao <chao.gao@intel.com>
Reviewed-by: Kai Huang <kai.huang@intel.com>
Link: https://lore.kernel.org/r/20250324140849.2099723-1-chao.gao@intel.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 1bee4838eb upstream.
When freeing a vCPU and thus its VMCB, clear current_vmcb for all possible
CPUs, not just online CPUs, as it's theoretically possible a CPU could go
offline and come back online in conjunction with KVM reusing the page for
a new VMCB.
Link: https://lore.kernel.org/all/20250320013759.3965869-1-yosry.ahmed@linux.dev
Fixes: fd65d3142f ("kvm: svm: Ensure an IBPB on all affected CPUs when freeing a vmcb")
Cc: stable@vger.kernel.org
Cc: Jim Mattson <jmattson@google.com>
Signed-off-by: Yosry Ahmed <yosry.ahmed@linux.dev>
[sean: split to separate patch, write changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit e34dbbc85d upstream.
Clear the software event flag in the augmented SS to prevent immediate
repeat of single step trap on return from SIGTRAP handler if the trap
flag (TF) is set without an external debugger attached.
Following is a typical single-stepping flow for a user process:
1) The user process is prepared for single-stepping by setting
RFLAGS.TF = 1.
2) When any instruction in user space completes, a #DB is triggered.
3) The kernel handles the #DB and returns to user space, invoking the
SIGTRAP handler with RFLAGS.TF = 0.
4) After the SIGTRAP handler finishes, the user process performs a
sigreturn syscall, restoring the original state, including
RFLAGS.TF = 1.
5) Goto step 2.
According to the FRED specification:
A) Bit 17 in the augmented SS is designated as the software event
flag, which is set to 1 for FRED event delivery of SYSCALL,
SYSENTER, or INT n.
B) If bit 17 of the augmented SS is 1 and ERETU would result in
RFLAGS.TF = 1, a single-step trap will be pending upon completion
of ERETU.
In step 4) above, the software event flag is set upon the sigreturn
syscall, and its corresponding ERETU would restore RFLAGS.TF = 1.
This combination causes a pending single-step trap upon completion of
ERETU. Therefore, another #DB is triggered before any user space
instruction is executed, which leads to an infinite loop in which the
SIGTRAP handler keeps being invoked on the same user space IP.
Fixes: 14619d912b ("x86/fred: FRED entry/exit and dispatch code")
Suggested-by: H. Peter Anvin (Intel) <hpa@zytor.com>
Signed-off-by: Xin Li (Intel) <xin@zytor.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Tested-by: Sohil Mehta <sohil.mehta@intel.com>
Cc:stable@vger.kernel.org
Link: https://lore.kernel.org/all/20250609084054.2083189-2-xin%40zytor.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8b68e97871 upstream.
io_bitmap_exit() is invoked from exit_thread() when a task exists or
when a fork fails. In the latter case the exit_thread() cleans up
resources which were allocated during fork().
io_bitmap_exit() invokes task_update_io_bitmap(), which in turn ends up
in tss_update_io_bitmap(). tss_update_io_bitmap() operates on the
current task. If current has TIF_IO_BITMAP set, but no bitmap installed,
tss_update_io_bitmap() crashes with a NULL pointer dereference.
There are two issues, which lead to that problem:
1) io_bitmap_exit() should not invoke task_update_io_bitmap() when
the task, which is cleaned up, is not the current task. That's a
clear indicator for a cleanup after a failed fork().
2) A task should not have TIF_IO_BITMAP set and neither a bitmap
installed nor IOPL emulation level 3 activated.
This happens when a kernel thread is created in the context of
a user space thread, which has TIF_IO_BITMAP set as the thread
flags are copied and the IO bitmap pointer is cleared.
Other than in the failed fork() case this has no impact because
kernel threads including IO workers never return to user space and
therefore never invoke tss_update_io_bitmap().
Cure this by adding the missing cleanups and checks:
1) Prevent io_bitmap_exit() to invoke task_update_io_bitmap() if
the to be cleaned up task is not the current task.
2) Clear TIF_IO_BITMAP in copy_thread() unconditionally. For user
space forks it is set later, when the IO bitmap is inherited in
io_bitmap_share().
For paranoia sake, add a warning into tss_update_io_bitmap() to catch
the case, when that code is invoked with inconsistent state.
Fixes: ea5f1cd7ab ("x86/ioperm: Remove bitmap if all permissions dropped")
Reported-by: syzbot+e2b1803445d236442e54@syzkaller.appspotmail.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/87wmdceom2.ffs@tglx
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 600e960604 ]
Ensure the PIR is read exactly once at the start of handle_pending_pir(),
to guarantee that checking for an outstanding posted interrupt in a given
chuck doesn't reload the chunk from the "real" PIR. Functionally, a reload
is benign, but it would defeat the purpose of pre-loading into a copy.
Fixes: 1b03d82ba1 ("x86/irq: Install posted MSI notification handler")
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20250401163447.846608-2-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 824c6384e8 ]
When suspending, save_processor_state() calls mtrr_save_fixed_ranges()
to save fixed-range MTRRs.
On platforms without fixed-range MTRRs like the ACRN hypervisor which
has removed fixed-range MTRR emulation, accessing these MSRs will
trigger an unchecked MSR access error. Make sure fixed-range MTRRs are
supported before access to prevent such error.
Since mtrr_state.have_fixed is only set when MTRRs are present and
enabled, checking the CPU feature flag in mtrr_save_fixed_ranges() is
unnecessary.
Fixes: 3ebad59056 ("[PATCH] x86: Save and restore the fixed-range MTRRs of the BSP when suspending")
Signed-off-by: Jiaqing Zhao <jiaqing.zhao@linux.intel.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/20250509170633.3411169-2-jiaqing.zhao@linux.intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit ca698ec2f0 ]
Commit:
159039af8c ("x86/insn: x86/insn: Add support for REX2 prefix to the instruction decoder opcode map")
added (!REX2) superscript with a space, but the correct format requires ','
for concatination with other superscript tags.
Add ',' to generate correct insn attribute tables.
I confirmed with following command:
arch/x86/lib/x86-opcode-map.txt | grep e8 | head -n 1
[0xe8] = INAT_MAKE_IMM(INAT_IMM_VWORD32) | INAT_FORCE64 | INAT_NO_REX2,
Fixes: 159039af8c ("x86/insn: x86/insn: Add support for REX2 prefix to the instruction decoder opcode map")
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/174580489027.388420.15539375184727726142.stgit@devnote2
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cc663ba3fe ]
CPUID(0x80000000).EAX returns the max extended CPUID leaf available. On
x86-32 machines without an extended CPUID range, a CPUID(0x80000000)
query will just repeat the output of the last valid standard CPUID leaf
on the CPU; i.e., a garbage values. Current tip:x86/cpu code protects against
this by doing:
eax = cpuid_eax(0x80000000);
c->extended_cpuid_level = eax;
if ((eax & 0xffff0000) == 0x80000000) {
// CPU has an extended CPUID range. Check for 0x80000001
if (eax >= 0x80000001) {
cpuid(0x80000001, ...);
}
}
This is correct so far. Afterwards though, the same possibly broken EAX
value is used to check the availability of other extended CPUID leaves:
if (c->extended_cpuid_level >= 0x80000007)
...
if (c->extended_cpuid_level >= 0x80000008)
...
if (c->extended_cpuid_level >= 0x8000000a)
...
if (c->extended_cpuid_level >= 0x8000001f)
...
which is invalid. Fix this by immediately setting the CPU's max extended
CPUID leaf to zero if CPUID(0x80000000).EAX doesn't indicate a valid
CPUID extended range.
While at it, add a comment, similar to kernel/head_32.S, clarifying the
CPUID(0x80000000) sanity check.
References: 8a50e5135a ("x86-32: Use symbolic constants, safer CPUID when enabling EFER.NX")
Fixes: 3da99c9776 ("x86: make (early)_identify_cpu more the same between 32bit and 64 bit")
Signed-off-by: Ahmed S. Darwish <darwi@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Andrew Cooper <andrew.cooper3@citrix.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: John Ogness <john.ogness@linutronix.de>
Cc: x86-cpuid@lists.linux.dev
Link: https://lore.kernel.org/r/20250506050437.10264-3-darwi@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit b43dc4ab09 ]
After
6f059e634dcd("x86/microcode: Clarify the late load logic"),
if the load is up-to-date, the AMD side returns UCODE_OK which leads to
load_late_locked() returning -EBADFD.
Handle UCODE_OK in the switch case to avoid this error.
[ bp: Massage commit message. ]
Fixes: 6f059e634d ("x86/microcode: Clarify the late load logic")
Signed-off-by: Annie Li <jiayanli@google.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/20250430053424.77438-1-jiayanli@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 2492e5aba2 ]
Unlike L3 and DF counters, UMC counters (PERF_CTRs) set the Overflow bit
(bit 48) and saturate on overflow. A subsequent pmu->read() of the event
reports an incorrect accumulated count as there is no difference between
the previous and the current values of the counter.
To avoid this, inspect the current counter value and proactively reset
the corresponding PERF_CTR register on every pmu->read(). Combined with
the periodic reads initiated by the hrtimer, the counters never get a
chance saturate but the resolution reduces to 47 bits.
Fixes: 25e5684782 ("perf/x86/amd/uncore: Add memory controller support")
Signed-off-by: Sandipan Das <sandipan.das@amd.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Song Liu <song@kernel.org>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/dee9c8af2c6d66814cf4c6224529c144c620cf2c.1744906694.git.sandipan.das@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 1f13c60d84 ]
The following commit, 12 years ago:
7e98b71920 ("x86, idle: Use static_cpu_has() for CLFLUSH workaround, add barriers")
added barriers around the CLFLUSH in mwait_idle_with_hints(), justified with:
... and add memory barriers around it since the documentation is explicit
that CLFLUSH is only ordered with respect to MFENCE.
This also triggered, 11 years ago, the same adjustment in:
f8e617f458 ("sched/idle/x86: Optimize unnecessary mwait_idle() resched IPIs")
during development, although it failed to get the static_cpu_has_bug() treatment.
X86_BUG_CLFLUSH_MONITOR (a.k.a the AAI65 errata) is specific to Intel CPUs,
and the SDM currently states:
Executions of the CLFLUSH instruction are ordered with respect to each
other and with respect to writes, locked read-modify-write instructions,
and fence instructions[1].
With footnote 1 reading:
Earlier versions of this manual specified that executions of the CLFLUSH
instruction were ordered only by the MFENCE instruction. All processors
implementing the CLFLUSH instruction also order it relative to the other
operations enumerated above.
i.e. The SDM was incorrect at the time, and barriers should not have been
inserted. Double checking the original AAI65 errata (not available from
intel.com any more) shows no mention of barriers either.
Note: If this were a general codepath, the MFENCEs would be needed, because
AMD CPUs of the same vintage do sport otherwise-unordered CLFLUSHs.
Remove the unnecessary barriers. Furthermore, use a plain alternative(),
rather than static_cpu_has_bug() and/or no optimisation. The workaround
is a single instruction.
Use an explicit %rax pointer rather than a general memory operand, because
MONITOR takes the pointer implicitly in the same way.
[ mingo: Cleaned up the commit a bit. ]
Fixes: 7e98b71920 ("x86, idle: Use static_cpu_has() for CLFLUSH workaround, add barriers")
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Dave Hansen <dave.hansen@intel.com>
Acked-by: Borislav Petkov (AMD) <bp@alien8.de>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@surriel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/20250402172458.1378112-1-andrew.cooper3@citrix.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit 7170130e4c upstream.
As Bert Karwatzki reported, the following recent commit causes a
performance regression on AMD iGPU and dGPU systems:
7ffb791423 ("x86/kaslr: Reduce KASLR entropy on most x86 systems")
It exposed a bug with nokaslr and zone device interaction.
The root cause of the bug is that, the GPU driver registers a zone
device private memory region. When KASLR is disabled or the above commit
is applied, the direct_map_physmem_end is set to much higher than 10 TiB
typically to the 64TiB address. When zone device private memory is added
to the system via add_pages(), it bumps up the max_pfn to the same
value. This causes dma_addressing_limited() to return true, since the
device cannot address memory all the way up to max_pfn.
This caused a regression for games played on the iGPU, as it resulted in
the DMA32 zone being used for GPU allocations.
Fix this by not bumping up max_pfn on x86 systems, when pgmap is passed
into add_pages(). The presence of pgmap is used to determine if device
private memory is being added via add_pages().
More details:
devm_request_mem_region() and request_free_mem_region() request for
device private memory. iomem_resource is passed as the base resource
with start and end parameters. iomem_resource's end depends on several
factors, including the platform and virtualization. On x86 for example
on bare metal, this value is set to boot_cpu_data.x86_phys_bits.
boot_cpu_data.x86_phys_bits can change depending on support for MKTME.
By default it is set to the same as log2(direct_map_physmem_end) which
is 46 to 52 bits depending on the number of levels in the page table.
The allocation routines used iomem_resource's end and
direct_map_physmem_end to figure out where to allocate the region.
[ arch/powerpc is also impacted by this problem, but this patch does not fix
the issue for PowerPC. ]
Testing:
1. Tested on a virtual machine with test_hmm for zone device inseration
2. A previous version of this patch was tested by Bert, please see:
https://lore.kernel.org/lkml/d87680bab997fdc9fb4e638983132af235d9a03a.camel@web.de/
[ mingo: Clarified the comments and the changelog. ]
Reported-by: Bert Karwatzki <spasswolf@web.de>
Tested-by: Bert Karwatzki <spasswolf@web.de>
Fixes: 7ffb791423 ("x86/kaslr: Reduce KASLR entropy on most x86 systems")
Signed-off-by: Balbir Singh <balbirs@nvidia.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: David Airlie <airlied@gmail.com>
Cc: Simona Vetter <simona@ffwll.ch>
Link: https://lore.kernel.org/r/20250401000752.249348-1-balbirs@nvidia.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[ Upstream commit 99bcd91fab ]
Currently, using PEBS-via-PT with a sample frequency instead of a sample
period, causes a segfault. For example:
BUG: kernel NULL pointer dereference, address: 0000000000000195
<NMI>
? __die_body.cold+0x19/0x27
? page_fault_oops+0xca/0x290
? exc_page_fault+0x7e/0x1b0
? asm_exc_page_fault+0x26/0x30
? intel_pmu_pebs_event_update_no_drain+0x40/0x60
? intel_pmu_pebs_event_update_no_drain+0x32/0x60
intel_pmu_drain_pebs_icl+0x333/0x350
handle_pmi_common+0x272/0x3c0
intel_pmu_handle_irq+0x10a/0x2e0
perf_event_nmi_handler+0x2a/0x50
That happens because intel_pmu_pebs_event_update_no_drain() assumes all the
pebs_enabled bits represent counter indexes, which is not always the case.
In this particular case, bits 60 and 61 are set for PEBS-via-PT purposes.
The behaviour of PEBS-via-PT with sample frequency is questionable because
although a PMI is generated (PEBS_PMI_AFTER_EACH_RECORD), the period is not
adjusted anyway.
Putting that aside, fix intel_pmu_pebs_event_update_no_drain() by passing
the mask of counter bits instead of 'size'. Note, prior to the Fixes
commit, 'size' would be limited to the maximum counter index, so the issue
was not hit.
Fixes: 722e42e45c ("perf/x86: Support counter mask")
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Ian Rogers <irogers@google.com>
Cc: linux-perf-users@vger.kernel.org
Link: https://lore.kernel.org/r/20250508134452.73960-1-adrian.hunter@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit f7387eff4b ]
The GHCB_MSR_VMPL_REQ_LEVEL macro lacked parentheses around the bitmask
expression, causing the shift operation to bind too early. As a result,
when requesting VMPL1 (e.g., GHCB_MSR_VMPL_REQ_LEVEL(1)), incorrect
values such as 0x000000016 were generated instead of the intended
0x100000016 (the requested VMPL level is specified in GHCBData[39:32]).
Fix the precedence issue by grouping the masked value before applying
the shift.
[ bp: Massage commit message. ]
Fixes: 34ff659017 ("x86/sev: Use kernel provided SVSM Calling Areas")
Signed-off-by: Seongman Lee <augustus92@kaist.ac.kr>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://lore.kernel.org/20250511092329.12680-1-cloudlee1719@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 5595c31c37 ]
Calling core::fmt::write() from rust code while FineIBT is enabled
results in a kernel panic:
[ 4614.199779] kernel BUG at arch/x86/kernel/cet.c:132!
[ 4614.205343] Oops: invalid opcode: 0000 [#1] PREEMPT SMP NOPTI
[ 4614.211781] CPU: 2 UID: 0 PID: 6057 Comm: dmabuf_dump Tainted: G U O 6.12.17-android16-0-g6ab38c534a43 #1 9da040f27673ec3945e23b998a0f8bd64c846599
[ 4614.227832] Tainted: [U]=USER, [O]=OOT_MODULE
[ 4614.241247] RIP: 0010:do_kernel_cp_fault+0xea/0xf0
...
[ 4614.398144] RIP: 0010:_RNvXs5_NtNtNtCs3o2tGsuHyou_4core3fmt3num3impyNtB9_7Display3fmt+0x0/0x20
[ 4614.407792] Code: 48 f7 df 48 0f 48 f9 48 89 f2 89 c6 5d e9 18 fd ff ff 0f 1f 84 00 00 00 00 00 f3 0f 1e fa 41 81 ea 14 61 af 2c 74 03 0f 0b 90 <66> 0f 1f 00 55 48 89 e5 48 89 f2 48 8b 3f be 01 00 00 00 5d e9 e7
[ 4614.428775] RSP: 0018:ffffb95acfa4ba68 EFLAGS: 00010246
[ 4614.434609] RAX: 0000000000000000 RBX: 0000000000000010 RCX: 0000000000000000
[ 4614.442587] RDX: 0000000000000007 RSI: ffffb95acfa4ba70 RDI: ffffb95acfa4bc88
[ 4614.450557] RBP: ffffb95acfa4bae0 R08: ffff0a00ffffff05 R09: 0000000000000070
[ 4614.458527] R10: 0000000000000000 R11: ffffffffab67eaf0 R12: ffffb95acfa4bcc8
[ 4614.466493] R13: ffffffffac5d50f0 R14: 0000000000000000 R15: 0000000000000000
[ 4614.474473] ? __cfi__RNvXs5_NtNtNtCs3o2tGsuHyou_4core3fmt3num3impyNtB9_7Display3fmt+0x10/0x10
[ 4614.484118] ? _RNvNtCs3o2tGsuHyou_4core3fmt5write+0x1d2/0x250
This happens because core::fmt::write() calls
core::fmt::rt::Argument::fmt(), which currently has CFI disabled:
library/core/src/fmt/rt.rs:
171 // FIXME: Transmuting formatter in new and indirectly branching to/calling
172 // it here is an explicit CFI violation.
173 #[allow(inline_no_sanitize)]
174 #[no_sanitize(cfi, kcfi)]
175 #[inline]
176 pub(super) unsafe fn fmt(&self, f: &mut Formatter<'_>) -> Result {
This causes a Control Protection exception, because FineIBT has sealed
off the original function's endbr64.
This makes rust currently incompatible with FineIBT. Add a Kconfig
dependency that prevents FineIBT from getting turned on by default
if rust is enabled.
[ Rust 1.88.0 (scheduled for 2025-06-26) should have this fixed [1],
and thus we relaxed the condition with Rust >= 1.88.
When `objtool` lands checking for this with e.g. [2], the plan is
to ideally run that in upstream Rust's CI to prevent regressions
early [3], since we do not control `core`'s source code.
Alice tested the Rust PR backported to an older compiler.
Peter would like that Rust provides a stable `core` which can be
pulled into the kernel: "Relying on that much out of tree code is
'unfortunate'".
- Miguel ]
Signed-off-by: Paweł Anikiel <panikiel@google.com>
Reviewed-by: Alice Ryhl <aliceryhl@google.com>
Acked-by: Peter Zijlstra <peterz@infradead.org>
Link: https://github.com/rust-lang/rust/pull/139632 [1]
Link: https://lore.kernel.org/rust-for-linux/20250410154556.GB9003@noisy.programming.kicks-ass.net/ [2]
Link: https://github.com/rust-lang/rust/pull/139632#issuecomment-2801950873 [3]
Link: https://lore.kernel.org/r/20250410115420.366349-1-panikiel@google.com
Link: https://lore.kernel.org/r/att0-CANiq72kjDM0cKALVy4POEzhfdT4nO7tqz0Pm7xM+3=_0+L1t=A@mail.gmail.com
[ Reduced splat. - Miguel ]
Signed-off-by: Miguel Ojeda <ojeda@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 598bdf4fef ]
Instead of using standard perf_event_attr->freq=0 and ->sample_period
fields, IBS event in 'sample period mode' can also be opened by setting
period value directly in perf_event_attr->config in a MaxCnt bit-field
format.
IBS OP MaxCnt bits are defined as:
(high bits) IbsOpCtl[26:20] = IbsOpMaxCnt[26:20]
(low bits) IbsOpCtl[15:0] = IbsOpMaxCnt[19:4]
Perf event sample period can be derived from MaxCnt bits as:
sample_period = (high bits) | ((low_bits) << 4);
However, current code just masks MaxCnt bits and shifts all of them,
including high bits, which is incorrect. Fix it.
Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lkml.kernel.org/r/20250115054438.1021-4-ravi.bangoria@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 46dcf85566 ]
IBS Op uses two counters: MaxCnt and CurCnt. MaxCnt is programmed with
the desired sample period. IBS hw generates sample when CurCnt reaches
to MaxCnt. The size of these counter used to be 20 bits but later they
were extended to 27 bits. The 7 bit extension is indicated by CPUID
Fn8000_001B_EAX[6 / OpCntExt].
perf_ibs->cnt_mask variable contains bit masks for MaxCnt and CurCnt.
But IBS driver does not set upper 7 bits of CurCnt in cnt_mask even
when OpCntExt CPUID bit is set. Fix this.
IBS driver uses cnt_mask[CurCnt] bits only while disabling an event.
Fortunately, CurCnt bits are not read from MSR while re-enabling the
event, instead MaxCnt is programmed with desired period and CurCnt is
set to 0. Hence, we did not see any issues so far.
Signed-off-by: Ravi Bangoria <ravi.bangoria@amd.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lkml.kernel.org/r/20250115054438.1021-5-ravi.bangoria@amd.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit c20ad96c9a ]
Notably, don't attempt to decode an immediate when MOD == 3.
Additionally have it return the instruction length, such that WARN
like bugs can more reliably skip to the correct instruction.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Sami Tolvanen <samitolvanen@google.com>
Link: https://lore.kernel.org/r/20250207122546.721120726@infradead.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 882b86fd4e ]
Sami reminded me that FineIBT failure does not hook into the regular
CFI failure case, and as such CFI_PERMISSIVE does not work.
Reported-by: Sami Tolvanen <samitolvanen@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Sami Tolvanen <samitolvanen@google.com>
Link: https://lkml.kernel.org/r/20250214092619.GB21726@noisy.programming.kicks-ass.net
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 4087e16b03 ]
percpu_{,try_}cmpxchg{64,128}() macros use CALL instruction inside
asm statement in one of their alternatives. Use ALT_OUTPUT_SP()
macro to add required dependence on %esp register.
ALT_OUTPUT_SP() implements the above dependence by adding
ASM_CALL_CONSTRAINT to its arguments. This constraint should be used
for any inline asm which has a CALL instruction, otherwise the
compiler may schedule the asm before the frame pointer gets set up
by the containing function, causing objtool to print a "call without
frame pointer save/setup" warning.
Signed-off-by: Uros Bizjak <ubizjak@gmail.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20250214150929.5780-1-ubizjak@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7ffb791423 ]
When CONFIG_PCI_P2PDMA=y (which is basically enabled on all
large x86 distros), it maps the PFN's via a ZONE_DEVICE
mapping using devm_memremap_pages(). The mapped virtual
address range corresponds to the pci_resource_start()
of the BAR address and size corresponding to the BAR length.
When KASLR is enabled, the direct map range of the kernel is
reduced to the size of physical memory plus additional padding.
If the BAR address is beyond this limit, PCI peer to peer DMA
mappings fail.
Fix this by not shrinking the size of the direct map when
CONFIG_PCI_P2PDMA=y.
This reduces the total available entropy, but it's better than
the current work around of having to disable KASLR completely.
[ mingo: Clarified the changelog to point out the broad impact ... ]
Signed-off-by: Balbir Singh <balbirs@nvidia.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Acked-by: Bjorn Helgaas <bhelgaas@google.com> # drivers/pci/Kconfig
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Andy Lutomirski <luto@kernel.org>
Link: https://lore.kernel.org/lkml/20250206023201.1481957-1-balbirs@nvidia.com/
Link: https://lore.kernel.org/r/20250206234234.1912585-1-balbirs@nvidia.com
--
arch/x86/mm/kaslr.c | 10 ++++++++--
drivers/pci/Kconfig | 6 ++++++
2 files changed, 14 insertions(+), 2 deletions(-)
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit fe37c699ae ]
Depending on the type of panics, it was found that the
__register_nmi_handler() function can be called in NMI context from
nmi_shootdown_cpus() leading to a lockdep splat:
WARNING: inconsistent lock state
inconsistent {INITIAL USE} -> {IN-NMI} usage.
lock(&nmi_desc[0].lock);
<Interrupt>
lock(&nmi_desc[0].lock);
Call Trace:
_raw_spin_lock_irqsave
__register_nmi_handler
nmi_shootdown_cpus
kdump_nmi_shootdown_cpus
native_machine_crash_shutdown
__crash_kexec
In this particular case, the following panic message was printed before:
Kernel panic - not syncing: Fatal hardware error!
This message seemed to be given out from __ghes_panic() running in
NMI context.
The __register_nmi_handler() function which takes the nmi_desc lock
with irq disabled shouldn't be called from NMI context as this can
lead to deadlock.
The nmi_shootdown_cpus() function can only be invoked once. After the
first invocation, all other CPUs should be stuck in the newly added
crash_nmi_callback() and cannot respond to a second NMI.
Fix it by adding a new emergency NMI handler to the nmi_desc
structure and provide a new set_emergency_nmi_handler() helper to set
crash_nmi_callback() in any context. The new emergency handler will
preempt other handlers in the linked list. That will eliminate the need
to take any lock and serve the panic in NMI use case.
Signed-off-by: Waiman Long <longman@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Rik van Riel <riel@surriel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/20250206191844.131700-1-longman@redhat.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e451630226 ]
Problem: Currently when running the "make isoimage" command there is an
error related to wrong parameters passed to the cp command:
"cp: missing destination file operand after 'arch/x86/boot/isoimage/'"
This is caused because FDINITRDS is an empty array.
Solution: Check if FDINITRDS is empty before executing the "cp" command,
similar to how it is done in the case of hdimage.
Signed-off-by: Nir Lichtman <nir@lichtman.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Michal Marek <michal.lkml@markovi.net>
Link: https://lore.kernel.org/r/20250110120500.GA923218@lichtman.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 98fdaeb296 ]
Change the default value of spectre v2 in user mode to respect the
CONFIG_MITIGATION_SPECTRE_V2 config option.
Currently, user mode spectre v2 is set to auto
(SPECTRE_V2_USER_CMD_AUTO) by default, even if
CONFIG_MITIGATION_SPECTRE_V2 is disabled.
Set the spectre_v2 value to auto (SPECTRE_V2_USER_CMD_AUTO) if the
Spectre v2 config (CONFIG_MITIGATION_SPECTRE_V2) is enabled, otherwise
set the value to none (SPECTRE_V2_USER_CMD_NONE).
Important to say the command line argument "spectre_v2_user" overwrites
the default value in both cases.
When CONFIG_MITIGATION_SPECTRE_V2 is not set, users have the flexibility
to opt-in for specific mitigations independently. In this scenario,
setting spectre_v2= will not enable spectre_v2_user=, and command line
options spectre_v2_user and spectre_v2 are independent when
CONFIG_MITIGATION_SPECTRE_V2=n.
Signed-off-by: Breno Leitao <leitao@debian.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: David Kaplan <David.Kaplan@amd.com>
Link: https://lore.kernel.org/r/20241031-x86_bugs_last_v2-v2-2-b7ff1dab840e@debian.org
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit cef721e0d5 ]
Doing this allows using registers as retrieved from an mcontext to be
pushed to a process using PTRACE_SETREGS.
It is not entirely clear to me why CSGSFS was masked. Doing so creates
issues when using the mcontext as process state in seccomp and simply
copying the register appears to work perfectly fine for ptrace.
Signed-off-by: Benjamin Berg <benjamin@sipsolutions.net>
Link: https://patch.msgid.link/20250224181827.647129-2-benjamin@sipsolutions.net
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 631ca8909f ]
At least with CONFIG_PHYSICAL_START=0x100000, if there is < 4 MiB of
contiguous free memory available at this point, the kernel will crash
and burn because memblock_phys_alloc_range() returns 0 on failure,
which leads memblock_phys_free() to throw the first 4 MiB of physical
memory to the wolves.
At a minimum it should fail gracefully with a meaningful diagnostic,
but in fact everything seems to work fine without the weird reserve
allocation.
Signed-off-by: Philip Redkin <me@rarity.fan>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Link: https://lore.kernel.org/r/94b3e98f-96a7-3560-1f76-349eb95ccf7f@rarity.fan
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7e6b0a2e41 ]
The Family model check to read the processor flag MSR is misleading and
potentially incorrect. It doesn't consider Family while comparing the
model number. The original check did have a Family number but it got
lost/moved during refactoring.
intel_collect_cpu_info() is called through multiple paths such as early
initialization, CPU hotplug as well as IFS image load. Some of these
flows would be error prone due to the ambiguous check.
Correct the processor flag scan check to use a Family number and update
it to a VFM based one to make it more readable.
Signed-off-by: Sohil Mehta <sohil.mehta@intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Link: https://lore.kernel.org/r/20250219184133.816753-4-sohil.mehta@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 7a2ad75274 ]
Some old crusty CPUs need an extra delay that slows down booting. See
the comment above 'init_udelay' for details. Newer CPUs don't need the
delay.
Right now, for Intel, Family 6 and only Family 6 skips the delay. That
leaves out both the Family 15 (Pentium 4s) and brand new Family 18/19
models.
The omission of Family 15 (Pentium 4s) seems like an oversight and 18/19
do not need the delay.
Skip the delay on all Intel processors Family 6 and beyond.
Signed-off-by: Sohil Mehta <sohil.mehta@intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: https://lore.kernel.org/r/20250219184133.816753-11-sohil.mehta@intel.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 91d5451d97 ]
The __ref_stack_chk_guard symbol doesn't exist on UP:
<stdin>:4:15: error: ‘__ref_stack_chk_guard’ undeclared here (not in a function)
Fix the #ifdef around the entry.S export.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Ard Biesheuvel <ardb@kernel.org>
Cc: Uros Bizjak <ubizjak@gmail.com>
Link: https://lore.kernel.org/r/20250123190747.745588-8-brgerst@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 8a141be323 ]
__ASSEMBLY__ is only defined by the Makefile of the kernel, so
this is not really useful for UAPI headers (unless the userspace
Makefile defines it, too). Let's switch to __ASSEMBLER__ which
gets set automatically by the compiler when compiling assembly
code.
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Brian Gerst <brgerst@gmail.com>
Link: https://lore.kernel.org/r/20250310104256.123527-1-thuth@redhat.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit e5f1e8af9c ]
Upon a wakeup from S4, the restore kernel starts and initializes the
FRED MSRs as needed from its perspective. It then loads a hibernation
image, including the image kernel, and attempts to load image pages
directly into their original page frames used before hibernation unless
those frames are currently in use. Once all pages are moved to their
original locations, it jumps to a "trampoline" page in the image kernel.
At this point, the image kernel takes control, but the FRED MSRs still
contain values set by the restore kernel, which may differ from those
set by the image kernel before hibernation. Therefore, the image kernel
must ensure the FRED MSRs have the same values as before hibernation.
Since these values depend only on the location of the kernel text and
data, they can be recomputed from scratch.
Reported-by: Xi Pardee <xi.pardee@intel.com>
Reported-by: Todd Brandt <todd.e.brandt@intel.com>
Tested-by: Todd Brandt <todd.e.brandt@intel.com>
Suggested-by: H. Peter Anvin (Intel) <hpa@zytor.com>
Signed-off-by: Xin Li (Intel) <xin@zytor.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Reviewed-by: H. Peter Anvin (Intel) <hpa@zytor.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Juergen Gross <jgross@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Link: https://lore.kernel.org/r/20250401075728.3626147-1-xin@zytor.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit 9129633d56 ]
When changing memory attributes on a subset of a potential hugepage, add
the hugepage to the invalidation range tracking to prevent installing a
hugepage until the attributes are fully updated. Like the actual hugepage
tracking updates in kvm_arch_post_set_memory_attributes(), process only
the head and tail pages, as any potential hugepages that are entirely
covered by the range will already be tracked.
Note, only hugepage chunks whose current attributes are NOT mixed need to
be added to the invalidation set, as mixed attributes already prevent
installing a hugepage, and it's perfectly safe to install a smaller
mapping for a gfn whose attributes aren't changing.
Fixes: 8dd2eee9d5 ("KVM: x86/mmu: Handle page fault for private memory")
Cc: stable@vger.kernel.org
Reported-by: Michael Roth <michael.roth@amd.com>
Tested-by: Michael Roth <michael.roth@amd.com>
Link: https://lore.kernel.org/r/20250430220954.522672-1-seanjc@google.com
Signed-off-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
[ Upstream commit dca6c88532 ]
Add new members to strut kvm_gfn_range to indicate which mapping
(private-vs-shared) to operate on: enum kvm_gfn_range_filter
attr_filter. Update the core zapping operations to set them appropriately.
TDX utilizes two GPA aliases for the same memslots, one for memory that is
for private memory and one that is for shared. For private memory, KVM
cannot always perform the same operations it does on memory for default
VMs, such as zapping pages and having them be faulted back in, as this
requires guest coordination. However, some operations such as guest driven
conversion of memory between private and shared should zap private memory.
Internally to the MMU, private and shared mappings are tracked on separate
roots. Mapping and zapping operations will operate on the respective GFN
alias for each root (private or shared). So zapping operations will by
default zap both aliases. Add fields in struct kvm_gfn_range to allow
callers to specify which aliases so they can only target the aliases
appropriate for their specific operation.
There was feedback that target aliases should be specified such that the
default value (0) is to operate on both aliases. Several options were
considered. Several variations of having separate bools defined such
that the default behavior was to process both aliases. They either allowed
nonsensical configurations, or were confusing for the caller. A simple
enum was also explored and was close, but was hard to process in the
caller. Instead, use an enum with the default value (0) reserved as a
disallowed value. Catch ranges that didn't have the target aliases
specified by looking for that specific value.
Set target alias with enum appropriately for these MMU operations:
- For KVM's mmu notifier callbacks, zap shared pages only because private
pages won't have a userspace mapping
- For setting memory attributes, kvm_arch_pre_set_memory_attributes()
chooses the aliases based on the attribute.
- For guest_memfd invalidations, zap private only.
Link: https://lore.kernel.org/kvm/ZivIF9vjKcuGie3s@google.com/
Signed-off-by: Isaku Yamahata <isaku.yamahata@intel.com>
Co-developed-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Signed-off-by: Rick Edgecombe <rick.p.edgecombe@intel.com>
Message-ID: <20240718211230.1492011-3-rick.p.edgecombe@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Stable-dep-of: 9129633d56 ("KVM: x86/mmu: Prevent installing hugepages when mem attributes are changing")
Signed-off-by: Sasha Levin <sashal@kernel.org>
commit e52c1dc745 upstream.
FineIBT-paranoid was using the retpoline bytes for the paranoid check,
disabling retpolines, because all parts that have IBT also have eIBRS
and thus don't need no stinking retpolines.
Except... ITS needs the retpolines for indirect calls must not be in
the first half of a cacheline :-/
So what was the paranoid call sequence:
<fineibt_paranoid_start>:
0: 41 ba 78 56 34 12 mov $0x12345678, %r10d
6: 45 3b 53 f7 cmp -0x9(%r11), %r10d
a: 4d 8d 5b <f0> lea -0x10(%r11), %r11
e: 75 fd jne d <fineibt_paranoid_start+0xd>
10: 41 ff d3 call *%r11
13: 90 nop
Now becomes:
<fineibt_paranoid_start>:
0: 41 ba 78 56 34 12 mov $0x12345678, %r10d
6: 45 3b 53 f7 cmp -0x9(%r11), %r10d
a: 4d 8d 5b f0 lea -0x10(%r11), %r11
e: 2e e8 XX XX XX XX cs call __x86_indirect_paranoid_thunk_r11
Where the paranoid_thunk looks like:
1d: <ea> (bad)
__x86_indirect_paranoid_thunk_r11:
1e: 75 fd jne 1d
__x86_indirect_its_thunk_r11:
20: 41 ff eb jmp *%r11
23: cc int3
[ dhansen: remove initialization to false ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
[ Just a portion of the original commit, in order to fix a build issue
in stable kernels due to backports ]
Tested-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Link: https://lore.kernel.org/r/20250514113952.GB16434@noisy.programming.kicks-ass.net
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 9f35e33144 upstream.
Fix several build errors when CONFIG_MODULES=n, including the following:
../arch/x86/kernel/alternative.c:195:25: error: incomplete definition of type 'struct module'
195 | for (int i = 0; i < mod->its_num_pages; i++) {
Fixes: 872df34d7c ("x86/its: Use dynamic thunks for indirect branches")
Cc: stable@vger.kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
Acked-by: Dave Hansen <dave.hansen@intel.com>
Tested-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 872df34d7c upstream.
ITS mitigation moves the unsafe indirect branches to a safe thunk. This
could degrade the prediction accuracy as the source address of indirect
branches becomes same for different execution paths.
To improve the predictions, and hence the performance, assign a separate
thunk for each indirect callsite. This is also a defense-in-depth measure
to avoid indirect branches aliasing with each other.
As an example, 5000 dynamic thunks would utilize around 16 bits of the
address space, thereby gaining entropy. For a BTB that uses
32 bits for indexing, dynamic thunks could provide better prediction
accuracy over fixed thunks.
Have ITS thunks be variable sized and use EXECMEM_MODULE_TEXT such that
they are both more flexible (got to extend them later) and live in 2M TLBs,
just like kernel code, avoiding undue TLB pressure.
[ pawan: CONFIG_EXECMEM_ROX is not supported on backport kernel, made
adjustments to set memory to RW and ROX ]
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit ebebe30794 upstream.
cfi_rewrite_callers() updates the fineIBT hash matching at the caller side,
but except for paranoid-mode it relies on apply_retpoline() and friends for
any ENDBR relocation. This could temporarily cause an indirect branch to
land on a poisoned ENDBR.
For instance, with para-virtualization enabled, a simple wrmsrl() could
have an indirect branch pointing to native_write_msr() who's ENDBR has been
relocated due to fineIBT:
<wrmsrl>:
push %rbp
mov %rsp,%rbp
mov %esi,%eax
mov %rsi,%rdx
shr $0x20,%rdx
mov %edi,%edi
mov %rax,%rsi
call *0x21e65d0(%rip) # <pv_ops+0xb8>
^^^^^^^^^^^^^^^^^^^^^^^
Such an indirect call during the alternative patching could #CP if the
caller is not *yet* adjusted for the new target ENDBR. To prevent a false
#CP, keep CET-IBT disabled until all callers are patched.
Patching during the module load does not need to be guarded by IBT-disable
because the module code is not executed until the patching is complete.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f0cd7091cc upstream.
The software mitigation for BHI is to execute BHB clear sequence at syscall
entry, and possibly after a cBPF program. ITS mitigation thunks RETs in the
lower half of the cacheline. This causes the RETs in the BHB clear sequence
to be thunked as well, adding unnecessary branches to the BHB clear
sequence.
Since the sequence is in hot path, align the RET instructions in the
sequence to avoid thunking.
This is how disassembly clear_bhb_loop() looks like after this change:
0x44 <+4>: mov $0x5,%ecx
0x49 <+9>: call 0xffffffff81001d9b <clear_bhb_loop+91>
0x4e <+14>: jmp 0xffffffff81001de5 <clear_bhb_loop+165>
0x53 <+19>: int3
...
0x9b <+91>: call 0xffffffff81001dce <clear_bhb_loop+142>
0xa0 <+96>: ret
0xa1 <+97>: int3
...
0xce <+142>: mov $0x5,%eax
0xd3 <+147>: jmp 0xffffffff81001dd6 <clear_bhb_loop+150>
0xd5 <+149>: nop
0xd6 <+150>: sub $0x1,%eax
0xd9 <+153>: jne 0xffffffff81001dd3 <clear_bhb_loop+147>
0xdb <+155>: sub $0x1,%ecx
0xde <+158>: jne 0xffffffff81001d9b <clear_bhb_loop+91>
0xe0 <+160>: ret
0xe1 <+161>: int3
0xe2 <+162>: int3
0xe3 <+163>: int3
0xe4 <+164>: int3
0xe5 <+165>: lfence
0xe8 <+168>: pop %rbp
0xe9 <+169>: ret
Suggested-by: Andrew Cooper <andrew.cooper3@citrix.com>
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit facd226f7e upstream.
When retpoline mitigation is enabled for spectre-v2, enabling
call-depth-tracking and RSB stuffing also mitigates ITS. Add cmdline option
indirect_target_selection=stuff to allow enabling RSB stuffing mitigation.
When retpoline mitigation is not enabled, =stuff option is ignored, and
default mitigation for ITS is deployed.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 2665281a07 upstream.
Ice Lake generation CPUs are not affected by guest/host isolation part of
ITS. If a user is only concerned about KVM guests, they can now choose a
new cmdline option "vmexit" that will not deploy the ITS mitigation when
CPU is not affected by guest/host isolation. This saves the performance
overhead of ITS mitigation on Ice Lake gen CPUs.
When "vmexit" option selected, if the CPU is affected by ITS guest/host
isolation, the default ITS mitigation is deployed.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit f4818881c4 upstream.
Indirect Target Selection (ITS) is a bug in some pre-ADL Intel CPUs with
eIBRS. It affects prediction of indirect branch and RETs in the
lower half of cacheline. Due to ITS such branches may get wrongly predicted
to a target of (direct or indirect) branch that is located in the upper
half of the cacheline.
Scope of impact
===============
Guest/host isolation
--------------------
When eIBRS is used for guest/host isolation, the indirect branches in the
VMM may still be predicted with targets corresponding to branches in the
guest.
Intra-mode
----------
cBPF or other native gadgets can be used for intra-mode training and
disclosure using ITS.
User/kernel isolation
---------------------
When eIBRS is enabled user/kernel isolation is not impacted.
Indirect Branch Prediction Barrier (IBPB)
-----------------------------------------
After an IBPB, indirect branches may be predicted with targets
corresponding to direct branches which were executed prior to IBPB. This is
mitigated by a microcode update.
Add cmdline parameter indirect_target_selection=off|on|force to control the
mitigation to relocate the affected branches to an ITS-safe thunk i.e.
located in the upper half of cacheline. Also add the sysfs reporting.
When retpoline mitigation is deployed, ITS safe-thunks are not needed,
because retpoline sequence is already ITS-safe. Similarly, when call depth
tracking (CDT) mitigation is deployed (retbleed=stuff), ITS safe return
thunk is not used, as CDT prevents RSB-underflow.
To not overcomplicate things, ITS mitigation is not supported with
spectre-v2 lfence;jmp mitigation. Moreover, it is less practical to deploy
lfence;jmp mitigation on ITS affected parts anyways.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit a75bf27fe4 upstream.
RETs in the lower half of cacheline may be affected by ITS bug,
specifically when the RSB-underflows. Use ITS-safe return thunk for such
RETs.
RETs that are not patched:
- RET in retpoline sequence does not need to be patched, because the
sequence itself fills an RSB before RET.
- RET in Call Depth Tracking (CDT) thunks __x86_indirect_{call|jump}_thunk
and call_depth_return_thunk are not patched because CDT by design
prevents RSB-underflow.
- RETs in .init section are not reachable after init.
- RETs that are explicitly marked safe with ANNOTATE_UNRET_SAFE.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 8754e67ad4 upstream.
Due to ITS, indirect branches in the lower half of a cacheline may be
vulnerable to branch target injection attack.
Introduce ITS-safe thunks to patch indirect branches in the lower half of
cacheline with the thunk. Also thunk any eBPF generated indirect branches
in emit_indirect_jump().
Below category of indirect branches are not mitigated:
- Indirect branches in the .init section are not mitigated because they are
discarded after boot.
- Indirect branches that are explicitly marked retpoline-safe.
Note that retpoline also mitigates the indirect branches against ITS. This
is because the retpoline sequence fills an RSB entry before RET, and it
does not suffer from RSB-underflow part of the ITS.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit 159013a7ca upstream.
ITS bug in some pre-Alderlake Intel CPUs may allow indirect branches in the
first half of a cache line get predicted to a target of a branch located in
the second half of the cache line.
Set X86_BUG_ITS on affected CPUs. Mitigation to follow in later commits.
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Reviewed-by: Alexandre Chartre <alexandre.chartre@oracle.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
commit c8c8145886 upstream.
Commit:
010c4a461c ("x86/speculation: Simplify and make CALL_NOSPEC consistent")
added an #ifdef CONFIG_MITIGATION_RETPOLINE around the CALL_NOSPEC definition.
This is not required as this code is already under a larger #ifdef.
Remove the extra #ifdef, no functional change.
vmlinux size remains same before and after this change:
CONFIG_MITIGATION_RETPOLINE=y:
text data bss dec hex filename
25434752 7342290 2301212 35078254 217406e vmlinux.before
25434752 7342290 2301212 35078254 217406e vmlinux.after
# CONFIG_MITIGATION_RETPOLINE is not set:
text data bss dec hex filename
22943094 6214994 1550152 30708240 1d49210 vmlinux.before
22943094 6214994 1550152 30708240 1d49210 vmlinux.after
Signed-off-by: Pawan Gupta <pawan.kumar.gupta@linux.intel.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Josh Poimboeuf <jpoimboe@kernel.org>
Link: https://lore.kernel.org/r/20250320-call-nospec-extra-ifdef-v1-1-d9b084d24820@linux.intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>