The make_xfence_request() function uses a shift operation to check if a
vCPU is in the hart mask:
if (!(hmask & (1UL << (vcpu->vcpu_id - hbase))))
However, when the difference between vcpu_id and hbase
is >= BITS_PER_LONG, the shift operation causes undefined behavior.
This was detected by UBSAN:
UBSAN: shift-out-of-bounds in arch/riscv/kvm/tlb.c:343:23
shift exponent 256 is too large for 64-bit type 'long unsigned int'
Fix this by adding a bounds check before the shift operation.
This bug was found by fuzzing the KVM RISC-V interface.
Fixes: 13acfec2db ("RISC-V: KVM: Add remote HFENCE functions based on VCPU requests")
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260403232011.2394966-1-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
Reuse KVM_CAP_VM_GPA_BITS to advertise and select the effective
G-stage GPA width for a VM.
KVM_CHECK_EXTENSION(KVM_CAP_VM_GPA_BITS) returns the effective GPA
bits for a VM, KVM_ENABLE_CAP(KVM_CAP_VM_GPA_BITS) allows userspace
to downsize the effective GPA width by selecting a smaller G-stage
page table format:
- gpa_bits <= 41 selects Sv39x4 (pgd_levels=3)
- gpa_bits <= 50 selects Sv48x4 (pgd_levels=4)
- gpa_bits <= 59 selects Sv57x4 (pgd_levels=5)
Reject the request with -EINVAL for unsupported values and with -EBUSY
if vCPUs have been created or any memslot is populated.
Signed-off-by: Fangyu Yu <fangyu.yu@linux.alibaba.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Reviewed-by: Guo Ren <guoren@kernel.org>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/20260403153019.9916-4-fangyu.yu@linux.alibaba.com
Signed-off-by: Anup Patel <anup@brainfault.org>
Gstage page-table helpers frequently chase gstage->kvm->arch to
fetch pgd_levels. This adds noise and repeats the same dereference
chain in hot paths.
Add pgd_levels to struct kvm_gstage and initialize it from kvm->arch
when setting up a gstage instance. Introduce kvm_riscv_gstage_init()
to centralize initialization and switch gstage code to use
gstage->pgd_levels.
Suggested-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Fangyu Yu <fangyu.yu@linux.alibaba.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Link: https://lore.kernel.org/r/20260403153019.9916-3-fangyu.yu@linux.alibaba.com
Signed-off-by: Anup Patel <anup@brainfault.org>
Introduces one per-VM architecture-specific fields to support runtime
configuration of the G-stage page table format:
- kvm->arch.pgd_levels: the corresponding number of page table levels
for the selected mode.
These fields replace the previous global variables
kvm_riscv_gstage_mode and kvm_riscv_gstage_pgd_levels, enabling different
virtual machines to independently select their G-stage page table format
instead of being forced to share the maximum mode detected by the kernel
at boot time.
Signed-off-by: Fangyu Yu <fangyu.yu@linux.alibaba.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Reviewed-by: Guo Ren <guoren@kernel.org>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Link: https://lore.kernel.org/r/20260403153019.9916-2-fangyu.yu@linux.alibaba.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The hstateen0 will be programmed differently for guest HS-mode
and guest VS/VU-mode so don't check hstateen0.SSTATEEN0 bit when
updating sstateen0 CSR in kvm_riscv_vcpu_swap_in_guest_state()
and kvm_riscv_vcpu_swap_in_host_state().
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-10-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The VCPU config deals with hideleg, hedeleg, henvcfg, and hstateenX
CSR configuration for each VCPU. Factor-out VCPU config into separate
sources so that VCPU config can do things differently for guest HS-mode
and guest VS/VU-mode.
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-9-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The hideleg CSR state when VCPU is running in guest VS/VU-mode will
be different from when it is running in guest HS-mode. To achieve
this, add hideleg to struct kvm_vcpu_config and re-program hideleg
CSR upon every kvm_arch_vcpu_load().
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-8-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The KVM_RISCV_TIMER_STATE_xyz defines specify possible values of the
"state" member in struct kvm_riscv_timer so move these defines closer
to struct kvm_riscv_timer in uapi/asm/kvm.h.
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-7-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The KVM ISA extension related checks are not VCPU specific and
should be factored out of vcpu_onereg.c into separate sources.
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-6-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
Rename kvm_riscv_vcpu_isa_check_host() to kvm_riscv_isa_check_host()
and use it as common function with KVM RISC-V to check isa extensions
supported by host.
Signed-off-by: Anup Patel <anup.patel@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260120080013.2153519-5-anup.patel@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
kvm_arch_has_default_irqchip is required for irqfd_test and returns
true if an in-kernel interrupt controller is supported.
Fixes: a133052666 ("KVM: selftests: Fix irqfd_test for non-x86 architectures")
Signed-off-by: Mayuresh Chitale <mayuresh.chitale@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260402101818.2982071-1-mayuresh.chitale@oss.qualcomm.com
Signed-off-by: Anup Patel <anup@brainfault.org>
Currently, kvm_arch_vcpu_load() unconditionally restores guest CSRs,
HGATP, and AIA state. However, when a VCPU is loaded back on the same
physical CPU, and no other KVM VCPU has run on this CPU since it was
last put, the hardware CSRs and AIA registers are still valid.
This patch optimizes the vcpu_load path by skipping the expensive CSR
and AIA writes if all the following conditions are met:
1. It is being reloaded on the same CPU (vcpu->arch.last_exit_cpu == cpu).
2. The CSRs are not dirty (!vcpu->arch.csr_dirty).
3. No other VCPU used this CPU (vcpu == __this_cpu_read(kvm_former_vcpu)).
To ensure this fast-path doesn't break corner cases:
- Live migration and VCPU reset are naturally safe. KVM initializes
last_exit_cpu to -1, which guarantees the fast-path won't trigger.
- The 'csr_dirty' flag tracks runtime userspace interventions. If
userspace modifies guest configurations (e.g., hedeleg via
KVM_SET_GUEST_DEBUG, or CSRs including AIA via KVM_SET_ONE_REG),
the flag is set to skip the fast path.
With the 'csr_dirty' safeguard proven effective, it is safe to
include kvm_riscv_vcpu_aia_load() inside the skip logic now.
Signed-off-by: Jinyu Tang <tjytimi@163.com>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Reviewed-by: Radim Krčmář <radim.krcmar@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260227121008.442241-1-tjytimi@163.com
Signed-off-by: Anup Patel <anup@brainfault.org>
During dirty logging, all huge pages are write-protected. When the guest
writes to a write-protected huge page, a page fault is triggered. Before
recovering the write permission, the huge page must be split into smaller
pages (e.g., 4K). After splitting, the normal mapping process proceeds,
allowing write permission to be restored at the smaller page granularity.
If dirty logging is disabled because migration failed or was cancelled,
only recover the write permission at the 4K level, and skip recovering the
huge page mapping at this time to avoid the overhead of freeing page tables.
The huge page mapping can be recovered in the ioctl context, similar to x86,
in a later patch.
Signed-off-by: Wang Yechao <wang.yechao255@zte.com.cn>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/202603301612587174XZ6QMCrymBqv30S6BN50@zte.com.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
When enabling dirty log in small chunks (e.g., QEMU default chunk
size of 256K), the chunk size is always smaller than the page size
of huge pages (1G or 2M) used in the gstage page tables. This caused
the write protection to be incorrectly skipped for huge PTEs because
the condition `(end - addr) >= page_size` was not satisfied.
Remove the size check in `kvm_riscv_gstage_wp_range()` to ensure huge
PTEs are always write-protected regardless of the chunk size. Additionally,
explicitly align the address down to the page size before invoking
`kvm_riscv_gstage_op_pte()` to guarantee that the address passed to the
operation function is page-aligned.
This fixes the issue where dirty pages might not be tracked correctly
when using huge pages.
Fixes: 9d05c1fee8 ("RISC-V: KVM: Implement stage2 page table programming")
Signed-off-by: Wang Yechao <wang.yechao255@zte.com.cn>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/202603301610527120YZ-pAJY6x9SBpSRo1Wg4@zte.com.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
When a guest initiates an SBI_EXT_PMU_COUNTER_CFG_MATCH call with
ctr_base=0xfffffffffffffffe, ctr_mask=0xeb5f and flags=0x1
(SBI_PMU_CFG_FLAG_SKIP_MATCH), kvm_riscv_vcpu_pmu_ctr_cfg_match()
first invokes kvm_pmu_validate_counter_mask() to verify whether
ctr_base and ctr_mask are valid, by evaluating:
!ctr_mask || (ctr_base + __fls(ctr_mask) >= kvm_pmu_num_counters(kvpmu))
With the above inputs, __fls(0xeb5f) equals 15, and adding 15 to
0xfffffffffffffffe causes an integer overflow, wrapping around to 13.
Since 13 is less than kvm_pmu_num_counters(), the validation wrongly
succeeds.
Thereafter, since flags & SBI_PMU_CFG_FLAG_SKIP_MATCH is satisfied,
the code evaluates:
!test_bit(ctr_base + __ffs(ctr_mask), kvpmu->pmc_in_use)
Here __ffs(0xeb5f) equals 0, so test_bit() receives 0xfffffffffffffffe
as the bit index and attempts to access the corresponding element of
the kvpmu->pmc_in_use, which results in an invalid memory access. This
triggers the following Oops:
Unable to handle kernel paging request at virtual address e3ebffff12abba89
generic_test_bit include/asm-generic/bitops/generic-non-atomic.h:128
kvm_riscv_vcpu_pmu_ctr_cfg_match arch/riscv/kvm/vcpu_pmu.c:758
kvm_sbi_ext_pmu_handler arch/riscv/kvm/vcpu_sbi_pmu.c:49
kvm_riscv_vcpu_sbi_ecall arch/riscv/kvm/vcpu_sbi.c:608
kvm_riscv_vcpu_exit arch/riscv/kvm/vcpu_exit.c:240
The root cause is that kvm_pmu_validate_counter_mask() does not account
for the case where ctr_base itself is out of range, allowing the
subsequent addition to silently overflow and bypass the check.
Fix this by explicitly validating ctr_base against kvm_pmu_num_counters()
before performing the addition.
This bug was found by fuzzing the KVM RISC-V PMU interface.
Fixes: 0cb74b65d2 ("RISC-V: KVM: Implement perf support without sampling")
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Reviewed-by: Atish Patra <atish.patra@linux.dev>
Link: https://lore.kernel.org/r/20260319035902.924661-1-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
In kvm_riscv_vcpu_pmu_snapshot_set_shmem(), when kvm_vcpu_write_guest()
fails, kvpmu->sdata is freed but not set to NULL. This leaves a dangling
pointer that will be freed again when kvm_pmu_clear_snapshot_area() is
called during vcpu teardown, triggering a KASAN double-free report.
First free occurs in kvm_riscv_vcpu_pmu_snapshot_set_shmem():
kvm_riscv_vcpu_pmu_snapshot_set_shmem arch/riscv/kvm/vcpu_pmu.c:443
kvm_sbi_ext_pmu_handler arch/riscv/kvm/vcpu_sbi_pmu.c:74
kvm_riscv_vcpu_sbi_ecall arch/riscv/kvm/vcpu_sbi.c:608
kvm_riscv_vcpu_exit arch/riscv/kvm/vcpu_exit.c:240
kvm_arch_vcpu_ioctl_run arch/riscv/kvm/vcpu.c:1008
kvm_vcpu_ioctl virt/kvm/kvm_main.c:4476
Second free (double-free) occurs in kvm_pmu_clear_snapshot_area():
kvm_pmu_clear_snapshot_area arch/riscv/kvm/vcpu_pmu.c:403 [inline]
kvm_riscv_vcpu_pmu_deinit.part arch/riscv/kvm/vcpu_pmu.c:905
kvm_riscv_vcpu_pmu_deinit arch/riscv/kvm/vcpu_pmu.c:893
kvm_arch_vcpu_destroy arch/riscv/kvm/vcpu.c:199
kvm_vcpu_destroy virt/kvm/kvm_main.c:469 [inline]
kvm_destroy_vcpus virt/kvm/kvm_main.c:489
kvm_arch_destroy_vm arch/riscv/kvm/vm.c:54
kvm_destroy_vm virt/kvm/kvm_main.c:1301 [inline]
kvm_put_kvm virt/kvm/kvm_main.c:1338
kvm_vm_release virt/kvm/kvm_main.c:1361
Fix it by setting kvpmu->sdata to NULL after kfree() in
kvm_riscv_vcpu_pmu_snapshot_set_shmem(), so that the subsequent
kfree(NULL) in kvm_pmu_clear_snapshot_area() becomes a safe no-op.
This bug was found by fuzzing the KVM RISC-V PMU interface.
Fixes: c2f41ddbcd ("RISC-V: KVM: Implement SBI PMU Snapshot feature")
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260318092956.708246-1-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
Add WARN_ON check before accessing cntx->vector.datap in
kvm_riscv_vcpu_vreg_addr() to detect potential null pointer
dereferences early, consistent with the pattern used in
kvm_riscv_vcpu_vector_reset().
This helps catch initialization issues where vector context
allocation may have failed.
Signed-off-by: Yufeng Wang <wangyufeng@kylinos.cn>
Reviewed-by: Anup Patel <anup@brainfault.org>
Link: https://lore.kernel.org/r/20260317114759.53165-1-r4o5m6e8o@163.com
Signed-off-by: Anup Patel <anup@brainfault.org>
The current sbi_pmu_test attempts to read firmware counters without
configuring them first with SBI_EXT_PMU_COUNTER_CFG_MATCH.
Previously this did not fail because KVM incorrectly allowed the read
and accessed fw_event[] with an out-of-bounds index when the counter
was unconfigured. After fixing that bug, the read now correctly returns
SBI_ERR_INVALID_PARAM, causing the selftest to fail.
Update the test to configure a firmware event before reading the
counter. Also add a negative test to ensure that attempting to read an
unconfigured firmware counter fails gracefully.
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Reviewed-by: Nutty Liu <nutty.liu@hotmail.com>
Link: https://lore.kernel.org/r/20260316014533.2312254-3-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
When a guest invokes SBI_EXT_PMU_COUNTER_FW_READ or
SBI_EXT_PMU_COUNTER_FW_READ_HI on a firmware counter that has not been
configured via SBI_EXT_PMU_COUNTER_CFG_MATCH, the pmc->event_idx remains
SBI_PMU_EVENT_IDX_INVALID (0xFFFFFFFF). get_event_code() extracts the
lower 16 bits, yielding 0xFFFF (65535), which is then used to index into
kvpmu->fw_event[]. Since fw_event is only RISCV_KVM_MAX_FW_CTRS (32)
entries, this triggers an array-index-out-of-bounds:
UBSAN: array-index-out-of-bounds in arch/riscv/kvm/vcpu_pmu.c:255:37
index 65535 is out of range for type 'kvm_fw_event [32]'
Add a check for the known unconfigured case (SBI_PMU_EVENT_IDX_INVALID)
and a WARN_ONCE guard for any unexpected out-of-bounds event codes,
returning -EINVAL in both cases.
Fixes: badc386869 ("RISC-V: KVM: Support firmware events")
Fixes: 08fb07d6dc ("RISC-V: KVM: Support 64 bit firmware counters on RV32")
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260316014533.2312254-2-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
When the second kzalloc (host_context.vector.datap) fails in
kvm_riscv_vcpu_alloc_vector_context, the first allocation
(guest_context.vector.datap) is leaked. Free it before returning.
Fixes: 0f4b825797 ("riscv: KVM: Add vector lazy save/restore support")
Cc: stable@vger.kernel.org
Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
Reviewed-by: Andy Chiu <andybnac@gmail.com>
Link: https://lore.kernel.org/r/20260316151612.13305-1-osama.abdelkader@gmail.com
Signed-off-by: Anup Patel <anup@brainfault.org>
When saddr_high != 0 on RV32, the goto out was unconditional, causing
valid 64-bit addresses to be rejected. Only goto out when the address
is invalid (64-bit host with saddr_high != 0).
Fixes: c2f41ddbcd ("RISC-V: KVM: Implement SBI PMU Snapshot feature")
Signed-off-by: Osama Abdelkader <osama.abdelkader@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260311231833.13189-1-osama.abdelkader@gmail.com
Signed-off-by: Anup Patel <anup@brainfault.org>
Add RISC-V KVM selftests to verify the SBI Steal-Time Accounting (STA)
shared memory alignment requirements.
The SBI specification requires the STA shared memory GPA to be 64-byte
aligned, or set to all-ones to explicitly disable steal-time accounting.
This test verifies that KVM enforces the expected behavior when
configuring the SBI STA shared memory via KVM_SET_ONE_REG.
Specifically, the test checks that:
- misaligned GPAs are rejected with -EINVAL
- 64-byte aligned GPAs are accepted
- all-ones GPA is accepted
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260303010859.1763177-4-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
Move steal time UAPI tests from steal_time_init() into a separate
check_steal_time_uapi() function for better code organization and
maintainability.
Previously, x86 and ARM64 architectures performed UAPI validation
tests within steal_time_init(), mixing initialization logic with
uapi tests.
Changes by architecture:
x86_64:
- Extract MSR reserved bits test from steal_time_init()
- Move to check_steal_time_uapi() which tests that setting
MSR_KVM_STEAL_TIME with KVM_STEAL_RESERVED_MASK fails
ARM64:
- Extract three UAPI tests from steal_time_init():
Device attribute support check
Misaligned IPA rejection (EINVAL)
Duplicate IPA setting rejection (EEXIST)
- Move all tests to check_steal_time_uapi()
RISC-V:
- Add empty check_steal_time_uapi() stub for future use
- No changes to steal_time_init() (had no tests to extract)
The new check_steal_time_uapi() function:
- Is called once before the per-VCPU test loop
No functional change intended.
Suggested-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260303010859.1763177-3-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
The RISC-V SBI Steal-Time Accounting (STA) extension requires the shared
memory physical address to be 64-byte aligned, or set to all-ones to
explicitly disable steal-time accounting.
KVM exposes the SBI STA shared memory configuration to userspace via
KVM_SET_ONE_REG. However, the current implementation of
kvm_sbi_ext_sta_set_reg() does not validate the alignment of the configured
shared memory address. As a result, userspace can install a misaligned
shared memory address that violates the SBI specification.
Such an invalid configuration may later reach runtime code paths that
assume a valid and properly aligned shared memory region. In particular,
KVM_RUN can trigger the following WARN_ON in
kvm_riscv_vcpu_record_steal_time():
WARNING: arch/riscv/kvm/vcpu_sbi_sta.c:49 at
kvm_riscv_vcpu_record_steal_time
WARN_ON paths are not expected to be reachable during normal runtime
execution, and may result in a kernel panic when panic_on_warn is enabled.
Fix this by validating the computed shared memory GPA at the
KVM_SET_ONE_REG boundary. A temporary GPA is constructed and checked
before committing it to vcpu->arch.sta.shmem. The validation allows
either a 64-byte aligned GPA or INVALID_GPA (all-ones), which disables
STA as defined by the SBI specification.
This prevents invalid userspace state from reaching runtime code paths
that assume SBI STA invariants and avoids unexpected WARN_ON behavior.
Fixes: f61ce890b1 ("RISC-V: KVM: Add support for SBI STA registers")
Signed-off-by: Jiakai Xu <xujiakai2025@iscas.ac.cn>
Signed-off-by: Jiakai Xu <jiakaiPeanut@gmail.com>
Reviewed-by: Andrew Jones <andrew.jones@oss.qualcomm.com>
Link: https://lore.kernel.org/r/20260303010859.1763177-2-xujiakai2025@iscas.ac.cn
Signed-off-by: Anup Patel <anup@brainfault.org>
I2C devices with associated pinctrl states (DPAUX I2C controllers)
will change pinctrl state during runtime PM. This requires taking
a mutex, so these devices cannot be marked as IRQ safe.
Add PINCTRL as dependency to avoid build errors.
Signed-off-by: Mikko Perttunen <mperttunen@nvidia.com>
Reported-by: Russell King <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/all/E1vsNBv-00000009nfA-27ZK@rmk-PC.armlinux.org.uk/
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-----BEGIN PGP SIGNATURE-----
iQIzBAABCAAdFiEE+soXsSLHKoYyzcli6rmadz2vbToFAmnAGisACgkQ6rmadz2v
bTqjsw/9GfHT/fdnjfA/q27TQH28ZdrZfq90BpI3m5BfTO8/l+Kt+g1HDGpku+C/
iWh66rg9t/P9nMvtdzvPsdT833UbwbY6fPEK3r7ANgf7SBb1DNvaGHBM6XNefvZV
j+VcykKUaEo8U1GeG+gI4TyAALSqvvMeBPYpAPZDUYguYLyE+YIl2Pl6tWt+A7yf
9V3JjCSz63t75qqnhY2SIBZv2pqWiMaCI8uPgaF7drhQM5Xc0l/R75CMPGeF9BrT
GRtTVJhY+6UyI2Q0ZRSRSVHZ1j2kYHI/eK3Kamxwal5hNh37BYHm3pT5TSHbZTe1
xO7c1AB0vds8kznRkclQfsMdjVwuBQj03ukLVNqnnaaE4Ir7JlXlXYgeG0KJbbfW
kQG8UyDD7tMWZkvaA0Z51FC88WJNLJoNAku519alcMtgAf1CrxzG9aUAYEWE4erh
E/FKKvFqQ6T0mOFSXlk1NFeMjNXcg5Tu2KKKKOjAWT6goUc4hw80IWydTyxMy32m
8/eLmdTZpAQovc2rS+5LSTigQ3DT082J950sxdQ3yRaLTWBGNC06gkA/WcRq2ZI+
hBdW6GI1XFwkXGw5+F9fN9Bt5FmE42v44i+RrlNZV1R5bVr0Za/ofkWP3dm1/SOg
QRSJk30hx9JveR9gD/xWawycYFuwmha/BL0tur2T32M67MneJpo=
=Ye1S
-----END PGP SIGNATURE-----
Merge tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Pull bpf fixes from Alexei Starovoitov:
- Fix how linked registers track zero extension of subregisters (Daniel
Borkmann)
- Fix unsound scalar fork for OR instructions (Daniel Wade)
- Fix exception exit lock check for subprogs (Ihor Solodrai)
- Fix undefined behavior in interpreter for SDIV/SMOD instructions
(Jenny Guanni Qu)
- Release module's BTF when module is unloaded (Kumar Kartikeya
Dwivedi)
- Fix constant blinding for PROBE_MEM32 instructions (Sachin Kumar)
- Reset register ID for END instructions to prevent incorrect value
tracking (Yazhou Tang)
* tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
selftests/bpf: Add a test cases for sync_linked_regs regarding zext propagation
bpf: Fix sync_linked_regs regarding BPF_ADD_CONST32 zext propagation
selftests/bpf: Add tests for maybe_fork_scalars() OR vs AND handling
bpf: Fix unsound scalar forking in maybe_fork_scalars() for BPF_OR
selftests/bpf: Add tests for sdiv32/smod32 with INT_MIN dividend
bpf: Fix undefined behavior in interpreter sdiv/smod for INT_MIN
selftests/bpf: Add tests for bpf_throw lock leak from subprogs
bpf: Fix exception exit lock checking for subprogs
bpf: Release module BTF IDR before module unload
selftests/bpf: Fix pkg-config call on static builds
bpf: Fix constant blinding for PROBE_MEM32 stores
selftests/bpf: Add test for BPF_END register ID reset
bpf: Reset register ID for BPF_END value tracking
- Revert "tracing: Remove pid in task_rename tracing output"
A change was made to remove the pid field from the task_rename event
because it was thought that it was always done for the current task and
recording the pid would be redundant. This turned out to be incorrect and
there are a few corner case where this is not true and caused some
regressions in tooling.
- Fix the reading from user space for migration
The reading of user space uses a seq lock type of logic where it uses a
per-cpu temporary buffer and disables migration, then enables preemption,
does the copy from user space, disables preemption, enables migration and
checks if there was any schedule switches while preemption was enabled. If
there was a context switch, then it is considered that the per-cpu buffer
could be corrupted and it tries again. There's a protection check that
tests if it takes a hundred tries, it issues a warning and exits out to
prevent a live lock.
This was triggered because the task was selected by the load balancer to
be migrated to another CPU, every time preemption is enabled the migration
task would schedule in try to migrate the task but can't because migration
is disabled and let it run again. This caused the scheduler to schedule out
the task every time it enabled preemption and made the loop never exit
(until the 100 iteration test triggered).
Fix this by enabling and disabling preemption and keeping migration
enabled if the reading from user space needs to be done again. This will
let the migration thread migrate the task and the copy from user space
will likely pass on the next iteration.
- Fix trace_marker copy option freeing
The "copy_trace_marker" option allows a tracing instance to get a copy of
a write to the trace_marker file of the top level instance. This is
managed by a link list protected by RCU. When an instance is removed, a
check is made if the option is set, and if so synchronized_rcu() is
called. The problem is that an iteration is made to reset all the flags to
what they were when the instance was created (to perform clean ups) was
done before the check of the copy_trace_marker option and that option was
cleared, so the synchronize_rcu() was never called.
Move the clearing of all the flags after the check of copy_trace_marker to
do synchronize_rcu() so that the option is still set if it was before and
the synchronization is performed.
- Fix entries setting when validating the persistent ring buffer
When validating the persistent ring buffer on boot up, the number of
events per sub-buffer is added to the sub-buffer meta page. The validator
was updating cpu_buffer->head_page (the first sub-buffer of the per-cpu
buffer) and not the "head_page" variable that was iterating the
sub-buffers. This was causing the first sub-buffer to be assigned the
entries for each sub-buffer and not the sub-buffer that was supposed to be
updated.
- Use "hash" value to update the direct callers
When updating the ftrace direct callers, it assigned a temporary callback
to all the callback functions of the ftrace ops and not just the
functions represented by the passed in hash. This causes an unnecessary
slow down of the functions of the ftrace_ops that is not being modified.
Only update the functions that are going to be modified to call the
ftrace loop function so that the update can be made on those functions.
-----BEGIN PGP SIGNATURE-----
iIoEABYKADIWIQRRSw7ePDh/lE+zeZMp5XQQmuv6qgUCacAMahQccm9zdGVkdEBn
b29kbWlzLm9yZwAKCRAp5XQQmuv6qr0sAQCoI4L3iAR5HU1z8dw2GWhOz9fTnzfw
9VPRZAsga9J5xgEA1Y0bvKBM0UPHFAL2POkaILYV1aT00lZ7aIVHPqfdYgA=
=OoGW
-----END PGP SIGNATURE-----
Merge tag 'trace-v7.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull tracing fixes from Steven Rostedt:
- Revert "tracing: Remove pid in task_rename tracing output"
A change was made to remove the pid field from the task_rename event
because it was thought that it was always done for the current task
and recording the pid would be redundant. This turned out to be
incorrect and there are a few corner case where this is not true and
caused some regressions in tooling.
- Fix the reading from user space for migration
The reading of user space uses a seq lock type of logic where it uses
a per-cpu temporary buffer and disables migration, then enables
preemption, does the copy from user space, disables preemption,
enables migration and checks if there was any schedule switches while
preemption was enabled. If there was a context switch, then it is
considered that the per-cpu buffer could be corrupted and it tries
again. There's a protection check that tests if it takes a hundred
tries, it issues a warning and exits out to prevent a live lock.
This was triggered because the task was selected by the load balancer
to be migrated to another CPU, every time preemption is enabled the
migration task would schedule in try to migrate the task but can't
because migration is disabled and let it run again. This caused the
scheduler to schedule out the task every time it enabled preemption
and made the loop never exit (until the 100 iteration test
triggered).
Fix this by enabling and disabling preemption and keeping migration
enabled if the reading from user space needs to be done again. This
will let the migration thread migrate the task and the copy from user
space will likely pass on the next iteration.
- Fix trace_marker copy option freeing
The "copy_trace_marker" option allows a tracing instance to get a
copy of a write to the trace_marker file of the top level instance.
This is managed by a link list protected by RCU. When an instance is
removed, a check is made if the option is set, and if so
synchronized_rcu() is called.
The problem is that an iteration is made to reset all the flags to
what they were when the instance was created (to perform clean ups)
was done before the check of the copy_trace_marker option and that
option was cleared, so the synchronize_rcu() was never called.
Move the clearing of all the flags after the check of
copy_trace_marker to do synchronize_rcu() so that the option is still
set if it was before and the synchronization is performed.
- Fix entries setting when validating the persistent ring buffer
When validating the persistent ring buffer on boot up, the number of
events per sub-buffer is added to the sub-buffer meta page. The
validator was updating cpu_buffer->head_page (the first sub-buffer of
the per-cpu buffer) and not the "head_page" variable that was
iterating the sub-buffers. This was causing the first sub-buffer to
be assigned the entries for each sub-buffer and not the sub-buffer
that was supposed to be updated.
- Use "hash" value to update the direct callers
When updating the ftrace direct callers, it assigned a temporary
callback to all the callback functions of the ftrace ops and not just
the functions represented by the passed in hash. This causes an
unnecessary slow down of the functions of the ftrace_ops that is not
being modified. Only update the functions that are going to be
modified to call the ftrace loop function so that the update can be
made on those functions.
* tag 'trace-v7.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
ftrace: Use hash argument for tmp_ops in update_ftrace_direct_mod
ring-buffer: Fix to update per-subbuf entries of persistent ring buffer
tracing: Fix trace_marker copy link list updates
tracing: Fix failure to read user space from system call trace events
tracing: Revert "tracing: Remove pid in task_rename tracing output"
pxa: fix broken I2C communication on Armada 3700 with recovery
fsi: fix device_node reference leak in probe
cp2615: fix NULL-deref when serial string is missing
-----BEGIN PGP SIGNATURE-----
iQIzBAABCgAdFiEEOZGx6rniZ1Gk92RdFA3kzBSgKbYFAmm/trAACgkQFA3kzBSg
KbaRxQ//a6yPhI6jQ8NDNnaO3c/RwvLbJsOyh7y7Q3cXl1IVbb6bPEY/PWJBGGeK
XbmtqdT8vR7sh/OoxwSJYH70ZtRi0aWpL5adabiOl8VJhnW3rABrWS5WIBgm9djL
k6p+lnTGNJHWw3PzwVb7lzkgkmyolc2g6qzpTJJd+it39G/ChPTt6vOU6wNgWSMK
auRY59FBRwMkKymMt1co77bmvi/ehFm/Ohsjs4/bGkPr/Ngq6wVtm7tcdu5drShR
VcexM+yu92pLBr7zR/p89E36W7aqWPGl8K7aYyDJjYXDHcOcRJAHjKi4lf+boCNa
ze4dg66Hj6ONVgxOF5lTFGP3ITPg/JP02kskRlVqzJsZdm8hIQraq5eJEUBF0rWZ
WsA/E8SuGeNe3hDck5YyoaNbpww0Ph1y+MgNglMXS/hKedEZL7ku5mU66wA1gjjm
ek18V0rAFiBP3Ch7vTK4Se1CE1TUUSNoPdEwdPU5/3T8ZZLbgRewr4z5FPL2+Yb+
NdGy6ZPV7rRYXxQVVbNTQ9c4h4cuBYZgnrRGfYXyXXyQ77m5mZlMBa4emPoNExtO
9U2fxbfHU2QBTG0lMOWkbEtMc1bVSvYtDwv1X2gFsdkOy+tmyCDkDHsp0hqao0oo
C5rlOYTl/9U6byrpBe8GiCrKHNNmtiuSDJzRUhdXHb8JqqPqp+c=
=m7EX
-----END PGP SIGNATURE-----
Merge tag 'i2c-for-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux
Pull i2c fixes from Wolfram Sang:
- fix broken I2C communication on Armada 3700 with recovery
- fix device_node reference leak in probe (fsi)
- fix NULL-deref when serial string is missing (cp2615)
* tag 'i2c-for-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
i2c: pxa: defer reset on Armada 3700 when recovery is used
i2c: fsi: Fix a potential leak in fsi_i2c_probe()
i2c: cp2615: fix serial string NULL-deref at probe
- Improve Qemu MCE-injection behavior by only using
AMD SMCA MSRs if the feature bit is set.
- Fix the relative path of gettimeofday.c inclusion
in vclock_gettime.c
- Fix a boot crash on UV clusters when a socket is
marked as 'deconfigured' which are mapped to the
SOCK_EMPTY (0xffff [short -1]) node ID by the UV firmware,
while Linux APIs expect NUMA_NO_NODE (0xffffffff [int -1]).
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=uIqH
-----END PGP SIGNATURE-----
Merge tag 'x86-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 fixes from Ingo Molnar:
- Improve Qemu MCE-injection behavior by only using AMD SMCA MSRs if
the feature bit is set
- Fix the relative path of gettimeofday.c inclusion in vclock_gettime.c
- Fix a boot crash on UV clusters when a socket is marked as
'deconfigured' which are mapped to the SOCK_EMPTY node ID by
the UV firmware, while Linux APIs expect NUMA_NO_NODE.
The difference being (0xffff [unsigned short ~0]) vs [int -1]
* tag 'x86-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/platform/uv: Handle deconfigured sockets
x86/entry/vdso: Fix path of included gettimeofday.c
x86/mce/amd: Check SMCA feature bit before accessing SMCA MSRs
- Fix a PMU driver crash on AMD EPYC systems, caused by
a race condition in x86_pmu_enable()
- Fix a possible counter-initialization bug in x86_pmu_enable()
- Fix a counter inheritance bug in inherit_event() and
__perf_event_read()
- Fix an Intel PMU driver branch constraints handling bug
found by UBSAN
- Fix the Intel PMU driver's new Off-Module Response (OMR)
support code for Diamond Rapids / Nova lake, to fix a snoop
information parsing bug
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=jp5H
-----END PGP SIGNATURE-----
Merge tag 'perf-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull perf fixes from Ingo Molnar:
- Fix a PMU driver crash on AMD EPYC systems, caused by
a race condition in x86_pmu_enable()
- Fix a possible counter-initialization bug in x86_pmu_enable()
- Fix a counter inheritance bug in inherit_event() and
__perf_event_read()
- Fix an Intel PMU driver branch constraints handling bug
found by UBSAN
- Fix the Intel PMU driver's new Off-Module Response (OMR)
support code for Diamond Rapids / Nova lake, to fix a snoop
information parsing bug
* tag 'perf-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
perf/x86/intel: Fix OMR snoop information parsing issues
perf/x86/intel: Add missing branch counters constraint apply
perf: Make sure to use pmu_ctx->pmu for groups
x86/perf: Make sure to program the counter value for stopped events on migration
perf/x86: Move event pointer setup earlier in x86_pmu_enable()
and fix a false positive warning with Clang jump tables.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=W/aM
-----END PGP SIGNATURE-----
Merge tag 'objtool-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull objtool fixes from Ingo Molnar:
"Fix three more livepatching related build environment bugs, and a
false positive warning with Clang jump tables"
* tag 'objtool-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
objtool: Fix Clang jump table detection
livepatch/klp-build: Fix inconsistent kernel version
objtool/klp: fix mkstemp() failure with long paths
objtool/klp: fix data alignment in __clone_symbol()
caused by the locking context-analysis changes.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
-----BEGIN PGP SIGNATURE-----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=o1hi
-----END PGP SIGNATURE-----
Merge tag 'locking-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull locking fix from Ingo Molnar:
"Fix a sparse build error regression in <linux/local_lock_internal.h>
caused by the locking context-analysis changes"
* tag 'locking-urgent-2026-03-22' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
include/linux/local_lock_internal.h: Make this header file again compatible with sparse
- Generalize driver_override in the driver core, providing a common
sysfs implementation and concurrency-safe accessors for bus
implementations
- Do not use driver_override as IRQ name in the hwmon axi-fan driver
- Remove an unnecessary driver_override check in sh platform_early
- Migrate the platform bus to use the generic driver_override
infrastructure, fixing a UAF condition caused by accessing the
driver_override field without proper locking in the platform_match()
callback
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQS2q/xV6QjXAdC7k+1FlHeO1qrKLgUCab6dRwAKCRBFlHeO1qrK
LnELAP9MkP5mnaIbAM5/kYWaG+nQhsBTdzSbLSqv0vca7bljlwD/RFzAXOaCKvNz
8GrnY7FHzGSMPQhT+L86idGRCPmujgI=
=dxv9
-----END PGP SIGNATURE-----
Merge tag 'driver-core-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core
Pull driver core fixes from Danilo Krummrich:
- Generalize driver_override in the driver core, providing a common
sysfs implementation and concurrency-safe accessors for bus
implementations
- Do not use driver_override as IRQ name in the hwmon axi-fan driver
- Remove an unnecessary driver_override check in sh platform_early
- Migrate the platform bus to use the generic driver_override
infrastructure, fixing a UAF condition caused by accessing the
driver_override field without proper locking in the platform_match()
callback
* tag 'driver-core-7.0-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core:
driver core: platform: use generic driver_override infrastructure
sh: platform_early: remove pdev->driver_override check
hwmon: axi-fan: don't use driver_override as IRQ name
docs: driver-model: document driver_override
driver core: generalize driver_override in struct device
The modify logic registers temporary ftrace_ops object (tmp_ops) to trigger
the slow path for all direct callers to be able to safely modify attached
addresses.
At the moment we use ops->func_hash for tmp_ops filter, which represents all
the systems attachments. It's faster to use just the passed hash filter, which
contains only the modified sites and is always a subset of the ops->func_hash.
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Andrii Nakryiko <andrii@kernel.org>
Cc: Menglong Dong <menglong8.dong@gmail.com>
Cc: Song Liu <song@kernel.org>
Link: https://patch.msgid.link/20260312123738.129926-1-jolsa@kernel.org
Fixes: e93672f770 ("ftrace: Add update_ftrace_direct_mod function")
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Since the validation loop in rb_meta_validate_events() updates the same
cpu_buffer->head_page->entries, the other subbuf entries are not updated.
Fix to use head_page to update the entries field, since it is the cursor
in this loop.
Cc: stable@vger.kernel.org
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Ian Rogers <irogers@google.com>
Fixes: 5f3b6e839f ("ring-buffer: Validate boot range memory events")
Link: https://patch.msgid.link/177391153882.193994.17158784065013676533.stgit@mhiramat.tok.corp.google.com
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
When the "copy_trace_marker" option is enabled for an instance, anything
written into /sys/kernel/tracing/trace_marker is also copied into that
instances buffer. When the option is set, that instance's trace_array
descriptor is added to the marker_copies link list. This list is protected
by RCU, as all iterations uses an RCU protected list traversal.
When the instance is deleted, all the flags that were enabled are cleared.
This also clears the copy_trace_marker flag and removes the trace_array
descriptor from the list.
The issue is after the flags are called, a direct call to
update_marker_trace() is performed to clear the flag. This function
returns true if the state of the flag changed and false otherwise. If it
returns true here, synchronize_rcu() is called to make sure all readers
see that its removed from the list.
But since the flag was already cleared, the state does not change and the
synchronization is never called, leaving a possible UAF bug.
Move the clearing of all flags below the updating of the copy_trace_marker
option which then makes sure the synchronization is performed.
Also use the flag for checking the state in update_marker_trace() instead
of looking at if the list is empty.
Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20260318185512.1b6c7db4@gandalf.local.home
Fixes: 7b382efd5e ("tracing: Allow the top level trace_marker to write into another instances")
Reported-by: Sasha Levin <sashal@kernel.org>
Closes: https://lore.kernel.org/all/20260225133122.237275-1-sashal@kernel.org/
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
The system call trace events call trace_user_fault_read() to read the user
space part of some system calls. This is done by grabbing a per-cpu
buffer, disabling migration, enabling preemption, calling
copy_from_user(), disabling preemption, enabling migration and checking if
the task was preempted while preemption was enabled. If it was, the buffer
is considered corrupted and it tries again.
There's a safety mechanism that will fail out of this loop if it fails 100
times (with a warning). That warning message was triggered in some
pi_futex stress tests. Enabling the sched_switch trace event and
traceoff_on_warning, showed the problem:
pi_mutex_hammer-1375 [006] d..21 138.981648: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981651: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981656: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981659: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981664: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981667: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981671: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981675: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981679: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981682: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981687: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981690: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981695: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981698: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981703: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981706: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981711: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981714: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981719: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981722: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981727: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981730: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
pi_mutex_hammer-1375 [006] d..21 138.981735: sched_switch: prev_comm=pi_mutex_hammer prev_pid=1375 prev_prio=95 prev_state=R+ ==> next_comm=migration/6 next_pid=47 next_prio=0
migration/6-47 [006] d..2. 138.981738: sched_switch: prev_comm=migration/6 prev_pid=47 prev_prio=0 prev_state=S ==> next_comm=pi_mutex_hammer next_pid=1375 next_prio=95
What happened was the task 1375 was flagged to be migrated. When
preemption was enabled, the migration thread woke up to migrate that task,
but failed because migration for that task was disabled. This caused the
loop to fail to exit because the task scheduled out while trying to read
user space.
Every time the task enabled preemption the migration thread would schedule
in, try to migrate the task, fail and let the task continue. But because
the loop would only enable preemption with migration disabled, it would
always fail because each time it enabled preemption to read user space,
the migration thread would try to migrate it.
To solve this, when the loop fails to read user space without being
scheduled out, enabled and disable preemption with migration enabled. This
will allow the migration task to successfully migrate the task and the
next loop should succeed to read user space without being scheduled out.
Cc: stable@vger.kernel.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Link: https://patch.msgid.link/20260316130734.1858a998@gandalf.local.home
Fixes: 64cf7d058a ("tracing: Have trace_marker use per-cpu data to read user space")
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
This reverts commit e3f6a42272.
The commit says that the tracepoint only deals with the current task,
however the following case is not current task:
comm_write() {
p = get_proc_task(inode);
if (!p)
return -ESRCH;
if (same_thread_group(current, p))
set_task_comm(p, buffer);
}
where set_task_comm() calls __set_task_comm() which records
the update of p and not current.
So revert the patch to show pid.
Cc: <mhiramat@kernel.org>
Cc: <mathieu.desnoyers@efficios.com>
Cc: <elver@google.com>
Cc: <kees@kernel.org>
Link: https://patch.msgid.link/20260306075954.4533-1-xuewen.yan@unisoc.com
Fixes: e3f6a42272 ("tracing: Remove pid in task_rename tracing output")
Reported-by: Guohua Yan <guohua.yan@unisoc.com>
Signed-off-by: Xuewen Yan <xuewen.yan@unisoc.com>
Reviewed-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Add multiple test cases for linked register tracking with alu32 ops:
- Add a test that checks sync_linked_regs() regarding reg->id (the linked
target register) for BPF_ADD_CONST32 rather than known_reg->id (the
branch register).
- Add a test case for linked register tracking that exposes the cross-type
sync_linked_regs() bug. One register uses alu32 (w7 += 1, BPF_ADD_CONST32)
and another uses alu64 (r8 += 2, BPF_ADD_CONST64), both linked to the
same base register.
- Add a test case that exercises regsafe() path pruning when two execution
paths reach the same program point with linked registers carrying
different ADD_CONST flags (BPF_ADD_CONST32 from alu32 vs BPF_ADD_CONST64
from alu64). This particular test passes with and without the fix since
the pruning will fail due to different ranges, but it would still be
useful to carry this one as a regression test for the unreachable div
by zero.
With the fix applied all the tests pass:
# LDLIBS=-static PKG_CONFIG='pkg-config --static' ./vmtest.sh -- ./test_progs -t verifier_linked_scalars
[...]
./test_progs -t verifier_linked_scalars
#602/1 verifier_linked_scalars/scalars: find linked scalars:OK
#602/2 verifier_linked_scalars/sync_linked_regs_preserves_id:OK
#602/3 verifier_linked_scalars/scalars_neg:OK
#602/4 verifier_linked_scalars/scalars_neg_sub:OK
#602/5 verifier_linked_scalars/scalars_neg_alu32_add:OK
#602/6 verifier_linked_scalars/scalars_neg_alu32_sub:OK
#602/7 verifier_linked_scalars/scalars_pos:OK
#602/8 verifier_linked_scalars/scalars_sub_neg_imm:OK
#602/9 verifier_linked_scalars/scalars_double_add:OK
#602/10 verifier_linked_scalars/scalars_sync_delta_overflow:OK
#602/11 verifier_linked_scalars/scalars_sync_delta_overflow_large_range:OK
#602/12 verifier_linked_scalars/scalars_alu32_big_offset:OK
#602/13 verifier_linked_scalars/scalars_alu32_basic:OK
#602/14 verifier_linked_scalars/scalars_alu32_wrap:OK
#602/15 verifier_linked_scalars/scalars_alu32_zext_linked_reg:OK
#602/16 verifier_linked_scalars/scalars_alu32_alu64_cross_type:OK
#602/17 verifier_linked_scalars/scalars_alu32_alu64_regsafe_pruning:OK
#602/18 verifier_linked_scalars/alu32_negative_offset:OK
#602/19 verifier_linked_scalars/spurious_precision_marks:OK
#602 verifier_linked_scalars:OK
Summary: 1/19 PASSED, 0 SKIPPED, 0 FAILED
Co-developed-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20260319211507.213816-2-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Jenny reported that in sync_linked_regs() the BPF_ADD_CONST32 flag is
checked on known_reg (the register narrowed by a conditional branch)
instead of reg (the linked target register created by an alu32 operation).
Example case with reg:
1. r6 = bpf_get_prandom_u32()
2. r7 = r6 (linked, same id)
3. w7 += 5 (alu32 -- r7 gets BPF_ADD_CONST32, zero-extended by CPU)
4. if w6 < 0xFFFFFFFC goto safe (narrows r6 to [0xFFFFFFFC, 0xFFFFFFFF])
5. sync_linked_regs() propagates to r7 but does NOT call zext_32_to_64()
6. Verifier thinks r7 is [0x100000001, 0x100000004] instead of [1, 4]
Since known_reg above does not have BPF_ADD_CONST32 set above, zext_32_to_64()
is never called on alu32-derived linked registers. This causes the verifier
to track incorrect 64-bit bounds, while the CPU correctly zero-extends the
32-bit result.
The code checking known_reg->id was correct however (see scalars_alu32_wrap
selftest case), but the real fix needs to handle both directions - zext
propagation should be done when either register has BPF_ADD_CONST32, since
the linked relationship involves a 32-bit operation regardless of which
side has the flag.
Example case with known_reg (exercised also by scalars_alu32_wrap):
1. r1 = r0; w1 += 0x100 (alu32 -- r1 gets BPF_ADD_CONST32)
2. if r1 > 0x80 - known_reg = r1 (has BPF_ADD_CONST32), reg = r0 (doesn't)
Hence, fix it by checking for (reg->id | known_reg->id) & BPF_ADD_CONST32.
Moreover, sync_linked_regs() also has a soundness issue when two linked
registers used different ALU widths: one with BPF_ADD_CONST32 and the
other with BPF_ADD_CONST64. The delta relationship between linked registers
assumes the same arithmetic width though. When one register went through
alu32 (CPU zero-extends the 32-bit result) and the other went through
alu64 (no zero-extension), the propagation produces incorrect bounds.
Example:
r6 = bpf_get_prandom_u32() // fully unknown
if r6 >= 0x100000000 goto out // constrain r6 to [0, U32_MAX]
r7 = r6
w7 += 1 // alu32: r7.id = N | BPF_ADD_CONST32
r8 = r6
r8 += 2 // alu64: r8.id = N | BPF_ADD_CONST64
if r7 < 0xFFFFFFFF goto out // narrows r7 to [0xFFFFFFFF, 0xFFFFFFFF]
At the branch on r7, sync_linked_regs() runs with known_reg=r7
(BPF_ADD_CONST32) and reg=r8 (BPF_ADD_CONST64). The delta path
computes:
r8 = r7 + (delta_r8 - delta_r7) = 0xFFFFFFFF + (2 - 1) = 0x100000000
Then, because known_reg->id has BPF_ADD_CONST32, zext_32_to_64(r8) is
called, truncating r8 to [0, 0]. But r8 used a 64-bit ALU op -- the
CPU does NOT zero-extend it. The actual CPU value of r8 is
0xFFFFFFFE + 2 = 0x100000000, not 0. The verifier now underestimates
r8's 64-bit bounds, which is a soundness violation.
Fix sync_linked_regs() by skipping propagation when the two registers
have mixed ALU widths (one BPF_ADD_CONST32, the other BPF_ADD_CONST64).
Lastly, fix regsafe() used for path pruning: the existing checks used
"& BPF_ADD_CONST" to test for offset linkage, which treated
BPF_ADD_CONST32 and BPF_ADD_CONST64 as equivalent.
Fixes: 7a433e5193 ("bpf: Support negative offsets, BPF_SUB, and alu32 for linked register tracking")
Reported-by: Jenny Guanni Qu <qguanni@gmail.com>
Co-developed-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20260319211507.213816-1-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Daniel Wade says:
====================
bpf: Fix unsound scalar forking for BPF_OR
maybe_fork_scalars() unconditionally sets the pushed path dst register
to 0 for both BPF_AND and BPF_OR. For AND this is correct (0 & K == 0),
but for OR it is wrong (0 | K == K, not 0). This causes the verifier to
track an incorrect value on the pushed path, leading to a verifier/runtime
divergence that allows out-of-bounds map value access.
v4: Use block comment style for multi-line comments in selftests (Amery Hung)
Add Reviewed-by/Acked-by tags
v3: Use single-line comment style in selftests (Alexei Starovoitov)
v2: Use push_stack(env, env->insn_idx, ...) to re-execute the insn
on the pushed path (Eduard Zingerman)
====================
Link: https://patch.msgid.link/20260314021521.128361-1-danjwade95@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add three test cases to verifier_bounds.c to verify that
maybe_fork_scalars() correctly tracks register values for BPF_OR
operations with constant source operands:
1. or_scalar_fork_rejects_oob: After ARSH 63 + OR 8, the pushed
path should have dst = 8. With value_size = 8, accessing
map_value + 8 is out of bounds and must be rejected.
2. and_scalar_fork_still_works: Regression test ensuring AND
forking continues to work. ARSH 63 + AND 4 produces pushed
dst = 0 and current dst = 4, both within value_size = 8.
3. or_scalar_fork_allows_inbounds: After ARSH 63 + OR 4, the
pushed path has dst = 4, which is within value_size = 8
and should be accepted.
These tests exercise the fix in the previous patch, which makes the
pushed path re-execute the ALU instruction so it computes the correct
result for BPF_OR.
Signed-off-by: Daniel Wade <danjwade95@gmail.com>
Reviewed-by: Amery Hung <ameryhung@gmail.com>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20260314021521.128361-3-danjwade95@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
maybe_fork_scalars() is called for both BPF_AND and BPF_OR when the
source operand is a constant. When dst has signed range [-1, 0], it
forks the verifier state: the pushed path gets dst = 0, the current
path gets dst = -1.
For BPF_AND this is correct: 0 & K == 0.
For BPF_OR this is wrong: 0 | K == K, not 0.
The pushed path therefore tracks dst as 0 when the runtime value is K,
producing an exploitable verifier/runtime divergence that allows
out-of-bounds map access.
Fix this by passing env->insn_idx (instead of env->insn_idx + 1) to
push_stack(), so the pushed path re-executes the ALU instruction with
dst = 0 and naturally computes the correct result for any opcode.
Fixes: bffacdb80b ("bpf: Recognize special arithmetic shift in the verifier")
Signed-off-by: Daniel Wade <danjwade95@gmail.com>
Reviewed-by: Amery Hung <ameryhung@gmail.com>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20260314021521.128361-2-danjwade95@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Jenny Guanni Qu says:
====================
bpf: Fix abs(INT_MIN) undefined behavior in interpreter sdiv/smod
The BPF interpreter's signed 32-bit division and modulo handlers use
abs() on s32 operands, which is undefined for S32_MIN. This causes
the interpreter to compute wrong results, creating a mismatch with
the verifier's range tracking.
For example, INT_MIN / 2 returns 0x40000000 instead of the correct
0xC0000000. The verifier tracks the correct range, so a crafted BPF
program can exploit the mismatch for out-of-bounds map value access
(confirmed by KASAN).
Patch 1 introduces abs_s32() which handles S32_MIN correctly and
replaces all 8 abs((s32)...) call sites. s32 is the only affected
case -- the s64 handlers do not use abs().
Patch 2 adds selftests covering sdiv32 and smod32 with INT_MIN
dividend to prevent regression.
Changes since v4:
- Renamed __safe_abs32() to abs_s32() and dropped inline keyword
per Alexei Starovoitov's feedback
Changes since v3:
- Fixed stray blank line deletion in the file header
- Improved comment per Yonghong Song's suggestion
- Added JIT vs interpreter context to selftest commit message
Changes since v2:
- Simplified to use -(u32)x per Mykyta Yatsenko's suggestion
Changes since v1:
- Moved helper above kerneldoc comment block to fix build warnings
====================
Link: https://patch.msgid.link/20260311011116.2108005-1-qguanni@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add tests to verify that signed 32-bit division and modulo operations
produce correct results when the dividend is INT_MIN (0x80000000).
The bug fixed in the previous commit only affects the BPF interpreter
path. When JIT is enabled (the default on most architectures), the
native CPU division instruction produces the correct result and these
tests pass regardless. With bpf_jit_enable=0, the interpreter is used
and without the previous fix, INT_MIN / 2 incorrectly returns
0x40000000 instead of 0xC0000000 due to abs(S32_MIN) undefined
behavior, causing these tests to fail.
Test cases:
- SDIV32 INT_MIN / 2 = -1073741824 (imm and reg divisor)
- SMOD32 INT_MIN % 2 = 0 (positive and negative divisor)
Reviewed-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Signed-off-by: Jenny Guanni Qu <qguanni@gmail.com>
Link: https://lore.kernel.org/r/20260311011116.2108005-3-qguanni@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The BPF interpreter's signed 32-bit division and modulo handlers use
the kernel abs() macro on s32 operands. The abs() macro documentation
(include/linux/math.h) explicitly states the result is undefined when
the input is the type minimum. When DST contains S32_MIN (0x80000000),
abs((s32)DST) triggers undefined behavior and returns S32_MIN unchanged
on arm64/x86. This value is then sign-extended to u64 as
0xFFFFFFFF80000000, causing do_div() to compute the wrong result.
The verifier's abstract interpretation (scalar32_min_max_sdiv) computes
the mathematically correct result for range tracking, creating a
verifier/interpreter mismatch that can be exploited for out-of-bounds
map value access.
Introduce abs_s32() which handles S32_MIN correctly by casting to u32
before negating, avoiding signed overflow entirely. Replace all 8
abs((s32)...) call sites in the interpreter's sdiv32/smod32 handlers.
s32 is the only affected case -- the s64 division/modulo handlers do
not use abs().
Fixes: ec0e2da95f ("bpf: Support new signed div/mod instructions.")
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Jenny Guanni Qu <qguanni@gmail.com>
Link: https://lore.kernel.org/r/20260311011116.2108005-2-qguanni@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add test cases to ensure the verifier correctly rejects bpf_throw from
subprogs when RCU, preempt, or IRQ locks are held:
* reject_subprog_rcu_lock_throw: subprog acquires bpf_rcu_read_lock and
then calls bpf_throw
* reject_subprog_throw_preempt_lock: always-throwing subprog called while
caller holds bpf_preempt_disable
* reject_subprog_throw_irq_lock: always-throwing subprog called while
caller holds bpf_local_irq_save
Assisted-by: Claude:claude-opus-4-6
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260320000809.643798-2-ihor.solodrai@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
process_bpf_exit_full() passes check_lock = !curframe to
check_resource_leak(), which is false in cases when bpf_throw() is
called from a static subprog. This makes check_resource_leak() to skip
validation of active_rcu_locks, active_preempt_locks, and
active_irq_id on exception exits from subprogs.
At runtime bpf_throw() unwinds the stack via ORC without releasing any
user-acquired locks, which may cause various issues as the result.
Fix by setting check_lock = true for exception exits regardless of
curframe, since exceptions bypass all intermediate frame
cleanup. Update the error message prefix to "bpf_throw" for exception
exits to distinguish them from normal BPF_EXIT.
Fix reject_subprog_with_rcu_read_lock test which was previously
passing for the wrong reason. Test program returned directly from the
subprog call without closing the RCU section, so the error was
triggered by the unclosed RCU lock on normal exit, not by
bpf_throw. Update __msg annotations for affected tests to match the
new "bpf_throw" error prefix.
The spin_lock case is not affected because they are already checked [1]
at the call site in do_check_insn() before bpf_throw can run.
[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/kernel/bpf/verifier.c?h=v7.0-rc4#n21098
Assisted-by: Claude:claude-opus-4-6
Fixes: f18b03faba ("bpf: Implement BPF exceptions")
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Acked-by: Yonghong Song <yonghong.song@linux.dev>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260320000809.643798-1-ihor.solodrai@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
