The test verifies attachment to various hooks in a kernel module,
however, everything is flattened into a single test. This makes it
impossible to run or skip test cases selectively.
Isolate each BPF program into a separate subtest. This is done by
disabling auto-loading of programs and loading and testing each program
separately.
At the same time, modernize the test to use ASSERT* instead of CHECK and
replace `return` by `goto cleanup` where necessary.
Signed-off-by: Viktor Malik <vmalik@redhat.com>
Link: https://lore.kernel.org/r/20260225120904.1529112-1-vmalik@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Emil Tsalapatis says:
====================
bpf: Allow void return type for global subprogs
Global subprogs are currently not allowed to return void. Adjust
verifier logic to allow global functions with a void return type.
Exception callbacks are excluded from this change, and still require
a scalar return type.
Adding additional logic for void globals significantly complicates
check_return_code(), mainly because of how the new feature interacts
with bpf_throw(). To avoid complicating the code, refactor the return
value range validation logic into two separate functions, one for
program exits through BPF_EXIT and one for program exits through
bpf_throw(). The changes are non-functional.
Also adjust existing selftests that were ensuring that global subprogs
with void return types fail to verify. These tests should now load
successfully.
Patches 1-3 refactor check_return_code into two separate functions, one
for BPF_EXIT and one for bpf_throw(). No functional changes.
Patches 4-5 enable void globals in the verifier and add additional
selftests.
v5 -> v6
(https://lore.kernel.org/bpf/20260227154616.6846-1-emil@etsalapatis.com/)
Patch 1:
- Removed unused variables (kernel test robot)
- Add break; on empty default to suppress warning (kernel test robot)
- Added explanation on why change in switch statement for
BPF_PROG_TYPE_TRACING in check_return_code is a no-op
Patches 2/3:
- Added From: tags to Eduard's patches and changed my Reviewed-by to
Signed-off-by to properly denote patch route (submitting-patches.rst)
- Renamed check_subprogram_return_code for into check_global_subprog_return_code
(Eduard)
Patch 4:
- Remove unnecessary subprog_is_global() check in check_return_code (Eduard)
- Fix indentation on patch 4 (AI)
Patch 5:
- Fix unused data in selftests (Eduard)
v4 -> v5
(https://lore.kernel.org/bpf/20260225033356.518313-1-emil@etsalapatis.com/)
- Split value validation logic from throw/nonthrow to main prog/subprog (Eduard)
- Turn return_retval_range() to bool and incorporate return_32bit in retval (Eduard)
- Adjust if-else chain in return_retval_range for cgroup programs into switch
statement to bring it in line with the other program types
v3 -> v4
(https://lore.kernel.org/bpf/20260223215046.1706110-1-emil@etsalapatis.com/)
- Factor out bpf_throw() return value validation logic into its own
function (Eduard)
- Validate the range of bpf_throw() return values from void globals (Eduard)
- Move new selftest to verifier_global_subprogs.c (Eduard)
- Remove unnecessary verifier_bug_if() statements (Eduard)
- Add new test to ensure void globals can call bpf_throw
v2 -> v3
(https://lore.kernel.org/bpf/20260210183257.217285-1-emil@etsalapatis.com/)
- Allow freplace for global programs that return void and add selftests (Kumar)
v1 -> v2
(https://lore.kernel.org/bpf/20260127231414.359283-1-emil@etsalapatis.com/)
- Only mark R0 as valid for non-void global functions (AI)
- Rename subprog_is_void to subprog_returns_void (Kumar)
- Add test that R0 is invalid after calling a void global (Alexei)
- Add check to ensure exception callbacks do not return void
and associated selftest
Cc: Kumar Kartikeya Dwivedi <memxor@gmail.com>
====================
Link: https://patch.msgid.link/20260228184759.108145-1-emil@etsalapatis.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
From: Eduard Zingerman <eddyz87@gmail.com>
Both main progs and subprogs use the same function in the verifier,
check_return_code, to verify the type and value range of the register
being returned. However, subprogs only need a subset of the logic in
check_return_code. this also goes the way - check_return_code explicitly
checks whether it is handling a subprogram in multiple places, complicating
the logic. Separate the handling of the two into separate functions.
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://lore.kernel.org/r/20260228184759.108145-4-emil@etsalapatis.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Factor the return value range calculation logic in check_return_code
out of the function in preparation for separating the return value
validation logic for BPF_EXIT and bpf_throw().
No functional changes. The change made in return_retval_code's handling
of PROG_TRACING program types (not error'ing on the default case) is a
no-op because the match on the program attach type is exhaustive.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://lore.kernel.org/r/20260228184759.108145-2-emil@etsalapatis.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The test_bpftool.sh script runs a python unittest script checking
bpftool json output on different commands. As part of the ongoing effort
to get rid of any standalone test, this script should either be
converted to test_progs or removed.
As validating bpftool json output does not bring much value to the test
base (and because it would need test_progs to bring in a json parser),
remove the standalone test script.
Signed-off-by: Alexis Lothoré (eBPF Foundation) <alexis.lothore@bootlin.com>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/r/20260227-bpftool_feature-v1-1-a25860fd52fb@bootlin.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Kumar Kartikeya Dwivedi says:
====================
Allow fixed offsets for PTR_TO_CTX
Enable pointer modification with constant offsets accumulated in the
register for PTR_TO_CTX for programs where it won't be rewritten. See
patches for details.
====================
Link: https://patch.msgid.link/20260227005725.1247305-1-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add tests to ensure PTR_TO_CTX supports fixed offsets for program types
that don't rewrite accesses to it. Ensure that variable offsets and
negative offsets are still rejected. An extra test also checks writing
into ctx with modified offset for syscall progs. Other program types do
not support writes (notably, writable tracepoints offer a pointer for a
writable buffer through ctx, but don't allow writing to the ctx itself).
Before the fix made in the previous commit, these tests do not succeed,
except the ones testing for failures regardless of the change.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://lore.kernel.org/r/20260227005725.1247305-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The limitation on fixed offsets stems from the fact that certain program
types rewrite the accesses to the context structure and translate them
to accesses to the real underlying type. Technically, in the past, we
could have stashed the register offset in insn_aux and made rewrites
work, but we've never needed it in the past since the offset for such
context structures easily fit in the s16 signed instruction offset.
Regardless, the consequence is that for program types where the program
type's verifier ops doesn't supply a convert_ctx_access callback, we
unnecessarily reject accesses with a modified ctx pointer (i.e., one
whose offset has been shifted) in check_ptr_off_reg. Make an exception
for such program types (like syscall, tracepoint, raw_tp, etc.).
Pass in fixed_off_ok as true to __check_ptr_off_reg for such cases, and
accumulate the register offset into insn->off passed to check_ctx_access.
In particular, the accumulation is critical since we need to correctly
track the max_ctx_offset which is used for bounds checking the buffer
for syscall programs at runtime.
Reported-by: Tejun Heo <tj@kernel.org>
Reported-by: Dan Schatzberg <dschatzberg@meta.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://lore.kernel.org/r/20260227005725.1247305-2-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The A64_MOV macro unconditionally uses ADD Rd, Rn, #0 to implement
register moves. While functionally correct, this is not the canonical
encoding when both operands are general-purpose registers.
On AArch64, MOV has two aliases depending on the operand registers:
- MOV <Xd|SP>, <Xn|SP> → ADD <Xd|SP>, <Xn|SP>, #0
- MOV <Xd>, <Xn> → ORR <Xd>, XZR, <Xn>
The ADD form is required when the stack pointer is involved (as ORR
does not accept SP), while the ORR form is the preferred encoding for
general-purpose registers.
The ORR encoding is also measurably faster on modern microarchitectures.
A microbenchmark [1] comparing dependent chains of MOV (ORR) vs ADD #0
on an ARM Neoverse-V2 (72-core, 3.4 GHz) shows:
=== mov (ORR Xd, XZR, Xn) ===
run1 cycles/op=0.749859456
run2 cycles/op=0.749991250
run3 cycles/op=0.749601847
avg cycles/op=0.749817518
=== add0 (ADD Xd, Xn, #0) ===
run1 cycles/op=1.004777689
run2 cycles/op=1.004558266
run3 cycles/op=1.004806559
avg cycles/op=1.004714171
The ORR form completes in ~0.75 cycles/op vs ~1.00 cycles/op for ADD #0,
a ~25% improvement. This is likely because the CPU's register renaming
hardware can eliminate ORR-based moves, while ADD #0 must go through the
ALU pipeline.
Update A64_MOV to select the appropriate encoding at JIT time:
use ADD when either register is A64_SP, and ORR (via
aarch64_insn_gen_move_reg()) otherwise.
Update verifier_private_stack selftests to expect "mov x7, x0" instead
of "add x7, x0, #0x0" in the JITed instruction checks, matching the
new ORR-based encoding.
[1] https://github.com/puranjaymohan/scripts/blob/main/arm64/bench/run_mov_vs_add0.sh
Signed-off-by: Puranjay Mohan <puranjay@kernel.org>
Acked-by: Xu Kuohai <xukuohai@huawei.com>
Link: https://lore.kernel.org/r/20260225134339.2723288-1-puranjay@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Jiri Olsa says:
====================
libbpf: Make optimized uprobes backward compatible
hi,
we can currently optimize uprobes on top of nop5 instructions,
so application can define USDT_NOP to nop5 and use USDT macro
to define optimized usdt probes.
This works fine on new kernels, but could have performance penalty
on older kernels, that do not have the support to optimize and to
emulate nop5 instruction.
This patchset adds support to workaround the performance penalty
on older kernels that do not support uprobe optimization, please
see detailed description in patch 2.
v1: https://lore.kernel.org/bpf/20251117083551.517393-1-jolsa@kernel.org/
v2: https://lore.kernel.org/bpf/20260210133649.524292-1-jolsa@kernel.org/
v3: https://lore.kernel.org/bpf/20260211084858.750950-1-jolsa@kernel.org/T/#t
v4: https://lore.kernel.org/bpf/20260220104220.634154-1-jolsa@kernel.org/
v5 changes:
- keep nop_combo on stack and levae buf uninitialized
in has_nop_combo function [David]
v4 changes:
- rebased on latest bpf-next/master
- use pread for nop combo read [Andrii]
- renamed usdt triger benchmark names [Andrii]
- added more ip address checks to tests [Andrii]
v3 changes:
- fix __x86_64 define and other typos [CI]
- add missing '?' to usdt trigger program [CI]
v2 changes:
- after more investigation we realized there are some versions of
bpftrace and stap that does not work with solution suggested in
version 1, so we decided to switch to following solution:
- change USDT macro [1] emits nop,nop5 instructions combo by
default
- libbpf detects nop,nop5 instructions combo for USDT probe,
if there is and if uprobe syscall is detected libbpf installs
usdt probe on top of nop5 instruction to get it optimized
- added usdt trigger benchmarks [Andrii]
- several small fixes on uprobe syscall detection, tests and other places [Andrii]
- true usdt.h source [1] updated [Andrii]
- compile usdt_* objects unconditionally [Andrii]
thanks,
jirka
[1] https://github.com/libbpf/usdt
---
====================
Link: https://patch.msgid.link/20260224103915.1369690-1-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Adding test that attaches bpf program on usdt probe in 2 scenarios;
- attach program on top of usdt_1, which is single nop instruction,
so the probe stays on nop instruction and is not optimized.
- attach program on top of usdt_2 which is probe defined on top
of nop,nop5 combo, so the probe is placed on top of nop5 and
is optimized.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20260224103915.1369690-5-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Adding support to detect nop,nop5 instructions combo for usdt probe
by checking on probe's following nop5 instruction.
When the nop,nop5 combo is detected together with uprobe syscall,
we can place the probe on top of nop5 and get it optimized.
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Link: https://lore.kernel.org/r/20260224103915.1369690-3-jolsa@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
cpu_map_bpf_prog_run_xdp() handles XDP_PASS, XDP_REDIRECT, and
XDP_DROP but is missing an XDP_ABORTED case. Without it, XDP_ABORTED
falls into the default case which logs a misleading "invalid XDP
action" warning instead of tracing the abort via trace_xdp_exception().
Add the missing XDP_ABORTED case with trace_xdp_exception(), matching
the handling already present in the skb path (cpu_map_bpf_prog_run_skb),
devmap (dev_map_bpf_prog_run), and the generic XDP path (do_xdp_generic).
Also pass xdpf->dev_rx instead of NULL to bpf_warn_invalid_xdp_action()
in the default case, so the warning includes the actual device name.
This aligns with the generic XDP path in net/core/dev.c which already
passes the real device.
Signed-off-by: Anand Kumar Shaw <anandkrshawheritage@gmail.com>
Link: https://lore.kernel.org/r/20260218042924.42931-1-anandkrshawheritage@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Ilya Leoshkevich says:
====================
bpf/s390: Implement get_preempt_count()
This series adds get_preempt_count() BPF implementation for s390.
On s390 preempt_count lives in lowcore: a per-cpu data structure
mapped at virtual address 0 or 0x70000 depending on the kernel
command line.
Patch 1 adds a kfunc to obtain lowcore address.
Patch 2 is the implementation that delegates to it.
====================
Link: https://patch.msgid.link/20260217160813.100855-1-iii@linux.ibm.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Pull bpf fixes from Alexei Starovoitov:
- Fix alignment of arm64 JIT buffer to prevent atomic tearing (Fuad
Tabba)
- Fix invariant violation for single value tnums in the verifier
(Harishankar Vishwanathan, Paul Chaignon)
- Fix a bunch of issues found by ASAN in selftests/bpf (Ihor Solodrai)
- Fix race in devmpa and cpumap on PREEMPT_RT (Jiayuan Chen)
- Fix show_fdinfo of kprobe_multi when cookies are not present (Jiri
Olsa)
- Fix race in freeing special fields in BPF maps to prevent memory
leaks (Kumar Kartikeya Dwivedi)
- Fix OOB read in dmabuf_collector (T.J. Mercier)
* tag 'bpf-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf: (36 commits)
selftests/bpf: Avoid simplification of crafted bounds test
selftests/bpf: Test refinement of single-value tnum
bpf: Improve bounds when tnum has a single possible value
bpf: Introduce tnum_step to step through tnum's members
bpf: Fix race in devmap on PREEMPT_RT
bpf: Fix race in cpumap on PREEMPT_RT
selftests/bpf: Add tests for special fields races
bpf: Retire rcu_trace_implies_rcu_gp() from local storage
bpf: Delay freeing fields in local storage
bpf: Lose const-ness of map in map_check_btf()
bpf: Register dtor for freeing special fields
selftests/bpf: Fix OOB read in dmabuf_collector
selftests/bpf: Fix a memory leak in xdp_flowtable test
bpf: Fix stack-out-of-bounds write in devmap
bpf: Fix kprobe_multi cookies access in show_fdinfo callback
bpf, arm64: Force 8-byte alignment for JIT buffer to prevent atomic tearing
selftests/bpf: Don't override SIGSEGV handler with ASAN
selftests/bpf: Check BPFTOOL env var in detect_bpftool_path()
selftests/bpf: Fix out-of-bounds array access bugs reported by ASAN
selftests/bpf: Fix array bounds warning in jit_disasm_helpers
...
Pull driver core fixes from Danilo Krummrich:
- Do not register imx_clk_scu_driver in imx8qxp_clk_probe(); besides
fixing two other issues, this avoids a deadlock in combination with
commit dc23806a7c ("driver core: enforce device_lock for
driver_match_device()")
- Move secondary node lookup from device_get_next_child_node() to
fwnode_get_next_child_node(); this avoids issues when users switch
from the device API to the fwnode API
- Export io_define_{read,write}!() to avoid unused import warnings when
CONFIG_PCI=n
* tag 'driver-core-7.0-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/driver-core/driver-core:
clk: scu/imx8qxp: do not register driver in probe()
rust: io: macro_export io_define_read!() and io_define_write!()
device property: Allow secondary lookup in fwnode_get_next_child_node()
Pull smb client fixes from Steve French:
- Two multichannel fixes
- Locking fix for superblock flags
- Fix to remove debug message that could log password
- Cleanup fix for setting credentials
* tag 'v7.0rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6:
smb: client: Use snprintf in cifs_set_cifscreds
smb: client: Don't log plaintext credentials in cifs_set_cifscreds
smb: client: fix broken multichannel with krb5+signing
smb: client: use atomic_t for mnt_cifs_flags
smb: client: fix cifs_pick_channel when channels are equally loaded
Pull spi fixes from Mark Brown:
"One fix for the stm32 driver which got broken for DMA chaining cases,
plus a removal of some straggling bindings for the Bikal SoC which has
been pulled out of the kernel"
* tag 'spi-fix-v7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: stm32: fix missing pointer assignment in case of dma chaining
spi: dt-bindings: snps,dw-abp-ssi: Remove unused bindings
Pull regulator fixes from Mark Brown:
"A small pile of fixes, none of which are super major - the code fixes
are improved error handling and fixing a leak of a device node.
We also have a typo fix and an improvement to make the binding example
for mt6359 more directly usable"
* tag 'regulator-fix-v7.0-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
regulator: Kconfig: fix a typo
regulator: bq257xx: Fix device node reference leak in bq257xx_reg_dt_parse_gpio()
regulator: fp9931: Fix PM runtime reference leak in fp9931_hwmon_read()
regulator: tps65185: check devm_kzalloc() result in probe
regulator: dt-bindings: mt6359: make regulator names unique
Pull s390 fixes from Vasily Gorbik:
- Fix guest pfault init to pass a physical address to DIAG 0x258,
restoring pfault interrupts and avoiding vCPU stalls during host
page-in
- Fix kexec/kdump hangs with stack protector by marking
s390_reset_system() __no_stack_protector; set_prefix(0) switches
lowcore and the canary no longer matches
- Fix idle/vtime cputime accounting (idle-exit ordering, vtimer
double-forwarding) and small cleanups
* tag 's390-7.0-3' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
s390/pfault: Fix virtual vs physical address confusion
s390/kexec: Disable stack protector in s390_reset_system()
s390/idle: Remove psw_idle() prototype
s390/vtime: Use lockdep_assert_irqs_disabled() instead of BUG_ON()
s390/vtime: Use __this_cpu_read() / get rid of READ_ONCE()
s390/irq/idle: Remove psw bits early
s390/idle: Inline update_timer_idle()
s390/idle: Slightly optimize idle time accounting
s390/idle: Add comment for non obvious code
s390/vtime: Fix virtual timer forwarding
s390/idle: Fix cpu idle exit cpu time accounting
Paul Chaignon says:
====================
Fix invariant violation for single-value tnums
We're hitting an invariant violation in Cilium that sometimes leads to
BPF programs being rejected and Cilium failing to start [1]. As far as
I know this is the first case of invariant violation found in a real
program (i.e., not by a fuzzer). The following extract from verifier
logs shows what's happening:
from 201 to 236: R1=0 R6=ctx() R7=1 R9=scalar(smin=umin=smin32=umin32=3584,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100)) R10=fp0
236: R1=0 R6=ctx() R7=1 R9=scalar(smin=umin=smin32=umin32=3584,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100)) R10=fp0
; if (magic == MARK_MAGIC_HOST || magic == MARK_MAGIC_OVERLAY || magic == MARK_MAGIC_ENCRYPT) @ bpf_host.c:1337
236: (16) if w9 == 0xe00 goto pc+45 ; R9=scalar(smin=umin=smin32=umin32=3585,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100))
237: (16) if w9 == 0xf00 goto pc+1
verifier bug: REG INVARIANTS VIOLATION (false_reg1): range bounds violation u64=[0xe01, 0xe00] s64=[0xe01, 0xe00] u32=[0xe01, 0xe00] s32=[0xe01, 0xe00] var_off=(0xe00, 0x0)
More details are given in the second patch, but in short, the verifier
should be able to detect that the false branch of instruction 237 is
never true. After instruction 236, the u64 range and the tnum overlap
in a single value, 0xf00.
The long-term solution to invariant violation is likely to rely on the
refinement + invariant violation check to detect dead branches, as
started by Eduard. To fix the current issue, we need something with
less refactoring that we can backport to affected kernels.
The solution implemented in the second patch is to improve the bounds
refinement to avoid this case. It relies on a new tnum helper,
tnum_step, first sent as an RFC in [2]. The last two patches extend and
update the selftests.
Link: https://github.com/cilium/cilium/issues/44216 [1]
Link: https://lore.kernel.org/bpf/20251107192328.2190680-2-harishankar.vishwanathan@gmail.com/ [2]
Changes in v3:
- Fix commit description error spotted by AI bot.
- Simplify constants in first two tests (Eduard).
- Rework comment on third test (Eduard).
- Add two new negative test cases (Eduard).
- Rebased.
Changes in v2:
- Add guard suggested by Hari in tnum_step, to avoid undefined
behavior spotted by AI code review.
- Add explanation diagrams in code as suggested by Eduard.
- Rework conditions for readability as suggested by Eduard.
- Updated reference to SMT formula.
- Rebased.
====================
Link: https://patch.msgid.link/cover.1772225741.git.paul.chaignon@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
The reg_bounds_crafted tests validate the verifier's range analysis
logic. They focus on the actual ranges and thus ignore the tnum. As a
consequence, they carry the assumption that the tested cases can be
reproduced in userspace without using the tnum information.
Unfortunately, the previous change the refinement logic breaks that
assumption for one test case:
(u64)2147483648 (u32)<op> [4294967294; 0x100000000]
The tested bytecode is shown below. Without our previous improvement, on
the false branch of the condition, R7 is only known to have u64 range
[0xfffffffe; 0x100000000]. With our improvement, and using the tnum
information, we can deduce that R7 equals 0x100000000.
19: (bc) w0 = w6 ; R6=0x80000000
20: (bc) w0 = w7 ; R7=scalar(smin=umin=0xfffffffe,smax=umax=0x100000000,smin32=-2,smax32=0,var_off=(0x0; 0x1ffffffff))
21: (be) if w6 <= w7 goto pc+3 ; R6=0x80000000 R7=0x100000000
R7's tnum is (0; 0x1ffffffff). On the false branch, regs_refine_cond_op
refines R7's u32 range to [0; 0x7fffffff]. Then, __reg32_deduce_bounds
refines the s32 range to 0 using u32 and finally also sets u32=0.
From this, __reg_bound_offset improves the tnum to (0; 0x100000000).
Finally, our previous patch uses this new tnum to deduce that it only
intersect with u64=[0xfffffffe; 0x100000000] in a single value:
0x100000000.
Because the verifier uses the tnum to reach this constant value, the
selftest is unable to reproduce it by only simulating ranges. The
solution implemented in this patch is to change the test case such that
there is more than one overlap value between u64 and the tnum. The max.
u64 value is thus changed from 0x100000000 to 0x300000000.
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Paul Chaignon <paul.chaignon@gmail.com>
Link: https://lore.kernel.org/r/50641c6a7ef39520595dcafa605692427c1006ec.1772225741.git.paul.chaignon@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This patch introduces selftests to cover the new bounds refinement
logic introduced in the previous patch. Without the previous patch,
the first two tests fail because of the invariant violation they
trigger. The last test fails because the R10 access is not detected as
dead code. In addition, all three tests fail because of R0 having a
non-constant value in the verifier logs.
In addition, the last two cases are covering the negative cases: when we
shouldn't refine the bounds because the u64 and tnum overlap in at least
two values.
Signed-off-by: Paul Chaignon <paul.chaignon@gmail.com>
Link: https://lore.kernel.org/r/90d880c8cf587b9f7dc715d8961cd1b8111d01a8.1772225741.git.paul.chaignon@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
We're hitting an invariant violation in Cilium that sometimes leads to
BPF programs being rejected and Cilium failing to start [1]. The
following extract from verifier logs shows what's happening:
from 201 to 236: R1=0 R6=ctx() R7=1 R9=scalar(smin=umin=smin32=umin32=3584,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100)) R10=fp0
236: R1=0 R6=ctx() R7=1 R9=scalar(smin=umin=smin32=umin32=3584,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100)) R10=fp0
; if (magic == MARK_MAGIC_HOST || magic == MARK_MAGIC_OVERLAY || magic == MARK_MAGIC_ENCRYPT) @ bpf_host.c:1337
236: (16) if w9 == 0xe00 goto pc+45 ; R9=scalar(smin=umin=smin32=umin32=3585,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100))
237: (16) if w9 == 0xf00 goto pc+1
verifier bug: REG INVARIANTS VIOLATION (false_reg1): range bounds violation u64=[0xe01, 0xe00] s64=[0xe01, 0xe00] u32=[0xe01, 0xe00] s32=[0xe01, 0xe00] var_off=(0xe00, 0x0)
We reach instruction 236 with two possible values for R9, 0xe00 and
0xf00. This is perfectly reflected in the tnum, but of course the ranges
are less accurate and cover [0xe00; 0xf00]. Taking the fallthrough path
at instruction 236 allows the verifier to reduce the range to
[0xe01; 0xf00]. The tnum is however not updated.
With these ranges, at instruction 237, the verifier is not able to
deduce that R9 is always equal to 0xf00. Hence the fallthrough pass is
explored first, the verifier refines the bounds using the assumption
that R9 != 0xf00, and ends up with an invariant violation.
This pattern of impossible branch + bounds refinement is common to all
invariant violations seen so far. The long-term solution is likely to
rely on the refinement + invariant violation check to detect dead
branches, as started by Eduard. To fix the current issue, we need
something with less refactoring that we can backport.
This patch uses the tnum_step helper introduced in the previous patch to
detect the above situation. In particular, three cases are now detected
in the bounds refinement:
1. The u64 range and the tnum only overlap in umin.
u64: ---[xxxxxx]-----
tnum: --xx----------x-
2. The u64 range and the tnum only overlap in the maximum value
represented by the tnum, called tmax.
u64: ---[xxxxxx]-----
tnum: xx-----x--------
3. The u64 range and the tnum only overlap in between umin (excluded)
and umax.
u64: ---[xxxxxx]-----
tnum: xx----x-------x-
To detect these three cases, we call tnum_step(tnum, umin), which
returns the smallest member of the tnum greater than umin, called
tnum_next here. We're in case (1) if umin is part of the tnum and
tnum_next is greater than umax. We're in case (2) if umin is not part of
the tnum and tnum_next is equal to tmax. Finally, we're in case (3) if
umin is not part of the tnum, tnum_next is inferior or equal to umax,
and calling tnum_step a second time gives us a value past umax.
This change implements these three cases. With it, the above bytecode
looks as follows:
0: (85) call bpf_get_prandom_u32#7 ; R0=scalar()
1: (47) r0 |= 3584 ; R0=scalar(smin=0x8000000000000e00,umin=umin32=3584,smin32=0x80000e00,var_off=(0xe00; 0xfffffffffffff1ff))
2: (57) r0 &= 3840 ; R0=scalar(smin=umin=smin32=umin32=3584,smax=umax=smax32=umax32=3840,var_off=(0xe00; 0x100))
3: (15) if r0 == 0xe00 goto pc+2 ; R0=3840
4: (15) if r0 == 0xf00 goto pc+1
4: R0=3840
6: (95) exit
In addition to the new selftests, this change was also verified with
Agni [3]. For the record, the raw SMT is available at [4]. The property
it verifies is that: If a concrete value x is contained in all input
abstract values, after __update_reg_bounds, it will continue to be
contained in all output abstract values.
Link: https://github.com/cilium/cilium/issues/44216 [1]
Link: https://pchaigno.github.io/test-verifier-complexity.html [2]
Link: https://github.com/bpfverif/agni [3]
Link: https://pastebin.com/raw/naCfaqNx [4]
Fixes: 0df1a55afa ("bpf: Warn on internal verifier errors")
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Tested-by: Marco Schirrmeister <mschirrmeister@gmail.com>
Co-developed-by: Harishankar Vishwanathan <harishankar.vishwanathan@gmail.com>
Signed-off-by: Harishankar Vishwanathan <harishankar.vishwanathan@gmail.com>
Signed-off-by: Paul Chaignon <paul.chaignon@gmail.com>
Link: https://lore.kernel.org/r/ef254c4f68be19bd393d450188946821c588565d.1772225741.git.paul.chaignon@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
This commit introduces tnum_step(), a function that, when given t, and a
number z returns the smallest member of t larger than z. The number z
must be greater or equal to the smallest member of t and less than the
largest member of t.
The first step is to compute j, a number that keeps all of t's known
bits, and matches all unknown bits to z's bits. Since j is a member of
the t, it is already a candidate for result. However, we want our result
to be (minimally) greater than z.
There are only two possible cases:
(1) Case j <= z. In this case, we want to increase the value of j and
make it > z.
(2) Case j > z. In this case, we want to decrease the value of j while
keeping it > z.
(Case 1) j <= z
t = xx11x0x0
z = 10111101 (189)
j = 10111000 (184)
^
k
(Case 1.1) Let's first consider the case where j < z. We will address j
== z later.
Since z > j, there had to be a bit position that was 1 in z and a 0 in
j, beyond which all positions of higher significance are equal in j and
z. Further, this position could not have been unknown in a, because the
unknown positions of a match z. This position had to be a 1 in z and
known 0 in t.
Let k be position of the most significant 1-to-0 flip. In our example, k
= 3 (starting the count at 1 at the least significant bit). Setting (to
1) the unknown bits of t in positions of significance smaller than
k will not produce a result > z. Hence, we must set/unset the unknown
bits at positions of significance higher than k. Specifically, we look
for the next larger combination of 1s and 0s to place in those
positions, relative to the combination that exists in z. We can achieve
this by concatenating bits at unknown positions of t into an integer,
adding 1, and writing the bits of that result back into the
corresponding bit positions previously extracted from z.
>From our example, considering only positions of significance greater
than k:
t = xx..x
z = 10..1
+ 1
-----
11..0
This is the exact combination 1s and 0s we need at the unknown bits of t
in positions of significance greater than k. Further, our result must
only increase the value minimally above z. Hence, unknown bits in
positions of significance smaller than k should remain 0. We finally
have,
result = 11110000 (240)
(Case 1.2) Now consider the case when j = z, for example
t = 1x1x0xxx
z = 10110100 (180)
j = 10110100 (180)
Matching the unknown bits of the t to the bits of z yielded exactly z.
To produce a number greater than z, we must set/unset the unknown bits
in t, and *all* the unknown bits of t candidates for being set/unset. We
can do this similar to Case 1.1, by adding 1 to the bits extracted from
the masked bit positions of z. Essentially, this case is equivalent to
Case 1.1, with k = 0.
t = 1x1x0xxx
z = .0.1.100
+ 1
---------
.0.1.101
This is the exact combination of bits needed in the unknown positions of
t. After recalling the known positions of t, we get
result = 10110101 (181)
(Case 2) j > z
t = x00010x1
z = 10000010 (130)
j = 10001011 (139)
^
k
Since j > z, there had to be a bit position which was 0 in z, and a 1 in
j, beyond which all positions of higher significance are equal in j and
z. This position had to be a 0 in z and known 1 in t. Let k be the
position of the most significant 0-to-1 flip. In our example, k = 4.
Because of the 0-to-1 flip at position k, a member of t can become
greater than z if the bits in positions greater than k are themselves >=
to z. To make that member *minimally* greater than z, the bits in
positions greater than k must be exactly = z. Hence, we simply match all
of t's unknown bits in positions more significant than k to z's bits. In
positions less significant than k, we set all t's unknown bits to 0
to retain minimality.
In our example, in positions of greater significance than k (=4),
t=x000. These positions are matched with z (1000) to produce 1000. In
positions of lower significance than k, t=10x1. All unknown bits are set
to 0 to produce 1001. The final result is:
result = 10001001 (137)
This concludes the computation for a result > z that is a member of t.
The procedure for tnum_step() in this commit implements the idea
described above. As a proof of correctness, we verified the algorithm
against a logical specification of tnum_step. The specification asserts
the following about the inputs t, z and output res that:
1. res is a member of t, and
2. res is strictly greater than z, and
3. there does not exist another value res2 such that
3a. res2 is also a member of t, and
3b. res2 is greater than z
3c. res2 is smaller than res
We checked the implementation against this logical specification using
an SMT solver. The verification formula in SMTLIB format is available
at [1]. The verification returned an "unsat": indicating that no input
assignment exists for which the implementation and the specification
produce different outputs.
In addition, we also automatically generated the logical encoding of the
C implementation using Agni [2] and verified it against the same
specification. This verification also returned an "unsat", confirming
that the implementation is equivalent to the specification. The formula
for this check is also available at [3].
Link: https://pastebin.com/raw/2eRWbiit [1]
Link: https://github.com/bpfverif/agni [2]
Link: https://pastebin.com/raw/EztVbBJ2 [3]
Co-developed-by: Srinivas Narayana <srinivas.narayana@rutgers.edu>
Signed-off-by: Srinivas Narayana <srinivas.narayana@rutgers.edu>
Co-developed-by: Santosh Nagarakatte <santosh.nagarakatte@rutgers.edu>
Signed-off-by: Santosh Nagarakatte <santosh.nagarakatte@rutgers.edu>
Signed-off-by: Harishankar Vishwanathan <harishankar.vishwanathan@gmail.com>
Link: https://lore.kernel.org/r/93fdf71910411c0f19e282ba6d03b4c65f9c5d73.1772225741.git.paul.chaignon@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Jiayuan Chen says:
====================
bpf: Fix per-CPU bulk queue races on PREEMPT_RT
On PREEMPT_RT kernels, local_bh_disable() only calls migrate_disable()
(when PREEMPT_RT_NEEDS_BH_LOCK is not set) and does not disable
preemption. This means CFS scheduling can preempt a task inside the
per-CPU bulk queue (bq) operations in cpumap and devmap, allowing
another task on the same CPU to concurrently access the same bq,
leading to use-after-free, list corruption, and kernel panics.
Patch 1 fixes the cpumap race in bq_flush_to_queue(), originally
reported by syzbot [1].
Patch 2 fixes the same class of race in devmap's bq_xmit_all(),
identified by code inspection after Sebastian Andrzej Siewior pointed
out that devmap has the same per-CPU bulk queue pattern [2].
Both patches use local_lock_nested_bh() to serialize access to the
per-CPU bq. On non-RT this is a pure lockdep annotation with no
overhead; on PREEMPT_RT it provides a per-CPU sleeping lock.
To reproduce the devmap race, insert an mdelay(100) in bq_xmit_all()
after "cnt = bq->count" and before the actual transmit loop. Then pin
two threads to the same CPU, each running BPF_PROG_TEST_RUN with an XDP
program that redirects to a DEVMAP entry (e.g. a veth pair). CFS
timeslicing during the mdelay window causes interleaving. Without the
fix, KASAN reports null-ptr-deref due to operating on freed frames:
BUG: KASAN: null-ptr-deref in __build_skb_around+0x22d/0x340
Write of size 32 at addr 0000000000000d50 by task devmap_race_rep/449
CPU: 0 UID: 0 PID: 449 Comm: devmap_race_rep Not tainted 6.19.0+ #31 PREEMPT_RT
Call Trace:
<TASK>
__build_skb_around+0x22d/0x340
build_skb_around+0x25/0x260
__xdp_build_skb_from_frame+0x103/0x860
veth_xdp_rcv_bulk_skb.isra.0+0x162/0x320
veth_xdp_rcv.constprop.0+0x61e/0xbb0
veth_poll+0x280/0xb50
__napi_poll.constprop.0+0xa5/0x590
net_rx_action+0x4b0/0xea0
handle_softirqs.isra.0+0x1b3/0x780
__local_bh_enable_ip+0x12a/0x240
xdp_test_run_batch.constprop.0+0xedd/0x1f60
bpf_test_run_xdp_live+0x304/0x640
bpf_prog_test_run_xdp+0xd24/0x1b70
__sys_bpf+0x61c/0x3e00
</TASK>
Kernel panic - not syncing: Fatal exception in interrupt
[1] https://lore.kernel.org/all/69369331.a70a0220.38f243.009d.GAE@google.com/T/
[2] https://lore.kernel.org/bpf/20260212023634.366343-1-jiayuan.chen@linux.dev/
v3 -> v4: https://lore.kernel.org/all/20260213034018.284146-1-jiayuan.chen@linux.dev/
- Move panic trace to cover letter. (Sebastian Andrzej Siewior)
- Add Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> to both patches
from cover letter.
v2 -> v3: https://lore.kernel.org/bpf/20260212023634.366343-1-jiayuan.chen@linux.dev/
- Fix commit message: remove incorrect "spin_lock() becomes rt_mutex"
claim, the per-CPU bq has no spin_lock at all. (Sebastian Andrzej Siewior)
- Fix commit message: accurately describe local_lock_nested_bh()
behavior instead of referencing local_lock(). (Sebastian Andrzej Siewior)
- Remove incomplete discussion of snapshot alternative.
(Sebastian Andrzej Siewior)
- Remove panic trace from commit message. (Sebastian Andrzej Siewior)
- Add patch 2/2 for devmap, same race pattern. (Sebastian Andrzej Siewior)
v1 -> v2: https://lore.kernel.org/bpf/20260211064417.196401-1-jiayuan.chen@linux.dev/
- Use local_lock_nested_bh()/local_unlock_nested_bh() instead of
local_lock()/local_unlock(), since these paths already run under
local_bh_disable(). (Sebastian Andrzej Siewior)
- Replace "Caller must hold bq->bq_lock" comment with
lockdep_assert_held() in bq_flush_to_queue(). (Sebastian Andrzej Siewior)
- Fix Fixes tag to 3253cb49cb ("softirq: Allow to drop the
softirq-BKL lock on PREEMPT_RT") which is the actual commit that
makes the race possible. (Sebastian Andrzej Siewior)
====================
Link: https://patch.msgid.link/20260225121459.183121-1-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
On PREEMPT_RT kernels, the per-CPU xdp_dev_bulk_queue (bq) can be
accessed concurrently by multiple preemptible tasks on the same CPU.
The original code assumes bq_enqueue() and __dev_flush() run atomically
with respect to each other on the same CPU, relying on
local_bh_disable() to prevent preemption. However, on PREEMPT_RT,
local_bh_disable() only calls migrate_disable() (when
PREEMPT_RT_NEEDS_BH_LOCK is not set) and does not disable
preemption, which allows CFS scheduling to preempt a task during
bq_xmit_all(), enabling another task on the same CPU to enter
bq_enqueue() and operate on the same per-CPU bq concurrently.
This leads to several races:
1. Double-free / use-after-free on bq->q[]: bq_xmit_all() snapshots
cnt = bq->count, then iterates bq->q[0..cnt-1] to transmit frames.
If preempted after the snapshot, a second task can call bq_enqueue()
-> bq_xmit_all() on the same bq, transmitting (and freeing) the
same frames. When the first task resumes, it operates on stale
pointers in bq->q[], causing use-after-free.
2. bq->count and bq->q[] corruption: concurrent bq_enqueue() modifying
bq->count and bq->q[] while bq_xmit_all() is reading them.
3. dev_rx/xdp_prog teardown race: __dev_flush() clears bq->dev_rx and
bq->xdp_prog after bq_xmit_all(). If preempted between
bq_xmit_all() return and bq->dev_rx = NULL, a preempting
bq_enqueue() sees dev_rx still set (non-NULL), skips adding bq to
the flush_list, and enqueues a frame. When __dev_flush() resumes,
it clears dev_rx and removes bq from the flush_list, orphaning the
newly enqueued frame.
4. __list_del_clearprev() on flush_node: similar to the cpumap race,
both tasks can call __list_del_clearprev() on the same flush_node,
the second dereferences the prev pointer already set to NULL.
The race between task A (__dev_flush -> bq_xmit_all) and task B
(bq_enqueue -> bq_xmit_all) on the same CPU:
Task A (xdp_do_flush) Task B (ndo_xdp_xmit redirect)
---------------------- --------------------------------
__dev_flush(flush_list)
bq_xmit_all(bq)
cnt = bq->count /* e.g. 16 */
/* start iterating bq->q[] */
<-- CFS preempts Task A -->
bq_enqueue(dev, xdpf)
bq->count == DEV_MAP_BULK_SIZE
bq_xmit_all(bq, 0)
cnt = bq->count /* same 16! */
ndo_xdp_xmit(bq->q[])
/* frames freed by driver */
bq->count = 0
<-- Task A resumes -->
ndo_xdp_xmit(bq->q[])
/* use-after-free: frames already freed! */
Fix this by adding a local_lock_t to xdp_dev_bulk_queue and acquiring
it in bq_enqueue() and __dev_flush(). These paths already run under
local_bh_disable(), so use local_lock_nested_bh() which on non-RT is
a pure annotation with no overhead, and on PREEMPT_RT provides a
per-CPU sleeping lock that serializes access to the bq.
Fixes: 3253cb49cb ("softirq: Allow to drop the softirq-BKL lock on PREEMPT_RT")
Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jiayuan Chen <jiayuan.chen@shopee.com>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Link: https://lore.kernel.org/r/20260225121459.183121-3-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
On PREEMPT_RT kernels, the per-CPU xdp_bulk_queue (bq) can be accessed
concurrently by multiple preemptible tasks on the same CPU.
The original code assumes bq_enqueue() and __cpu_map_flush() run
atomically with respect to each other on the same CPU, relying on
local_bh_disable() to prevent preemption. However, on PREEMPT_RT,
local_bh_disable() only calls migrate_disable() (when
PREEMPT_RT_NEEDS_BH_LOCK is not set) and does not disable
preemption, which allows CFS scheduling to preempt a task during
bq_flush_to_queue(), enabling another task on the same CPU to enter
bq_enqueue() and operate on the same per-CPU bq concurrently.
This leads to several races:
1. Double __list_del_clearprev(): after bq->count is reset in
bq_flush_to_queue(), a preempting task can call bq_enqueue() ->
bq_flush_to_queue() on the same bq when bq->count reaches
CPU_MAP_BULK_SIZE. Both tasks then call __list_del_clearprev()
on the same bq->flush_node, the second call dereferences the
prev pointer that was already set to NULL by the first.
2. bq->count and bq->q[] races: concurrent bq_enqueue() can corrupt
the packet queue while bq_flush_to_queue() is processing it.
The race between task A (__cpu_map_flush -> bq_flush_to_queue) and
task B (bq_enqueue -> bq_flush_to_queue) on the same CPU:
Task A (xdp_do_flush) Task B (cpu_map_enqueue)
---------------------- ------------------------
bq_flush_to_queue(bq)
spin_lock(&q->producer_lock)
/* flush bq->q[] to ptr_ring */
bq->count = 0
spin_unlock(&q->producer_lock)
bq_enqueue(rcpu, xdpf)
<-- CFS preempts Task A --> bq->q[bq->count++] = xdpf
/* ... more enqueues until full ... */
bq_flush_to_queue(bq)
spin_lock(&q->producer_lock)
/* flush to ptr_ring */
spin_unlock(&q->producer_lock)
__list_del_clearprev(flush_node)
/* sets flush_node.prev = NULL */
<-- Task A resumes -->
__list_del_clearprev(flush_node)
flush_node.prev->next = ...
/* prev is NULL -> kernel oops */
Fix this by adding a local_lock_t to xdp_bulk_queue and acquiring it
in bq_enqueue() and __cpu_map_flush(). These paths already run under
local_bh_disable(), so use local_lock_nested_bh() which on non-RT is
a pure annotation with no overhead, and on PREEMPT_RT provides a
per-CPU sleeping lock that serializes access to the bq.
To reproduce, insert an mdelay(100) between bq->count = 0 and
__list_del_clearprev() in bq_flush_to_queue(), then run reproducer
provided by syzkaller.
Fixes: 3253cb49cb ("softirq: Allow to drop the softirq-BKL lock on PREEMPT_RT")
Reported-by: syzbot+2b3391f44313b3983e91@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/69369331.a70a0220.38f243.009d.GAE@google.com/T/
Reviewed-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jiayuan Chen <jiayuan.chen@shopee.com>
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Link: https://lore.kernel.org/r/20260225121459.183121-2-jiayuan.chen@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Kumar Kartikeya Dwivedi says:
====================
Close race in freeing special fields and map value
There exists a race across various map types where the freeing of
special fields (tw, timer, wq, kptr, etc.) can be done eagerly when a
logical delete operation is done on a map value, such that the program
which continues to have access to such a map value can recreate the
fields and cause them to leak.
The set contains fixes for this case. It is a continuation of Mykyta's
previous attempt in [0], but applies to all fields. A test is included
which reproduces the bug reliably in absence of the fixes.
Local Storage Benchmarks
------------------------
Evaluation Setup: Benchmarked on a dual-socket Intel Xeon Gold 6348 (Ice
Lake) @ 2.60GHz (56 cores / 112 threads), with the CPU governor set to
performance. Bench was pinned to a single NUMA node throughout the test.
Benchmark comes from [1] using the following command:
./bench -p 1 local-storage-create --storage-type <socket,task> --batch-size <16,32,64>
Before the test, 10 runs of all cases ([socket|task] x 3 batch sizes x 7
iterations per batch size) are done to warm up and prime the machine.
Then, 3 runs of all cases are done (with and without the patch, across
reboots).
For each comparison, we have 21 samples, i.e. per batch size (e.g.
socket 16) of a given local storage, we have 3 runs x 7 iterations.
The statistics (mean, median, stddev) and t-test is done for each
scenario (local storage and batch size pair) individually (21 samples
for either case). All values are for local storage creations in thousand
creations / sec (k/s).
Baseline (without patch) With patch Delta
Case Median Mean Std. Dev. Median Mean Std. Dev. Median %
---------------------------------------------------------------------------------------------------
socket 16 432.026 431.941 1.047 431.347 431.953 1.635 -0.679 -0.16%
socket 32 432.641 432.818 1.535 432.488 432.302 1.508 -0.153 -0.04%
socket 64 431.504 431.996 1.337 429.145 430.326 2.469 -2.359 -0.55%
task 16 38.816 39.382 1.456 39.657 39.337 1.831 +0.841 +2.17%
task 32 38.815 39.644 2.690 38.721 39.122 1.636 -0.094 -0.24%
task 64 37.562 38.080 1.701 39.554 38.563 1.689 +1.992 +5.30%
The cases for socket are within the range of noise, and improvements in task
local storage are due to high variance (CV ~4%-6% across batch sizes). The only
statistically significant case worth mentioning is socket with batch size 64
with p-value from t-test < 0.05, but the absolute difference is small (~2k/s).
TL;DR there doesn't appear to be any significant regression or improvement.
[0]: https://lore.kernel.org/bpf/20260216131341.1285427-1-mykyta.yatsenko5@gmail.com
[1]: https://lore.kernel.org/bpf/20260205222916.1788211-1-ameryhung@gmail.com
Changelog:
----------
v2 -> v3
v2: https://lore.kernel.org/bpf/20260227052031.3988575-1-memxor@gmail.com
* Add syzbot Tested-by.
* Add Amery's Reviewed-by.
* Fix missing rcu_dereference_check() in __bpf_selem_free_rcu. (BPF CI Bot)
* Remove migrate_disable() in bpf_selem_free_rcu. (Alexei)
v1 -> v2
v1: https://lore.kernel.org/bpf/20260225185121.2057388-1-memxor@gmail.com
* Add Paul's Reviewed-by.
* Fix use-after-free in accessing bpf_mem_alloc embedded in map. (syzbot CI)
* Add benchmark numbers for local storage.
* Add extra test case for per-cpu hashmap coverage with up to 16 refcount leaks.
* Target bpf tree.
====================
Link: https://patch.msgid.link/20260227224806.646888-1-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add a couple of tests to ensure that the refcount drops to zero when we
exercise the race where creation of a special field succeeds the logical
bpf_obj_free_fields done when deleting an element. Prior to previous
changes, the fields would be freed eagerly and repopulate and end up
leaking, causing the reference to not drop down correctly. Running this
test on a kernel without fixes will cause a hang in delete_module, since
the module reference stays active due to the leaked kptr not dropping
it. After the fixes tests succeed as expected.
Reviewed-by: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260227224806.646888-6-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Currently, when use_kmalloc_nolock is false, the freeing of fields for a
local storage selem is done eagerly before waiting for the RCU or RCU
tasks trace grace period to elapse. This opens up a window where the
program which has access to the selem can recreate the fields after the
freeing of fields is done eagerly, causing memory leaks when the element
is finally freed and returned to the kernel.
Make a few changes to address this. First, delay the freeing of fields
until after the grace periods have expired using a __bpf_selem_free_rcu
wrapper which is eventually invoked after transitioning through the
necessary number of grace period waits. Replace usage of the kfree_rcu
with call_rcu to be able to take a custom callback. Finally, care needs
to be taken to extend the rcu barriers for all cases, and not just when
use_kmalloc_nolock is true, as RCU and RCU tasks trace callbacks can be
in flight for either case and access the smap field, which is used to
obtain the BTF record to walk over special fields in the map value.
While we're at it, drop migrate_disable() from bpf_selem_free_rcu, since
migration should be disabled for RCU callbacks already.
Fixes: 9bac675e63 ("bpf: Postpone bpf_obj_free_fields to the rcu callback")
Reviewed-by: Amery Hung <ameryhung@gmail.com>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260227224806.646888-4-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
BPF hash map may now use the map_check_btf() callback to decide whether
to set a dtor on its bpf_mem_alloc or not. Unlike C++ where members can
opt out of const-ness using mutable, we must lose the const qualifier on
the callback such that we can avoid the ugly cast. Make the change and
adjust all existing users, and lose the comment in hashtab.c.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260227224806.646888-3-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
There is a race window where BPF hash map elements can leak special
fields if the program with access to the map value recreates these
special fields between the check_and_free_fields done on the map value
and its eventual return to the memory allocator.
Several ways were explored prior to this patch, most notably [0] tried
to use a poison value to reject attempts to recreate special fields for
map values that have been logically deleted but still accessible to BPF
programs (either while sitting in the free list or when reused). While
this approach works well for task work, timers, wq, etc., it is harder
to apply the idea to kptrs, which have a similar race and failure mode.
Instead, we change bpf_mem_alloc to allow registering destructor for
allocated elements, such that when they are returned to the allocator,
any special fields created while they were accessible to programs in the
mean time will be freed. If these values get reused, we do not free the
fields again before handing the element back. The special fields thus
may remain initialized while the map value sits in a free list.
When bpf_mem_alloc is retired in the future, a similar concept can be
introduced to kmalloc_nolock-backed kmem_cache, paired with the existing
idea of a constructor.
Note that the destructor registration happens in map_check_btf, after
the BTF record is populated and (at that point) avaiable for inspection
and duplication. Duplication is necessary since the freeing of embedded
bpf_mem_alloc can be decoupled from actual map lifetime due to logic
introduced to reduce the cost of rcu_barrier()s in mem alloc free path in
9f2c6e96c6 ("bpf: Optimize rcu_barrier usage between hash map and bpf_mem_alloc.").
As such, once all callbacks are done, we must also free the duplicated
record. To remove dependency on the bpf_map itself, also stash the key
size of the map to obtain value from htab_elem long after the map is
gone.
[0]: https://lore.kernel.org/bpf/20260216131341.1285427-1-mykyta.yatsenko5@gmail.com
Fixes: 14a324f6a6 ("bpf: Wire up freeing of referenced kptr")
Fixes: 1bfbc267ec ("bpf: Enable bpf_timer and bpf_wq in any context")
Reported-by: Alexei Starovoitov <ast@kernel.org>
Tested-by: syzbot@syzkaller.appspotmail.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://lore.kernel.org/r/20260227224806.646888-2-memxor@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Pull arm64 fixes from Will Deacon:
"The diffstat is dominated by changes to our TLB invalidation errata
handling and the introduction of a new GCS selftest to catch one of
the issues that is fixed here relating to PROT_NONE mappings.
- Fix cpufreq warning due to attempting a cross-call with interrupts
masked when reading local AMU counters
- Fix DEBUG_PREEMPT warning from the delay loop when it tries to
access per-cpu errata workaround state for the virtual counter
- Re-jig and optimise our TLB invalidation errata workarounds in
preparation for more hardware brokenness
- Fix GCS mappings to interact properly with PROT_NONE and to avoid
corrupting the pte on CPUs with FEAT_LPA2
- Fix ioremap_prot() to extract only the memory attributes from the
user pte and ignore all the other 'prot' bits"
* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
arm64: topology: Fix false warning in counters_read_on_cpu() for same-CPU reads
arm64: Fix sampling the "stable" virtual counter in preemptible section
arm64: tlb: Optimize ARM64_WORKAROUND_REPEAT_TLBI
arm64: tlb: Allow XZR argument to TLBI ops
kselftest: arm64: Check access to GCS after mprotect(PROT_NONE)
arm64: gcs: Honour mprotect(PROT_NONE) on shadow stack mappings
arm64: gcs: Do not set PTE_SHARED on GCS mappings if FEAT_LPA2 is enabled
arm64: io: Extract user memory type in ioremap_prot()
arm64: io: Rename ioremap_prot() to __ioremap_prot()
