Remove bpf_asm, bpf_dbg and bpf_jit_disasm, along with the assembler's
lexer and parser, as suggested by Daniel. These tools are old classic BPF
and these days there is no need to keep them in kernel anymore.
Drop their build and install rules, feature checks and ignore entries.
Remove the tool-specific documentation and update the PTP classifier and
packet socket test comments. Keep the classic BPF instruction descriptions
and assembly examples.
Keep the bpftool and resolve_btfids build and clean targets, and the
bpftool install target.
Suggested-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Xia Yan <yanxia@cmss.chinamobile.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/61f7315e-c835-425e-a99b-d2470ba9a875@iogearbox.net
Link: https://lore.kernel.org/bpf/20261020095124.928021-1-yanxia@cmss.chinamobile.com
"bpftool map dump" walks the map with bpf_map_get_next_key() and treats
ENOENT as the end of the map. Map types that can't be iterated fail that
call on the very first key: local storage and ringbuf with ENOTSUPP,
bloom filter and arena with EOPNOTSUPP, queue and stack with EINVAL.
bpftool then ends the walk without a word and prints what it prints for
an empty map, "[]" with BTF or "Found 0 elements" without, and only the
exit status tells that the map was not read. On a socket storage map
that holds storage for a socket:
# bpftool map dump pinned /sys/fs/bpf/sk_storage
[]
Ask for the first key before printing anything, and when that fails with
anything but ENOENT, fail with the map type and the error instead of a
dump:
# bpftool map dump pinned /sys/fs/bpf/sk_storage
Error: can't dump sk_storage map: Operation not supported
# bpftool -j map dump pinned /sys/fs/bpf/sk_storage
{"error":"can't dump sk_storage map: Operation not supported"}
An error later in the walk is now reported too, as "bpftool map getnext"
reports it, and the element count is printed only after a complete walk.
Signed-off-by: Maxim Skokov <skokovmaksimevg@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://patch.msgid.link/20261016194825.230433-3-skokovmaksimevg@gmail.com
Some BPF commands return ENOTSUPP (524) to user space: iterating local
storage maps, element operations on ringbuf maps, creating an LRU hash
map without preallocation, freezing a struct_ops map, or running a
program type that has no test_run. ENOTSUPP is kernel internal, so
strerror() only gives "Unknown error 524":
# bpftool map getnext pinned /sys/fs/bpf/sk_storage
Error: can't get next key: Unknown error 524
Add a small helper that knows ENOTSUPP and use it where bpftool prints
the error of map create, map element operations, map freeze and prog
run.
Suggested-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: Maxim Skokov <skokovmaksimevg@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://patch.msgid.link/20261016194825.230433-2-skokovmaksimevg@gmail.com
The /proc/mounts fallback in get_tracefs_pipe() accepts only non-tracefs
mount points, so a tracefs mounted outside the default locations is
ignored: bpftool prog tracelog then mounts a second tracefs, or exits
silently with --nomount. Negate the condition.
Fixes: be3245e22d ("tools: bpftool: attempt to mount tracefs if required for tracelog cmd")
Signed-off-by: Yuan Chen <chenyuan@kylinos.cn>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20261016015057.127061-1-chenyuan_fl@163.com
The instruction tests of the test_bpf module live in the BPF selftests
now, as the jit_insns test of test_progs, which BPF CI runs on every
architecture it covers, and its skb_segment tests have a KUnit successor
in net/core/net_test.c.
Remove the module together with its Kconfig entry, the test_kmod.sh and
test_bpf.sh wrappers of the bpf and net selftests, the config fragments
and defconfigs that enabled it, as well as all other references to it and
symbol exports in core kernel.
For archs wanting to support BPF JITs, it's time they move over to the BPF
selftests for proper coverage.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20261013171230.523677-19-daniel@iogearbox.net
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
The resolve_btfids --fatal_warnings option is intended to make the tool
exit with a non-zero status if any warnings are reported. main()
initializes err to -1 and uses 'goto out' to return this status when
a sub-operation fails. If no error occurs,
'if (!(fatal_warnings && warnings)) err = 0;' modifies the final status
depending on whether any warnings have been reported.
Commit 522397d05e ("resolve_btfids: Change in-place update with raw
binary output") broke this logic by assigning the return values of newly
added sub-operations to err. As a result, successful sub-operations
overwrite the initial err value of -1 with 0, causing the tool to exit
successfully even when --fatal_warnings is enabled and warnings have been
reported.
Fix the issue by setting the final value of err unconditionally, so the
code doesn't depend on its previous value.
Fixes: 522397d05e ("resolve_btfids: Change in-place update with raw binary output")
Signed-off-by: Petr Pavlu <petr.pavlu@suse.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Link: https://lore.kernel.org/bpf/20261008-resolve_btfids-warnings-v1-1-0d4e613ffa6e@suse.com
show_kprobe_multi_{plain,json}() walk the sorted kallsyms array against
the sorted addr/cookie array with j as a cursor into the latter. The
loop terminating condition is checked after the cursor has already been
used to access the array:
if (j++ == info->kprobe_multi.count)
break;
After all count entries have been printed, j has been bumped to count,
but the loop keeps running and dereferences data[count] -- one element
past the end of the calloc(count, sizeof(*data)) allocation -- while
comparing every remaining kallsyms symbol. If the out-of-bounds value
happens to match a symbol address, a bogus entry with a garbage
addr/cookie pair is printed to boot.
Fix the off-by-one by advancing the cursor before the comparison, so the
loop terminates as soon as the last entry has been printed.
Verified with an ASAN build of bpftool against a live kprobe_multi link
of 3 symbols:
before: ERROR: AddressSanitizer: heap-buffer-overflow READ of size 8
at link.c:860 (show_kprobe_multi_plain) and link.c:352
(show_kprobe_multi_json), 0 bytes after the 48-byte
get_addr_cookie_array() allocation
after: clean output, no ASAN report
Fixes: edd7f49bb8 ("bpftool: Show kprobe_multi link info")
Signed-off-by: Yuan Chen <chenyuan@kylinos.cn>
Acked-by: Yafang Shao <laoar.shao@gmail.com>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20261010074435.1415234-1-chenyuan_fl@163.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Allow the final BTF patch pass to populate one or more BTF link
records. A link identifies the BTF section to be delivered, its
carrier module, and the raw finalized BTF payload:
--btf_link <section>:<module>:<raw-btf-file>
Use the raw file size and SHA-256 digest to populate the corresponding
<section>.link record. This avoids reparsing BTF during final ELF
patching and supports inline vmlinux BTF linking.
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20261009112324.1442319-3-alan.maguire@oracle.com
Extract .BTF.inline via the following logic:
- Traverse all types reachable from LOC_PARAM, LOC_PROTO, LOCSEC
entries and mark them as having LOC color.
- Traverse all types reachable from the remaining types
and mark them as having MAIN or SHARED.
- Create a new base BTF object and copy all MAIN/SHARED types there.
- Create a new .BTF.inline object with base set to the new base BTF
object, and copy all LOC types there.
This way .BTF.inline can refer to types and strings in the new base,
and strings used only by .BTF.inline remain in it.
Additionally, all BTF_KIND_FUNC types that map to ELF symbols with the
same name are forced to be in the base BTF. Same for types referenced
from BTF id sets.
Regular types referenced only from .BTF.inline remain in the base BTF,
and sizes are 5 and 9.5Mb for vmlinux and vmlinux.inline respectively.
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20261009112324.1442319-2-alan.maguire@oracle.com
The kernel accepts a variable in DATASEC that is smaller than its type.
It's a piece of a variable that the compiler split, with the type of
the whole variable. bpftool prints a variable by its type, so it reads
past the piece and, when the piece is the last in the section, past
the buffer with the value of the map:
$ bpftool map dump pinned /sys/fs/bpf/piece
".bss": [{
"whole": 67305985
},{
"piece.0": {
"a": [134678021,1994496712,32592,32,0,833,0,414764080
Only the first number is in the map. Skip the pieces in
btf_dumper_datasec(), like btf_datasec_show() in the kernel does, and
in show_prog_metadata().
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://lore.kernel.org/bpf/20261002124714.180012-11-alexei.starovoitov@gmail.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
bpftool prog tracelog -w ends from its SIGINT, SIGHUP and SIGTERM handler,
which calls exit(). exit() is not async-signal-safe: it runs atexit
handlers and flushes stdio streams, and the signal may land while the read
loop is inside fwrite() or fflush() on the stream being flushed. Whether
that deadlocks or flushes a half-updated buffer depends on the C library.
Call _exit() instead. The loop flushes every chunk as soon as it is
printed, so the stdio buffer is empty except while a chunk is being
written, and skipping the exit-time flush loses nothing.
While here, note in the manual page that opening a stream as a file
descriptor arrives in Linux 7.4, and restrict the -w/--wait description to
the stdout/stderr form of the tracelog command, since the trace pipe form
does not take the option.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/d64c6402-175f-4ea6-be6d-1a544a95c97c@qmon.net
Link: https://lore.kernel.org/bpf/20260926154515.191689-1-memxor@gmail.com
bpftool prog tracelog { stdout | stderr } PROG dumps the output buffered
in a program stream and exits, which is all BPF_PROG_STREAM_READ_BY_FD
allows. Waiting for further output means rerunning the command.
Add a -w/--wait option that opens the stream with bpf_prog_stream_open()
in its default blocking mode and keeps printing output as the program
produces it. Flush each chunk as it arrives so redirected output is not
held in stdio buffers while the next read blocks. Waiting is opt-in: the
default dump keeps using BPF_PROG_STREAM_READ_BY_FD and exits once the
buffered output is drained, so existing scripts behave the same on old and
new kernels.
A stream descriptor does not keep its program alive, so bpftool drops the
program descriptor once the stream is open. Waiting then ends with EOF
when the program is unloaded and otherwise runs until interrupted, as
bpftool prog tracelog already does for the trace pipe. Exit from the
SIGINT, SIGHUP and SIGTERM handlers like that command does. A flag set by
the handler and checked before each read would miss a signal that lands
between the check and the blocking read(), leaving bpftool asleep until the
next print.
Waiting requires BPF_PROG_STREAM_OPEN. The bpf() syscall fails with EINVAL
for an unknown command, which bpftool reports as missing kernel support
rather than degrading into a dump. Document the option and offer it in the
bash completion.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://patch.msgid.link/20260925045536.1480933-5-memxor@gmail.com
Commit 4b82b181a2 ("bpf: Allow pre-ordering for bpf cgroup progs")
introduced BPF_F_PREORDER to request pre-order execution across the
cgroup hierarchy. Furthermore, attachments legitimately use combinations
such as BPF_F_ALLOW_MULTI | BPF_F_PREORDER or
BPF_F_ALLOW_OVERRIDE | BPF_F_PREORDER.
With BPF_PROG_QUERY reporting the per-program BPF_F_PREORDER attribute,
bpftool's exact-match formatter falls back to "unknown(40)" when
BPF_F_PREORDER is present alone, or "unknown(41)" / "unknown(42)" when
combined with BPF_F_ALLOW_OVERRIDE or BPF_F_ALLOW_MULTI. Additionally,
do_attach() only accepts "multi" and "override", rejecting "preorder"
with "unknown option".
Before:
$ bpftool cgroup show <cg>
1234 cgroup_inet_ingress unknown(42) test_prog
$ bpftool cgroup attach <cg> cgroup_inet_ingress id 5678 multi preorder
Error: unknown option: preorder
After:
$ bpftool cgroup show <cg>
1234 cgroup_inet_ingress multi,preorder test_prog
$ bpftool cgroup attach <cg> cgroup_inet_ingress id 5678 multi preorder
(attaches successfully)
Refactor the attach flags formatter into a bitmask formatter that outputs
comma-separated flag names for plain text while preserving unrecognized
bits as "unknown(...)". Render attach flags as an array of flag names in
JSON output. Accept "preorder" in do_attach(), update the cgroup
documentation and synopsis to express valid flag combinations, and teach
bash completion about them ("multi" or "override" optionally combined
with "preorder").
Signed-off-by: Hui Su <sh_def@163.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://patch.msgid.link/20260922025442.3176057-4-sh_def@163.com
In raw mode ensure we can dump new BTF kinds in normal/json format.
BTF_KIND_LOC_PARAMs are rendered as strings, for example a
const value of 0x2a and a dereference of %r10 + 0x20:
[12] LOC_PARAM '(anon)' size=4 flags=0x2 vlen=1 values='0x2a'
[13] LOC_PARAM '(anon)' size=8 flags=0x38 vlen=2 values='*(reg10 + 0x20)'
LOC_PROTOs render the associated values of each of their
LOC_PARAMs for easier readability:
[14] LOC_PROTO '(anon)' vlen=2
type_id=12 value='0x2a'
type_id=13 value='*(reg10 + 0x20)'
and LOCSEC shows function name associated with site:
[15] LOCSEC 'inline.text' vlen=1
name='foo' func_type_id=5 loc_proto_type_id=14 offset=64
Registers are displayed in an architecture-neutral form;
'regN' where N is the DWARF register number, or 'fbreg'
for the stack frame base.
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20260924111428.75957-10-alan.maguire@oracle.com
For bpftool to be able to dump .BTF.inline data in
/sys/kernel/btf/foo.inline for module foo, it needs to support
multi-split BTF because the parent-child relationship of BTF
inline data for modules is
vmlinux BTF data
module BTF data
module BTF inline data
So for example to dump BTF inline info for xfs we would run
$ bpftool btf dump -B /sys/kernel/btf/vmlinux -B /sys/kernel/btf/xfs file /sys/kernel/btf/xfs.inline
Multiple bases are specified with the vmlinux base BTF first (parent)
followed by the xfs BTF (child), and finally the XFS BTF inline info.
Update help text accordingly to reflect the ability to specify multiple
in-order root-to-branch bases.
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Link: https://patch.msgid.link/20260924111428.75957-8-alan.maguire@oracle.com
bpftool prog profile currently treats the number of possible CPUs as
both the logical CPU ID range and the stride of the perf event array.
That misses valid logical CPUs when the possible CPU mask is sparse,
such as 0,2-3, and can use incorrect PERF_EVENT_ARRAY keys.
Keep the compact possible CPU count for per-CPU result buffers, while
enumerating the actual logical CPU IDs when creating per-CPU perf event
groups. Use the maximum logical CPU ID plus one as the metric stride in
the PERF_EVENT_ARRAY.
This keeps the current per-CPU event grouping intact while separating
the compact per-CPU buffer index from the logical CPU ID and event-array
key. Report the logical CPU ID when a per-CPU event was not counted.
Fixes: 47c09d6a9f ("bpftool: Introduce "prog profile" command")
Signed-off-by: Hui Su <sh_def@163.com>
Link: https://lore.kernel.org/bpf/20260813160858.1042834-3-sh_def@163.com/
Link: https://lore.kernel.org/bpf/20260923004243.969919-3-sh_def@163.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
bpftool uses dense per-CPU value-buffer slots when printing per-CPU map
values. It also uses the slot index as the CPU ID, which produces
incorrect labels when the possible CPU mask is sparse, such as 0,2-3.
Parse the possible CPU mask and use the corresponding logical CPU ID in
plain, JSON, and BTF-formatted output. Keep the dense slot index for
accessing the per-CPU value buffer.
Pass the parsed CPU count together with the CPU ID array so that the
printed CPU IDs and loop bounds come from the same possible-CPU mask.
Propagate CPU-ID lookup and map output errors through the shared output
path to its callers.
Fixes: 71bb428fe2 ("tools: bpf: add bpftool")
Signed-off-by: Hui Su <sh_def@163.com>
Link: https://lore.kernel.org/bpf/20260813155131.1022745-3-sh_def@163.com/
Link: https://lore.kernel.org/bpf/20260923004243.969919-2-sh_def@163.com
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
map event_pipe only accepts perf event arrays, leaving no built-in way
to inspect records produced through the BPF ring buffer API. Extend it
to consume queued and live BPF_MAP_TYPE_RINGBUF records in plain or
JSON output, retaining the existing perf event array path.
Ring buffers have a shared consumer position and no implicit CPU or
timestamp. Document that this command consumes records rather than
observing them passively, and reject perf-only CPU/index selectors.
A producer can keep ring_buffer__poll() busy after a stop signal, so
return -EINTR from the record callback when stopping. Keep stdio out of
the shared signal handler and preserve callback output errors.
Signed-off-by: Tianyi Chen <hi@tychen.cc>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Tested-by: Quentin Monnet <qmo@kernel.org>
Link: https://github.com/libbpf/bpftool/issues/54
Link: https://lore.kernel.org/bpf/20260921073556.99421-2-hi@tychen.cc
bpftool rounds memory-mapped data map sizes to the host page size when
generating a light skeleton. The generated code therefore uses a 64K
mapping size when bpftool runs on a 64K-page host, even if the skeleton
runs on a 4K-page target. The target rejects the oversized map mmap(),
causing failure of loading the light skeleton.
When try to run 64K-page selftests on 4K-page VM, the error message does
not provide the reason about page size.
test_atomics:PASS:atomics skeleton open 0 nsec
test_atomics:FAIL:atomics skeleton load unexpected error: -12 (errno 22)
#15 atomics:FAIL
Pass the original map value size and max entries to the generated code and
round mmap size to the runtime page size in the user-space light
skeleton helpers. This keeps generated light skeletons independent of the
build host page size.
Fixes: d510296d33 ("bpftool: Use syscall/loader program in "prog load" and "gen skeleton" command.")
Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20260915161450.96249-1-leon.hwang@linux.dev
Perf counters undercount when the PMU multiplexes them. Scale each
per-CPU value before aggregation. Per-CPU event groups make PMU ratios
use matching scheduling intervals.
Keep run_cnt as the total number of program executions, independent of
counter scheduling.
Keep JSON value raw, report the scaled value in value_scaled, and
document the output.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260916-bpftool_cyles_per_run-v4-3-f2de02c480d5@meta.com
Independently scheduled perf events can cover different intervals,
making derived ratios inconsistent.
Open one event group per CPU and enable it only after all selected
metrics have joined.
Fail the profile setup when a selected event cannot join its group.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260916-bpftool_cyles_per_run-v4-2-f2de02c480d5@meta.com
A perf counter can be zero at fentry when PMU multiplexing has
not scheduled the event. Do not reject that sample.
Use enabled time to distinguish a successful snapshot. Read
directly into the per-CPU map value.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260916-bpftool_cyles_per_run-v4-1-f2de02c480d5@meta.com
Cross-merge BPF and other fixes after downstream PR.
Conflicts:
kernel/bpf/backtrack.c
include/linux/bpf_verifier.h
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
When setting the XDP hints ifname, bpftool would set the
BPF_F_XDP_DEV_BOUND_ONLY without looking at the existing program flags,
overriding any other flag values. This was always a destructive action,
but after we change libbpf to carry the frags section flag in
prog_flags, this can impact bpftool loading of XDP frags programs.
Change the flag setting to be non-destructive by OR'ing it with the
existing flags.
Fixes: f46392ee3d ("bpftool: Specify XDP Hints ifname when loading program")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Larysa Zaremba <larysa.zaremba@intel.com>
Link: https://lore.kernel.org/bpf/20260901-libbpf-frags-flags-v3-1-4eb6f14968b0@redhat.com
The C dump sorts types by default, so that generated headers are
diffable. The sorted dump emits one type fewer than the unsorted dump
of the same BTF.
dump_btf_c() starts its loop at index 1 to skip the void type at BTF
type ID 0. That holds for the unsorted dump, where the array index is
the type ID, but not after qsort(): position 0 is then the lowest
ranked type, and btf_type_rank() ranks an anonymous enum 0 while void
takes the default rank of 10. So the enum is skipped, and void is
emitted instead as a no-op.
Fixes: 94133cf24b ("bpftool: Introduce btf c dump sorting")
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Link: https://lore.kernel.org/r/20260828215207.3105313-7-ihor.solodrai@linux.dev
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Add bpftool support for ML-DSA program signing and drop the flag for
ML-DSA keys on affected OpenSSL versions, the same way as commit
0ad9a71933 ("modsign: Enable ML-DSA module signing").
Also, an ML-DSA-87 signature is 4627 bytes on its own, so the blob does
not fit into the 4 KiB of MAX_SIG_SIZE anymore and signing would fail
otherwise. Bump to 16 KiB. MAX_SIG_SIZE only sizes the buffer for
what bpftool itself emits (unrelated to BPF_PROG_MAX_SIGNATURE_SIZE).
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/r/20260828175227.1537793-5-daniel@iogearbox.net
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
do_batch() strips trailing comments by truncating the line at '#'
before checking whether fgets() filled the buffer. If a batch line
longer than the buffer contains a '#' within the first
sizeof(buf) - 1 bytes, the truncation makes strlen(buf) smaller and the
line-length check is bypassed. The unread remainder of the line then
stays in the file stream and is parsed and executed as a separate
command on the next loop iteration.
Continuation lines handled below are affected the same way: an overlong
continuation line containing '#' bypasses the "command is too long"
check, and its unread remainder is executed as a separate command.
Move the line-length checks before the comment is stripped, so they see
the full line as read from the file and overlong lines are rejected
regardless of comments. A line that fills the buffer exactly is now
rejected as well, which is fine: batch command lines are not expected
to come anywhere near the buffer limit.
Fixes: 71bb428fe2 ("tools: bpf: add bpftool")
Signed-off-by: Yuan Chen <chenyuan@kylinos.cn>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260824092657.1789956-3-chenyuan_fl@163.com
do_batch() checks errno after the read loop to detect read failures,
but fgets() does not clear errno on success, so a stale errno left by
a previously executed command (e.g. map dump's EBADF from a double
close) makes bpftool report a batch file read failure and exit with an
error even though every command succeeded.
Clear errno before each fgets() call, so the post-loop check only
sees the outcome of the last read: zero on success or EOF, E2BIG for
an overlong line, and a genuine errno when fgets() fails.
Since errno is now reset before every read in batch mode, drop the
USE_LIBCAP errno reset in main() that existed only to keep errno clean
for the batch mode.
Fixes: 71bb428fe2 ("tools: bpf: add bpftool")
Signed-off-by: Yuan Chen <chenyuan@kylinos.cn>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260824092657.1789956-2-chenyuan_fl@163.com
perf c2c:
- Add 'function view' in perf c2c report TUI (switched by pressing 'TAB'
in the cacheline view) to organize samples around functions rather than
cachelines in 3-level hierarchy:
Level 1: Read-side function (sorted by estimated Cycles %)
Level 2: Contending writer functions (sorted by Store count)
Level 3: Shared cacheline addresses
Users can navigate the entries and fold/unfold using 'e' key. An example
output would look like below:
Shared Data Functions Table (19 entries, sorted on Cycles %)
Cycles Store
% count Function / Contending function / Cacheline
----------------------------------------------------------------------
+ 35.67% 876 + [k] cpupri_set
+ 24.31% 424 + [k] pull_rt_task
- 16.53% 555 - [k] dequeue_pushable_task
145 - [k] pull_rt_task
145 0xff2d0082809da080
139 - [k] enqueue_pushable_task
70 0xff2d00a2071f9640
69 0xff2d0082809da000
python module support:
- Extend "perf" python module so that it can be fully functional. The goal
is to run scripts directly, not by 'perf script' command. This would give
better performance as well as more control to build standalone programs
with UI.
- Add LiveSession helper (perf_live.py) to enable live event collection
directly from Python using perf.evlist and perf.parse_events.
perf stat:
- Add --hide-zero-events option to suppress zero-count events.
- Reject conflicting --field-separator and --json-output options.
- Fix duplicate event output with --for-each-cgroup.
perf sched latency:
- Add -H/--histogram and --hist-mode (log|linear) options to show scheduler
wait latency histograms.
- Add --time option to filter analysis by time span in 'perf sched latency'.
ARM CoreSight:
- Synthesize callchains for instruction samples from CoreSight trace
using thread stack ('--itrace=g...').
- Support call indentation ('perf script -F +callindent') to display call
depth hierarchy on branch samples.
- Decode ETE (Embedded Trace Extension) exception packets.
Build system:
- Add 'make install-build-deps' target to install required packages.
- Parallelize JSON and metric pre-computation in jevents.py for faster builds.
Vendor event/metric updates:
- Add Intel Nova Lake events and update tables for existing models.
- Update AMD Zen 5 and Zen 6 core events.
- Update Arm64 Tegra410 metrics and PowerPC hcalls.
Internal changes and fixes:
- Harden trace-event and synthetic event parsing against corrupted data.
- Fix unwinding of multi-threaded processes in libdw unwinder.
- Fix memory leaks in various commands and python bindings.
- Speed up 'perf test' shell tests.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
-----BEGIN PGP SIGNATURE-----
iHUEABYKAB0WIQSo2x5BnqMqsoHtzsmMstVUGiXMgwUCaojCVQAKCRCMstVUGiXM
gxZLAP9u7Bqhfcrc1KKS9RcxBaiYwsPOIgbtdrDDCSiLtdz0EwEAk1UCSXodUsaR
45nbx01h+h7cUWk2HM735f4RXRfO/AE=
=GLJH
-----END PGP SIGNATURE-----
Merge tag 'perf-tools-for-v7.3-2026-08-21' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools
Pull perf tools updates from Namhyung Kim:
"perf c2c:
- Add 'function view' in perf c2c report TUI (switched by pressing
'TAB' in the cacheline view) to organize samples around functions
rather than cachelines in 3-level hierarchy:
Level 1: Read-side function (sorted by estimated Cycles %)
Level 2: Contending writer functions (sorted by Store count)
Level 3: Shared cacheline addresses
Users can navigate the entries and fold/unfold using 'e' key. An
example output would look like below:
Shared Data Functions Table (19 entries, sorted on Cycles %)
Cycles Store
% count Function / Contending function / Cacheline
----------------------------------------------------------------------
+ 35.67% 876 + [k] cpupri_set
+ 24.31% 424 + [k] pull_rt_task
- 16.53% 555 - [k] dequeue_pushable_task
145 - [k] pull_rt_task
145 0xff2d0082809da080
139 - [k] enqueue_pushable_task
70 0xff2d00a2071f9640
69 0xff2d0082809da000
python module support:
- Extend "perf" python module so that it can be fully functional. The
goal is to run scripts directly, not by 'perf script' command. This
would give better performance as well as more control to build
standalone programs with UI.
- Add LiveSession helper (perf_live.py) to enable live event
collection directly from Python using perf.evlist and
perf.parse_events.
perf stat:
- Add --hide-zero-events option to suppress zero-count events
- Reject conflicting --field-separator and --json-output options
- Fix duplicate event output with --for-each-cgroup
perf sched latency:
- Add -H/--histogram and --hist-mode (log|linear) options to show
scheduler wait latency histograms
- Add --time option to filter analysis by time span in 'perf sched
latency'
ARM CoreSight:
- Synthesize callchains for instruction samples from CoreSight trace
using thread stack ('--itrace=g...')
- Support call indentation ('perf script -F +callindent') to display
call depth hierarchy on branch samples
- Decode ETE (Embedded Trace Extension) exception packets
Build system:
- Add 'make install-build-deps' target to install required packages
- Parallelize JSON and metric pre-computation in jevents.py for
faster builds
Vendor event/metric updates:
- Add Intel Nova Lake events and update tables for existing models
- Update AMD Zen 5 and Zen 6 core events
- Update Arm64 Tegra410 metrics and PowerPC hcalls
Internal changes and fixes:
- Harden trace-event and synthetic event parsing against corrupted
data
- Fix unwinding of multi-threaded processes in libdw unwinder
- Fix memory leaks in various commands and python bindings
- Speed up 'perf test' shell tests"
* tag 'perf-tools-for-v7.3-2026-08-21' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools: (232 commits)
perf vendor events arm64: Fix Tegra410 Olympus event 0x0197
perf vendor events arm64: fix swapped MetricGroup for Tegra410 L1 prefetcher metrics
perf evlist: Warn when 'sleep' workload is used without system-wide (-a) option
perf c2c: document function view in perf-c2c man page
perf c2c: add function view browser UI and cacheline detail
perf c2c: build and finalize the function view hierarchy
perf c2c: add function view hierarchy entry creation
perf c2c: add function view stats merge and memory management
perf c2c: add HPP list parsing for function view columns
perf c2c: add column rendering for function view
perf c2c: add function view model skeleton
perf c2c: extract shared data structures into util/c2c.h
perf test sample-parsing: Validate PERF_FORMAT_GROUP values without LOST
perf dso: Replace assert with runtime check in dso__read_symbol()
perf dso: Guard against cache underflow on short reads in dso_cache__memcpy()
perf dso: Use stored fd error instead of stale errno in file_read() and file_size()
perf dso: Guard close() against invalid fd in dso__decompress_kmodule_path()
perf dso: Guard against errno==0 when dso__get_filename() returns NULL
perf build: install-build-deps: add RHEL family devel package mapping
perf build: Remove leftover feature tests for removed cxx and clang support
...
map_dump() closes the map fd in its error path, and do_dump() then
closes the same fd again after a successful dump. Closing an already
closed fd leaves errno set to EBADF, which poisons later errno checks
such as the batch file read check in do_batch(). Let do_dump() own the
fd and remove the close from map_dump().
The same double-close pattern exists in do_show_subset(): both
show_map_close_json() and show_map_close_plain() already close the fd,
so drop the extra close() there as well.
Also propagate the error when bpf_map_get_info_by_fd() fails on a
subsequent map in do_dump(): set err = -1 before breaking out of the
loop, so a later failure is not silently hidden after an earlier
iteration succeeded.
Fixes: 99f9863a0c ("bpftool: Match maps by name")
Signed-off-by: Yuan Chen <chenyuan@kylinos.cn>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20260810142224.2907373-2-chenyuan_fl@163.com
Enhance bpftool to generate skeletons that properly handle global percpu
variables. The generated skeleton now includes a dedicated structure for
percpu data, allowing users to initialize and access percpu variables more
efficiently.
For global percpu variables, the skeleton now includes a nested
structure, e.g.:
struct test_global_percpu_data {
struct bpf_object_skeleton *skeleton;
struct bpf_object *obj;
struct {
struct bpf_map *percpu;
} maps;
// ...
struct test_global_percpu_data__percpu {
int data;
char run;
struct {
char set;
int i;
int nums[7];
} struct_data;
int nums[7];
} *percpu;
// ...
};
* The "struct test_global_percpu_data__percpu *percpu" points to
initialized data, which is actually "maps.percpu->mmaped".
* Before loading the skeleton, updating the
"struct test_global_percpu_data__percpu *percpu" modifies the initial
value of the corresponding global percpu variables.
* After loading the skeleton, "maps.percpu->mmaped" has been marked as
read-only in libbpf. If users want to update the global percpu
variables, they have to update the "maps.percpu" map instead.
* For lightweight skeleton, "lskel->percpu" will be protected by
"mprotect(p, sz, PROT_READ)".
* For subskeleton, those variables of global percpu data will be
skipped.
Assisted-by: Codex:gpt-5.5-xhigh
Signed-off-by: Leon Hwang <leon.hwang@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20260813152324.97937-7-leon.hwang@linux.dev
Kfunc declarations can identify arena arguments through parameter name
suffixes without repeating KF_ARENA_ARG flags in their BTF ID sets.
resolve_btfids currently misses those arguments when synthesizing the
address_space(1) attributes used by generated vmlinux.h files.
Teach the arena prototype rewrite to recognize __arena and
__arena__nullable directly on each parameter. Keep KF_ARENA_ARG1 and
KF_ARENA_ARG2 handling for explicitly flagged kfuncs, while allowing
suffixes on any argument without synthesizing kfunc flags.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Link: https://patch.msgid.link/20260812193842.2879226-2-memxor@gmail.com
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
At the moment there are more callsites that want bpf_verbose_insn() to
not print a newline after the instruction, than callsites that want a
newline. Drop '\n' from disasm.c. Non-functional change.
The changes in bpftool are verified by writing a bpf program using a
variety of instructions and comparing `prog dump xlated` output in the
following modes: plain, opcodes, visual, visual opcodes. The output
before and after the changes is identical.
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <qmo@kernel.org>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20260807-static-zext-v4-1-b6c270013c77@gmail.com
resolve_btfids now emits the bpf_kfunc and bpf_fastcall BTF decl tags and
the arena address_space(1) type attribute for kfuncs, which were
previously produced by pahole.
Reflect this in the in-tree comments and documentation.
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Link: https://patch.msgid.link/20260807032029.78092-7-ihor.solodrai@linux.dev
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Emit the bpf_kfunc decl tag for every discovered kfunc, and bpf_fastcall
for kfuncs flagged KF_FASTCALL. These were previously produced by pahole
under --btf_features=decl_tag_kfuncs.
resolve_btfids now discovers kfuncs from the BTF ID sets [1] and
becomes the source of truth for their annotations.
Drop decl_tag_kfuncs pahole feature flag from scripts/Makefile.btf
[1] https://lore.kernel.org/all/20260722233518.778854-1-ihor.solodrai@linux.dev/
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Acked-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://patch.msgid.link/20260807032029.78092-5-ihor.solodrai@linux.dev
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
For kfuncs flagged with KF_ARENA_RET, KF_ARENA_ARG1 or KF_ARENA_ARG2,
the address_space(1) attribute (a type tag with kflag=1) must be
emitted for the corresponding type in BTF. This was previously done by
pahole via the "attributes" BTF feature [1].
Implement the emission of the arena attributes in resolve_btfids: for
flagged kfuncs create a new function prototype with updated BTF types.
The original proto may be shared with sibling FUNCs, so it is not
modified in place.
Emission is unconditional: kbuild controls the pahole flags, so the
input BTF is expected to not have these attributes. Invalid
declarations are reported as errors.
Drop the "attributes" pahole feature from scripts/Makefile.btf
resolve_btfids now emits them for all supported pahole versions.
[1] https://lore.kernel.org/dwarves/20250228194654.1022535-1-ihor.solodrai@linux.dev/
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Link: https://patch.msgid.link/20260807032029.78092-3-ihor.solodrai@linux.dev
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
btf2btf() adds new types to the BTF: the KF_IMPLICIT_ARGS transform
synthesizes an _impl FUNC together with its FUNC_PROTO and copies of the
kfunc's decl tags. Nothing deduplicates them afterwards. pahole runs
btf__dedup() on its own output, but that happens before resolve_btfids
sees the BTF, so any type the tool itself creates is emitted as-is, even
when a structurally identical type is already present.
Call btf__dedup() at the start of finalize_btf(), so that base
distillation and the by-name sort both operate on the canonical set of
types.
On an x86_64 build with the BPF selftests config this removes 17
duplicate FUNC_PROTOs from vmlinux BTF.
The dedup call increases runtime of resolve_btfids on vmlinux by 30-40%.
The performance hit is an acceptable cost to keep kernel BTF deduped [1].
[1] https://lore.kernel.org/bpf/986e6f4e-4b51-4440-a37c-9624906d7370@linux.dev/
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Link: https://patch.msgid.link/20260807032029.78092-2-ihor.solodrai@linux.dev
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
A kfunc may be listed in several BTF ID sets, which is expected
because different kfuncs are available to BPF programs depending on
their type.
However kfunc flags across different BTF ID sets must be consistent [1].
The flags should be considered a part of the kfunc declaration,
because they influence its BTF representation and verifier handling.
Enforce the kfunc flag consistency in resolve_btifds by hard failing
on error and blocking kernel (or module) build.
[1] https://lore.kernel.org/bpf/9b2196dd-443b-4632-ae11-030cdbdc59b4@linux.dev/
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260722233518.778854-9-ihor.solodrai@linux.dev
collect_kfuncs() currently uses bpf_kfunc decl tags to identify the
list of kfuncs. The decl tags are generated by pahole, which makes
current implementation implicitly rely on those tags being generated.
The authoritative source, used by the the BPF verifier for kfunc
registration, of functions being BPF kfuncs are
BTF_KFUNCS_START()/END() declarations. These are BTF_ID_SET8 under the
hood. Currently resolve_btfids reads kfunc flags from these sets, and
populates them with BTF IDs.
Implement kfunc discovery from BTF_ID_SET8 symbols in resolve_btfids,
removing the dependency on pahole's emmission of decl tags.
Walk BTF_ID_KIND_SET8 sets, and use the address-to-symbol index to
look up set entry's BTF_ID symbol name (before .BTF_ids is patched),
recording the paired flags directly. This makes find_kfunc_flags()
helper unnecessary, so it's removed.
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260722233518.778854-8-ihor.solodrai@linux.dev
process_kfunc_with_implicit_args() skips generating <name>_impl
function when one already exists for backwards compatibility.
It uses btf__find_by_name_kind(), which searches the base BTF before
the split BTF. When resolve_btfids processes a module, a same-named
_impl in vmlinux would be found and the module's own counterpart would
not be created.
Fix by using btf__find_by_name_kind_own() for the lookup.
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260722233518.778854-6-ihor.solodrai@linux.dev
Keep an address-sorted index of parsed .BTF_ids symbols so that the
original BTF_ID symbol name can be recovered from an entry address.
Use the index in find_kfunc_flags() to scan BTF_SET8_KFUNCS entries
directly and match each entry back to the requested kfunc.
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Emil Tsalapatis <emil@etsalapatis.com>
Reviewed-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260722233518.778854-3-ihor.solodrai@linux.dev
push_*() helpers in resolve_btfids contain copy-pasted array growth
logic. Factor it out into ensure_mem() - a simplified variant of
libbpf_ensure_mem(), and use it in the helpers.
Signed-off-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Reviewed-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/bpf/20260722233518.778854-2-ihor.solodrai@linux.dev
Try to remove some, but not all duplicate toolchain definitions. In
these instances, their makefiles already include
tools/scripts/Makefile.include which defines these in a consistent way.
STRIP is the only one that was set with an '=', but I don't think it
was significant so that difference can be dropped.
Signed-off-by: James Clark <james.clark@linaro.org>
Reviewed-by: Ian Rogers <irogers@google.com>
Acked-by: Ihor Solodrai <ihor.solodrai@linux.dev>
Acked-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Looking up a prog or map by name walks the whole id space. There is a
window between bpf_prog_get_next_id()/bpf_map_get_next_id() and getting
an fd for that id in which an unrelated object can be freed, and the
lookup then fails with ENOENT and aborts the whole command.
Skip such ids and keep walking, the same way do_show() already does.
Signed-off-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Link: https://lore.kernel.org/bpf/20260720071520.396363-1-jiayuan.chen@linux.dev
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
bpftool_prog_sign() signed only the loader instructions. The metadata
blob the loader installs was left to an in-loader hash check, which
the kernel now performs at load time over insns || metadata.
Sign that same concatenation: pass the metadata blob (gen_loader_opts
data) through to bpftool_prog_sign() and feed insns || metadata to
CMS_final(). The excl_prog_hash stays a digest of the instructions
alone; it binds the metadata map to the loader and is matched against
prog->digest by the verifier, independent of what the signature covers.
The signed artifact is now plain data: both bytes the signature
covers are embedded verbatim in the generated skeleton, so signing
and verifying an lskel is an ordinary CMS operation that a signer or
auditor can perform (or reproduce) offline, without analyzing loader
bytecode to establish what the signature actually attests to.
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20260708075343.358712-6-daniel@iogearbox.net
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
bpftool_prog_sign() ignores the return value of EVP_Digest(). If the
digest computation fails (context allocation failure, or a digest
fetch failure under OpenSSL), EVP_Digest() returns 0 and leaves the
output buffer untouched, but the function still reports success.
Fixes: 40863f4d6e ("bpftool: Add support for signing BPF programs")
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Quentin Monnet <qmo@kernel.org>
Link: https://lore.kernel.org/bpf/20260708075343.358712-5-daniel@iogearbox.net
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
