From: Vitaly Chikunov <vt@altlinux.org>
To: dwarves@vger.kernel.org, Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: bpf@vger.kernel.org, "Dmitry V. Levin" <ldv@altlinux.org>
Subject: pahole 1.25 SIGSEGV when building kernel (-j when cores > 81)
Date: Wed, 5 Jul 2023 05:00:40 +0300 [thread overview]
Message-ID: <20230705020040.bqhtd3yxpntk44q5@altlinux.org> (raw)
Hi,
After updating Dwarves to v1.25 we are getting SIGSEGV while building
kernel on 32-bit x86 architecture. It looks like this:
BTF .btf.vmlinux.bin.o
scripts/link-vmlinux.sh: line 111: 395728 Segmentation fault LLVM_OBJCOPY="${OBJCOPY}" ${PAHOLE} -J ${PAHOLE_FLAGS} ${1}
LD .tmp_vmlinux.kallsyms1
NM .tmp_vmlinux.kallsyms1.syms
KSYMS .tmp_vmlinux.kallsyms1.S
AS .tmp_vmlinux.kallsyms1.S
LD .tmp_vmlinux.kallsyms2
NM .tmp_vmlinux.kallsyms2.syms
KSYMS .tmp_vmlinux.kallsyms2.S
AS .tmp_vmlinux.kallsyms2.S
LD vmlinux
BTFIDS vmlinux
libbpf: failed to find '.BTF' ELF section in vmlinux
FAILED: load BTF from vmlinux: No data available
What crashes is this command:
pahole -J --btf_gen_floats -j --lang_exclude=rust .tmp_vmlinux.btf
I found that cause of the crash is that build box having 128 cores. By
experiment I found that with -j81 pahole works OK, but with -j82 or
greater it crashes.
$ gdb -q --args pahole -J --btf_gen_floats -j111 --lang_exclude=rust .tmp_vmlinux.btf
Thread 15 "pahole" received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0xda5ffb40 (LWP 3102466)]
0xf7f3c944 in btf_encoder__btf (encoder=0x0) at /usr/src/debug/dwarves-1.25/btf_encoder.c:1890
1890 return encoder->btf;
(gdb) bt
#0 0xf7f3c944 in btf_encoder__btf (encoder=0x0) at /usr/src/debug/dwarves-1.25/btf_encoder.c:1890
#1 0x5655c25d in pahole_stealer (cu=0xd9a01f80, conf_load=0x565640c0 <conf_load>, thr_data=0x56567c18) at /usr/src/debug/dwarves-1.25/pahole.c:3100
#2 0xf7f452d7 in cu__finalize (cu=cu@entry=0xd9a01f80, conf=0x565640c0 <conf_load>, thr_data=thr_data@entry=0x56567c18)
at /usr/src/debug/dwarves-1.25/dwarf_loader.c:3001
#3 0xf7f4541d in cus__finalize (thr_data=0x56567c18, conf=<optimized out>, cu=0xd9a01f80, cus=0x565651c0) at /usr/src/debug/dwarves-1.25/dwarf_loader.c:3008
#4 dwarf_cus__create_and_process_cu (dcus=dcus@entry=0xffffd19c, cu_die=cu_die@entry=0xda5ff38c, pointer_size=<optimized out>, thr_data=0x56567c18)
at /usr/src/debug/dwarves-1.25/dwarf_loader.c:3207
#5 0xf7f461af in dwarf_cus__process_cu_thread (arg=0xffffcbf8) at /usr/src/debug/dwarves-1.25/dwarf_loader.c:3250
#6 0xf7db4258 in start_thread (arg=<optimized out>) at pthread_create.c:444
#7 0xf7e3a878 in clone3 () from /lib/libc.so.6
(gdb) p encoder
$1 = (struct btf_encoder *) 0x0
(gdb) f 1
#1 0x5655c25d in pahole_stealer (cu=0xd9a01f80, conf_load=0x565640c0 <conf_load>, thr_data=0x56567c18) at /usr/src/debug/dwarves-1.25/pahole.c:3100
3100 thread->btf = btf_encoder__btf(thread->encoder);
(gdb) list -2
3093 thread->encoder =
3094 btf_encoder__new(cu, detached_btf_filename,
3095 NULL,
3096 skip_encoding_btf_vars,
3097 btf_encode_force,
3098 btf_gen_floats,
3099 global_verbose);
3100 thread->btf = btf_encoder__btf(thread->encoder);
3101 }
3102 encoder = thread->encoder;
I think that return value of btf_encoder__new is not checked. But did
not investigate further why is this happening. It would be great to have
this fixed.
Thanks,
next reply other threads:[~2023-07-05 2:17 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-07-05 2:00 Vitaly Chikunov [this message]
2023-07-10 13:42 ` pahole 1.25 SIGSEGV when building kernel (-j when cores > 81) Arnaldo Carvalho de Melo
2023-07-10 20:00 ` Vitaly Chikunov
2023-09-14 3:37 ` Vitaly Chikunov
2023-09-15 2:02 ` Vitaly Chikunov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230705020040.bqhtd3yxpntk44q5@altlinux.org \
--to=vt@altlinux.org \
--cc=acme@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=dwarves@vger.kernel.org \
--cc=ldv@altlinux.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).