Date: Thu, 15 Dec 2016 07:51:24 +0900 (JST) From: Hiroki Sato <hrs@FreeBSD.org> To: freebsd-dtrace@FreeBSD.org Subject: malformed symbol Message-ID: <20161215.075124.1459885758696268380.hrs@allbsd.org>
next in thread | raw e-mail | index | archive | help
----Security_Multipart0(Thu_Dec_15_07_51_24_2016_391)--
Content-Type: Multipart/Mixed;
boundary="--Next_Part(Thu_Dec_15_07_51_24_2016_877)--"
Content-Transfer-Encoding: 7bit
----Next_Part(Thu_Dec_15_07_51_24_2016_877)--
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Hi,
I noticed that a simple USDT definition caused an odd error. The
source files in question are as follows (a complete tarball is
attached to this email):
sample.c:
| #include <sys/param.h>
| #include <stdlib.h>
| #include <unistd.h>
|
| #include "sample_probes.h"
|
| int
| main(void)
| {
| ssize_t i;
| char buf[80];
|
| i = read(STDIN_FILENO, buf, sizeof(buf));
| SAMPLE_FIRST_READ(i, buf);
|
| return (0);
| }
sample_probes.d:
| provider sample {
| probe first__read(ssize_t, char *);
| };
Makefile:
| PROG= sample
| MAN=
| SRCS= sample.c sample_probes.d
|
|.include <bsd.prog.mk>
Note that the program itself is useless because this is a minimal
working example to reproduce the symptom. It causes the following
error:
| dtrace -C -x nolibs -h -s sample_probes.d
| rm -f .depend
| echo sample.full: /usr/lib/libc.a >> .depend
| cc -O2 -pipe -g -I/var/home/hrs/sample -MD -MF.depend.sample.o -MTsample.o -std=gnu99 -fstack-protector-strong -Qunused-arguments -c sample.c -o sample.o
| dtrace -C -x nolibs -G -o sample_probes.o -s sample_probes.d sample.o
| cc -O2 -pipe -g -I/var/home/hrs/sample -std=gnu99 -fstack-protector-strong -Qunused-arguments -o sample.full sample.o sample_probes.o
| sample.o: In function `main':
| /var/home/hrs/sample/sample.c:13: undefined reference to `ead'
| cc: error: linker command failed with exit code 1 (use -v to see invocation)
| *** Error code 1
A symbol "read" in sample.o seems to be rewritten in the "dtrace -G"
line. Sample.o just after compilation of sample.c had the correct
symbol name like the following:
% nm sample.o
U __dtrace_sample___first__read
U __stack_chk_fail
U __stack_chk_guard
0000000000000000 T main
U read
% dtrace -C -x nolibs -G -o sample_probes.o -s sample_probes.d sample.o
% nm sample.o
0000000000000000 A __dtrace_sample___first-read
U __stack_chk_fail
U __stack_chk_guard
U ead
0000000000000000 T main
And it worked when I changed the probe name with "first__readx".
This was reproducible on 11.x and 12.x, not on 10.x. Could anyone
try this and let me know if this is reproducible on your 11.x or 12.x
box? I guess this is a regression of symbol rewrite routine such as
s/__/-/ in the dtrace utility while I have not investigated the
details yet. Or am I missing something here?
-- Hiroki
----Next_Part(Thu_Dec_15_07_51_24_2016_877)--
Content-Type: Application/Octet-Stream
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="sample-20161215-1.tar.gz"
H4sIAPnHUVgAA+2VW2vCMBSA+9r8iuBe2iE1vdmBU5DphuAN69sYpdo4g1olaQfb2H9fqvUKW2Gs
Dlm+l6anSZP05Dtl/mI1xyUpTxBCjmNDCVkOsvSD6x6om8iwTMt0bBMi3UDIlqCd66pSYhb5lC9l
Stm3/bKep/vYXS8Etsl/x5/hCZnjXObg36Nctr7Ov64b+/xbFs9/0pAgymU1J/zz/PcHvYeqvDkF
oFPvVmXgDu7cbUgbw03DW9HlCDMtAEAj4XgeBxjejlig8fiztpjVwF/vRPATUv+32c5ljiz/Teug
/pvOxn9d+H8OrnYys1dWWvnUX2jTGjgIR8GcjI5jcUh4OIntg4XjOjEtAEDCCCx8EiovSxKo4B3I
jJE37EWQVIA8nvoUjuLJ4w16qgAgE1iFFPuB4g4bra5332o3u71i0qMIk2HLicLbqsqHuvVOv93k
XQbu0Bs06w2FrDuqyXsojmIaQgXxuw9RlTI48n9X5H93jiz/kWOc+G/zEcL/c8Az/kICTNPfPOSS
rg8BnBDKIs9bG5lqW4RrZ68TsSrCLIFAILhoPgGJ7BmNABIAAA==
----Next_Part(Thu_Dec_15_07_51_24_2016_877)----
----Security_Multipart0(Thu_Dec_15_07_51_24_2016_391)--
Content-Type: application/pgp-signature
Content-Transfer-Encoding: 7bit
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iEYEABECAAYFAlhRzOwACgkQTyzT2CeTzy2XtQCcDwMUh4c/6mH1PKeikadIGmE/
yE4AnRHx+4kh1/jFsxvigfS2wDRkAQd/
=i9jC
-----END PGP SIGNATURE-----
----Security_Multipart0(Thu_Dec_15_07_51_24_2016_391)----
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20161215.075124.1459885758696268380.hrs>