fche changed the topic of #systemtap to: http://sourceware.org/systemtap; email systemtap@sourceware.org if answers here not timely, conversations may be logged
hpt has joined #systemtap
<agentzh>
hi guys, i've noted that stap may sometimes emit garbaled stdout output with lots of successive \0 even for simplest printf()/println() scripts. it's especially common when the machine is under load. is it a known issue?
<agentzh>
i guess there might be some race or buffer mishandling in staprun's relayfs reader thread. any hints to debug such hard-to-reproduce issues in staprun?
gromero has quit [Ping timeout: 272 seconds]
gromero has joined #systemtap
derek0883 has joined #systemtap
_whitelogger has joined #systemtap
derek0883 has quit [Remote host closed the connection]
derek0883 has joined #systemtap
orivej has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]
derek0883 has quit [Remote host closed the connection]
orivej has joined #systemtap
orivej has quit [Ping timeout: 265 seconds]
orivej has joined #systemtap
khaled_ has joined #systemtap
_whitelogger has joined #systemtap
orivej has quit [Read error: Connection reset by peer]
orivej_ has joined #systemtap
orivej_ has quit [Ping timeout: 256 seconds]
orivej has joined #systemtap
hassan64 has joined #systemtap
orivej has quit [Ping timeout: 258 seconds]
<hassan64>
I need *.stp file, for which I can get syscalls, PID, Argument being parsed, return value etc, when I ran a sample on a target machine
orivej has joined #systemtap
<hassan64>
I am using a command on target machine as "staprun -c /root/test.sh /root/test.stp"
<hassan64>
in test.sh
<hassan64>
Any help
<hassan64>
"/root/analyzed_bin > /root/prog.log"
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
<hassan64>
Cam someone answer my query
<hassan64>
*Can \
orivej has quit [Ping timeout: 260 seconds]
orivej has joined #systemtap
<hassan64>
can anybody help. I got an error while translating to *.ko module
<agentzh>
it should be defined in such locations in your kernel tree.
<agentzh>
maybe it's your strace.stp script referencing the syscall.
<agentzh>
you can always check.
<agentzh>
seems like probe nd_syscall.* or probe nd_syscall.*.return introduced the fadvise64 syscall.
<agentzh>
maybe you can explicitly list all the syscalls your arm kernel actually supports via syntax like "probe syscall.open, syscall.openat, ... { ... }"
<agentzh>
for some reason, nd_syscall.* includes advise64 syscall but it is absent in your arm kernel tree.
<agentzh>
sorry, end of day for me. gotta get some sleep. good luck.
orivej_ has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
orivej has quit [Ping timeout: 264 seconds]
orivej_ has joined #systemtap
<hassan64>
agentzh: I am compiling latest systemtap (master version) with command as "./configure --with-elfutils=$(pwd)/../elfutils-0.180
<hassan64>
make &&make install (with sudo)
<hassan64>
But when I use stap -V , It comes as "Systemtap translator/driver (version 4.3/0.170, commit release-4.3-2-g7a7016a121c0 + changes)"
<hassan64>
Okay with 4.3 but what about elfutils. It shoul dhave been 0.180. isn't it ?
<hassan64>
Please guide me
<mjw>
--with-elfutils is silently ignored
<hassan64>
Please help me compiling this .
<hassan64>
Or systemtap compiled with 0.180 elfutils infact but not showing
<mjw>
You'll have to build and install elfutils yourself and then configure systemtap with CFLAGS="-g -O2 -I.../where/you/installed elfutils headers" LDFLAGS="-L.../where/you/installed elfutils libs"
<hassan64>
Can you please send me complete command. Thanks
<mjw>
hassan64, how did you configure/build/install elfutils?
<mjw>
ok, then for systemtap ./configure CFLAGS="-g -O2 -I/home/hassan/Desktop/sandbox_design/elf_dir/include" LDFLAGS="-L/home/hassan/Desktop/sandbox_design/elf_dir/lib" should work
<mjw>
although it still claims to have been build against 0.170
<mjw>
which is odd, so maybe my configure instructions were wrong.
<mjw>
But you have a systemtap stap command that uses elfutils 0.180 there.
<hassan64>
Btw, this export command should be given before configuration , right ?
<mjw>
export LD_LIBRARY_PATH= says "use the runtime libraries found in this directory"
<mjw>
hassan64, maybe it would help the build to realize it needs to compile/link against what is in that dir, but in theory the -I and -L flags should have done that.
<mjw>
basically you have a somewhat unusual setup that nobody has tested. Normally the distro installed elfutils is fine/new enough.
<hassan64>
so am I using updated versin of systemtap compiled with latest elfutils ?
<mjw>
It certainly looks so :)
<hassan64>
as per "Systemtap translator/driver (version 4.3/0.180/0.170, commit release-4.3-2-g7a7016a121c0 + changes)" says
<hassan64>
But then why, when I use to convert *.stp to *.ko, I got an error
orivej has quit [Ping timeout: 258 seconds]
<hassan64>
"error: invalid application of ‘sizeof’ to incomplete type ‘struct old_timex32’"
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
orivej_ has joined #systemtap
orivej has quit [Ping timeout: 265 seconds]
orivej_ has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
sapatel has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
hassan64 has quit [Remote host closed the connection]
orivej has quit [Ping timeout: 246 seconds]
orivej has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
orivej has quit [Ping timeout: 240 seconds]
orivej has joined #systemtap
derek0883 has joined #systemtap
orivej has quit [Ping timeout: 256 seconds]
derek0883 has quit [Remote host closed the connection]
orivej has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]
orivej has joined #systemtap
derek0883 has joined #systemtap
<agentzh>
fche: will you please backlog and see my messages above (sent late yesterday)? many thanks!
<fche>
agentzh, looking
<fche>
the first issue, fix-vma-leak, I wouldn't bother with a NEWS blurb for it as a mere bug fix, but rather a sourceware.org/bugzilla entry you could open & close
<fche>
the comment looks good, so feel free to comment runtime/vma.c
<fche>
commit I mean
<agentzh>
k
<agentzh>
thanks for your reply.
<agentzh>
for the garbaled output with lots of \0 in stap's output, it seems more likely in the kernel client side of relayfs instead of on the staprun side. the latter looks so simple it's hard to have issues.
<fche>
could it be a length mismatch problem?
<fche>
have a reproducer? (I know you said it's worse under heavy load but still)
<agentzh>
what do you mean by "length mismatch problem"?
<fche>
meaning we may be copying some sort of string-terminating \0
<fche>
length off-by-one kind of thing
orivej_ has joined #systemtap
orivej has quit [Ping timeout: 264 seconds]
orivej_ has quit [Ping timeout: 256 seconds]
orivej has joined #systemtap
orivej has quit [Quit: No Ping reply in 180 seconds.]