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
ego has joined #systemtap
hkshaw has joined #systemtap
eichiro has quit [Ping timeout: 244 seconds]
eichiro has joined #systemtap
nkambo has joined #systemtap
hchiramm has quit [Ping timeout: 244 seconds]
ravi_ has joined #systemtap
Empyrium has joined #systemtap
naveen has joined #systemtap
naveen has quit [Client Quit]
naveen has joined #systemtap
srikar_away is now known as srikar
irker916 has quit [Quit: transmission timeout]
hchiramm has joined #systemtap
naveen has quit [Ping timeout: 248 seconds]
ravi_ has quit [Ping timeout: 264 seconds]
ravi_ has joined #systemtap
naveen has joined #systemtap
naveen1 has joined #systemtap
ravi_ has quit [Ping timeout: 246 seconds]
naveen has quit [Ping timeout: 244 seconds]
ravi_ has joined #systemtap
lorddoskias has left #systemtap [#systemtap]
ton31337 has quit [Ping timeout: 268 seconds]
ton31337 has joined #systemtap
naveen has joined #systemtap
naveen1 has quit [Ping timeout: 246 seconds]
mjw has joined #systemtap
naveen has quit [Ping timeout: 244 seconds]
scox has quit [Ping timeout: 246 seconds]
naveen has joined #systemtap
nkambo has quit [Ping timeout: 248 seconds]
hpt has quit [Ping timeout: 250 seconds]
modem has quit [Read error: Connection reset by peer]
modem has joined #systemtap
hkshaw has quit [Ping timeout: 244 seconds]
naveen has quit [Quit: Leaving.]
ego has quit [Ping timeout: 244 seconds]
hkshaw has joined #systemtap
nkambo has joined #systemtap
atomlin has joined #systemtap
atomlin has left #systemtap [#systemtap]
srikar is now known as srikar_away
ravi_ has quit [Remote host closed the connection]
<irker745>
systemtap: flu systemtap.git:refs/heads/master * release-2.9-438-g8fb73e4 / elaborate.cxx staprun/monitor.c: Monitor mode: display number of script probes http://tinyurl.com/hqmlqch
darvon has quit [Ping timeout: 248 seconds]
darvon has joined #systemtap
<jemarch>
mjw: ok
<mjw>
jemarch, sorry, have to go out for a while
<jemarch>
mjw: anyway, when skipping prologues the kernel is giving me a EINVAL in register_kprobe, trying to determine why :)
<mjw>
but I am sure fche knows everything you ever want to know about skipping prologues, -P, dwarf_entry_breakpoints and dwflpp::resolve_prologue_endings
<jemarch>
mjw: excellent, I will ask him :)
<jemarch>
mjw: at the end of the week I will finally be back in CET timezone
<fche>
jemarch, the kernel is picky as to the instructions it's willing to kprobe
<fche>
e.g. if the first post-prologue instruction is a lock prefix --- na'ah
<jemarch>
I checked kernel/kprobes.c, and I can't find any obvious reason why the addresses are rejected
<jemarch>
The address is not already registerd, it is not blacklisted, it doesn't collide with a jtrace location...
<jemarch>
I don't know what jump_label_text_reserved checks for
<jemarch>
but I don't see any static_key used in the function I am probing, oh well...
hpt has quit [Ping timeout: 276 seconds]
ego has joined #systemtap
<fche>
if you don't see any reason
<fche>
and if you can reproduce the failure with a 'perf probe'
<fche>
then it's worth filing a bug against the kernel about it
pfallenop has quit [Ping timeout: 250 seconds]
naveen has quit [Ping timeout: 248 seconds]
naveen has joined #systemtap
<jemarch>
ah, a perf probe
naveen has quit [Ping timeout: 244 seconds]
naveen has joined #systemtap
ego has quit [Quit: Leaving]
pfallenop has joined #systemtap
naveen has quit [Quit: Leaving.]
pfallenop has quit [Remote host closed the connection]
pfallenop has joined #systemtap
<jemarch>
fche: found the problem. It is the ftrace nops, generated by GCC -pg.
<jemarch>
fche: so, at least in sparc, we must check for CONFIG_FTRACE and skip the nops if it is defined
<jemarch>
fche: with -P
<jemarch>
but only if CONFIG_HAVE_FENTRY is not defined, hmmm...
hchiramm has joined #systemtap
<fche>
modern kernels let one kprobe right on top of the nops
<fche>
and implement kprobes via a branch rather than trap
<fche>
at least on x86
nkambo has quit [Ping timeout: 276 seconds]
nkambo has joined #systemtap
nkambo has quit [Ping timeout: 244 seconds]
nkambo has joined #systemtap
<jemarch>
we don't have hw breakpointing in sparc
<jemarch>
kprobes are implemented using a trap
<jemarch>
oh, I see all the recent changes in that area.