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
philip_ has quit [Quit: philip_]
hpt has quit [Quit: Lost terminal]
hpt has joined #systemtap
hpt has quit [Quit: Lost terminal]
scox has joined #systemtap
hkshaw has quit [Ping timeout: 248 seconds]
hkshaw has joined #systemtap
irker777 has quit [Quit: transmission timeout]
hchiramm has quit [Ping timeout: 258 seconds]
hpt has joined #systemtap
hchiramm has joined #systemtap
hkshaw has quit [Quit: Leaving.]
hkshaw has joined #systemtap
hpt has quit [Ping timeout: 265 seconds]
zodbot has quit [Read error: Connection reset by peer]
zodbot has joined #systemtap
pwithnall_ has joined #systemtap
hkshaw has quit [Ping timeout: 250 seconds]
hkshaw has joined #systemtap
hkshaw has quit [Read error: Connection reset by peer]
hkshaw has joined #systemtap
modem has joined #systemtap
<modem>
hey
<modem>
I am using systemtap to alter a function control flow
<modem>
so I put on hook on it and call some embeded C-code
<modem>
but then I would like the systemtap probe to return
<modem>
basically I am trying a new version of the function for a specific scenario, I do not want the original code to be executed at all
<modem>
everytime i have an "if" match I am expecting my code to be called instead of the embeded code
<modem>
do you guy can let me know how I could achieve that using systemtap ? Thanks!
<modem>
I found a way to let the orig function return triggering an error path
<modem>
but it is somehow hacky
tromey has joined #systemtap
<fche>
modem, it certainly is somehow hacky
<fche>
but lacking any way to dramatically & portably modify -control- flow, we're left with changing -data- flow so to speak
<fche>
(I mean one could in theory adjust the pt_regs PC value to jump over things, but that's far more hacky)
sabayonuser has joined #systemtap
<modem>
fche: hey
<modem>
ok let's do it hacky way
<modem>
thx :-)
<fche>
yeah. that's the way we do security band-aids too
<fche>
though I think by the time I presented the concept at FOSDEM last year, some ideas crystallized better
<modem>
I have seen such patch recently
<modem>
rh bugzilla iirc
<modem>
didn't know you presented the concept previous to that, tought it was a good idee
<modem>
i defintely would patch my system with systemtap instead of rebooting from time to time to give me opportunity to migrate properly without exposing meanwhile
<fche>
that's the idea & attraction
<fche>
it's definitely not a turn-the-crank easy operation though
<fche>
harder than taking someone's patch and applying/rebooting etc.
<fche>
but it's a neat puzzle
hkshaw has quit [Ping timeout: 252 seconds]
wcohen has quit [Remote host closed the connection]
<Guest56127>
Hey folks, i have compiled systemtap version 3.0. Unfortunately it is not running. The compilation of generated C code fails (Pass 4). E.g. running stap -ve 'probe begin { log("hello world") exit () }' yields in an error like this http://pastebin.com/iqz3LnwK. Do you have any clue what might be wrong?
<fche>
Guest56127, hi, use git systemtap for recent kernels
<fche>
the 3.0 release is now many months old, and newer kernels break interfaces
<fche>
a stap 3.1 release will come out in a few weeks
<Guest56127>
Ok, I will clone the git repo and try again. I am using 4.7.2 is that already too recent for version 3.0?
<Guest56127>
I tried with recent version. Now systemtap prints hello world but i receive some strange error about a module. Here is the error http://pastebin.com/U2r4Cw3h.