fche changed the topic of #systemtap to: http://sourceware.org/systemtap; email systemtap@sourceware.org if answers here not timely, conversations may be logged
ciupicri has quit [Quit: Leaving.]
<fche>
yeah
<fche>
introom, ftrace has no knowledge about inlined functions, parameters such as variables, statements,
<fche>
that's what stap uses debuginfo for. without debuginfo, it can still do about the same things as ftrace can
<introom>
fche: wondering how stap got inlined right.
<fche>
we get about as right as the compiler can tell us about
<fche>
gcc8 should be about the best ever; it gets us even more info than we currently know how to handle, which should make a big difference with inlines
hpt has joined #systemtap
orivej has joined #systemtap
fche has quit [Quit: Leaving]
naveen has joined #systemtap
naveen has quit [Quit: WeeChat 2.1]
sanoj has joined #systemtap
sanoj has quit [Ping timeout: 256 seconds]
sanoj has joined #systemtap
naveen has joined #systemtap
ciupicri has joined #systemtap
sanoj has quit [Ping timeout: 256 seconds]
sanoj has joined #systemtap
naveen_ has joined #systemtap
naveen has quit [Ping timeout: 256 seconds]
sanoj has quit [Ping timeout: 264 seconds]
naveen has joined #systemtap
sanoj has joined #systemtap
naveen_ has quit [Ping timeout: 255 seconds]
hpt has quit [Quit: Lost terminal]
sanoj has quit [Ping timeout: 240 seconds]
sanoj has joined #systemtap
naveen has quit [Quit: WeeChat 2.1]
sanoj has quit [Ping timeout: 240 seconds]
naveen has joined #systemtap
orivej has quit [Ping timeout: 264 seconds]
fche has joined #systemtap
fche has quit [Changing host]
fche has joined #systemtap
orivej has joined #systemtap
orivej has quit [Ping timeout: 268 seconds]
naveen has quit [Quit: WeeChat 2.1]
wcohen has quit [Remote host closed the connection]
drsmith_ is now known as drsmith
nkambo has joined #systemtap
wcohen has joined #systemtap
drsmith has left #systemtap [#systemtap]
drsmith has joined #systemtap
gromero has joined #systemtap
sanoj has joined #systemtap
naveen has joined #systemtap
tromey has joined #systemtap
naveen_ has joined #systemtap
naveen has quit [Ping timeout: 260 seconds]
<ciupicri>
print_backtrace prints only "0xffffffffae8f6b52 : ipt_do_table+0x172/0x660 [kernel] | 0x0 (inexact)". Could this be because the probe is for an inline function?
naveen_ has quit [Quit: WeeChat 2.1]
<fche>
it should be able to do better
<fche>
we have had a few reports recently of stap becoming less able to backtrace the kernel
<ciupicri>
if it matters, I'm running 4.15.14-300.fc27.x86_64+debug
<ciupicri>
I'm probing a function that uses struct for one of it's parameters. What should I use table->name, $table->name, or ...?
<fche>
$table->name
<ciupicri>
the definition say $table:struct xt_table*
sanoj has quit [Ping timeout: 260 seconds]
<ciupicri>
and if (kernel_string($table->name) != @1) ... makes sense, right?
<fche>
sure
<fche>
if you want that @1 parameter to be a compile-time parameter, vs. a script-level global variable or something like that
<ciupicri>
it's a constant
<ciupicri>
I couldn't figure out a better way to specify it
<fche>
could make it a literal "string"
<fche>
could do global name = "foo" and then kernel_string($table->name) != name
<fche>
etc.
<ciupicri>
it's used in more than one place, so the first option fell off. I thought of using global, but got scared because of stupid reasons.
<fche>
haha, be brave :)
<fche>
but sure, whatever works for you is fine
<ciupicri>
got scared of the fact that it's using memory for the whole duration, but now I realize that a literal string would use the same memory anyway
<fche>
yea
<ciupicri>
and the second reason was of a possible name conflict
<ciupicri>
if I were to name it table, would $table-> still work?
<fche>
yeah, different namespace
<ciupicri>
how good is calling get_cycles in call and return for profiling to see if a function is faster than another one?
<fche>
the probe effect (overheads of just calling into stap) may cancel out little timing differences
<ciupicri>
so a black box approach might be better for comparison purposes sometimes