-
jbk
has anyone done anonymous dtrace on smartos? I'm guessing I probably need to make the resulting dtrace.conf file (from dtrace -A) into a boot module, but not sure if there's more to it or some other trick
-
rmustacc
That should be it. And a forceload.
-
jbk
ahh.. so also make /etc/system a boot module too..
-
jbk
(w/ the forceload statements)
-
rmustacc
I don't believe you need anything in /etc/system.
-
rmustacc
Forceload usually goes in the module's .conf.
-
jbk
oh hrm.. i always remember them being in /etc/system
-
jbk
but it's been a long time since i've had to mess w/ them...
-
rmustacc
No, generally it's a 'ddi-forceattach=1' in the .conf.
-
rmustacc
That's what happens in all the USB HCDs, etc.
-
andyf
jbk I have followed the instructions in
wesolows.dtrace.org/2013/12/28/anonymous-tracing-on-smartos before. Which reminds me that I need to start an IPD that will help with a mechanism to make this transparent.
-
jbk
ahh nice... looks like i was pretty much on the right path..
-
jbk
didn't have enough time last night to start on it, but hopefully will be able to get to it sometime this week
-
jbk
i had tried to use kmdb to track down my kcf issue, but it's happening far enough in the boot process where a bunch of kernel threads are running (and I think at least some userland), so while tracing one thread, a different thread causes the panic (and the context I'm trying to look at is gone by then)
-
jbk
but not so late enough I can login to the console and run dtrace
-
jbk
so anonymous dtrace seems like a better option
-
jbk
(and probably more fun too, and since this is just a personal project... at least for now)
-
rmustacc
That is the kind of thing I'd look to anonymous DTrace for.
-
gitomat
[illumos-gate] 16070 Drivers for ATTO Celerity FC-162E Gen 5 and Celerity FC-162P Gen 6 16GB FC cards support -- Hans Rosenfeld <hans.rosenfeld⊙nc>
-
gitomat
[illumos-gate] 16078 `fcadm force-lip wwn`: Failed to reinitialize the link of Emulex HBA -- Rick McNeal <rick.mcneal⊙nc>
-
danmcd
@Woodstock do these need driver_aliases entries in packaging? Or does that come in a subsequent push?
-
Woodstock
i have no idea :)
-
Woodstock
i don't see any new driver_aliases for those in omnios, which is were these commits were coming from
-
danmcd
Ahh beautiful. Thank you.
-
danmcd
-
fenix
→ CODE REVIEW 3154: 7288 qlt/qlc integration & cleanup (NEW) |
illumos.org/issues/7288
-
Woodstock
yes, but thats an entirely different driver :)
-
Woodstock
hm. i just found the missing driver_aliases entries.
-
Woodstock
i guess whoever used these cards with omnios added them themselves.
-
danmcd
Wow.
-
danmcd
File a followup, if nobody noticed with Omni, nobody will notice with -gate.
-
danmcd
And don't make it a high-priority.