Hi! This post is long, but this is because I attach logs. I have to profile vxWorks.
I have:
Tornado 2.2.1
vxWorks for PCC 603
SunOS 5.9

What I did:
Establish connection with target from host. Enable symbol synchronization, log:
tgtsvr -V -s -c vxWorks_debug_image 10.74.115.29 -m 50000000
tgtsvr.ex (10.74.115.29@selix045): Fri Sep 2 14:42:46 2011
Checking License ...OK
Connecting to target agent... succeeded.
Attaching C++ interface... succeeded.
Attaching elf OMF reader for PPC CPU family... succeeded.
symbol synchronization: Giving "150.132.35.29" as registry address to the target
Added target_modules to target-server.....done

That looks fine. Now I use launch and then I can use all the tools from Tornado Launcher. I mean I can execute WindSh command on target, I can use "Browser" to check memory and with Spy I can check task CPU usage (all task view correctly) and other such information. MemScope works fine. When i run TraceScope i get all symbols (for VxWorks OS and loaded programs on VxWorks) and i can trace functions. Those tools works fine.

Now, I need to Profile, so i click ProfileScope. I keep default values. When I run Profiler from this place, i get error. I forgot to wrote it down, but it's very similar to error i get in manual profiler starting. I will edit this post soon to fill log error.
I got error, ProfileScope run anyway, but I don't get any information. Main windows is almost always empty. Rarely i got one sample: tLogTask. When i click for task profiling i have only tIdle and tKernel processes. When i click update button I got:
There are 0 unanalyzed samples in the buffer.
There have been 12280 total samples analyzed.
There have been 1075 untraceable samples.
The sampling rate is 107 Hz. The analysis rate is once per 1.0 seconds.
The sampling is being done by the aux clock.
StethoScope sampling is not active.
There have been 0 samples skipped due to full buffers
The maximum stack depth encountered is 0.
The program counter was outside of the memory limits 0 times.
Routine profile:
A total of 0 different routines have been encountered.
Subroutine utilization is summed with the parent function.
Local VxWorks symbols are used.
The routines profile digital filter pole is at 0.950.
Task profile:
A total of 0 different tasks have been encountered.
The task profile digital filter pole is at 0.800.
Tasks are not installed to StethoScope as they are encountered.

Of course this "untraceable samples" to "analyzed samples" proportion changes if i run some programs on vxWorks.

Anyway if automatic GUI loading failed, I was forced to run manual ProfileScope, first i load libraries in WindSh:

-> ld 1 < /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/libansi.so

Loading /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/libansi.so |
value = 1155584 = 0x11a200 = wtxTsInfoGet + 0xc

-> ld 1 < /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/libdummyscope.so

Loading /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/libdummyscope.so |
value = 1150072 = 0x118c78 = wtxSymListByModuleIdGet + 0x288

-> ld 1 < /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/scopeutils.so

Loading /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/scopeutils.so |
Undefined symbols:
evtBufferId
wvEvtLogStop
Warning: object module may not be usable because of undefined symbols.
value = 1148688 = 0x118710 = wtxSymAdd + 0xe4

-> ld 1 < /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/profile.so

Loading /proj/tss/Tools/Tornado_2.2.1/rti/scopetools.4.0c/lib/ppc603Vx5.5gcc/profile.so /
value = 1293872 = 0x13be30 = windTickAnnounce + 0x94

On server console, i got:

Added libansi.so to target............done
Added libdummyscope.so to target............done
Added scopeutils.so to target............done
Added profile.so to target............done

Y, i notice two missing symbols, but I think that, they are not very important.

No i run sampler in WindSh:
-> ProfileScopeInit(5000, 3, 32, 0, sysInit, sysBootLine)
value = 1 = 0x1

On target console I got :
-> 0x1e8bd90 (t1): ProfileFreeze: Bad ProfileInfo value `0x1ea6578'
0x1e8bd90 (t1): Last valid access was to `0x1ea6578'

What that mean?

Sampler process start (t1), so now I try to run profile scope (not from tornado launcher, but from sun console):
profile 10.74.115.29@selix -verbosity 3

On target console I got :
0x1f62d78 (tWdbTask): ProfileFreeze: Bad ProfileInfo value `0x1ea6578'
0x1f62d78 (tWdbTask): Last valid access was to `0x1ea6578'

Well, now situation is the same as when i run ProfileScope from Tornado Launcher. See above. (lot of untraceable samples, empty tree view windows and two tasks).

In desperation i try to run target base profiling. I load same libraries as above, and then (WindSh):
ProfileInit(1,0,0,-1,1024,32,0,0)
ProfilePrompt

Output:
-> 0x1dee230 (t4): ProfileFreeze: Bad ProfileInfo value `0x1f519f0'
0x1dee230 (t4): Last valid access was to `0x1f519f0'
0x1d257d0 (t5): ProfileSetTaskIgnore: Bad ProfileInfo value `0x1f519f0'
0x1d257d0 (t5): Last valid access was to `0x1f519f0'

Next, i try:
show routines

i got:
0x1d257d0 (t5): ProfileShowTree: Bad ProfileInfo value `0x1f519f0'
0x1d257d0 (t5): Last valid access was to `0x1f519f0'

I ask for any advise. Can I test something else? Should i configure my tools in other way? I used different parameters, but it didn't help.
Sorry for poor English skills.
Lerhes