HP / Agilent 16700A / 16700B / 16702A / 16702B root password

[gslick]

If you do a clean install of the OS from scratch on an HP / Agilent 16700A / 16700B / 16702A / 16702B logic analyzer mainframe you should end up with an /etc/passwd file that looks like this:

Code: [Select]

root:J8HyPr7uddHCc:0:3::/:/sbin/sh
daemon:*:1:5::/:/sbin/sh
bin:*:2:2::/usr/bin:/sbin/sh
sys:*:3:3::/:
adm:*:4:4::/var/adm:/sbin/sh
uucp:*:5:3::/var/spool/uucppublic:/usr/lbin/uucp/uucico
lp:*:9:7::/var/spool/lp:/sbin/sh
nobody:*:-2:-24::/:
logic::16700:16700:Logic default login:/logic:/usr/sprockets/bin/sessionWrapper
hplogic::16700:16700:Logic default login:/logic:/usr/sprockets/bin/sessionWrapper
hplogicz:*:16700:16700:Logic default login:/home/hplogic:/bin/sh
ftp:*:16700:16700:Anonymous FTP:/logic:/usr/bin/false
www:*:15:15:World Wide Web:/home/www:/usr/bin/false
It turns out that the default password for that root entry is: "uh,uhuh"

I made a few unsuccessful attempts to crack that myself before I was recently given this tip from someone else.

So if you want to do some hacking on your 16700A / 16700B / 16702A / 16702B and assuming it still has the default root password, you don't have to boot into single user mode to change it there. Just use the Control+Alt+Shift+F8 keyboard shortcut to bring up an xterm window and then you can su to root using the default password and hack away.

And of course if you have the system configured on an open network you might want to change the default root password if you haven't already done so.

Have fun.

[smkb]

Hi,

Excellent tip. Seems that it is possible to open a shell from PS2 keyboard only. I could not open shell window from vnc session.
And in the opposite to e.g. linux passwd, the hpux passwd must be given root user name explicitly.

Thanks.

[gslick]

Hi,

Excellent tip. Seems that it is possible to open a shell from PS2 keyboard only. I could not open shell window from vnc session.
And in the opposite to e.g. linux passwd, the hpux passwd must be given root user name explicitly.

Thanks.

I just tried this now and I was able to use the Control+Alt+Shift+F8 keyboard shortcut to bring up an xterm window when using VNC-Viewer-5.2.3-Windows-64bit.exe running on a Windows 8.1 desktop.

Looks like the current version now is 5.3.3  I haven't updated yet.
https://www.realvnc.com/download/viewer/

When I use the su command I don't have to explicitly specify the user root. This is after a clean install of the 16700 OS from scratch. Maybe that can change if the configuration is altered, such as creating additional user accounts. I haven't tried that.

If the the Control+Alt+Shift+F8 keyboard shortcut doesn't work maybe there are configuration options in the VNC viewer to control whether or not F8 and other keys are passed through. By default PrintScreen was passed through with my VNC viewer and I had to disable that in order to be able to use Alt+PrintScreen on my Windows desktop to capture the VNC viewer window image.

[smkb]

I see. I can't confirm because I do not use windows at all instead of I work with openSuse very long time.
I couldn't open the xterm with KRDC  4.6 (kde vnc/rdp client) which is rather old tool. There are many more vnc clients. It is poss these do better.
Maybe it is just  a matter of enabling passing over some keys combinations to vnc server.

I meant that after I logged in as root I had to use "passwd root" because it turned out that simple "passwd" changes password for "logic" user despite of that it had been used from root user context.
That was something confusing me thus I wrote about it.

[PaulAm]

I just got gcc running on my 16702.  I'm remembering why I'm glad sysadmin is not my primary job.

There are still HPUX 10.20 binaries available at 
    http://mirrors.develooper.com/hpux/downloads.html.
There may be others, but that's the only repository I can find that's still online.  Everything else starts with 11.0

Also Dave Partridge's hp16700 page has lots of useful information
     http://www.perdrix.co.uk/hp16700/

Once you get rid of that overbearing sessionMgr and can run the logic analyzer (vp) natively over a remote X session, it's pretty nice.  I'm working at turning my system from a logic analyzer with an embedded unix processor to a Unix system with a logic analyzer application.  That makes it much more useful, to me at least.

[smkb]

Well done!,

Did you install the gcc just to give a try ? I am just curious and
 I wonder if you can e.g. recompile the vp, can you ? (means: do you have soure code?)

I use 16702 as a measurement tool mainly. Haven't done any funny HPUX pa-risc stress tests yet.

and yes, Dave's page is full of usefull information regarding 16700s

I see. Seems it is interesting idea. I recall I tried once to setup native X-session but noticed some protocol miscompatibility with my linux server (client ..) so gave up.
It was long time ago.
VNC client in KRDC works fine. I noticed only that sometimes a scope window loops on scrolling a plot if time base is changed too frequently.
It is possible to restore normal operation by tearing down current VNC session. New one can be opened later.

[PaulAm]

My goal is to make the system usable for a programmer.  It also makes it easier to integrate the system into my network.

The primary immediate reason for gcc is to get my version of emacs compiled so I have a decent editor available.  Beyond that, I want to get at least tcl up.  It might be harder to get tk up, I don't think the X header files are there and I'm not sure I want to go through a build of X11R4.  Being able to build native tools to interact with the data on the system seems like a plus to me.  Don't know if I'll ever need that, but I like my Unix systems to have similar environments and tools.

Getting bash installed is a big help though.

I ran a couple of vnc sesisons and the update is pretty jerky.  I found that annoying, but the native X works nicely on a local network.  Maybe I'll post a screen shot of it running remote.

There's no source available for the vp app, as far as I know.

[smkb]

I see. Splendid.
Good luck with X11 devel package. I saw recently there are plenty HP servers on eBay (ePay available e.g. C3700 runnig e.g. HPUX 10.20 maybe this will help.

Regarding post-processing on-site tools: do you have any information regarding native LA file format ? e.g. 16717 and I guess that  other LA cards should use same or nearly same like e.g. SCSI is flexible due to structured design which means every data field follows preceding simple header of type and length.

These can be done also remotely (x86 side) with help of builtin RPI and maybe this project would be something interesting to you:

https://sourceforge.net/projects/tv16k7/

It is only Linux project.
I found that 16700's RPI implementation is not perfect (limited) however it allows still basic postprocessing and data preservation including as many settings as possible but
 e.g. it is not possible to read back trigger settings which can be very sophisticated (multilevelled) and this is what I really enjoy in 16700

I made this project available this year and actually I look for someone who would like to give a try and post a feedback regarding this project, maybe even a code review which I would appreciate.

[PaulAm]

That looks pretty interesting.  I'll put that on my take-a-look-at list 

I'm not a bit fan of Qt though.  Applications tend to be pretty version specific so I have to have multiple Qt versions hanging around to build them.

It's a little out of vogue now, but I prefer to build window apps in Tk; it's simple, portable and even cross platform without too much effort.

[smkb]

You are welcome enjoy

I see. One day I found Qt usefull and started using it. Seems Qt is multi-os supported as well but some #ifdefs will be inevitable I guess. (almost sure of that)
I would not agree with such thesis that Qt becomes less popular.
I reckon Nokia does the best they can do to promote Qt. Latest is 5.6 I have never tried.
For example for QNX Qt is the be-or-not-to-be choice.

[Patrick]

Can anyone help me figure out how I can use XMing on a PC to connect up with my hp 16700A? I can PUTTY to my LA okay, but I just can't get it to connect to the PC via an X-session. I've read through Matt's and Dave Partridge's web site pages, but I seem to be missing something. I'm not interested so much in unlocking it as Dave Partridge details so much as having the ability to simply remotely connect to it via an X-Term program on a PC. Please advise.

[smkb]

I was not successful in setting X session either.
Many claim that X session is faster than vnc (so called "shared session") but I reckon that this is negligible.
Consider this: a trace captured by 16717 in "time transitional", 2M samples, several signals gives local LA file containing configuration and data of size of abt. 21MB.
Loading it back to 16717 digital acquisition card takes half a minute or so.
Moreover if there is present "view" window with big enough time base which results in drawing e.g. 1M samples in the given time <delay - 4*TB, delay + 4*TB> then redrawing complete plot takes a couple of minutes.

I am almost sure it is not a matter of vnc connection. Such operation times are imposed due to limitation of mainframe cpu performance. It is pa-risc clocked at 150MHz which was decent in 2000's but now. We expect much faster responses nowadays.
(and e.g. my tool is a remedy for mainframe slow response)

So be patient. Anyway many acquisition cards are still superb hardware.
I reckon vnc is good enough.

[PaulAm]

When I was playing with this, it seemed the session manager wanted to run a VNC session for an external display.  When I ran the app without the session manager, it worked much better.  Not only was it faster, but I could resize the individual windows up to the size of my display.

You need an X server along with Putty (something like Xming).  Check this link
http://www.geo.mtu.edu/geoschem/docs/putty_install.html

An easy way to test X11 setups is using xclock.  Log in with the root id, set the DISPLAY variable and then use the xclock command.  If it appears on your display, you're good to go.  It's a lot faster than waiting for the main app to spin up.

[PaulAm]

Here's a screenshot using native X.  Response is excellent and who doesn't like a 20 inch DSO screen? 

Response is as good as using a local screen, but you're not limited to the predefined graphics resolutions in the 16700.

That mess in the lower corner is my desktop and the white overlayed window is on my workstation.

[smkb]

well done !.

What will you say about this ? (see screenshot)
I guess I am the only one who can do this by now.
vxfs support has been broken in Linux kernel for damn long time.
I have fixed this, so RO mounts are possible.

If anyone wants to give a try then go here:
https://sourceforge.net/projects/linux-vxfs/

[2x2l]

That "uh,uhuh" was a Jurrassic Park reference I'm pretty sure that the fatman used to protect his SGI boxen.

VNC has to bind to a specific X interface (:0 by default usually), but especially in the "old days" there was a lot of overhead because it was multi-platform. If you terminal in (turn off encryption in a local environment, those RISC rs/6k's machines didn't have any hardware-offloaded encryption so you're wasting tons of cycles) with something like -- ssh -p 8400 -l username -X -v <ip> , then from that terminal, fire up an instance of which ever application. What will end up happening is that all of the X instructions will be passed (a way lower payload of data size-wise) to your own instance of X, which will suffer the "heavy brunt" of rendering/updates. It should be faster than in native.

There's still some silent tk love out there since what it was good at, it was really good at. (Just like there's a semi-healthy Delphi community still making VCLs since it was really good at making line of business applications). Modern GUI frameworks (QT5 + QtQuick + QML) and modern languages (C++11 has threads part of the standard, so that portability headache of win32 threads vs POSIX pthreads is taken care of) makes things a lot easier to write these days.

I haven't used this logic analyser myself but I know it was fast enough to develop PCI-e and do in-circuit non-invasive probing for Pentium 4's and Xeons (you had to max out all them expansion slots and get a specific FPGA aux component from Xilinx or someone, but it was possible). I'm sure there's a way to increase the DAQ such that 21 megabytes doesn't hurt

[smkb]

re movie: I can't recall

Whatever one would do, use native X or stick to vnc, that mainframe 1670x is still too slow.

re Qt, I enjoy using it too and prefer native pthreads then the Qt's encapsulation. Windows ? I don't care.

right. these DAQs have outstanding performance in many ways like operating speed like ~700MHz .. 1200MHz, trace memory capacity and features like trigerring and setting input level voltages.
There are adaptors like e.g. N4220 or FS4120 which allow tracing PCIe or USB busses.

The problem is, that more advanced DAQ is - mainly bigger trace buffer, it will produce more data obviously (full buffer capture) while performance of mainframe cpu is still same and not enugh.