Agilent E7495 linux root account

[DogP]

I picked up an E7495A described as "LOCKS UP ON POWER UP." I suspected that the firmware was corrupted and it would be easy enough to flash the latest A.06.25 version. I've tried using both a Sandisk 512MB and a 8Gb Compact Flash formatted to FAT32. With no luck

Did you try doing a fresh firmware install?  When you put the CF card in with the 6.25 update, and are promted to install... press the bottom right button, across from NO.  I believe that does a rebuild of the drive.  That error does look more like a Linux/disk kinda error than a general hardware error (other than bad flash maybe).

Other than that, if you can connect into it over a serial port, maybe you can check the logs for more specifics on the crash?

DogP

[kirill_ka]

Hello Everybody,

I picked up an E7495A described as "LOCKS UP ON POWER UP." I suspected that the firmware was corrupted and it would be easy enough to flash the latest A.06.25 version. I've tried using both a Sandisk 512MB and a 8Gb Compact Flash formatted to FAT32. With no luck

So, I followed the "Update via Load Firmware Utility" procedure described in the Service Guide. I removed the RF section and set the DIP Switch as instructed. I tried again with both Sandisk CFs, but my E7495A would only turn on with a blank display. Is it possible this factory reset requires a PCMCIA instead of CF?

That's the right thing to do. I don't know if it's supposed to work from CF. If it doesn't boot kernel anymore, "utility" was able to erase it at least...

I'm beginning to suspect there might be a hardware problem. I also noticed that there was a red LED lit on the controller board. I'm hoping for any suggestions before I start doing a component level debug tomorrow.

I think that "load firmware utility" should write something to the screen or serial port. Maybe you should try it again and check the output?

[lectrohamlincoln]

I was able to get serial port 2 working. It says all DRAM tests have passed at least. If anyone's curious the baud rate is 38400 and it expects a return CR on transmitted text.

I also tried the A.06.25 firmware on a CF to PCMCIA adapter to no avail. I know its loaded too because without the PCMCIA inserted it says "There is no PCMCIA card on Socket 0"

With DIP switch set to 1&2 I get the same error message on the bottom of the screen: "Kernel Panic: No init found. Try passing init= option to kernel"

With DIP switch set to 2&3, 1&3 or 1&2&3, I get the following menu options from the serial terminal:

--
[1] Display Register
[2] Set Register
[3] Dump Memory
[4] Set Memory
[5] Erase Blocks of Flash
[E] Set Ethernet MAC Address

Select Menu & Hit Enter Key:
--

The "Dump Memory" returns real looking values. I'm unsure how to use the "Set Memory" and I feel a little uneasy using the "Erase Blocks of Flash."

I wish we had a board level Maintenance Manual. I just have a feeling that Red LED means somethings gone bad. I was able to find this brief BOOT ROM manual that seems to correspond to the E7495 http://support.eurotech-inc.com/developers/linux/files/boot/bootrom.pdf

I'll do some more investigating this weekend.

[DogP]

Did you try the full fresh reinstall (not just the standard firmware upgrade)?  IMO, it sounds like a file system problem... you can read a bit about the "No init found" message here: http://lxr.free-electrons.com/source/Documentation/init.txt?v=3.13 .

DogP

[lectrohamlincoln]

I believe I am doing a full reinstall. The boot process never gets to stage where I am prompted with a Yes or No about firmware upgrades and the buttons are all unresponsive.

The only sign of life I have gotten out the E7495 is through the Serial Terminal after following what the Service Guide calls a "Update via Load Firmware Utility." Which is "used to reload firmware on an E7495A/B just as it is loaded at the factory," according to the Guide.

Is there another way to initiate a full fresh reinstall? I scoured the manual and can't find any reference. I fear my only option might be to Block Erase the Flash and hope for the best. But, I think I'm first going to write a program to try and download the flash before I do anything irreversible.

[DogP]

Ah... if it never actually gets to the yes/no, then it's probably not actually doing any sort of reinstall.

I think you want to do switch 1 and 4 off, switch 2 and 3 on, as described in the manual.  I believe that will reconfigure the board to boot Linux from the PCMCIA card, rather than internal flash.  With the firmware upgade CF card in a PCMCIA adapter in the PCMCIA slot, it should boot.  You mentioned that you get a debug menu over the serial port... does nothing show up on the screen when you do that?  And the debug menu is all you can do... there's no way to log into a terminal over the serial port?

One thing you might want to try is the OLD firmware version (1.60) that's still available on the Keysight website: http://www.keysight.com/main/software.jspx?ckey=sw411&lc=eng&cc=US&nid=-11143.0.00&id=sw411 .  In the root of that zip is a script called reloadFFS, which does a reinstall.  In version 6.25, that file exists, but is empty.  This might be the key difference between the instructions and what you're seeing.  There's also readme describes how to use it.

This contains directions on how to recover a /flash file system that
has been corrupted.

WARNING: DO NOT DO THIS UNLESS ABSOLUTELY NECESSARY!  IF YOU'RE NOT
SURE, DON'T DO IT.

To use reloadFFS:

- Boot the elgato
- Log in as root through the serial port
  - This assumes that you can log in.  If you can't then this won't work.
  - If you must log in through telnet follow "Log in through telnet"
    instructions below.
- Install PCMCIA card
- "mount /xdrive"
- "sh /xdrive/reloadFFS"
- Follow instructions given by script



Log in through telnet:

- Log in as root
- mv /etc/passwd.old /etc/passwd
- mv /flash/linux /flash/linux-
- reboot -f
- After reboot is complete (30 seconds or more) log back in as root
- Continue instructions above

Pat

[TheSteve]

I don't have an E7495A but I do have the close cousin the N1996A. Firmware updates on it are generally done via USB. If you have a failure during flashing there is an internal PCMCIA slot that can be used to recover the unit. The instructions do say a CF card is supported via an adapter. You do need a different image compared to the USB image though which has the update and a boot image with a kernel etc.
Is there a special PCMCIA image for the E7495? And if so does it also have cmdline.txt, ramdisk.gz, register.txt, version and zImage?

If you don't have those files check out the USB and PCMCIA updates from the N1996A - maybe you can borrow the needed boot files.

[kirill_ka]

With DIP switch set to 1&2 I get the same error message on the bottom of the screen: "Kernel Panic: No init found. Try passing init= option to kernel"

It means that flash filesystem is corrupt (because of software error, or flash is failing). I can't imagine how other hardware problem can prevent linux from finding init executable.

I wish we had a board level Maintenance Manual. I just have a feeling that Red LED means somethings gone bad. I was able to find this brief BOOT ROM manual that seems to correspond to the E7495 http://support.eurotech-inc.com/developers/linux/files/boot/bootrom.pdf

"Then 2 & 3 ON Copy any/all of zImage, ramdisk.gz and flashfs1 file to Flash ROM and
copy zImage & ramdisk.gz from Flash ROM to RAM and Boot from RAM"
It's clear that it expects zImage and ramdisk.gz on the card. Note that bootloaders a quite picky about filesystem format. Some require fat16 for example. Hopefully it would say something if it can't read the files.

[lectrohamlincoln]

The filesystem is definitely corrupted. Most likely the ram disk is corrupted in a way that the kernel can't load it. The start-up log has the following, "RAMDISK: Couldn't find valid RAM disk image starting at 0"  (I can copy the entire boot log if anyone is curious)

The 1.60 firmware did not work. It definitely detects the PCMCIA when DIP 2&3 are set, but does not boot from it. It just gets to Serial Terminal with a the menu options I described in the previous post.

I'm a little confused by the 1.60 directions. What does it mean by "To use reloadFFS: - Boot the elgato" Is "elgato" just referring to the E7495?

Does anyone know how to recover from a corrupted filesystem? The only ideas I have right now are to use the Serial Terminal menu to load in data at 38400 baud, or to find a JTAG header for the SA1100 chip. In either case I'm not sure what I would load, or even what address I would load it to 

[DogP]

Yeah, I don't think loading memory locations is gonna help, since it's very unlikely that the flash is simply fully memory mapped to the CPU.

Have you tried plugging it into the network to see if you can telnet into it when you boot from the PCMCIA slot?  Obviously you'll need to check your router to see if DHCP gave out an IP address.

And if it is booting from PCMCIA, then another possibility is modifying the ramdisk file on the card.  You should be able to un-gzip the ramdisk.gz and mount the ramdisk (on Linux of course).  Then make changes (i.e. automatically run upgrade scripts, give you a terminal over the serial port instead of the debug menu, etc.), re-gzip the file and put it back on the card.

When you run from the PCMCIA slot, does you see any of the boot process on the screen, or does it stay blank?

And yeah, elgato is their name for the E7495.

Pat

[lectrohamlincoln]

I tried plugging it into the network again no luck. AFAIK it never boots from the PCMCIA.

If DIP switch 2&3 are set, it looks like it just boots from the 1Mib flash ROM (the AT29LV1024 chip). The screen is blank and I only get a Serial Menu with a few crappy option and no way of booting from the PCMCIA or getting an actual command line. Although, it does detect when a PCMCIA card is not inserted (it prints an extra line over Serial saying the PCMCIA card is missing).

If DIP 1&2 are set, it gets part way through the Kernel boot then hangs. No command line, no option for firmware upgrade, no difference if the PCMCIA card is inserted.

Is there some special formatting of the PCMCIA necessary for recovery?

[kirill_ka]

The filesystem is definitely corrupted. Most likely the ram disk is corrupted in a way that the kernel can't load it. The start-up log has the following, "RAMDISK: Couldn't find valid RAM disk image starting at 0"  (I can copy the entire boot log if anyone is curious)

Does it mean something's changed from your first attempt? There's "no init" error on the screen shot you posted earlier.

[kirill_ka]

Is there some special formatting of the PCMCIA necessary for recovery?

I'd try different cards formatted as FAT16.

[kirill_ka]

So this is quite odd, since Linux itself has no trouble with the current date, but the egServer process seems to be rolling it over zero to 'less than the minimum'. I guess egServer must be internally representing the time in a nonstandard way. Maybe just patching the comparison would make it work again?

No. What we see here is "GPS week rollover" problem.
http://www.npl.co.uk/reference/faqs/when-and-what-is-the-gps-week-rollover-problem-(faq-time)
The message doesn't show the time which is "less than minimum". I verified that it's current time minus 1024 weeks.
So GPS subsystem adjusts system clock and then it's immediately reset to "minimum time".

[kirill_ka]

Apply following patch to elgato binary to fix GPS time synchronisation:
0x0122880: replace 4b2585e2152882e2 with de2585e23f2882e2

[technogeeky]

Apply following patch to elgato binary to fix GPS time synchronisation:
0x0122880: replace 4b2585e2152882e2 with de2585e23f2882e2

This is why, as a citizen of the United States, I can't stay mad at Russians. The dopest hackers on the planet.

See below, finally we can calibrate out the expected timebase drift! Thanks kirill_ka. We all owe you one.

[DogP]

Apply following patch to elgato binary to fix GPS time synchronisation:
0x0122880: replace 4b2585e2152882e2 with de2585e23f2882e2

Awesome... I'll give this a try shortly!  And we get to look forward to this problem again in 2 years (got a patch ready for the 2019 rollover?).

Thanks,
Pat

[kirill_ka]

Awesome... I'll give this a try shortly!  And we get to look forward to this problem again in 2 years (got a patch ready for the 2019 rollover?).

I think the receiver was designed to operate before 1999. So it's thrown backwards not the same time GPS week count wraps. I.e. it should work for another 20 years.
In 2038 old *NIX stuff will die anyway

[DogP]

Just tested it, and it worked great!  For anyone keeping track, I let the unit warm up for 15 minutes (with the GPS antenna connected), like the instructions say.  From the calibration, my DAC setting went from 2162 (factory) to 2121... so a similar change to what technogeeky had.  I left it on the adjustment screen and clicked the repeat button a few times over the last 20 minutes, and it has continued to climb (it's up to 2190 now).  I'm gonna leave it on overnight and check it in the morning.

Thanks again!

Pat

[DogP]

And, a final update... after letting it run overnight, it seems to have landed at a DAC setting of 2223 (it bounces around +/- a few values, but hasn't drifted over the last 45 minutes).

Pat

[alpha89]

Apply following patch to elgato binary to fix GPS time synchronisation:
0x0122880: replace 4b2585e2152882e2 with de2585e23f2882e2

Where is the elgato binary file located?  Thanks for figuring this out as my box is needing a frequency
calibration.

[technogeeky]

I spent some time the other day looking for any spurs, and I only found one. I've attached two screenshots; the first is at RBW of 50 Hz with 10 averages and the second is at RBW of 10 Hz.

Can anyone else confirm this spur? Does anyone know of other spurs?

[usagi]

I spent some time the other day looking for any spurs, and I only found one. I've attached two screenshots; the first is at RBW of 50 Hz with 10 averages and the second is at RBW of 10 Hz.

Can anyone else confirm this spur? Does anyone know of other spurs?

spur at 396.8 is probably the LO.

[hawaii596]

New to forum. I have a couple of Agilent E7495B units with lots of accessory, the backpack, and transit case that I am preparing to issue out. They have a problem that they are not able to sinchronize wih GPS clock. According to one of the technicians who looked at them, it is because the onboard dates are too old (more than 12 years), and they can't synch to GPS if there is more than 12 years between today (May 8, 2017) and the onboard date.

I am not a programmer or engineer; just a metrologist in an instrument lab.  Due in part to things I read at this site, I figured out how to connect to the E7495B with Telnet. I was able to log in as "root," and using LINUX commands, check and reset both the normal date and hardware clock. But once I rebooted the system, both reset to some sort of base date. The Hardware clock, back to 1970, and the system date back to 2008. I am guessing the 1970 is a base date for the hardware, and 2008 is the initial release date of the installed firmware.

This leads me to think perhaps there may be a bad CPU battery.  However, I have been unable to physically find one in the instrument. And at least one user on a Keysight (Agilent) thread said there is no CPU battery. 

If anyone has experience or understanding about this, I would sincerely appreciate any of your thoughts. Is there a CPU battery?  If not, is there a reason why the dates are reverting?

Also, could it be that once I manually update both of these dates in LINUX, do I need to do a GPS synch to get them to take more permanently (could this be the need)?  Thanks one and all for any inputs.

[usagi]

If not, is there a reason why the dates are reverting?

read this entire thread, especially this post.