Products > Test Equipment

Sniffing the Rigol's internal I2C bus

<< < (545/899) > >>

bluesmoke:

--- Quote from: true on January 25, 2014, 02:33:04 am ---I would say during - get it started then go take a dump.
--- End quote ---
... It is hard to resist! :)

tirulerbach:

--- Quote from: granz on January 24, 2014, 02:57:52 pm ---
--- Quote from: neslekkim on January 24, 2014, 02:52:01 pm ---good to know, all my computers use 64bit os, so I guess I'll try from an virtualmachine then

--- End quote ---

Good thing to try.  If it works out for you, definitely post about your success, because I'm sure that would make it easier for lots of folks.

--- End quote ---

I made my dumps on a 64 bit systems with no problems at all. I even don't think that 32/64 bit host makes any differences for a cross-toolchain. I suppose that the mentioned problems on that particular case had another cause and the conclusion that it is a 64 bit issue is simply wrong.

granz:

--- Quote from: tirulerbach on January 25, 2014, 12:26:58 pm ---
I made my dumps on a 64 bit systems with no problems at all. I even don't think that 32/64 bit host makes any differences for a cross-toolchain. I suppose that the mentioned problems on that particular case had another cause and the conclusion that it is a 64 bit issue is simply wrong.

--- End quote ---

Well, I completely agree that it *shouldn't* make any difference at all.  The blackfin tools are literally the same pre-compiled binaries no matter what host you are on (they are 32-bit binaries).  The USB kernel-space driver could be very slightly different or something stupid with ioctl() is going on (this is my guess).  bfin-jtag/bfin-gdbproxy gets to the USB JTAG-cable device in this order: libftdi(user-space library)--->libusb(user-space library)--->usbfs(kernel-space driver).  Examining the bfin-jtag and bfin-gdbproxy binaries also shows that they have statically linked libusb and libftdi and they may have a problem such as this in the versions they used:

http://stackoverflow.com/questions/9258280/whats-different-between-compiling-in-32bit-mode-and-64bit-mode-on-64bit-os-abou

I've had these types of issues in the past with my own software at work.  I guess it's not really fair to say it's a 64/32 bit issue per se, probably more like a kernel compatibility configuration issue.  It's just easier to say 64 or 32-bit.

I'm curious: what variety/version of Linux did you use?

Co6aka:

--- Quote from: tirulerbach on January 25, 2014, 12:26:58 pm ---I suppose that the mentioned problems...had another cause...
--- End quote ---

FWIW, most of the 32-on-64 issues I've encountered turned out to be multiarch-related, and tracking them down and fixing them was not often straightforward... |O At least one was an upstream issue, which eventually got fixed; stuff just started working one day, immediately following an update. (Sometimes it's the other way around too... |O) Typically stuff just didn't work or worked incorrectly, with no or little indication of why. Sometimes running GUI apps from the terminal provides a hint, as does pouring over the logs.

(One of my favorite source code comment-gems is "Error reporting could be better." :-DD No, REALLY?)


--- Quote from: granz on January 25, 2014, 02:48:07 pm ---...statically linked libusb and libftdi...
--- End quote ---

"DLL Hell" of the Linux world.

thinwire:
Rigol DS2000(A) series oscilloscopes: When installing an option with usbtmc :SYSTem:OPTion:INSTall or uninstall all options with :SYSTem:OPTion:UNINSTall I detected that all system settings are reset to the default  :-- . To avoid this I wrote a linux command line utility to install/uninstall options with restoring system settings. It is based on libusb-1.0 and does not need a usbtmc kernel driver.

Install an option:

--- Code: ---rigoltmc -o XXXXXXX-XXXXXXX-XXXXXXX-XXXXXXX

--- End code ---

Uninstall all options:

--- Code: ---rigoltmc -u

--- End code ---

Identify the device:

--- Code: ---rigoltmc -i

--- End code ---
This should return something like:

RIGOL TECHNOLOGIES,DS2072A,DS2D15XXXXXXX,00.02.00


Have fun!

Navigation

[0] Message Index

[#] Next page

[*] Previous page

There was an error while thanking
Thanking...
Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod