Reading Kernel ram after a crash...

Jan 10, 2006 4 Replies

I have a need to read the kernel's ram following a crash. I am running on an embedded platform and am using an nfs filesystem.



What tools are available to do that? Any hints or links would be appreciated.



Thanks! Bill


-- Posted via NewsDemon.com - Premium Uncensored Newsgroup Service ------->>>>>>

formatting link


There are many different embedded platforms.

Maybe JTAG can be used if that is supported on the embedded platform.

Regards Rune

I think I need to be more specific.

I am on what appears to the kernel to be a fairly generic x86 platform. It is connected to the host machine via serial port and ethernet. It both network boots (BOOTP) and has a network file system (NFS). I am running MontaVista embedded linux. KDB is not supplied with this kernel. KGDB doesn't seem to fit - at least not without a lot of tweaking.

I have a two level driver, lower level talks to a FPGA that looks like a serial port. The data is then passed to the upper lever which handles the protocol.

I am crashing in interrupt state. The system no longer responds. I need access to memory as I have a trace buffer there that will point me to the problem. I need a way to access the memory, possibly through some simple command processor.

so - my questions are:

1) does code like that exist - either for built in or module. 2) how can I get control to it after a crash. 3) How can I force a dump of core. 4) If I force a dump, is there a way to read it by address as it existed in memory?

I hope this is more meaningful.

Thanks! Bill

-- Posted via NewsDemon.com - Premium Uncensored Newsgroup Service ------->>>>>>

formatting link

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required