This is the mail archive of the
mailing list for the GDB project.
Re: Multiprocessor remote debugging
- From: Paul Koning <pkoning at equallogic dot com>
- To: hoogerbrugge at hotmail dot com
- Cc: gdb at sources dot redhat dot com
- Date: Fri, 14 Feb 2003 10:05:17 -0500
- Subject: Re: Multiprocessor remote debugging
- References: <F130n2Yf0scL9NtDEJm00022ffe@hotmail.com>
>>>>> "Jan" == Jan Hoogerbrugge <firstname.lastname@example.org> writes:
Jan> Hi, I want to use remote debugging to debug a (shared memory)
Jan> multiprocessor. Does anyone have experience with this? My first
Jan> idea would be to protect the stub with a spinlock so that only
Jan> one processor will be talking to gdb at a time. Any other things
Jan> that I should do?
We've done something slightly different: a separate serial port for
each of the two CPUs. But yes, there is also a lock mechanism to
stall the "other" CPU while one is talking to gdb.
Part of the reason for the two consoles is that the two CPUs run
different code, so we have one instance of GDB for each image.