Re: Assorted Bus Errors - common problem?
"Assorted"? There's more than one kind? Perhaps "myriad" or "unusual"?
On Thu, Mar 14, 2002 at 07:19:00AM -0500, MIke Culbertson wrote:
> The general question I am asking is whether or not these bus errors are a
> common sparc/linux related problem, or if it is something uncommon that may
> be related to a specific configuration/software version/kernel. I would
> just as soon see if this is a core problem with sparc/linux before I clutter
> the lists with 100 strace outputs. I will, however, be more than happy to
> provide any info if someone wants to see it. Thanks in advance.
I'm fairly new to sparc, but I hope my advice isn't completely useless.
First, I'll assume you don't have any hardware problems (bad memory,
overheating, et cetera).
No, it's not a problem that is endemic to all architectures, and I haven't
seen it personally in a while. I don't know what word-size your CPU has,
but I'll assume it's 64-bit for the moment. I suggest that the programs
you're using (or libraries that the programs use) arent 64-bit safe, and
they might be asking the system for memory at an offset that doesn't fall
along a word boundry. Thus, a SIGBUS.
Finally, debugging information would help, but strace is useless in this
respect -- it will only show system calls, and if one of those were causing
a SIGBUS, you'd be in deeper trouble. Instead, find a situation in which
you can reproduce an error, and apt-get the source of it, confirm you can
reproduce it there, with debuging symbols turned on. Then, run it from
'gdb'. When it fails, list a 'backtrace' and tell us about it. If it
fails in a library function, get the source of that library, too. Rince.
There may be some architecture-specific voodoo that I don't know, which
will help, about which I'll defer to anyone else here.
Chad Miller <email@example.com>
``Having a smoking section in a restaurant is
like having a peeing section in a pool.''