[nSLUG] Machine Freezes on PCMCIA Card Insertion

George N. White III gnwiii at gmail.com
Thu Feb 22 08:22:47 AST 2007


On 2/22/07, D G Teed <donald.teed at gmail.com> wrote:

> Interesting what you say about stack size.  I've experienced a particular
> notebook and driver that would hang quietly very easily, while
> NDISwrapper was being used.  I asked the developer why he didn't
> code it so it produced a kernel stack dump.  Is the symptom of
> a stack overflow that it will not produce a stack trace when
> the kernel hangs?  I had kinda assumed the NDISwrapper developer
> wasn't using the proper error handler for kernel drivers.

I gather (from discussions on the XFS lists) that the overhead of
detecting stack overflow is too much to be practical for i386 hardware
(because i386 has few registers, function calls use the stack, so
every function call would need to ensure enough free stack to hold the
arguments).  This wasn't a problem on RISC machines where XFS
originated.

-- 
George N. White III <aa056 at chebucto.ns.ca>
Head of St. Margarets Bay, Nova Scotia

!DSPAM:45dd8b1c121966004818286!




More information about the nSLUG mailing list