Resend, [email protected] no longer exists, trying [email protected]. On Mon, 5 Jan 2009 15:24:13 -0600, Martin Hicks <[email protected]> wrote: > >On Mon, Dec 22, 2008 at 04:16:50PM -0500, Christoph Hellwig wrote: >> The kdb arch code for i386 and x86_64 is to 98% essentially the same >> modulo a few tiniy differences that crept in over the years. The real >> differences are: >> >> - the register handling due to the additional registers and slightly >> different ABI on x86_64 >> - setjmp / longjmp is different assembly code >> - i386 has an stackdepth command. I don't really see why it >> doesn't exist for x86_64, but at least right now it's not there > >I'll stick this item on my todo list.
Stackdepth was added to i386 back in the days of 8K stacks which would overflow when using nested filesystems. Nobody asked for that command on x86_64, mainly because x86_64 started off with separate stacks for almost everything. Mind you, given the recent patches to removed the separate x86_64 patches, adding stackdepth to x86_64 might be worth it. >> - x86_64 has a cpu_pda command. AFAIK the PDA doesn't exist on i386 >> in that form, so it's one to stay >> - x86_64 registers a die notifier. This looks like it would apply >> to i386 too, but I'd like to leave it to an expert. > >Any experts around who may know why this is only on x86_64? AFAICR when Andi Kleen was adding AMD64 support to Linux, he defined some die notifiers for AMD64. At the time those die notifiers were not in i386. I took advantage of the notifier to simplify kdb for AMD64. There is no reason not to use the i386 die notifiers now, as long as the semantics are the same. At one stage i386 and x86_64 used the same notifier names but with slightly different meanings. _______________________________________________ kdb mailing list [email protected] http://oss.sgi.com/mailman/listinfo/kdb
