Re: [Xen-devel] 2.6.7-xen0 hang on boot

[Rik van Riel <riel@xxxxxxxxxx>]

On Wed, 11 Aug 2004, Rik van Riel wrote:

> #8697 0x0091fecf in syslog () from /lib/tls/libc.so.6

Init's userspace EIP is 0x0091fef3, which I assume is
also inside syslog(), since /lib/tls/libc-2.3.3.so is
mapped at the same address in each program (prelink).

I can't attach a debugger to it, but alt-sysrq-p works.

Somehow the majority of the time appears to be spent
in kernel space, either page_fault and friends (especially
expand_stack is interesting) or Xen hypervisor calls
(work_notifysig -> do_notify_resume -> evtchn_do_upcall ->
 do_signal -> get_signal_to_deliver -> forc_evtchn_callback).

I guess this is explained by line 1800 of signal.c:

                /* Init gets no signals it doesn't want.  */
                if (current->pid == 1)
                        continue;

My guess is that init should also segfault, but the kernel
doesn't deliver the signal and dives straight back into the
pagefault handler, which doesn't know how to handle the fault,
etc...

The "all time spent in userspace" I reported before is an
artifact of Xen not reporting system time.  It really is
spending its time in kernel space.

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan



-------------------------------------------------------
SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media
100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33
Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift.
http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.sourceforge.net/lists/listinfo/xen-devel