Re: [XenPPC] fixed bad swap of console and store event channels

[hollisb@xxxxxxxxxxxxx]

Quoting Jimi Xenidis <jimix@xxxxxxxxxxxxxx>:

> I cannot how embarrassed I am by this.
> watching evtchn's fly everywhere seen the wrong one get delivered and
> the right one go no where and the whole time I was swapping the 2
> channels.
Glad you caught it. :)

> anyway, using the two files from:
>     
> http://lists.xensource.com/archives/html/xen-ppc-devel/2006-08/msg00006.html
> Using the command:
>   $ xm create -c dom1
> you should be able to boot a DomU and see its output, all the way until
> it says it cannot find its rootfs.
Unfortunately this isn't working for me: my output stops after
    PID hash table entries: 512 (order: 9, 4096 bytes)

When I bang keys, I can see events going on dom0 evtchn 10/domU evtchn  
2: dom0/10 sends an event to domU/2 (presumably to notify domU of the  
input), and domU/2 sends an event to dom0/10 (presumably to notify  
dom0 that there is new console output to display). Strace on the "xm  
console" process seems to confirm that the incoming event does not  
wake xm console from its select() call.
I'm pretty sure I've rebuilt/reinstalled the various components  
cleanly, twice. Could there be an outstanding diff still?
--
Hollis Blanchard
IBM Linux Technology Center

_______________________________________________
Xen-ppc-devel mailing list
Xen-ppc-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ppc-devel