RE: [Xen-ia64-devel] [PATCH] console

["Magenheimer, Dan (HP Labs Fort Collins)" <dan.magenheimer@xxxxxx>]

Did you intend to add a patch for this or was this
just for discussion?

> -----Original Message-----
> From: Tristan Gingold [mailto:Tristan.Gingold@xxxxxxxx] 
> Sent: Thursday, November 03, 2005 4:35 AM
> To: Tian, Kevin; Magenheimer, Dan (HP Labs Fort Collins)
> Cc: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> Subject: Re: [Xen-ia64-devel] [PATCH] console
> 
> Le Jeudi 03 Novembre 2005 04:56, Tian, Kevin a Ãcrit :
> > From: Tristan Gingold
> >
> > >Sent: 2005å11æ3æ 0:46
> > >
> > >Hi,
> > >
> > >here is a small patch which fixes the rejected patch during kernel
> > > compilation and a bug in evtchn_ia64.c
> > >
> > >I am now able to use xencons:
> > >* do not config serial port in linux
> > >* use 'console=ttyS0' in xen-linux command line.
> >
> > This is absolutely a good patch. Previously I also saw 
> "WARNING: failed to
> Thanks !
> [...]
> > >Note: without more modifs, xenconsole is output only.
> 
> > Actually up to this point, I begin to understand your 
> concern about irq
> > virtualization raised before. ;-) IMO, first we need change 
> Xen to register
> > interrupt handler for serial input interrupt like adding a
> > serial_init_postirq. Regarding xenlinux side, maybe we 
> don't need change
> > anything. Currently the VIRQ_CONSOLE is bound to event. And 
> once Xen gets
> > control upon serial port, current serial driver in xen will send
> > VIRQ_CONSOLE to dom0 in receive handler. In that case, 
> xenconsole can still
> > get input by event notification. How do you think?
> Locally I can now use xenconsole as input/output.
> 
> Here are my changes:
> * iosapic.[ch] copied from Linux to Xen; acpi.c was modified 
> (#if 0 previously 
> added removed) to init interrupts.  [This part is really safe].
> 
> * xensetup.c modified to call initialize_keytable and 
> serial_init_postirq.
> 
> * xen_do_IRQ modified to call xen interrupts handlers if the 
> handler is local.
> 
> * Linux modified to *not* initialize isa IRQ 3.
> 
> The first three points are clean, the last point is a real 
> hack.  We have to 
> handle interrupts in HV to make the last point disappear.
> 
> Tristan.
> 
> 

_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel