[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [Xen-ia64-devel] [PATCH] About fault handle in nested interrupt


  • To: "Tian, Kevin" <kevin.tian@xxxxxxxxx>, <xen-ia64-devel@xxxxxxxxxxxxxxxxxxx>
  • From: "Magenheimer, Dan (HP Labs Fort Collins)" <dan.magenheimer@xxxxxx>
  • Date: Mon, 28 Nov 2005 15:47:29 -0800
  • Delivery-date: Mon, 28 Nov 2005 23:47:18 +0000
  • List-id: Discussion of the ia64 port of Xen <xen-ia64-devel.lists.xensource.com>
  • Thread-index: AcXwvvwWKB0inxTAQIupbPqx5u57XgDtvUxw
  • Thread-topic: [Xen-ia64-devel] [PATCH] About fault handle in nested interrupt

Agreed, faults in Xen should never be reflected to a
guest, so this is a bug.  Should we use BUG_ON or
panic, instead of panic_domain?  

> -----Original Message-----
> From: xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx 
> [mailto:xen-ia64-devel-bounces@xxxxxxxxxxxxxxxxxxx] On Behalf 
> Of Tian, Kevin
> Sent: Wednesday, November 23, 2005 11:19 PM
> To: xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
> Subject: [Xen-ia64-devel] [PATCH] About fault handle in 
> nested interrupt
> 
> Hi, Dan,
>       When looking at the unaligned issue raised by Kanno, I think
> there's an issue to handle nested fault happened in the hypervisor.
> 
>       Take above unaligned fault for example. One bogus structure is
> allocated unaligned, and then Xen tries to access that unaligned
> structure. Unaligned fault happens within Xen, and finally 
> flow goes to
> ia64_handle_reflection. There's a check upon ipsr.cpl to see whether
> fault from Xen itself. If true, only a warning is printed out and
> subsequent lines will still be executed just as fault from guest.
> Finally reflect_interruption is called and pt_regs.iip is changed to
> unaligned offset within guest IVT table.
> 
>       Actually you can see from this flow that everything will get
> messed after fault is injected into guest, since incorrect register
> contents of Xen are recovered to guest. This will bring developer much
> confusion about when/where things go wrong exactly, since more faults
> will be raised due to above incorrect injection.
> 
>       Normally the faults reflected back to guest directly shouldn't
> happen within Xen, or else that means a bug and we need panic and dump
> the stack at that point immediately. This can make people easier to
> track bugs. ;-)
> 
> Signed-off-by Kevin Tian <Kevin.tian@xxxxxxxxx>
> 
> Thanks,
> Kevin
> 

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


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.