Description of problem:
Every HVM guest run spawns a qemu-dm process to manage the device model. This
process saves logs into a file /var/log/xen/qemu-dm-<PID>.log. There are two
problems with this. Firfst there is an unbounded number of log files created &
never cleaned, a minor point. More importantly & the point of this ticket, is
that there's no way to correlate the log file back to the domain being run.

If a HVM guest crashes or otherwise misbehaves it would be desirable to capture
the qemu-dm logfile, however, there's no easy way find out which for many log
files corresponds to the domain being run. This problem would be trivially
solved if the log file was named based on the guest name, rather than PID of the
qemu-dm process. This would also cut down on the number of log files created,
since the same log would be used each time a domain was booted.

Version-Release number of selected component (if applicable):
xen-3.0.2-33

How reproducible:
Always

Steps to Reproduce:
1. Create an HVM guest called 'demo'
2.
3.
  
Actual results:
Log file is /var/log/xen/qemu-dm-<some random PID number>.log

Expected results:
Log file is /var/log/xen/qemu-dm-demo.log

Additional info:
The logfile naming scheme appears to be set in the file tools/ioemu/vl.c