Re: port-xen/52016: Mismatch between mem_clusters and pmap_enter bounds check

[John Nemeth <jnemeth%cue.bc.ca@localhost>]

On Mar 1, 11:15am, Manuel Bouyer wrote:
}
} The following reply was made to PR port-xen/52016; it has been noted by GNATS.
} 
} From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
} To: coypu%SDF.ORG@localhost
} Cc: gnats-bugs%NetBSD.org@localhost, port-xen-maintainer%netbsd.org@localhost
} Date: Wed, 1 Mar 2017 12:13:10 +0100
} 
}  On Wed, Mar 01, 2017 at 09:55:40AM +0000, coypu%SDF.ORG@localhost wrote:
}  > On Wed, Mar 01, 2017 at 09:35:01AM +0000, Manuel Bouyer wrote:
}  > >  AFAIK X11 uses /dev/mem to access the hardware.
}  > >  
}  > 
}  > Huh, so it does.
}  > I guess X doesn't currently fail because it's already mapped
}  > by the boot process, so no fault and no trip to pmap_enter?
}  
}  No, it works because the hypervisor allows mapping the video
}  harware address. I don't know why it rejects the addresses used by dmidecode

     I haven't done a deep dive, but based on comments here and
elsewhere, it is likely that dmidecode is trying to access the
"BIOS ROM" and in PV mode there is no BIOS.  We need a way for
mappings done via /dev/mem to fail gracefully (i.e. not panic the
system).

}-- End of excerpt from Manuel Bouyer