Regression observations & misc errata

[Q]

To anyone who might be interested,

I have moved my Dell Latitude L400 from running 4.7-STABLE several weeks
ago to 5.0-CURRENT (after a very brief stopover at 5.0-RELEASE)

This is what I have observed after a few weeks of testing:

When using APM on 4.7-STABLE the internal cooling didn't seem to
function as it did under windows, resulting in the case getting VERY
warm at times. Suspend worked regardless of video mode, although X11
would need "flicking" to a virtual console to get the mouse pointer
back. Everything seemed to function fine, although I did experience
random reboots periodically when the CPU was hot.

Under 5.0-CURRENT/RELEASE the default behaviour to use ACPI appears to
cause problems with thermal control, and causes random panics. On
several occasions running ACPI I have seen warnings of the CPU reaching
99deg C, followed by an immediate ungraceful power off.  Suspending does
not work properly when using ACPI, and attempting to use it usually
results in a system hang.  I have attributed this to be a buggy ACPI
BIOS implementation in this machine, but would be happy to help anyone
who thinks they can fix these problems.

Switching to APM under 5.0-CURRENT seems to resolve most of these
issues, with the thermal fan functioning more as you would expect. The
case still gets quite warm, but I have not observed any heat related
incidents so far. However suspend will NOT work if X11 is running
without first switching to a vty. Attempting to suspend in X11 will
result in the console becoming unresponsive, and causing an incessant
beep when attempting to switch to a vty, requiring a reboot.

Other than a slight reduction in power management functionality (ie.
suspending in X no longer works), everything else seems to work fine.

A couple of other random things I have observed/found:

o Particularly bad "crashes" can leave the filesystem in such a state
  that continuing without a foreground fsck will cause a kernel panic
  when the effected files are accessed.
  I had a /var that exhibited this problem, only solution to panic was
  single user mode boot and 'fsck -y'  (no tracelog sorry)
o Using DHCP will not work with a PCMCIA network card if you also run
  DHCP on a fixed interface because dhclient is already running and
  pccard_ether (called by devd) will not attempt to kill it first. 
  (This is not a new problem)
o omshell(1) has now been included, however OMAPI does not appear to be
  functional within dhclient(8) at this time.
o The dhclient(8) manpage incorrectly lists omshell(1) as omshell(8)
  when detailing the '-w' flag
o There is no way to tell smbutil/smbfs which interface to use for
  NetBIOS broadcasts if you have more than one interface. Downing the
  undesired interfaces results in errors.

-- 
Seeya...Q

               -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
                        
                          _____  /   Quinton Dolan q_dolan at yahoo.com.au
  __  __/  /   /   __/   /      /          
     /    __  /   _/    /      /         Gold Coast, QLD, Australia
  __/  __/ __/ ____/   /   -  /             Ph: +61 419 729 806
                    _______  /              
                            _\