Subject: 1.5ALPHA on ThinkPad 310E
To: Port-i386 <port-i386@NetBSD.ORG>
From: Jukka Marin <jmarin@pyy.jmp.fi>
List: port-i386
Date: 07/14/2000 11:02:46
Hello,

Some info about 1.5 ALPHA on ThinkPad 310E:

APM:

Same problem as with 1.4.x: Battery status is not updated unless the
AC adapter status changes (ie. I plug it in).  If no AC adapter status
changes happen, battery status remains at 100% (or whatever it was when
the adapter was disconnected).

Haven't tested suspend etc. yet (and the machine just locked up, see
below).


PCMCIA/CardBus:

After enabling CardBus in BIOS, I was able to access CompactFLASH cards
(yippee!).  Also, my NE2000 compatible Ethernet card and WebGear Aviator
work (even when CardBus is enabled).  When CardBus was disabled, NetBSD
whined about "no irq being set for CardBus" and when inserting a Compact-
FLASH card, the system said "ready never occurred" and locked up (just
line in the 1.4.x times).

If I mount the CompactFLASH card and then eject it, I get messages like
these on console:

Jul 14 09:09:59 orava /netbsd: cbb1: bad Vcc request. sock_ctrl 0x0, sock_status 0x30000a19
Jul 14 09:10:12 orava /netbsd: cbb1: bad Vcc request. sock_ctrl 0x0, sock_status 0x30000a10
Jul 14 09:10:12 orava /netbsd: cbb1: bad Vcc request. sock_ctrl 0x33, sock_status 0x30000a10

And if I reinsert the card, I get more of those.  If I try to ls the disk
while the card is not present, but it was left mounted, I get various
interesting effects and error messages.  Would it be possible to handle
"removal of removable media while it is mounted" better?  I guess automatic
unmount is impossible, but how about returning a sensible error code to
processes trying to access the media until it has been unmounted?  At the
moment (using 000713 kernel sources), reinserting the card without umount
and mount and running ls -l gives a list of files named "???????????" which
is wrong (the media has no files at all).


MISC:

Trying to use the IBM Fn key to access the Info menu (kind of Help screen),
the system locks up completely.  This didn't happen with 1.4.2.

  -jm