Bug#248855: Useful clues for non-working USB keyboard on SunBlade 150 / sparc64

[Vincent.McIntyre@csiro.au]

> Please note that the keyboard was working correctly at least once.
> I've got a proof for it:
> http://ics.p.lodz.pl/~wiktorw/sparc_d-i/sarge-sparc-businesscard_20040511.iso
>

Things were also kind-of-working breaking at 20040619 (see bug 255296).

Here is a comparison of the versions for what I see as the relevant
packages on the iso above vs rc2. Probably there are more.
I'm open to suggestions about where to look...

I tried but was unable to copy the udebs off the 20040511 iso and unpack
them, there was some weird I/O error when reading (ls worked fine).


old  (sparc-businesscard_20040511)
console-keymaps-sun_2002.12.04dbs-38_all.udeb
console-keymaps-usb_2002.12.04dbs-38_all.udeb
discover1-data-udeb_1.2004.02.08-6_all.udeb
kbd-chooser_0.52_sparc.udeb
prebaseconfig_0.65_all.udeb

middle (sparc-businesscard_20040619)
console-keymaps-sun_2002.12.04dbs-40_all.udeb
console-keymaps-usb_2002.12.04dbs-40_all.udeb
discover1-data-udeb_1.2004.02.08-7_all.udeb
kbd-chooser_0.54_sparc.udeb
prebaseconfig_0.66_all.udeb

new(sparc-businesscard_rc2)
console-keymaps-sun_2002.12.04dbs-46_all.udeb
console-keymaps-usb_2002.12.04dbs-46_all.udeb
discover1-data-udeb_1.2004.11.09_all.udeb
kbd-chooser_1.08_sparc.udeb
prebaseconfig_1.06_all.udeb

The differences I see between rc2 and 20040619:
 * kbd-chooser now calls kbd-mode -u

buscard-rc2/pool/main/k/kbd-chooser/lib/debian-installer-startup.d/S55kbd-chooser
--- netinst-20040619/pool/main/k/kbd-chooser/etc/rcS.d/S55kbd-chooser
Tue May 18 07:29:10 2004
+++
buscard-rc2/pool/main/k/kbd-chooser/lib/debian-installer-startup.d/S55kbd-chooser
Thu Nov 11 05:24:05 2004
@@ -1,7 +1,13 @@
 #!/bin/sh
 kbdchooser=/usr/bin/kbd-chooser
+kbdmode=/usr/bin/kbd-mode

 set -e
+
+# Set console mode to unicode
+if [ -x "$kbdmode" ] ; then
+       $kbdmode -u || true
+fi

 # Set the kbd early, if necessary
 if [ -x "$kbdchooser" -a "$bootkbd" ] ; then


* small language-dependent differences in
    /usr/lib/prebaseconfig.d/70kbd-chooser
  which are irrelevant

* some updates to /usr/share/discover/pci.lst
 108e Sun Microsystems Computer Corp.
-       108e1001        unknown sunhme  Happy Meal
+       108e1001        ethernet        sunhme  Happy Meal
        108e1101        ethernet        sungem  RIO GEM
-       108e1102        unknown ohci1394        RIO 1394
+       108e1102        bridge  ohci1394        RIO 1394
        108e1103        usb     usb-ohci        RIO USB
        108e2bad        ethernet        sungem  GEM

it seems unlikely but could the problem be some stupid interaction with
ohci1394?


next I'm going to figure out if I can svn diff the d-i tree for the
dates of interest...