Laptops - Install Tricks & Release Status &

Runs fine, recognises & configs ep0 & xe0 pcmcia ether & can access pcmcia cdrom, need no special /boot/ magic.

This disk matches the following BIOS disk
BIOS    # cylinders heads   sectors total sectors   GB
0x80    992 128 63  0   0
a This is correct geometry
a Edit the MBR partition table

Even with the 10/100 cards, do not expect more than ~450KB/s.

Starting with FreeBSD-7.0BETA3

Message-Id: <200712041614.lB4GEiOw073415@@fire.js.berklix.net>
To: mobile@@freebsd.org
From: "Julian H. Stacey" <jhs@@berklix.org>
Date: Tue, 04 Dec 2007 17:14:44 +0100
Subject: 7.0BETA3 + Dell Latitude XPi : Disk problem

freebsd-mobile@@,
( bcc'd re@@freebsd.org for info )
Suggestions please to help escape nightmare (below) of 7.0BETA3 installer
repeatedly failing on a Dell Latitude XPi, (which runs 4.11 no problem).
It seems the 5,6,7 FreeBSD Installer wont work on this laptop.
All I can think of now for this laptop is:
   Reinstall 4.11
   Abandon FreeBSD installer for this laptop
   Hand allocate an Fdisk S2
   Hand partition disklabel ad0s2
   Hand install boot sectors
   Hand newfs ad0s2a & f (but UFS2 or 1 ?)
   Hand cross install a complete 7.0BETA3 operational tree
   Hand edit /boot/loader.conf
   Test. & tune boot & loader.conf till disk & pcmcia ep0 work.
   Report back to re@ more conservative values
   (by which time they may well have shipped release).
Other ideas ?

Note I have installed 7.0BETA3 on other hosts.

Detail with 7.0BETA3 below (previous history with 6.2 here:
    http://www.berklix.com/~jhs/hardware/laptops/#dell   )
===========================

   After continuously failing to persuade 7 installer to do any of
    - install to hard disk OK,
    - recognise & use pcmcia ep0    OR
    - recognise & use pcmcia cdrom card & avoid deathly timeouts
   (& each failure each time needing 7 floppy changes. )
   
   So I gave up on pcmcia ep0 & pcmcia cdrom to install from, and
   Deleted all fdsik partions
   Installed 4.11 fresh on first 1.5G partition,
   copied 4.11 CD to /usr/cd
   Booted 7 floppies (yet Again)
   set hw.ata.ata_dma=0
   boot
    ( About 5 mins wait after it says toot@logan ... sys/GENERIC
       before "real memory" )
   `A' utomatic disklabel fails, (thinks 2.5G insufficient for /usr),
   Manually
    300G root
    leave 90M swap found from 4.11
    Rest of space for /usr
    Mount old partions as /4 /4/var /4/tmp /4/usr
   Install from filesystem /4/usr/cd
   Finished with no errors.
   Reboot
   F2 refused, so push F1 to boot 4.11 to look:
   # disklabel ad0s2
   
   # cd /dev ; ./MAKEDEV ad0s2a
   fsck ad0s2a
    BAD SUPER BLOCK: MAGIC NUMBER WRONG
    LOOK FOR ALTERNATE SUPERBLOCKS ? [yn]
   y
   
   # fsck -b 32 /dev/ad0s2a
   Alternate super block location: 32
   ** /dev/ad0s2a
   BAD SUPER BLOCK: MAGIC NUMBER WRONG
   
   # uname -a
   FreeBSD norb.js.berklix.net 4.11-RELEASE FreeBSD 4.11-RELEASE
   #0: Fri Jan 21 17:21:22 GMT 2005
   root@perseus.cse.buffalo.edu:/usr/obj/usr/src/sys/GENERIC  i386
   
   # dmesg
   
   # fdisk
   ******* Working on device /dev/ad0 *******
   parameters extracted from in-core disklabel are:
   cylinders=7944 heads=16 sectors/track=63 (1008 blks/cyl)
   
   Figures below won't work with BIOS for partitions not in cyl 1
   
   parameters to be used for BIOS calculations are:
   cylinders=7944 heads=16 sectors/track=63 (1008 blks/cyl)
   
   Media sector size is 512
   Warning: BIOS sector numbering starts with sector 1
   Information from DOS bootblock is:
   The data for partition 1 is:
   sysid 165,(FreeBSD/NetBSD/386BSD)
       start 63, size 2999745 (1464 Meg), flag 80 (active)
           beg: cyl 0/ head 1/ sector 1;
           end: cyl 1023/ head 15/ sector 63
   The data for partition 2 is:
   sysid 165,(FreeBSD/NetBSD/386BSD)
       start 2999808, size 5007744 (2445 Meg), flag 80 (active)
           beg: cyl 1023/ head 255/ sector 63;
           end: cyl 1023/ head 15/ sector 63
   The data for partition 3 is:
   <UNUSED>
   The data for partition 4 is:
   <UNUSED>
   
   Next Attempt Failed too, as above, but with changes:
      Not manually typing   set hw.ata.ata_dma=0
        (else as on other laptops, 7 thinks whole disk is free)
      but more comprehensively
        "Safe Boot" (also does hw.ata.wc
        Ref my http://www.berklix.com/~jhs/hardware/laptops/#loader.conf )
      In fdisk delete S2 & 'C' recreate & 'S' mark bootable & 'Q'
      This time not "Leave MBR untouched" but "Install the FreeBSD Boot Manager"
   
      After succesfull install with no errors, reboot, & again both F1 & F2
      boots fail, beeping.

Ideas ?

Julian

6 Escape to loader
ed /boot/loader.conf # Add hw.ata.ata_dma=0 This should be added automatically, a send-pr for that & other things will be via here

I didnt have to try this ultimate fallback as slip worked. I'm happy I didnt have to try as some PC BIOS uses disk in different ways, LBA native & what not, so would have had to try up to 3 times.

Installed at 9600 Without -h in slattach. I've sent a send-pr to remove -h as some cables do not support it.
Kernel hang a long time as usual calibrating clocks
After "Timecounters tick" Many ad0: FAILURE - READ_DMA Mount of ad0s1a fails.
power off,
vi /etc/start_if.sl0 # /sbin/slattach -a -l -s 115200 /dev/cuad0
chmod a+rx /etc/start_if.sl0
EUREKA ! A maximal speed (I suppose, check later if faster possible) SLIP con with TCP-IP after months on & off trying.
Uploaded 250M of cdrom distribs over 115200 SLIP, built src with make world, & upgraded to 7 Stable

hostb0@pci0:0:0:0:      class=0x060000 card=0x00000000 chip=0x00011066 rev=0x04 
hdr=0x00
    vendor     = 'Picopower Technology (A division of National)'
    device     = 'PT86C521 Vesuvius V1-LS System Controller'
    class      = bridge
    subclass   = HOST-PCI
isab0@pci0:0:6:0:       class=0x060100 card=0x00000000 chip=0x80021066 rev=0x00 
hdr=0x00
    vendor     = 'Picopower Technology (A division of National)'
    device     = 'PT86C523 Vesuvius V3-LS ISA Bridge'
    class      = bridge
    subclass   = PCI-ISA
vgapci0@pci0:0:7:0:     class=0x030000 card=0x00000000 chip=0x000110c8 rev=0x01 
hdr=0x00
    vendor     = 'Neomagic Corporation'
    device     = 'NM2070 MagicGraph 128'
    class      = display
    subclass   = VGA
atapci0@pci0:0:8:0:     class=0x010180 card=0x00000000 chip=0x06431095 rev=0x00 
hdr=0x00
    vendor     = 'Silicon Image Inc (Was: CMD Technology Inc)'
    device     = 'PCI0643 PCI EIDE controller'
    class      = mass storage
    subclass   = ATA

interrupt                          total       rate
irq0: clk                         803370        999
irq1: atkbd0                           7          0
irq4: sio0                         18351         22
irq6: fdc0                            11          0
irq8: rtc                         102815        127
irq14: ata0                         3669          4
Total                             928223       1154

Finaly Works ! - Well Half Works, Not very well. Driver is flakey: does not erase old used pixels, so there's a sort of memory efefct.
How I did it:

I had previously compiled the base on host=lapd,
Then discovered startx can do an auto config :-)
Saw startx exiting on some xterm syntax error,
not sure where from /var/log/Xorg.0.log did not say where it was importing it from,
Also saw modules not loaded, so
cd /host/fire/usra/ftp/pri/lapd/7/usr/ports/x11-drivers/xf86-video-neomagic
su; make reinstall
cd /host/fire/usra/ftp/pri/lapd/7/usr/ports/x11-drivers/xf86-video-fbdev 
su; make reinstall
There was also in log some fbdev.hw module missing I couldnt see in ports/x11-driver
s/
Tried xinit, also auto configs (or inherited a config from startx ?
Mouse works,

su; scanpci

pci bus 0x0000 cardnum 0x00 function 0x00: vendor 0x1066 device 0x0001
 PicoPower Technology PT86C521 [Vesuvius v1] Host Bridge

pci bus 0x0000 cardnum 0x06 function 0x00: vendor 0x1066 device 0x8002
 PicoPower Technology PT86C523 [Vesuvius v3] PCI-ISA Bridge Slave

pci bus 0x0000 cardnum 0x07 function 0x00: vendor 0x10c8 device 0x0001
 Neomagic Corporation NM2070 [MagicGraph 128]

pci bus 0x0000 cardnum 0x08 function 0x00: vendor 0x1095 device 0x0643
 Silicon Image, Inc. PCI0643

Works, but judders horribly, just like it does on host=lapd

ep0 was not recognised, but is now. Why ? Now one can jack out the card, changed to a Xircom (xe0) & the ep0 is still seen & the xe0 is not seen.

Maybe that ep0 above was the expansion port ? 'Cos...

20 Jan 2009 I got the expansion port/ docking bay working with Ethernet ! it appears as ep0 ! (I got this working with my LAPN.small custom kernel, having turned off various other things in BIOS that might have conflicted for IRQs etc.

Dell Latitude XPi P133ST

Referred to by http://www.berklix.org/~jhs/hardware/laptops/#hosts

http://www.berklix.org/~jhs/hardware/laptops/dell_latitude_xpi_p133st

By Julian H. Stacey

INDEX

NetBSD-3.1 Up & Running.

FreeBSD-7

I suspect there's something sick with this BIOS or MBR, So I did a test:

DMA OFF

PLIP

Disk Transplant

SLIP Serial Line Internet Protocol

dmesg

pciconf -l -v

vmstat -i

X Windows

scanpci

SOUND

ETHERNET

TO DO LATER

Domains