[sf-lug] Linux CDs

Sat Mar 3 14:10:48 PST 2007

A list-member wrote me offlist to ask me what laptop I would recommend
for Linux use, that ordinary humans can afford, doesn't weigh too much,
and has reasonable battery life.

(No offence taken, but discussion threads should remain on-list unless
you have some reason for privacy.  If you have the latter, please begin
by explaining _why_ you've gone offlist into private discussion.  Most
of us participate in these public forums to benefit the community.  If
you want private help, it's called "consulting":  You should offer to
pay hourly consulting rates, and not expect professional work from
strangers for free.)

I'm not going to give a "buy this" endorsement, but I can help teach how
to select hardware (and how not to).  My guiding star in this area is
what I call Moen's Law of Hardware:  "Use what the programmers use."
http://linuxmafia.com/~rick/lexicon.html#moenslaw-hardware  (Cited URL 
elaborates on that point.)

You are unlikely to _literally_ know what Alan Cox, Dave Jones, Jeff
Garzik, et alii are using today -- but you can learn enough about
hardware to learn what they probably _would_ consider junk to be
avoided (and why), and that is what one should do, generally speaking.

For reasons cited at that essay, slightly _older_ gear (especially where
laptops are concerned) is likely to be less problematic than new gear.
I'd generally aim for a 1-2 year old model.  (To be more clear about
that, buying a spanking-new laptop model means you're either a Linux
hardware expert, or a masochist, or feeling really lucky.  Smart, lazy
people buy used.)

You can also benefit from other people's write-ups, based on their
experience:

http://www.linux-laptop.net/
http://tuxmobil.org/

It's useful to bring a notepad, pen, and Knoppix disc with you, as you
look at laptops.  Boot Knoppix, and jot down significant chipset
identities.  What's a chipset?  Let me illustrate using the server that
SF-LUG's mailing list runs on.  (Only selected return values are shown
for commands below.)

# dmesg | more

CPU: Intel Pentium III (Katmai) stepping 02

   OK, it's a single-proc PIII.

..... CPU clock speed is 498.7724 MHz.

   Running at 500 Mhz.

Serial driver version 5.05c (2001-07-08) with HUB-6 MANY_PORTS MULTIPORT SHARE_I
RQ SERIAL_PCI enabled
ttyS00 at 0x03f8 (irq = 4) is a 16550A
ttyS01 at 0x02f8 (irq = 3) is a 16550A

  Two serial ports.

SCSI subsystem driver Revision: 1.00
sym.0.13.0: setting PCI_COMMAND_PARITY...
sym.0.13.1: setting PCI_COMMAND_PARITY...
sym0: <875> rev 0x37 on pci bus 0 device 13 function 0 irq 10
sym0: No NVRAM, ID 7, Fast-20, SE, parity checking
sym0: SCSI BUS has been reset.
sym1: <875> rev 0x37 on pci bus 0 device 13 function 1 irq 5
sym1: No NVRAM, ID 7, Fast-20, SE, parity checking
sym1: SCSI BUS has been reset.
scsi0 : sym-2.1.17a
scsi1 : sym-2.1.17a

  Symbios model 875 SCSI chip (later revealed to have full model
designation 53c875).

blk: queue cfe38174, I/O limit 4095Mb (mask 0xffffffff)
  Vendor: QUANTUM   Model: QM39100TD-SW      Rev: N491
  Type:   Direct-Access                      ANSI SCSI revision: 02
blk: queue cfe38274, I/O limit 4095Mb (mask 0xffffffff)
  Vendor: QUANTUM   Model: QM39100TD-SW      Rev: N491
  Type:   Direct-Access                      ANSI SCSI revision: 02
blk: queue cfe38374, I/O limit 4095Mb (mask 0xffffffff)
sym0:3:0: tagged command queuing enabled, command queue depth 16.
sym0:4:0: tagged command queuing enabled, command queue depth 16.
Attached scsi disk sda at scsi0, channel 0, id 3, lun 0
Attached scsi disk sdb at scsi0, channel 0, id 4, lun 0
sym0:3: FAST-20 WIDE SCSI 40.0 MB/s ST (50.0 ns, offset 16)
SCSI device sda: 17783250 512-byte hdwr sectors (9105 MB)
Partition check:
 /dev/scsi/host0/bus0/target3/lun0: p1 p2 < p5 p6 p7 p8 p9 >
sym0:4: FAST-20 WIDE SCSI 40.0 MB/s ST (50.0 ns, offset 16)
SCSI device sdb: 17783250 512-byte hdwr sectors (9105 MB)
 /dev/scsi/host0/bus0/target4/lun0: p1 p2 < p5 p6 p7 p8 >

  Two Quantum model QM39100TD-SW SCSI hard drives.  And, oooh!  They're
  each a full 9 GB -- the very pinnacle of 1997 technology.

usb.c: registered new driver usbdevfs
usb.c: registered new driver hub
uhci.c: USB Universal Host Controller Interface driver v1.1
uhci.c: USB UHCI at I/O 0x10c0, IRQ 5
usb.c: new USB bus registered, assigned bus number 1
hub.c: USB hub found
hub.c: 2 ports detected

  Two UHCI-type USB ports.

Intel(R) PRO/100 Network Driver - version 2.3.43-k1
Copyright (c) 2004 Intel Corporation
e100: selftest OK.
e100: eth0: Intel(R) PRO/100 Network Connection
e100: selftest OK.
e100: eth1: Intel(R) PRO/100 Network Connection
  Hardware receive checksums enabled
  cpu cycle saver enabled

  A pair of Intel e100-compatible PRO/100 ethernet ports.

# lspci | more
0000:00:00.0 Host bridge: Intel Corporation 440BX/ZX/DX - 82443BX/ZX/DX Host bridge (AGP disabled) (rev 03)
0000:00:0d.0 SCSI storage controller: LSI Logic / Symbios Logic 53c875 (rev 37)
0000:00:0d.1 SCSI storage controller: LSI Logic / Symbios Logic 53c875 (rev 37)
0000:00:0f.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100]
 (rev 05)
0000:00:10.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100]
 (rev 08)
0000:00:12.0 ISA bridge: Intel Corporation 82371AB/EB/MB PIIX4 ISA (rev 02)
0000:00:12.1 IDE interface: Intel Corporation 82371AB/EB/MB PIIX4 IDE (rev 01)
0000:00:12.2 USB Controller: Intel Corporation 82371AB/EB/MB PIIX4 USB (rev 01)
0000:00:12.3 Bridge: Intel Corporation 82371AB/EB/MB PIIX4 ACPI (rev 02)
0000:00:14.0 VGA compatible controller: Cirrus Logic GD 5480 (rev 23)

   This confirms what we knew before, and gives more detail:  It's an
Intel 440BX motherboard (dating from 1998[1]).  We also find out that
the integrated video is Cirrus Logic GD 5480.  

The 82371AB chip is what's termed a "southbridge" chip, the motherboard
chip serving most of[2] the I/O ports.  The 82443BX "northbridge" chip is where
the CPU and system RAM attach.  (Together, they comprise the bulk of the
440 BX motherboard design.[3])

If this had been, say, a laptop you were interested in buying, as
opposed to just an old server, you would now be armed with the
identities of the constituent chips, and could ask around (or read on
the sites mentioned) about likely problem areas.

Also, Knoppix's hardware auto-probing is advanced enough that it can
tell you a tremendous amount just from seeing what it does and does not
support properly, upon bootup.

[1] http://en.wikipedia.org/wiki/Intel_440BX
[2] In later Intel designs, _all_ of the I/O ports moved to the
southbridge.  In 1998, the northbridge still connected to video,
network, and other very high-speed devices.
[3] http://arstechnica.com/articles/paedia/hardware/mobo-guide-1.ars/3