[LAU] Thinkpad R60 for Audio Update - Firewire Conflicts with Audio

[Robin Gareus]

Geoff Beasley wrote:
>> Maybe some deeper way of
>> manipulating the IRQs that I'm not aware of?
> 
> nope.looks like it's hardwired.
> 
indeed. - they "might" be firmware-programmable, but usually they are
hardcoded in the PCB. - in a few years your kid might find DIY
instructions to how to tweak that old R60.. ;)

>> I'm planning on getting a PCMCIA sound card at some point, but after
>> dropping 1,400 on this laptop my extra funds are a little tight for a
>> while.

PCMCIA is ok, but not great for low latency. (adds one extra interface
between PCI bus & sound-card). - the cheap solution is to get a PCMCIA
card for your hard-disk - instead of using the built-in firewire for the
disk. - maybe you want to do that anyway. use the built-in firewire to
an external audio-device and the PCMCIA for storage.

> go usb. better supported and cheaper too probably.i use a m-audio transit. 
> works well. 
> 
usb will get you only 2 channels (full-duplex @48kHz, or half-duplex at
96kSPS, due to USB1.1 bandwidth limitations. IIRC there's no USB2-audio
standard.

The hardwired thinkpad PCI-IRQ's are a bit annoying for pro-audio, but
after all it's possible! - thinkpad X60s (guess it's similar to the
R60): the build in hda soundcard works fine with
 `jackd -R -P70 -dalsa -dhw:0,0 -r48000 -p64 -n3` !!!

but the sound quality is.. mmmh great for telephone & entertaining music.

If I use the wireless (ipw3945 - same IRQ) extensively I get xruns.. a
Firewire drive might cause similar issues, I have not tested that. - The
easiest solution is to connect your firewire disk to a 2nd machine and
export it via eth-wire (or USB?)

If you are into hacks&tweaks you might be able to get a useable system,
though: tweak the PCI bus settings (try the inverse of what paul's
setpci script does: If the firewire device blocks the bus for too long,
the audio on the same IRQ won't have a chance.), disable DMA for the FW
harddisk (is that possible? -  maybe try different IO schdulers or limit
the disk-i/o bandwidth in favor of polling it nicely?, does the
fw-device driver have a PCI/split-transaction on/off option?) - IIRC
there's a way to re-nice kernel threads, but nothing that any sane
audio-engineer would want to try...


As for Sound: I rarely use the SD-reader (sdhci:slot0 - BTW. unloading
the module saves power!) so I connect a USB UA-25 to either IRQ22 or
IRQ23. even if it's the only device on that bus I get rare xruns on 64
frames/period. but a robust setup with  `jackd -R -P 70 -d alsa -d hw:1
-p  256 -r 48000` makes this Thinkpad a great machine for
stereo/binaural audio work and fun! - (recap: the Dell D820 seems
somewhat better suited for pro-audio, and a Vaio sz3 is similarly good,
but sony&linux requires a little more some devotion..)

thinkpad x60s:

           CPU0       CPU1
  0:  352029528          0   IO-APIC-edge      timer
  1:    1251013          0   IO-APIC-edge      i8042
  8:          0          0   IO-APIC-edge      rtc
  9:   65043220          0   IO-APIC-fasteoi   acpi
 12:   10126775          0   IO-APIC-edge      i8042
 20:   35583442   20485440   IO-APIC-fasteoi   yenta,
i915 at pci:0000:00:02.0, uhci_hcd:usb1
 21:  200293604  250337407   IO-APIC-fasteoi   ohci1394, ipw3945, HDA
Intel, uhci_hcd:usb2
 22:   65491123          0   IO-APIC-fasteoi   sdhci:slot0, uhci_hcd:usb3
 23:     745166    1065462   IO-APIC-fasteoi   ehci_hcd:usb5, uhci_hcd:usb4
218:    1670650        179   PCI-MSI-edge      eth1
219:     810037    3174745   PCI-MSI-edge      libata
NMI:          0          0
LOC:  107326912  241228138
ERR:          0
MIS:          0