Re: VAIO problems still

mccrobie2000 writes:
> <rant - please note this applies to the GRX-670 -
> YMMV>
> 
> UGH! UGH! UGH!  Please see:
> 
> http://w3.mivlmd.cablespeed.com/~mccrobie
> 
> Let me know if this fixes your problem - or not.
R505ES...
Does.  (the first chunk (RCS Id) fails but that doesn't matter)
Sound works fine with this patch.

Your notes about video don't apply to this R505 as the R505Ex uses
Intel 830M.  I have had to use HEAD of XF86 in order to get a working
driver but it works well (now - there were some loose ends for a while,
especially involving alt-ctl-f1 to get back to a text screen; it broke
things in a visually cute (but operationally disgusting) way.

> I have diagnosed this problem on the Sony Vaio GRX-670
> six months ago.
> 
> There is nothing you can do to get devices off of
> IRQ9, that I know of.  If its any consolation, Windows
> 2000 lists ethernet, firewire, usb, cardbus ALL on
> IRQ9.  Since FreeBSD shares interrupts, I have not
> noted any performance issues with so many devices on
> the same interrupt, although I don't generally have
> USB, firewire, sound, and ethernet all going at the
> time.
> 
> The problem with audio is that FreeBSD does NOT route
> the existing interrupt.  The patch on the above web
> site hacks the routing of the interrupt.
> 
> I can play sound WITHOUT moving my mouse, without
> network traffic, etc.
Now also.

> Because the interrupt is shared, the ich3 sound driver
> gets its interrupt routine called when anything that
> gets IRQ9 is used - including USB mouse.  The sound
> driver than finds work to do and does it - the result
> is you get sound.
I had figured out that much; just not why...


> FreeBSD does not seem to setup the interrupt mechanism
> to enable delivery of the IRQ9 for sound, because
> sound already has an assigned interrupt.
> 
> I've been waiting for someone to fix this.  I posted
> this patch several months ago, but the response was
> "this does nothing for IO APIC" - I don't know what
> that is or how it affects this issue, but if it fixes
> the problem, it would greatly be appreciated.

I don't know how IO APIC is relevant to the mobile chipset;
it normally applies to SMP systems...  (though some VIA single-proc
chipsets contain APIC functions too.)

My next project for this vaio is to try to get a Sierra 555 working 
in freebsd.  Stable recognizes it as a serial device but hangs the 
whole system if it gets used.  Current recognizes it as a network 
device (which is true, though it normally uses ppp in operation so 
should look serial), then ignores it due to lack of a driver.  I'm 
starting to figure out how to use the pccard_quirks stuff :-)

-- Pete