Baslinux mailing list

[sindi keesan <keesan AT sdf.lonestar.org>]

On Wed, 18 May 2005, James Miller wrote:

> On Thu, 19 May 2005, sindi keesan wrote:
>
> > How do I determine, in linux, why irq 5 7 9 and 10 are all busy?
>
> Try cat /proc/interrupts
>
> James
> _______________________________________________

I have done this and 10 is assigned to the scsi card (and 11 to the 
ethernet card).  At one point 7 was 'soundblaster' but 5 and 9 were free.

I removed the scsi and ethernet and parport (MGA) cards.

The first two made no difference but the MGA/parport (MGP) card appears to 
cause conflicts with the standard sb settings.

Syschk (before I disabled the onboard parport) said the card was using 
IRQ7.  (After I disabled the onboard port it said it was using ?).  It 
also told me 2, 5, 7 and 10 were free.  2 is cascade.

The MGP card, judging from experiments with it in and out of the computer, 
is using address 330 (default for MPU), IRQ 5 and DMA 1 and 0.  This 
explains why linux would not play sounds at all with standard settings 
until I took the card out.  I could get linux to play again with
insmod sb io=220 irq=7 dma=3, which also worked for DOS (diagnose).

First I had to use Win98 to change the settings on the plugnplay card to 
irq7 and dma3,5 and address 300.

The awe64 card won't play midi files in DOS but it plays wav and mp3.  In 
linux it still won't play mp3 files with play or mplayer, though I can 
insmod sb without any complaints. Mplayer gets stuck, mpg123 suggests 
IRQ/DRQ conflicts.  I compiled this kernel to support awe - perhaps it is 
not doing a good job of that?  I probably need an awe module, but would 
that affect mp3 playing?

The awe32 card is not being recognized at all now by any OS.  Before it 
was working in DOS and in linux without the parport card in there, despite 
Diagnose not being able to find any free IO address for it.  (Linux 
/proc/ioports seems to think all four possibilities 200 220 240 260 are 
free).  There may be something wrong with this card or else I have not 
managed to plug it in properly.  I pushed hard.  It is a very long card 
with two 30-pin SIMMs to hold libraries of sounds.

Our BL kernel (2.2.16) is said to be buggy with regard to soundblaster 
cards and this may be what they mean.  Perhaps I was just lucky with all 
but three computers (the newer ones).

Windows can be used to change io mpu irq and dma settings, which persist
in DOS and linux, as if you had moved jumpers (or used isapnp tools?) so I 
set the card back to standard 220 330 5 1 5 for use in other computers and 
I will try some non-awe SB16's and then ESS (ISA and pci) and Trident pci.
(For which I hope I compiled the kernel correctly).  I might also try 
bare.i kernel.

And if that fails, SW81 and SW91 kernels and modules with sb16.  I presume 
I can compile my own SW81 kernel using the SW71 compiler and SW81 kernel 
source code.  Luckily this computer has the larger hard drive and will 
compile faster.  It even loads Win98 in under 2 minutes, which explains 
why so many people were persuaded to buy faster computers in 1999.