Re: [9fans] Weird bug.

Re: [9fans] Weird bug.

[jmk]

There are a couple of issues here. First, for historical reasons which
I won't go into for fear of hurting feelings, the defaults used by the
SB16 driver bear no relation to the defaults on a card when you take
it out the box, so it's almost always a mistake to not specify the IRQ
and DMA in plan9.ini; for example, the default driver IRQ is 7 which
will clash with the lpt IRQ. The plan9.ini configuration for an SB16
with default hardware settings should be something like

	audio0=type=sb16 port=0x220 irq=5 dma=1

Secondly, the SB16 doesn't show up in the iomap output because the driver
doesn't call the routine to register the space it uses. That's a bug we
should fix but shouldn't affect the operation of the driver (assuming there
is no actual clash with another device).

--jim

Re: [9fans] Weird bug.

[Douglas A. Gwyn]

jmk@plan9.bell-labs.com wrote:
> Secondly, the SB16 doesn't show up in the iomap output because the
> driver doesn't call the routine to register the space it uses.
> That's a bug we should fix but shouldn't affect the operation of the
> driver (assuming there is no actual clash with another device).

Hm, with a PCI-only motherboard, which is about all one can buy
these days, the nearest thing to a Plan 9-suppored SoundBlaster
seems to be the SoundBlaster-16 PCI, which like everything else
for PCI these days is plug-and-play.  Is there some way to make
this work on Plan 9?

Also, does the 3C905C-TM (PCI) work a a Plan 9 Ethernet card?

It's getting harder and harder to build an ISA PC..

Re: [9fans] Weird bug.

[Eric Dorman]

jmk@plan9.bell-labs.com wrote:
 
>         audio0=type=sb16 port=0x220 irq=5 dma=1

I'll stick these in and see what color smoke comes out.

> Secondly, the SB16 doesn't show up in the iomap output because the driver
> doesn't call the routine to register the space it uses. That's a bug we
> should fix but shouldn't affect the operation of the driver (assuming there
> is no actual clash with another device).

sounds like a 1-liner.
 
> --jim

thanks much,

--eric

PS: what's weird though is why a failed SB16 configuration should
frob up IP checksumming...  that one is extremely vexing.

Re: [9fans] Weird bug.

[Julian Back]

In article <20010312154951.5F453199ED@mail.cse.psu.edu>, 
jmk@plan9.bell-labs.com () wrote:

> On Mon Mar 12 08:27:55 EST 2001, DAGwyn@null.net wrote:
> > jmk@plan9.bell-labs.com wrote:
> > > Secondly, the SB16 doesn't show up in the iomap output because the
> > > driver doesn't call the routine to register the space it uses.
> > > That's a bug we should fix but shouldn't affect the operation of the
> > > driver (assuming there is no actual clash with another device).
> > 
> > Hm, with a PCI-only motherboard, which is about all one can buy
> > these days, the nearest thing to a Plan 9-suppored SoundBlaster
> > seems to be the SoundBlaster-16 PCI, which like everything else
> > for PCI these days is plug-and-play.  Is there some way to make
> > this work on Plan 9?
> > 
> > Also, does the 3C905C-TM (PCI) work a a Plan 9 Ethernet card?
> > 
> > It's getting harder and harder to build an ISA PC..
> 
> If it is truly an SB-16 compatible then it's a matter of adding a
> handful of lines to scan the PCI bus for the card to make it work.
> But it can't really be compatible because the DMA must be completely
> different.

AFAIK the SB16-PCI is not SB16 compatible.  It's based on an ES1371 chip. 
The Linux driver for this card is completely different to the SB16 ISA 
driver.

Julian

> 
> The 3Com driver in the distribution requires some small fixes
> to work with the 905C, Jean Mehat <jm@ai.univ-paris8.fr> has them.

Re: [9fans] Weird bug.

[jmk]

On Mon Mar 12 08:27:55 EST 2001, DAGwyn@null.net wrote:
> jmk@plan9.bell-labs.com wrote:
> > Secondly, the SB16 doesn't show up in the iomap output because the
> > driver doesn't call the routine to register the space it uses.
> > That's a bug we should fix but shouldn't affect the operation of the
> > driver (assuming there is no actual clash with another device).
> 
> Hm, with a PCI-only motherboard, which is about all one can buy
> these days, the nearest thing to a Plan 9-suppored SoundBlaster
> seems to be the SoundBlaster-16 PCI, which like everything else
> for PCI these days is plug-and-play.  Is there some way to make
> this work on Plan 9?
> 
> Also, does the 3C905C-TM (PCI) work a a Plan 9 Ethernet card?
> 
> It's getting harder and harder to build an ISA PC..

If it is truly an SB-16 compatible then it's a matter of adding a
handful of lines to scan the PCI bus for the card to make it work.
But it can't really be compatible because the DMA must be completely
different.

The 3Com driver in the distribution requires some small fixes
to work with the 905C, Jean Mehat <jm@ai.univ-paris8.fr> has them.

Re: [9fans] Weird bug.

[Eric Dorman]

jmk@plan9.bell-labs.com wrote:
> 
> let's see the contents of
>         plan9.ini
>         #P/irqalloc
>         #P/ioalloc

plan9.ini:
bootdisk=local!#S/sdC0/fs
*nomp=1
ether0=type=elnk3
monitor=multisync75
vgasize=1024x768x8
mouseport=0
audio0=type=sb16

ioalloc:
       0        f dma
      20       21 i8259.0
      40       43 i8253
      60       60 kbd
      61       61 i8253.cntr2c
      64       64 kbd
      70       71 rtc/nvr
      80       8f dma
      a0       a1 i8259.1
      d0       df dma
     170      177 atacmd
     1f0      1f7 atacmd
     2b0      2df vga
     2f8      2ff eia1
     300      30f tcm509isa
     376      376 atactl
     378      37a lpt0
     3c0      3da vga
     3f0      3f5 floppy
     3f6      3f6 atactl
     3f7      3f7 floppy
     3f8      3ff eia0

irqalloc:
          3           0 debugpt
          7           0 mathemu
          9           0 mathover
         14           0 fault386
         16           0 matherror
         32           0 clock
         33           1 kbd
         35           3 eia1
         36           4 eia0
         38           6 floppy
         39           7 lpt
         42          10 ether0
         46          14 sdC (ata)
         47          15 sdD (ata)

i'm not sure where audio cards typically iomap, i guess
around 0x220.

--eric

Re: [9fans] Weird bug.

[jmk]

let's see the contents of
	plan9.ini
	#P/irqalloc
	#P/ioalloc

Re: [9fans] Weird bug.

[Eric Dorman]

Russ Cox wrote:
> cat '#P/irqalloc' -- is your ethernet card on IRQ 2 as well?

nope.  irqalloc says ethernet is 10.
isin't 2 the daisy-chain interrupt?

--eric

Re: [9fans] Weird bug.

[Russ Cox]

cat '#P/irqalloc' -- is your ethernet card on IRQ 2 as well?

[9fans] Weird bug.

[Eric Dorman]

My soundblaster clone is a Creative CT4170, supposedly
a Soundblaster 16.  Anyway saying 'audio0=type=sb16' gives
me a diagnostic:

i8259enable: irq 2 shared but not level
intrenable: couldn't enable irq 2 tbdf 0xffffffff for audio

This isin't particularly shocking, but what's weird is that
DHCP then won't work!  Examining dhcp packets with snoopy
shows that the 3rd octet in the Ip address is munged, resulting
in an IP checksum error.  Cool, huh?

--eric