errors with new (irq-sharing) code

G.J. dragyn@smaug.bitstreet.net
Thu Apr 1 12:35:33 1999


Thanks for the quick reply!

a few answers to your questions and update on what's happened so far...

On Thu, 1 Apr 1999, Donald Becker wrote:

> My PCI drivers have supported IRQ sharing longer than any others, since
> 1.1.70 or so. ( I wrote the first version of the shared IRQ code.)
> 
> > My Creative Labs Ensonic PCI64D sound card uses IRQ11 as does my
> > fancy-dancy 3com 3c905B Cyclone.
> > 
> > finally got my kernel compiled to do sound... but at the expense of my
> > network.
> 
> Read
>    http://cesdis.gsfc.nasa.gov/linux/misc/irq-conflict.html

thanks very much for that document... some how missed it on the page i
read. in case it's important, the machine is a fairly new Gateway machine.
PII450, 100Mhz bus, ECC RAM, AGP video.


> > went to the website and found the new 3c59x.c code that "also supports the
> > shared interrupts."
> 
> Which website?

the website given during boot-up right after it finds my NIC.

	3c59x.c:v0.99H-WOL 2/24/99 Donald Becker
	http://cesdis.gsfc.nasa.gov/linux/drivers/vortex.html

that's also where i found the mention of the "new driver that does
irq-sharing." maybe i misunderstood what it was saying. for all i know,
that part of the page has been there ever since you first added the
irq-sharing stuff. =]

> 
> > gcc -D__KERNEL__ -I/usr/src/linux-2.0.36/include -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strength-reduce -pipe -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=686  -c -o 3c59x.o 3c59x.c
> > 3c59x.c:94: parse error before `<'
> 
> Your web browser corrupted the source code.  Use FTP instead
>   ftp://cesdis.gsfc.nasa.gov/pub/linux/drivers/test/3c59x.c

yep. compiled fine with the ftp'ed source!

but ran into the same problem. once booted, i get the following errors
repeatedly (at about the rate of one every ten seconds at first, then
one each minute).

eth0: transmit timed out, tx_status 00 status e000.
eth0: transmit timed out, tx_status 00 status e000.

i'm going to try a new IRQ setting in BIOS as was mentioned in the
irq-conflict document.

.... all i could do was "reserve" irq 11 so it would not be assigned to
PCI devices. when i did that, the system simply gave them both irq 10.

es1371: version v0.2 time 10:44:13 Apr  1 1999
es1371: found adapter at io 0x1080 irq 10

eth0: 3Com 3c905B Cyclone 100baseTx at 0x1000,  00:10:4b:0a:51:f1, IRQ 10
  8K byte-wide RAM 5:3 Rx:Tx split, Autonegotiate interface.
  MII transceiver found at address 24, status 786d.
  MII transceiver found at address 0, status 786d.
  Enabling bus-master transmits and whole-frame receives.


thx for any feedback or suggestions!

--gareth

ps. i tried to see if i could change/set the IRQ under windows for either
the sound card or NIC, but since they are PCI... that's all done at boot
time (if i understand correctly).


--
/*
 * Gareth Jenkins
 * Linux hacker, "Mr. Sensitive", armchair philosopher,
 * PERL lover, child of God
 *
 * http://endor.hsutx.edu/~locutus http://www.wtlug.org
 */