Date: Fri, 21 Jan 2000 00:52:50 -0500 (EST) From: Bill Paul <wpaul@skynet.ctr.columbia.edu> To: arnee@geocities.com (arnee) Cc: dillon@apollo.backplane.com, current@FreeBSD.ORG, gerald@manhattanprojects.com Subject: Re: ...(file transfer crashes system) ethernet driver or IP stack bug? Message-ID: <200001210552.AAA11760@skynet.ctr.columbia.edu> In-Reply-To: <3887BDC4.A3AEBD35@geocities.com> from "arnee" at Jan 20, 2000 06:00:36 pm
next in thread | previous in thread | raw e-mail | index | archive | help
Of all the gin joints in all the towns in all the world, arnee had to walk
into mine and say:
> FreeBSD 4.0-CURRENT-20000118-08:18 #4: Tue Jan 18 13:17:49 PST 2000
> root@groovinit:/usr/src/sys/compile/GROOVINIT
> Timecounter "i8254" frequency 1193182 Hz
> Timecounter "TSC" frequency 249919443 Hz
> CPU: Cyrix 6x86MX (249.92-MHz 686-class CPU)
Ack...
> Origin = "CyrixInstead" Id = 0x601 Stepping = 1 DIR=0x1453
> Features=0x80a135<FPU,DE,TSC,MSR,CX8,PGE,CMOV,MMX>
> real memory = 29360128 (28672K bytes)
> avail memory = 25792512 (25188K bytes)
> Preloaded elf kernel "kernel" at 0xc02c9000.
> md0: Malloc disk
> npx0: <math processor> on motherboard
> npx0: INT 16 interface
> pcib0: <Host to PCI bridge> on motherboard
No, please...
> pci0: <PCI bus> on pcib0
> isab0: <SiS 85c503 PCI-ISA bridge> at device 1.0 on pci0
> isa0: <ISA bus> on isab0
> ata-pci0: <SiS 5591 ATA-33 controller> port
> 0x4000-0x400f,0xa8c42268-0xa8c4226b,0x35d8a0c0-0x35d8a0c7,0xc7cc0-0xc7cc3,0x42140728-0x4214072f
> irq 0 at device 1.1 on pci0
> ata-pci0: Busmastering DMA supported
> ata-pci1: <Promise ATA-66 controller> port
> 0xf080-0xf0bf,0xf7e0-0xf7e3,0xf790-0xf797,0xf7f0-0xf7f3,0xf7a0-0xf7a7 mem
> 0xffac0000-0xffadffff irq 11 at device 13.0 on pci0
> ata-pci1: Busmastering DMA supported
Stop, you're killing me...
> ata2 at 0xf7a0 irq 11 on ata-pci1
> dc0: <82c169 PNIC 10/100BaseTX> port 0xf200-0xf2ff mem 0xffaaff00-0xffaaffff irq 9 at
> device 15.0 on pci0
Aieeeee!
Oh for the love of *god*. Rip this thing out of your machine now. Put it
on the ground. Stomp on it. Repeatedly. Then set it on fire. Bury the
remains in the back yard. Then run - don't walk! - to your local computer
store, put a crowbar in your wallet, and buy a better ethernet card.
Don't whine, don't bitch, don't moan, don't complain that it works with
Windows (actually, I bet it doesn't; not on that hardware, using bus
master DMA for both controllers). Just do it.
You have a Cyrix CPU, and I'm willing to bet you've overclocked it. ("What?
You mean that might have some effect on the situation?") You have a PCI
chipset which the probe messages don't even identify, and you're using
not one but *two* bus master devices on it, one of which is a PNIC ethernet
controller that can barely do bus master DMA correctly on a good day. Not
only that, but you have two ATA controllers in this machine, one of which has
no devices on it, and you have the damn hard disk on the *second* controller.
Why do you have two controllers? Why not just use the built-in one!
I'm also missing some information here since I arrived in the middle of
this exchange. You didn't say if you're using 10Mbps or 100Mbps with this
card. And you don't say what it's connected to (hub? switch? vendor?
model?)
Does this have something to do with the PNIC? Oh, probably. But I'm not going
to even attempt to debug this. There's no way in hell I could duplicate this
hardware configuration locally, and even if I did, I probably couldn't
duplicate the exact same problem. And even if I could do *that*, I
still wouldn't know how to fix it. I'm sorry, but I've pulled out enough
hair over this damn chip. I'm sure I've wasted at least a couple months
of my life trying to make this thing work reliably. I've added several
software workarounds, I defaulted the stupid transmit threshold setting
to store and forward mode. Well I've had it: no more. I'm through trying
to bitchslap this rotten piece of silicon to its senses. I'm not going to
go nuts every time somebody comes up with yet another oddball hardware
combination. There's a limit to my patience and this chip has reached it.
Now it's somebody else's turn to lose their sanity. I have the datasheet for
the PNIC at http://www.freebsd.org/~wpaul/PNIC. You have the driver source.
Somebody *else* try and figure it out, and then tell me then answer when
you have it.
That said, if you have overclocked this machine, then un-overclock it *right*
*now* and never, ever do it again! PCI bus master DMA is goofy enough without
people sticking their fingers in the works.
Man, why does PC hardware have to suck so hard.
-Bill
--
=============================================================================
-Bill Paul (212) 854-6020 | System Manager, Master of Unix-Fu
Work: wpaul@ctr.columbia.edu | Center for Telecommunications Research
Home: wpaul@skynet.ctr.columbia.edu | Columbia University, New York City
=============================================================================
"It is not I who am crazy; it is I who am mad!" - Ren Hoek, "Space Madness"
=============================================================================
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-current" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200001210552.AAA11760>