Messages in this thread | | | Date | Tue, 26 May 1998 10:03:06 +0100 (BST) | From | Riley Williams <> | Subject | Re: NE2000 PCI problems FIXED! |
| |
Hi Linus.
>>> Hi, all. I've been pounding on 2.1.103 and I'm happy to say that >>> the lost interrupt problems with my PCI NE2000 card seem to have >>> gone away (Tyan Titan Pro, dual PPro 180). Lovely!
>> Good. We still have a case of level interrupts sometimes >> duplicating, but it's harmless. And we know why it happens and how >> to avoid it. So there's going to be some very slight work on the >> irq code still, but I'm beginning to consider it pretty much done >> other than that.
>> Anybody that still sees interrupt-related problems, speak now or >> forever hold your peace..
As a follow-up to the above, I noted:
> I have yet to try this on 2.1.103 (just downloaded it), so can't yet > comment thereon, but I recently came across an interrupt problem with > both 2.0.33 and 2.1.89 affecting the NE2k cards, which is NOT in the > 2.0.27 kernel supplied with RH 4.1 and originally installed on the > said system...
> The system in question has a PCI-based NE2k clone and a PCI-based TV > card, and the BIOS in auto mode puts both on IRQ 9, sharing the same > interrupt, which I understand to be OK with PCI. However, although the > TV card works happily under Linux, the NE2k card does not in that > configuration...
> I have done some investigation, and this appears to be a problem that > only occurs when the NE2k is sharing IRQ 9 with some other device, as > it doesn't occur with the NE2k on its own on IRQ 9, nor when the NE2k > is sharing any other interrupt with another PCI device (including the > said TV card). Beyond this, I wouldn't know where to start looking, so > if somebody out there can give me some pointers...
> For reference: P166/MMX processor, 48M RAM, 2.5G HD, networked using > the said NE2k with coax leads...
Follow-up to the above: 2.1.103 does NOT suffer this bug.
Best wishes from Riley.
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu
| |