Hi,
I was hoping that someone would be able to suggest a solution for some
hardware problems I'm having.
i) I have two eisa network cards: 3c507, and ni65 (ancient lance chip:)
These both work fine (500kB/s) when used on a 486DX100 M/B (with VLB)
However, I also have access to two more recent M/B's with PCI buses.
One has a k6-2 on an MVP4 chipset. The other has a k6 on an SIS chipset.
Whenever I attempt to use either of the NICs on the recent M/B's, the
cards lock up with Rx and Tx unit stopped messages.
The 3c507 behaves slightly better and will only lock up if I do a
ping flood or try to send a number of packets in quick succession -
it is able to limp along under a light load of ip-masquerading for
a dialup connection. The ni65 seems to be worse though.
Now, when I put an ne2000 eisa card on the PCI m/b's, it works fine.
I U'dTSL (3c507.c) and noticed Donald B's note about clearing the Tx
buffer etc. if the Tx unit ever stopped, however I believe it is more
of an eisa bus timing problem or conflict issue (i have disabled on
board sound and any unnecesary stuff like that). I also upgraded the
BIOS but that did not help (but did give me an option to enable an
onboard modem I never knew I had?:)
ii) My hard disk works fine on the old board, but starts up with the
following error on either of the PCI boards:
ide0 at 0x1f0-0x1f7,0x3f6 on irq 14
hda: ST33221A, 3077MB w/128kB Cache, CHS=781/128/63, UDMA
Partition check:
hda:hda: set_multmode: status=0x51 { DriveReady SeekComplete Error }
hda: set_multmode: error=0x04 { DriveStatusError }
hda1 hda2 hda3 hda4
(the ready/seek errors do not occur during operation so I am only mildly
concerned)
Thanking anybody who has made it this far in the message,
Mark.
(gcc2.72 kernel, pci quirks enabled, happens with either 486 or
k6 optiisations, also happens with gcc2.95 kernel)
