> It's kinda hard to work around a bug in the Bridge. There just isn't
People do
> all that much to program. You just write 0x07 to the command register to
> turn it on. Whatever bits stick are the ones supported by the device.
Get the full manuals for the chipsets. There are lots of features you can
control (streaming, peer concurrency, prefetch etc) and on quite a few chipsets
you have to turn some of them off to make it work.
> enabled. It either works or it doesn't. So as I said, it's unlikely that
> they would market a bad chip.
Triton has bugs, VIA have bugs, SIS have bugs, everyone has bugs
> PCI Interface Chips that support complex devices like a 'Mass storage
> controller' can be filthy with bugs, however the bridge chip itself
> either works or it doesn't.
Wrong.
Some examples
o SIS 486 chipsets. PCI dma to another PCI target directly doesnt work
o VIA VP* PCI/ISA bridge gets ISA bus mastering DMA disable wrong and
locks the board
o Triton you have to turn off streaming and peer concurrency for some
configurations
There is a reason we have pci-quirks.c
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Mon Jul 31 2000 - 21:00:35 EST