Dave then discovered he was wrong. Please read all DaveM's mail on
the subject. That was a PIO3/4 blacklist - which isnt DMA related.
No, the final clarification on what I know is:
1) _Our_ CMD640 driver uses the bogus "40" pci bus speed value to
determine it's workaround PIO timings for this broken chipset,
what you are remembering it my correction to Linus where he thought
this bogus pci bus speed value was used to calculate DMA timings,
when in fact it does not and never has
(I made this correction because Linus was pulling farts out of his
ass and not even looking at the code he was commenting on, this
was flat out ranting and I wanted to put a cork to it)
2) The CMD646 binary windows driver I disassembled supports DMA (and
on CMD646+ models, UDMA due to bugs in the non-plus UDMA hw) and
had a blacklist (in this driver) for drives to not perform DMA to,
or drives to use more restrictive DMA timings with.
And furthermore all README's I have ever taken a look at for vendor
supplied windows drivers for any modern IDE chipset uses DMA and
UltraDMA where it can, this is a fact. I would heavily encourage
people to go grab various CMD, PIIX, etc. windows driver zip archives
(most can be obtained on the net at some PeeCee archive or directly
fro the vendors webpages), and scan the README's in there quickly,
there are probably some pseudo black lists in there which would be
helpful to us.
Later,
David S. Miller
davem@dm.cobaltmicro.com
-
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.altern.org/andrebalsa/doc/lkml-faq.html