Re: PATCH, IDE corruption, 2.4.18

[Bartlomiej Zolnierkiewicz]

> Explanation: some code now differs in the code path concerned, and
> ide_register_subdriver now only calls ide_dma_check for UDMA drives
> (previously all DMA drives), but ultimately ide_dma_check still ends up
> in ide_config_drive_speed, and that's still fuctionally the same as
> 2.4.

You've got been mistaken by unfortunate name (Martin changed
name dmaproc() to udma() in 2.5.12).
Code calls ide_dma_check for chipsets which registerered udma()
handler (formerly dmaproc()), I think the same 2.4 does.

btw. udma() name is really misleading,
     it should be read (u)dma() not udma() :)

--
bkz