Re: IT8212/ITE RAID

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

On Sul, 2005-08-14 at 17:56 +0200, Bartlomiej Zolnierkiewicz wrote:
> * your stuff was accepted after all (and some stuff like ide-cd
>   fixes was never splitted from the -ac patchset and submitted)

They were.

> * you've never provided any technical details on "the stuff I broke"

I did, several times. I had some detailed locking discussions with
Manfred and others on it as a result. The locking in the base IDE is
still broken, in fact its become worse - the random locking around
timing changes now causes some PIIX users to see double spinlock debug
with the base kernel as an example.


> > Would make sense, but I thought I had the right bits masked. Will take a
> 
> WIN_RESTORE is send unconditionally (as it always was),
> 
> This is not the right thing, somebody should go over all ATA/ATAPI
> drafts and come with the correct strategy of handling WIN_RESTORE.

Ok that would make sense. Matthew Garrett also reported some problems in
that area with suspend/resume (BIOS restoring its idea of things...)