Andrea Arcangeli (firstname.lastname@example.org)
Mon, 31 Aug 1998 10:20:24 +0200 (CEST)
On Sun, 30 Aug 1998, Philip Blundell wrote:
>>It' s a code cleanup not a new feature. Anyway I think I' ll implement it
>Yes, but the point of code-freeze (as opposed to just feature-freeze) is that
>only real bugfixes go in.
I consider the spinlock thing a bugfix. I consider the new irq code very
more robust than the last one.
>>So they have to hack every other Linux driver that use an interrupt
>>handler out there?
>The Amiga doesn't share many drivers. Parport is the only one where the
>high-level drivers are exposed to so many details of the hardware. The only
So they are really lucky.
>>My thought was not allowing the lowlevel parport code to own the irq
>>hanlder but instead allowing the _common_ parport_share code to own the
>>irq so it would not help m68k guys at all.
>Allowing parport_share to own the irq isn't much different to what we have now
Infact is not very different but is cleaner more robust and simpler and
this is the reason I think it should be put in.
>(aside from the switching details). I think that having the lowlevel driver
>in control is definitely the right way to go in the long term.
If it' s really needed yes. The point is why not to put the wrapper in the
irq arch specific code of Amiga port? Is 2.1.119 uptodate with the latest
Amiga stuff? I am going to take a look to the Amiga irq stuff to
understand if it would be very tricky to put the wrapper there.
-- To unsubscribe, send mail to: email@example.com --
-- with the single word "unsubscribe" in the body of the message. --
This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:18:10 EST