Subject: Re: bus_space_barrier() [Re: CVS commit: syssrc]
To: None <email@example.com>
From: Jason R Thorpe <firstname.lastname@example.org>
Date: 04/27/2000 10:00:25
On Thu, Apr 27, 2000 at 06:52:55PM +0200, Wolfgang Solfrank wrote:
> Hmm, are you saying that we should fool around with the bus_space_*
> implementations in the ports to work around bugs in the drivers?
> That way we'd probably never get the drivers fixed...
Well, for a lot of devices, you'd need a barrier after every single
operation. That can become tedious to type, and could be more overhead
in terms of function calling, if the barrier can be in-lined.
What I'm saying is "redefine part of the interface" to match what is
basically current practice, and provide a hook to enable the the more
efficient mechanism for devices which can reasonably take advantage of it.
-- Jason R. Thorpe <email@example.com>