Date: Wed, 02 Jan 2008 13:19:18 -0800 From: Nate Lawson <nate@root.org> To: Poul-Henning Kamp <phk@phk.freebsd.dk> Cc: cvs-src@freebsd.org, src-committers@freebsd.org, cvs-all@freebsd.org Subject: Re: cvs commit: src/sys/dev/ata ata-all.h Message-ID: <477BFFD6.1080502@root.org> In-Reply-To: <1429.1199308501@critter.freebsd.dk> References: <1429.1199308501@critter.freebsd.dk>
next in thread | previous in thread | raw e-mail | index | archive | help
Poul-Henning Kamp wrote:
> In message <477BFB9D.8070201@root.org>, Nate Lawson writes:
>> Poul-Henning Kamp wrote:
>
>> Ok, but this still doesn't address the performance loss fixed by this
>> 3-year-old patch:
>
> I think your patch is wrong because it moves the responsibility for
> the performance in the wrong direction: We shouldn't have to care
> about this in every single device driver.
>
> When I created bus_{read|write}_* that was mostly to reduce the
> risk of repetive strain injuries for device driver writers, but
> it was also an attempt to move the responsibility in the right
> direction.
>
> rman and bus_space was sort of conceived separately, but it is
> time they realize that they are indeed a couple so everybody don't
> have to pretend how surprised they are to see them together at
> this instant.
I'm happy to see this encapsulated, as long as it gets rid of the
unnecessary overhead of what should be merely a structure member
dereference.
Thanks,
--
Nate
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?477BFFD6.1080502>