lkml.org 
[lkml]   [1998]   [Dec]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: PATCH: Raw device IO for 2.1.131

Nathan Hand writes:
> On 15 Dec 1998, Harald Milz wrote:

>> As a service professional, I learnt to a) listen to the customer's
>> wishes and b) not arguing against them of, and c) to talk to
>> development if my product doesn't fulfil them. This is proven to work.
>
> Sure, but unlike commercial systems here the developers have all of
> the power, because no pay cheques can be withheld or jobs lost.
>
> In other words, without meaning to seem harsh, nobody has to listen
> to you talk. Nobody has to care about the customer.

Please have pity on the geeks who are forced to use other operating
systems because Linux doesn't satisfy the PHBs. It is cruel to turn
down a patch that might free someone from a crummy OS without open
source code. Your fellow hackers are suffering.

Surely you don't want to see your friends stuck with NT and SCO.

> It's not even clear that raw IO is the proper solution. It may have
> been what some systems have implemented, but that doesn't mean it's
> the only way. As I see, raw IO basically says "this problem is much
> too hard so let the application figure this out". It's a cop out.

Well, yes. The kernel is general-purpose. There will always be
applications that are poorly supported by the normal assumptions.
If the kernel gets hacked to support one such odd application,
performance may suffer for other applications and the kernel
will be bloated by code for the abnormal cases.

> And I'm personally enjoying the discussions about alternatives such
> as copyfd() or zero-copy write()/read() calls. If there is a better
> solution, then it makes sense to say "NO" now to raw IO, because it
> would make it harder to adopt the better solution in the future.

It looked very well done. I don't know of any other unix clone that
lets you mark a normal file for unbuffered IO. Maybe Irix does.

> Plus you always have the option of creating raw IO in a new tree or
> as a standalone kernel module.

You can't expect Oracle to support unofficial kernel patches.
I doubt many people would want to run a video server on unofficial
kernel patches either. It's not a safe way to run your business.
(you might need to make use of the GPL, with your own developers)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:46    [W:0.058 / U:0.256 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site