Re: [PATCH 2.6.25.1] Add scsi_execute_async_fifo()

[Vladislav Bolkhovitin <vst@vlnb.net>]

James Bottomley wrote:
> On Fri, 2008-05-02 at 18:06 +0200, Bart Van Assche wrote:
>> On Fri, May 2, 2008 at 5:55 PM, Christoph Hellwig <hch@infradead.org> wrote:
>>> On Fri, May 02, 2008 at 05:53:22PM +0200, Bart Van Assche wrote:
>>>> Regarding out-of-tree modules: this is just a preparatory step before
>>>> submitting SCST for inclusion in the mainstream kernel.
>>> And what crackpipe did you smoke to thing we'd put duplicated target
>>> framework in?
>> Why are you so aggressive ? I didn't insult you in any way.
>>
>> Regarding inclusion of SCSI target code in the mainline, this subject
>> has already been discussed extensively in the past
>> (http://lkml.org/lkml/2008/1/23/134). The conclusion was clear: SCST
>> is faster than any other existing iSCSI target for Linux (IET, STGT,
>> LIO), stable, well maintained and the most standards compliant target.
>> Why do you want to reopen this discussion ?
> 
> That's an interesting rewrite of history.  The evidence you presented
> showed fairly identical results apart from on one contrived IB benchmark
> that couldn't directly compare the two.
> 
> I'm also on record in the thread saying that was insufficient proof for
> me to justify throwing STGT out and replacing it with SCST.

James, why do you keep ignoring important points, written by me in that 
e-mail: http://lkml.org/lkml/2008/1/30/178?

Namely:

1. Solid architecture of SCST is inherently more simple than distributed 
user/kernel space processing, when kernel behaves under control of user 
space, used in STGT, and allows to get better results with less effort. 
Better in all aspects: simplicity (hence, maintainability), reliability 
and performance. Linux once made step away from microkernel based design 
and that was for really good reasons.

2. Zero-copy operations with page cache will halve processing latency on 
high speed links, like InfiniBand, and it is impossible to implement 
that in a sane way with STGT approach, while for SCST it can be 
implemented simply and naturally.

Vlad