[Devel] posix locks to kernel

[Pavel Shilovsky]

Здравствуйте.

Было получено следующее сообщение.

Нам рекомендуют в вайне реализовывать именно оригинальное мандатори поведение 
блокировок. Видимо они не сильно представляю себе проблему работы вайна - он 
ведь может работать и на шаре и без неё... Идётся в таком случае различать, 
что довольно неудобно. Примерно в этом ключе я и хочу им ответить. Я прав?

----------  Forwarded Message  ----------

Subject: Re: [linux-cifs-client] [PATCH 0/2] posix locks behaviour on Windows 
server
Date: Thursday 05 of March 2009
From: "Steve French (smfltc)" <smfltc на us.ibm.com>
To: Pavel Shilovsky <piastry на etersoft.ru>

Pavel Shilovsky wrote:
> What is about my patch?
>
>
> --
> Best regards,
> Pavel Shilovsky.
>   
My first reaction was that it added complexity to a mainline path, for a 
case in which posix behavior was unlikely to be understood by developers 
(and this odd quirk of posix locking behavior is unlikely to be an app 
dependency) so would like more convincing evidence that it actually 
breaks an app before turning it on by default (I am not arguing against 
implementing posix behaviors in general, but just that "Linux 
developers" are just as likely to assume the opposite behavior as posix 
in this case, so this is riskier than it looks to turn on by default).

IIRC you noted that Wine depends on this behavior- but was hoping a 
better way to solve Wine's needs was by passing through the mandatory 
locking behavior that Windows apps would need (I realize that this has 
obstacles ... in getting the linux-fsdevel community to understand why 
headers or minor vfs changes would be needed for this)


Jeff,
Thoughts?

-------------------------------------------------------
-- 
Best regards,
Pavel Shilovsky.