Re: [PATCH] 2.5: push BKL out of llseek

[Anton Blanchard]

> This patch pushes the BKL out of llseek() and into the individual llseek
> methods.  For generic_file_llseek, I replaced it with the inode
> semaphore.  The lock contention is noticeable even on 2-way systems. 
> Since we simply push the BKL further down the call chain (its the llseek
> method's responsibilities now) we aren't doing anything hackish or
> unsafe.

A great idea, I wonder why someone didnt think of it before?

http://banyan.dlut.edu.cn/news/112801/0186.html

Wow someone did! And the patch basically matches yours including whitespace.

Robert _please_ attribute your sources more carefully. 

Anton