View Single Post
Old 23rd November 2010
TerryP's Avatar
TerryP TerryP is offline
Arp Constable
 
Join Date: May 2008
Location: USofA
Posts: 1,547
Default

Quote:
Originally Posted by vermaden View Post
Why You sent that link as response to my post?
You could say I'm one of those freakish there is a right way and a wrong way to solve problems sort of guy. You know, the kind who thinks using chewing gum is supposed to be a temporary solution, not a fix for problems coming straight out of the factory.


If you can improve performance: should it be done at the level of the thing responsible for it (hint: not the user) and better yet, maybe have it become tunable by the system administrator (BSD via a sysctl, Linux via the file system) Or should it be a cobbled together on a per user basis, sooner or later to fall out of sync with reality, and probably be so much more likely to see random bugs slip in?

While I don't always agree with Linus (and sometimes curse at him for one thing or another), I think in that posting, he was quite right. The most a user space implementation should be, is a configurable daemon, and it would still be faster done in kernel. It's called "Monolithic" for a reason after all.
__________________
My Journal

Thou shalt check the array bounds of all strings (indeed, all arrays), for surely where thou typest ``foo'' someone someday shall type ``supercalifragilisticexpialidocious''.
Reply With Quote