DragonFly kernel List (threaded) for 2005-06
Re: CALL TO COMMITTERS! HELPING WITH SPL REMOVAL!
On Thu, Jun 02, 2005 at 10:33:06AM -0700, Matthew Dillon wrote:
> And, again, I completely disagree with your performance concerns. I
> challenge you to find a case where performance would be degraded.
> The most widely used long-held SPLs in the system are SPLBIO and SPLVM,
> and guess what? They pretty much mask ALL MAJOR DEVICES anyway. So
> what's the point of having a variable mask in the first place? No
> point at all...
Let me find a different way for the first step. I want to measure (and feel)
the system before the change all of the tree. I have an idea how I can
detect the spl() abuse (which has to be fixed anyway) and how we can
replace the rest of the spls with critical sections to investigate the
effect. The second step becomes mostly mechanical and I want to write
a script to handle that first. Give me a few days to get it sorted out,
since I also have to learn for a written exam on Monday :)
This is not so much about whether the change is good or bad, but about
*rushing* it in without thinking about the results first. We are not