[prev in list] [next in list] [prev in thread] [next in thread] 

List:       eros-arch
Subject:    last bits on malloc
From:       Jonathan S. Shapiro jsshapiro () earthlink ! net
Date:       1998-05-01 14:46:12
[Download RAW message or body]

Well, coming off of yesterday's non-drug-induced haze (oh well), the
right short-term solution becomes pretty obvious.

I'll allocate a fixed-size malloc pool within the kernel and allocate
out of that.

One would like that malloc pool to be able to grow and shrink, trading 
pages back and forth with the page pool.

The first problem is that kernel malloc mostly needs to be prompt, and
therefore cannot wait for cleaning.  For reasons discussed long ago,
stealing only clean pages is not a desirable solution.


shap

[prev in list] [next in list] [prev in thread] [next in thread]