Mailinglist Archive: opensuse (4344 mails)

["Greg Wallace" <suse-linux-e@xxxxxxxx>]

On Monday, August 01, 2005 @ 4:51 Am, Jerry Feldman wrote:

>On Friday 29 July 2005 8:51 pm, Greg Wallace wrote:
>> Ok, I follow you now.  If the last process accessing the "text section of
>> a program" ... "instructions (and some initialized read-only data, such
>> as strings)" closes and then you re-start the program, is that part of
>> the program still in RAM, or would it have to be loaded from swap or from
>> the file?
>The text section is never copied over to backing store (swap). The load 
>process simply "maps" that section in from where it resides. 

>I don't know the specific algorithm the kernel uses these days for 
>persistence, but I would assume that if the pages are still resident in RAM

>that they would be reused. This has been a common technique used in Unix 
>systems for years since Unix/Linux commands are generally small 
>executables. The same thing applies to libraries. The core libraries, such 
>as libc.so are used by daemons, so it should be almost continuously 
>resident. 
>-- 
>Jerry Feldman <gaf@xxxxxxx>
>Boston Linux and Unix user group
>http://www.blu.org PGP key id:C5061EA9
>PGP Key fingerprint:053C 73EC 3AC1 5C44 3E14 9245 FB00 3ED5 C506 1EA9

So, taking the numbers from Carlos Robinson's Friday post --

SWAP
1,052,216K TOTAL
224,240K USED
827,976K FREE
404,580K CACHED

Maybe CACHED is the 224,240K USED plus 404,580K - 224,240K = 180,340K
previously used; i. e., modules that were earlier called but are no longer
running.  If you invoke one of those modules again, it gets copied from
CACHED to USED instead of from the medium to USED.

Greg Wallace