Mailing List ArchiveSupport open source code!
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: tlug: Caldera Japanese version (more comments)
- To: tlug@example.com
- Subject: Re: tlug: Caldera Japanese version (more comments)
- From: Jens-Ulrik Petersen <jens-ulrik.petersen@example.com>
- Date: 19 Nov 1999 13:26:53 +0900
- Content-Type: text/plain; charset=us-ascii
- In-Reply-To: "Stephen J. Turnbull"'s message of "Thu, 18 Nov 1999 19:48:04 +0900 (JST)"
- References: <Pine.LNX.4.10.9911170830400.11828-100000@example.com> <m3so245i8z.fsf@example.com> <m3g0y4xim5.fsf@example.com> <14387.55652.230170.788357@example.com>
- Reply-To: tlug@example.com
- Sender: owner-tlug@example.com
- User-Agent: Gnus/5.070096 (Pterodactyl Gnus v0.96) XEmacs/21.1 (Arches)
Stephen J. Turnbull <turnbull@example.com> writes: > >>>>> "juhp" == Jens-Ulrik Petersen <jens-ulrik.petersen@example.com> writes: > > juhp> basically "memory used" as reported by `free` or `cat > juhp> /proc/meminfo` is a high water mark. > > ah, verrrry interrrresting. > > How 'bout that "shared" number? Presumably libc.so never leaves (on > Debian even init is dynamically linked!), but suppose you shut down > all your X11 apps ... do libX11.so.6 et al get flushed from memory? My guess (to know the exact mechanism there is only one thing to do) is that libraries and executables that are no longer being used stay cached in memory until that memory is needed for something else (in which case they are then dropped from the cache on a memory need basis) or if the library (or executable) is still in memory and hasn't been modified on the disk (this part is probably more subtle) its memory registration is changed back from "cached" to "in use by a process". [Hey, this speculating is getting quite fun: I guess it's finally time to go and read "The Linux Kernel" from the LDP <http://www.linuxdoc.org/>.] > I assume they would (just keep a reference count), but I can imagine > that things could get hosed if you're on a development system, and > dynamic libs might not get flushed in case of say a program crash. Yep, this might happen if a process zombies but doesn't die? > That would account for Scott's eventual slows, maybe. Perhaps. Jens ------------------------------------------------------------------- Next Technical Meeting: TBA, January, 2000. Place: Temple Univ.
- References:
- Re: tlug: Caldera Japanese version (more comments)
- From: "Scott M. Stone" <sstone@example.com>
- Re: tlug: Caldera Japanese version (more comments)
- From: Mike Fabian <mike.fabian@example.com>
- Re: tlug: Caldera Japanese version (more comments)
- From: Jens-Ulrik Petersen <jens-ulrik.petersen@example.com>
- Re: tlug: Caldera Japanese version (more comments)
- From: "Stephen J. Turnbull" <turnbull@example.com>
Home | Main Index | Thread Index
- Prev by Date: Re: tlug: Caldera Japanese version (more comments)
- Next by Date: tlug: monitor settings
- Prev by thread: Re: tlug: Caldera Japanese version (more comments)
- Next by thread: Re: tlug: Caldera Japanese version (more comments)
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links