home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.os.linux
- Path: sparky!uunet!nih-csl!midway.dcrt.nih.gov!seidl
- From: seidl@midway.dcrt.nih.gov (Edward Seidl)
- Subject: Re: Swap partition != available swap space
- Message-ID: <1993Jan29.023338.5213@alw.nih.gov>
- Keywords: swap
- Sender: postman@alw.nih.gov (AMDS Postmaster)
- Organization: National Institutes of Health, Bethesda, MD
- References: <1993Jan26.124338.8238@infograph.com> <1993Jan28.170305.15436@cbfsb.cb.att.com>
- Date: Fri, 29 Jan 1993 02:33:38 GMT
- Lines: 39
-
- In article <1993Jan28.170305.15436@cbfsb.cb.att.com> wto@cbnewsg.cb.att.com (William T. O Connell) writes:
- >> [ wants 32Mb swap partition ]
- >>
- >
- >No, it's not related to the kernel config question, thats referring
- >to phy. RAM. - someone correct me if I'm wrong but, linux will use up
- >to 16MB of swap space, even though in your case you allocated 32MBs.
- >If you want to have more then 16MB of swap space you'll have to modify
- >the kernel to support the larger swap space. Someone did post a while
- >back that they got info from linus on how to modify the kernel to
- >do this, I don't believe the change was that big of a deal.
- >
-
- Correct on all counts. The changes to the kernel are not major at all.
- The only file that needed to be changed was linux/mm/swap.c. I also had
- to modify mkswap a bit, as well as free (I only have the kernel version,
- not the proc fs one). I've been running with 32Mb of swap for ~2 weeks
- now...no problems so far. I sent the patches to Linus so that he could
- verify that the changes are neither stupid nor dangerous. Considering that
- 2 kernel patches have been released since then makes me wonder :)
-
- So far I've been loath to give the patches to anyone until I hear back from
- Linus, or, better yet, he sticks the changes (or a better solution) into
- the kernel. If anyone out there can't wait, I'd be willing to send them
- the necessary patches with the understanding that they have been tested
- on exactly one machine, and I make NO promises that catastrophe does not
- await any who apply them to their machine. If enough people are interested,
- I might put them on tsx-11, but I'd really rather wait until I hear
- something from Linus.
-
- >- Bill O'Connell
-
- Ed Seidl
-
- --
- Edward Seidl DCRT/CSL Bldg 12A, Rm 2033
- seidl@alw.nih.gov NIH, Bethesda, MD 20892
-
- I'm not paid enough to have an opinion, much less express one for my employer.
-