Return to BSD News archive
Received: by minnie.vk1xwt.ampr.org with NNTP
id AA2355 ; Mon, 01 Mar 93 10:55:33 EST
Path: sserve!manuel.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!saimiri.primate.wisc.edu!zaphod.mps.ohio-state.edu!howland.reston.ans.net!paladin.american.edu!news.univie.ac.at!hp4at!mcsun!sun4nl!tuegate.tue.nl!rwb.urc.tue.nl!rcpt
From: rcpt@rwb.urc.tue.nl (Piet Tutelaers)
Newsgroups: comp.os.386bsd.bugs
Subject: Re: reproducable swap_pager_finish() problem
Message-ID: <rcpt.730840140@rwb.urc.tue.nl>
Date: 27 Feb 93 19:09:00 GMT
References: <rcpt.730677846@rw8.urc.tue.nl>
Sender: root@tuegate.tue.nl
Reply-To: rcpt@urc.tue.nl
Lines: 39
rcpt@rw8.urc.tue.nl (Piet Tutelaers) writes:
>I get a reproducable kernel error message when I try to compile
>gcc-2.3.3. The problem occurs in the next step:
> cc -c -DIN_GCC -g -I. -I./config cp-parse.c
>I get a huge amount of messages like:
> Feb 25 21:56:20 r06 /386bsd: error 5 blkno 9704 sz 4096
> swap_pager_finish: clean of page 51c000 failed
>I have a 486DX33 system with 8MB memory and use a `self-made' kernel
>including all patches from the 0.2.1 patchkit except for 504 and 600.
>I don not use a separate swap partition at the moment. I have located
>the swap_pager_finish() function in /sys/vm. Patches in this area are:
>75 unmatched spl() call in VM system
>74 pages lost in VM system
>63 kernel speedup
>07 Greenman bugfix
>02 big memeory hang
>Which patch could cause this trouble? I must admit that the system has
>hardware (?) problems with MSWindows3.1 (spontaneous reboots) but with
>UNIX it worked stable so far. Which wizzard can enlighten me?
New facts:
1) the problem does not occur whith my old kernel (pre-patchkit 0.2.1)
including patch00002.
2) I have a kernel without patches 75, 74, 63 and 07 (plus the patches it
depends on: 42 and 26) with the same problem. So the problem is caused
by another patch and probably not by hardware errors.
3) It would be interesting to make a separate swap partition and see if
the problem will vanish.
4) other suggestions are welcomed.
--Piet