Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site rochester.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!mhuxn!mhuxr!ulysses!allegra!mit-eddie!genrad!panda!talcott!harvard!seismo!rochester!bukys
From: bukys@rochester.UUCP (Liudvikas Bukys)
Newsgroups: net.bugs.4bsd
Subject: update() still writes out all inodes, because of parenthesis error
Message-ID: <7281@rochester.UUCP>
Date: Thu, 14-Mar-85 07:43:58 EST
Article-I.D.: rocheste.7281
Posted: Thu Mar 14 07:43:58 1985
Date-Received: Sat, 16-Mar-85 05:38:34 EST
Distribution: net
Organization: U. of Rochester, CS Dept.
Lines: 24
Keywords: sync, update, "4.3"
Index: 	sys/ufs_subr.c 4.2 Fix
To: 4bsd-bugs@berkeley.arpa

Description:
	Certain souped-up versions of 4.2 include an extra test in
	update() which will cause only modified inodes to be written
	out by sync.  The test is parenthesized wrong, though, so is
	effectively missing; behaviour will be the same as before (all
	inodes get written).

Repeat-By:
	Lint the kernel.  Notice the message
	"../sys/ufs_subr.c(72): warning: constant in conditional context".
	Look at Kernighan&Ritchie: '==' has higher precedence than '&'.

		if ((ip->i_flag & ILOCKED) != 0 || ip->i_count == 0 ||
		    (ip->i_flag & (IACC|IUPD|ICHG) == 0))
			continue;

Fix:
	Put parentheses around the '&' expression.
	--> I'll leave that to someone else to test. <--

Brought-to-you-by:
	Liudvikas Bukys
	rochester!bukys (uucp) via allegra, decvax, seismo
	bukys@rochester (arpa)