NetBSD Problem Report #55916

From www@netbsd.org  Sat Jan  9 10:58:50 2021
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits))
	(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 248441A9245
	for <gnats-bugs@gnats.NetBSD.org>; Sat,  9 Jan 2021 10:58:50 +0000 (UTC)
Message-Id: <20210109105848.6A6971A9246@mollari.NetBSD.org>
Date: Sat,  9 Jan 2021 10:58:48 +0000 (UTC)
From: william@25thandClement.com
Reply-To: william@25thandClement.com
To: gnats-bugs@NetBSD.org
Subject: cmp -s + regular files + skipping is broken
X-Send-Pr-Version: www-1.0

>Number:         55916
>Category:       bin
>Synopsis:       cmp -s + regular files + skipping is broken
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    bin-bug-people
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 09 11:00:01 +0000 2021
>Closed-Date:    Thu Jun 10 05:23:40 +0000 2021
>Last-Modified:  Thu Jun 10 05:23:40 +0000 2021
>Originator:     William Ahern
>Release:        MAIN
>Organization:
>Environment:
>Description:
NetBSD failed to pick-up a June 2000 (20-year-old) FreeBSD bugfix to the short-circuit in regular.c:c_regular of cmp(1). See https://svnweb.freebsd.org/base/head/usr.bin/cmp/regular.c?revision=61883&view=markup

OpenBSD had the same problem but the fix was just committed today, January 9, 2021. See misc@openbsd.org thread at https://marc.info/?t=161016193400003&r=1&w=2

Note that FreeBSD accidentally reintroduced the same bug a month later (July 2000) when they added -z support. See my report at https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=252542 Be careful if resync'ing to FreeBSD. DragonflyBSD inherited the FreeBSD regression.

>How-To-Repeat:
Run and compare

$ cmp file1 file2 skip1 skip2

with

$ cmp -s file1 file2 skip1 skip2

where file1 and file2 are regular files, skip1 and skip2 are different, and cmp should be expected to succeed with exit status 0 because the trailing data is identical. The former succeeds but the latter (cmp -s) will fail.

See example code snippet in original OpenBSD report at https://marc.info/?l=openbsd-misc&m=161016188503894&w=2

>Fix:
Move the short-circuit exit in regular.c:c_regular to after len1 and len2 are reduced by skip1 and skip2. See https://svnweb.freebsd.org/base/head/usr.bin/cmp/regular.c?revision=61883&view=markup and https://marc.info/?l=openbsd-misc&m=161017560806516&w=2

>Release-Note:

>Audit-Trail:
From: "Christos Zoulas" <christos@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc: 
Subject: PR/55916 CVS commit: src/usr.bin/cmp
Date: Sat, 9 Jan 2021 10:16:28 -0500

 Module Name:	src
 Committed By:	christos
 Date:		Sat Jan  9 15:16:28 UTC 2021

 Modified Files:
 	src/usr.bin/cmp: regular.c

 Log Message:
 PR/55916: William Ahern: cmp -s + regular files + skipping is broken
 Move test after length is adjusted


 To generate a diff of this commit:
 cvs rdiff -u -r1.24 -r1.25 src/usr.bin/cmp/regular.c

 Please note that diffs are not public domain; they are subject to the
 copyright notices on the relevant files.

State-Changed-From-To: open->feedback
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Wed, 09 Jun 2021 01:34:08 +0000
State-Changed-Why:
Did the commit in January fix it?


From: William Ahern <william@25thandclement.com>
To: gnats-bugs@netbsd.org
Cc: dholland@netbsd.org
Subject: Re: bin/55916 (cmp -s + regular files + skipping is broken)
Date: Wed, 9 Jun 2021 18:45:12 -0700

 On Wed, Jun 09, 2021 at 01:34:08AM +0000, dholland@NetBSD.org wrote:
 > Synopsis: cmp -s + regular files + skipping is broken
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: dholland@NetBSD.org
 > State-Changed-When: Wed, 09 Jun 2021 01:34:08 +0000
 > State-Changed-Why:
 > Did the commit in January fix it?

 It's been awhile so I've the lost context and whatever little scripts I
 might have had lying around debugging it. But it seems to be fixed as
 running the new code against the test case reported to openbsd-misc produces
 the correct result, unlike the cmp built into NetBSD 8.2 (my most recent
 VM).

State-Changed-From-To: feedback->closed
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Thu, 10 Jun 2021 05:23:40 +0000
State-Changed-Why:
fixed, thanks for checking


>Unformatted:

Home	PR Database Search