Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 15 Oct 2001 08:30:01 -0700 (PDT)
From:      Bruce Evans <bde@zeta.org.au>
To:        freebsd-bugs@FreeBSD.org
Subject:   Re: bin/31199: tunefs error is incorrect when enabling softupdates
Message-ID:  <200110151530.f9FFU1F89614@freefall.freebsd.org>

next in thread | raw e-mail | index | archive | help
The following reply was made to PR bin/31199; it has been noted by GNATS.

From: Bruce Evans <bde@zeta.org.au>
To: "Crist J. Clark" <cristjc@earthlink.net>
Cc: <freebsd-gnats-submit@FreeBSD.ORG>
Subject: Re: bin/31199: tunefs error is incorrect when enabling softupdates
Date: Tue, 16 Oct 2001 01:22:37 +1000 (EST)

 On Mon, 15 Oct 2001, Crist J. Clark wrote:
 
 >  On Wed, Oct 10, 2001 at 11:46:07AM -0400, Rob Simmons wrote:
 >  ...
 >  > >How-To-Repeat:
 >  > run "tunefs -n enable <fs>" on a mounted filesystem.
 >  >
 >  > You will get the following output:
 >  >
 >  > bash-2.05$ tunefs -n enable /dev/ad0s1a
 >  > tunefs: soft updates set
 >  > tunefs: cannot open /dev/ad0s1a: Permission denied
 >
 >  Looking at the code, I think the least obtrusive way to fix this is to
 >  slightly reword the messages a little more clear. tunefs(8) goes
 >  through all of the options before it actually tries to modify the file
 >  system.
 
 This was broken in rev.1.12.  tunefs previously opened the device with
 mode O_RDWR (spelled as "2") up front in getsb().  This was bogus for
 the -p case since it required write access to display the settings, but
 is needed for all other cases since tunefs writes even null changes.
 
 >  The messages should reflect that it is collecting the values
 >  before it actually tries to write them. Something more like,
 >
 >    $ tunefs -n enable /dev/ad0s1a
 >    tunefs: setting soft updates...
 >    tunefs: cannot open /dev/ad0s1a: Permission denied
 
 The "setting" message used to be "soft updates set".  I changed this to
 "soft updates changes from disabled to enabled" (and similarly for
 "soft updates cleared" so that the wording is similar for all the messages.
 
 >  And in a successful run (since tunefs(8) is already chatty on
 >  success),
 >
 >     $ tunefs -n enable /dev/ad0s1a
 >    tunefs: setting soft updates...
 >    tunefs: changes on /dev/ad0s1 done
 
 The "changes done" message is silly if tunefs has only printed "remains
 unchanged" messages.
 
 >  This patch look good?
 >
 >  Index: tunefs.c
 >  ===================================================================
 >  RCS file: /export/ncvs/src/sbin/tunefs/tunefs.c,v
 >  retrieving revision 1.20
 >  diff -u -r1.20 tunefs.c
 >  --- tunefs.c	2001/09/30 14:57:08	1.20
 >  +++ tunefs.c	2001/10/15 09:43:07
 >  @@ -231,7 +231,7 @@
 >   			warnx("%s remains unchanged as %d", name, avalue);
 >   		}
 >   		else {
 >  -			warnx("%s changes from %d to %d",
 >  +			warnx("changing %s from %d to %d...",
 >   					name, sblock.fs_maxcontig, avalue);
 >   			sblock.fs_maxcontig = avalue;
 >   		}
 
 I don't think this is any better.  The unusual wording in the original was
 apparently chosen to put the subject first.  Neither version claims to have
 completed the changes (that would be "%s changed").
 
 Bruce
 

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200110151530.f9FFU1F89614>