Fw: Returned mail: see transcript for details

Fw: Returned mail: see transcript for details

[David S. Miller]

[-- Attachment #1: Type: text/plain, Size: 708 bytes --]


Sorry I'm retarted and mis-typed the list address.

Begin forwarded message:

Date: Thu, 27 May 2004 16:29:57 -0400
From: Mail Delivery Subsystem <MAILER-DAEMON@redhat.com>
To: <davem@redhat.com>
Subject: Returned mail: see transcript for details


The original message was received at Thu, 27 May 2004 16:29:54 -0400
from int-mx1.corp.redhat.com [172.16.52.254]

   ----- The following addresses had permanent fatal errors -----
<linux-arch@osdl.org>
    (reason: 550 5.1.1 <linux-arch@osdl.org>... User unknown)

   ----- Transcript of session follows -----
... while talking to smtp.osdl.org.:
>>> DATA
<<< 550 5.1.1 <linux-arch@osdl.org>... User unknown
550 5.1.1 <linux-arch@osdl.org>... User unknown


[-- Attachment #2: 00000000.mimetmp --]
[-- Type: message/delivery-status, Size: 358 bytes --]

[-- Attachment #3: Type: message/rfc822, Size: 1523 bytes --]

From: "David S. Miller" <davem@redhat.com>
To: torvalds@osdl.org
Cc: linux-arch@osdl.org, jparadis@redhat.com
Subject: ffs() inconsistency
Date: Thu, 27 May 2004 13:29:50 -0700
Message-ID: <20040527132950.3ce28f41.davem@redhat.com>


It seems that x86 and currently x86_64 both return -1 from
ffs() if the word passed in is zero.

In contrast, linux/bitops.h:generic_ffs() returns 0 in this case.
Both sparc ports both return 0 as well.

Aparently XFS depends upon the "-1" return value behavior.
We need to clear things up.

The ffs man page, refernced above all of these implementations
as the specification, says that "0" should be returned.