GNU bug report logs - #8294
Minor nit with dd

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 8294 in the body.
You can then email your comments to 8294 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: John Morris <jmorris <at> beau.org>
To: bug-coreutils <at> gnu.org
Subject: Minor nit with dd
Date: Sat, 19 Mar 2011 13:26:04 -0500

[Message part 1 (text/plain, inline)]

Found an odd power of two size limit in dd. :)

On a 64bit platform shouldn't the following command have done the
expected thing instead of quietly clipping to 2^31?

[jmorris <at> mjolnir ~]$ dd if=/dev/zero of=/dev/null bs=3221225472 count=1
0+1 records in
0+1 records out
2147479552 bytes (2.1 GB) copied, 1.94223 s, 1.1 GB/s
[jmorris <at> mjolnir ~]$ rpm -qf `which dd`
coreutils-7.6-11.fc12.x86_64

[signature.asc (application/pgp-signature, inline)]

Message #8 received at 8294 <at> debbugs.gnu.org (full text, mbox):

From: Paul Eggert <eggert <at> cs.ucla.edu>
To: John Morris <jmorris <at> beau.org>
Cc: 8294 <at> debbugs.gnu.org
Subject: Re: bug#8294: Minor nit with dd
Date: Sat, 19 Mar 2011 19:03:48 -0700

On 03/19/2011 11:26 AM, John Morris wrote:
> On a 64bit platform shouldn't the following command have done the
> expected thing instead of quietly clipping to 231?

It depends on what your platform's 64-bit 'read'
system call does.  My guess is that your platform's
'read' syscall refuses to read more than
2**31 - 2**12 bytes at a time from /dev/zero,
so the problem (if any :-) is with 'read', not
with 'dd'.

Message #11 received at 8294 <at> debbugs.gnu.org (full text, mbox):

From: Jim Meyering <jim <at> meyering.net>
To: Paul Eggert <eggert <at> cs.ucla.edu>
Cc: John Morris <jmorris <at> beau.org>, 8294 <at> debbugs.gnu.org
Subject: Re: bug#8294: Minor nit with dd
Date: Sun, 17 Apr 2011 09:05:52 +0200

tags 8294 moreinfo
close 8294
thanks

Paul Eggert wrote:
> On 03/19/2011 11:26 AM, John Morris wrote:
>> On a 64bit platform shouldn't the following command have done the
>> expected thing instead of quietly clipping to 231?
>
> It depends on what your platform's 64-bit 'read'
> system call does.  My guess is that your platform's
> 'read' syscall refuses to read more than
> 2**31 - 2**12 bytes at a time from /dev/zero,
> so the problem (if any :-) is with 'read', not
> with 'dd'.

Thanks for replying, Paul.

John,
If the read system call were failing and dd were ignoring that failure,
that would be a bug.  However, dd is careful to check for that sort of
thing.  Here, read is returning after readling less than a full block,
and that's fine.  It's a so-called short read.  If you want dd to
accumulate full blocks, use GNU dd's iflag=fullblock option.

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.