Debian Bug report logs - #533360
libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Display info messages

Message #5 received at submit@bugs.debian.org (full text, mbox, reply):

From: Nick Black <dank@qemfd.net>

To: Debian Bug Tracking System <submit@bugs.debian.org>

Subject: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Tue, 16 Jun 2009 16:06:47 -0400

Package: libc6
Version: 2.9-13
Severity: important


Howdy!

This morning, I upgraded libc6 et al from 2.9-13 to 2.9-15. Immediately,
one of my unit tests began failing with a -1 return from signalfd():

14|***** Error (Invalid argument) in Signalfd(): Couldn't generate signalfd from -1/526336

with libc2.9-13, signalfd() returns a new file descriptor as expected.

I'm passing SFD_CLOEXEC|SFD_NONBLOCK (0x2004000), but the same behavior
is seen with either SFD_CLOEXEC or SFD_NONBLOCK (these are the only two
signalfd() flags described in the man page, last updated 2009-01-13).
With no flags passed, 2.9-15 works and creates the new signalfd.

-- System Information:
Debian Release: squeeze/sid
  APT prefers unstable
  APT policy: (990, 'unstable')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.30 (SMP w/2 CPU cores)
Locale: LANG=en_US.UTF-8, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages libc6 depends on:
ii  libgcc1                       1:4.4.0-6  GCC support library

libc6 recommends no packages.

Versions of packages libc6 suggests:
ii  glibc-doc                     2.9-15     GNU C Library: Documentation
ii  locales                       2.9-15     GNU C Library: National Language (

-- debconf information:
  glibc/upgrade: true
* glibc/disable-screensaver:
  glibc/restart-failed:
* glibc/restart-services: spamassassin snmpd rsync postfix cups cron

Message #10 received at 533360@bugs.debian.org (full text, mbox, reply):

From: Aurelien Jarno <aurelien@aurel32.net>

To: Nick Black <dank@qemfd.net>, 533360@bugs.debian.org

Cc: control@bugs.debian.org

Subject: Re: Bug#533360: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Tue, 16 Jun 2009 22:39:11 +0200

notfound 533360 2.9-13
found 533360 2.9-15
thanks

On Tue, Jun 16, 2009 at 04:06:47PM -0400, Nick Black wrote:
> Package: libc6
> Version: 2.9-13

Given your description that's clearly the wrong version.

> Severity: important
> 
> 
> Howdy!
> 
> This morning, I upgraded libc6 et al from 2.9-13 to 2.9-15. Immediately,
> one of my unit tests began failing with a -1 return from signalfd():
> 
> 14|***** Error (Invalid argument) in Signalfd(): Couldn't generate signalfd from -1/526336
> 
> with libc2.9-13, signalfd() returns a new file descriptor as expected.
> 
> I'm passing SFD_CLOEXEC|SFD_NONBLOCK (0x2004000), but the same behavior
> is seen with either SFD_CLOEXEC or SFD_NONBLOCK (these are the only two
> signalfd() flags described in the man page, last updated 2009-01-13).
> With no flags passed, 2.9-15 works and creates the new signalfd.
> 

I guess it is due to the fact that I built glibc 2.9-15 in squeeze, as
it is not buildable anymore in unstable due to the /lib32 transition.

As it only affects amd64, I'll do an binNMU using a (hackish) build
environment made of parts of squeeze and parts of sid.

-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Message #19 received at 533360@bugs.debian.org (full text, mbox, reply):

From: Nick Black <dank@qemfd.net>

To: Aurelien Jarno <aurelien@aurel32.net>

Cc: 533360@bugs.debian.org

Subject: Re: Bug#533360: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Tue, 16 Jun 2009 16:43:29 -0400

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

Aurelien Jarno left as an exercise for the reader:
> notfound 533360 2.9-13
> found 533360 2.9-15
> thanks
> 
> On Tue, Jun 16, 2009 at 04:06:47PM -0400, Nick Black wrote:
> > Package: libc6
> > Version: 2.9-13
> 
> Given your description that's clearly the wrong version.

I reported the bug from a machine I'd downgraded the package on; sorry for
the confusion. If I upgrade to 2.6-15 right now, I can't run my apps =].
The version should indeed reflect 2.9-15.

> I guess it is due to the fact that I built glibc 2.9-15 in squeeze, as
> it is not buildable anymore in unstable due to the /lib32 transition.

Indeed, see bug 533362. If that bug was filed foolishly, I ought go note
that. Thanks!

> As it only affects amd64, I'll do an binNMU using a (hackish) build
> environment made of parts of squeeze and parts of sid.

I appreciate it, and will be able to confirm the fix whenever it rolls out.

--nick

-- 
                                      Nick Black <dankamongmen@acm.org>
                                             Principal Engineer, McAfee
                                  Grad student, GT College of Computing
                                  Khalepa ta kala; per ardua, ad astra.

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

Message #24 received at 533360@bugs.debian.org (full text, mbox, reply):

From: Aurelien Jarno <aurelien@aurel32.net>

To: Nick Black <dank@qemfd.net>, 533360@bugs.debian.org

Subject: Re: Bug#533360: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Wed, 17 Jun 2009 09:09:18 +0200

Nick Black a écrit :
> Aurelien Jarno left as an exercise for the reader:
>> notfound 533360 2.9-13
>> found 533360 2.9-15
>> thanks
>>
>> On Tue, Jun 16, 2009 at 04:06:47PM -0400, Nick Black wrote:
>>> Package: libc6
>>> Version: 2.9-13
>> Given your description that's clearly the wrong version.
> 
> I reported the bug from a machine I'd downgraded the package on; sorry for
> the confusion. If I upgrade to 2.6-15 right now, I can't run my apps =].
> The version should indeed reflect 2.9-15.
> 
>> I guess it is due to the fact that I built glibc 2.9-15 in squeeze, as
>> it is not buildable anymore in unstable due to the /lib32 transition.
> 
> Indeed, see bug 533362. If that bug was filed foolishly, I ought go note
> that. Thanks!
> 
>> As it only affects amd64, I'll do an binNMU using a (hackish) build
>> environment made of parts of squeeze and parts of sid.
> 
> I appreciate it, and will be able to confirm the fix whenever it rolls out.

This should now be fixed in version 2.9.15+b1 which should already be on
the mirrors.

-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Message #29 received at 533360@bugs.debian.org (full text, mbox, reply):

From: Nick Black <dank@qemfd.net>

To: Aurelien Jarno <aurelien@aurel32.net>

Cc: 533360@bugs.debian.org

Subject: Re: Bug#533360: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Wed, 17 Jun 2009 04:08:13 -0400

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

Aurelien Jarno left as an exercise for the reader:
> Nick Black a écrit :
> This should now be fixed in version 2.9.15+b1 which should already be on
> the mirrors.

I can confirm a fix on amd64 2.6.30 with 2.9-15+b1. Thanks, Aurelien!

-- 
                                      Nick Black <dankamongmen@acm.org>
                                             Principal Engineer, McAfee
                                  Grad student, GT College of Computing
                                  Khalepa ta kala; per ardua, ad astra.

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

Message #34 received at 533360-done@bugs.debian.org (full text, mbox, reply):

From: Aurelien Jarno <aurelien@aurel32.net>

To: Nick Black <dank@qemfd.net>

Cc: 533360-done@bugs.debian.org

Subject: Re: Bug#533360: libc6: signalfd creation with SFD_* flags breaks under 2.9-15

Date: Wed, 17 Jun 2009 20:37:47 +0200

Version: 2.9.15+b1

On Wed, Jun 17, 2009 at 04:08:13AM -0400, Nick Black wrote:
> Aurelien Jarno left as an exercise for the reader:
> > Nick Black a écrit :
> > This should now be fixed in version 2.9.15+b1 which should already be on
> > the mirrors.
> 
> I can confirm a fix on amd64 2.6.30 with 2.9-15+b1. Thanks, Aurelien!
> 

Thanks, for your confirmation, I am therefore closing the bug.


-- 
Aurelien Jarno	                        GPG: 1024D/F1BCDB73
aurelien@aurel32.net                 http://www.aurel32.net

Send a report that this bug log contains spam.