Report information
The Basics
Id:
31992
Status:
resolved
Estimated:
24 hours (1,440 minutes)
Worked:
4 minutes
Users:
mcnally: 2 minutes
Left:
24 hours (1,440 minutes)
Priority:
Low/Low
Queue:

People
Owner:
Nobody in particular
Cc:
AdminCc:

BugTracker
Version Fixed:
4.4.0 4.3.6 4.1-ESV-R15
Version Found:
(no value)
Versions Affected:
(no value)
Versions Planned:
4.3.6
Priority:
P1 High
Severity:
S2 Normal
CVSS Score:
(no value)
CVE ID:
(no value)
Component:
(no value)
Area:
(no value)

Dates
Created:Tue, 27 Nov 2012 12:32:12 -0500
Updated:Fri, 28 Jul 2017 16:11:30 -0400
Closed:Wed, 26 Apr 2017 09:21:48 -0400



This bug tracker is no longer active.

Please go to our Gitlab to submit issues (both feature requests and bug reports) for active projects maintained by Internet Systems Consortium (ISC).

Due to security and confidentiality requirements, full access is limited to the primary maintainers.

Subject: linux interface discovery fix using getifaddrs
Date: Tue, 27 Nov 2012 18:31:57 +0100
To: dhcp-bugs@isc.org
From: Marius Tomaschewski <mt@suse.de>
Unlike dhcp 3.x, dhcp 4.x scans interfaces from /proc/net/dev, which provides only true interface names. When the address set on the interface has a label assigned (linux 2.0 alias interface compatibility), then the SIOCGIFADDR requires the label / alias name as argument instead of the interface name to return this address. When this is the only address assigned to an interface, dhcp-server is unable to find any address and fails to start. Changed to use getifaddrs() function, which retrieves all IP addresses on linux systems and is available since GLIBC 2.3. Gruesse / Regards, Marius Tomaschewski <mt@suse.de>, <mt@suse.com> -- SUSE LINUX Products GmbH, HRB 16746 (AG Nürnberg), GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, Maxfeldstraße 5, 90409 Nürnberg, Germany

Message body is not shown because sender requested not to inline it.

On Tue Nov 27 17:32:12 2012, mt@suse.de wrote:
> Unlike dhcp 3.x, dhcp 4.x scans interfaces from /proc/net/dev,
> which provides only true interface names. When the address set
> on the interface has a label assigned (linux 2.0 alias interface
> compatibility), then the SIOCGIFADDR requires the label / alias
> name as argument instead of the interface name to return this
> address. When this is the only address assigned to an interface,
> dhcp-server is unable to find any address and fails to start.
>
> Changed to use getifaddrs() function, which retrieves all IP
> addresses on linux systems and is available since GLIBC 2.3.
>
> Gruesse / Regards,
> Marius Tomaschewski <mt@suse.de>, <mt@suse.com>

Marius --

Thank you for your suggested fix.  I believe we have at least one other
similar proposed fix in the pipeline but I will make sure that the developers
see your suggestion.

Michael McNally
ISC Support


I've just pushed on the review queue a similar propose patch to use getifaddrs() for Linux.
Hello Marius: You'll be pleased to hear that we have corrected this issue in our upcoming releases 4.3.6 and 4.1-ESV-R15 both due out in May/2017 as well as 4.4.0 (date TBD). You were among others who reported the issue and contribute solutions. We'll cite your contribution as we have in the past in our release notes. Thanks for the patch and your continued interest and contributions. Sincerely, Thomas Markwalder ISC Software Engineering
Subject: Re: [ISC-Bugs #31992] linux interface discovery fix using getifaddrs [patch]
Date: Wed, 22 Feb 2017 00:35:52 +0100
To: dhcp-review@isc.org, "Nirmoy Das" <ndas@suse.de>
From: "Marius Tomaschewski" <mt@suse.de>
Hello Thomas! Thanks, this is very appreciated! Nirmoy (in Cc) took over the maintenance of our ISC dhcp package, but I continue to act as his fallback & reviewer. I've attached a ticked mail and a patch I'd like to incorporate upstream. Would be nice to hear, what you'd need to apply it. Short summary about infiniband: On a raw socket, the link-layer infiniband packet provides only some "useless" things, basically just the protocol number (there was an RFC about, but I don't remember it now). There is no source hardware address visible like on ethernet, so client-identifier is all what you have and infiniband dhcp RFC wants the new 0xff + iaid + duid at this place instead of hwtype + mac, zeroed and zero-len chaddr as the infiniband hwaddr is too long for chaddr and the client enables broadcast response flag in the dhcp header, because there is no source hwaddr visible on incoming packets the server could use for responses. Am 21.02.2017 um 16:32 schrieb Thomas Markwalder via RT: > Hello Marius: > > You'll be pleased to hear that we have corrected this issue in our upcoming releases 4.3.6 and 4.1-ESV-R15 both due out in May/2017 as well > as 4.4.0 (date TBD). You were among others who reported the issue and contribute solutions. > > We'll cite your contribution as we have in the past in our release notes. Thanks for the patch and your continued interest and contributions. > > Sincerely, > > Thomas Markwalder > ISC Software Engineering Gruesse / Regards, Marius Tomaschewski <mt@suse.de>, <mt@suse.com> -- SUSE LINUX GmbH, GF: Felix Imendörffer, Jane Smithard, Dilip Upmanyu, Graham Norton, HRB 21284 (AG Nürnberg), Maxfeldstraße 5, 90409 Nürnberg, Germany

Message body is not shown because sender requested not to inline it.

Message body is not shown because sender requested not to inline it.

Subject: [ISC-Bugs #44738] AutoReply: Infiniband support patch
Date: Tue, 21 Feb 2017 23:16:24 +0000
To: mt@suse.de
From: "DHCP Bugs via RT" <dhcp-bugs@isc.org>
Greetings, This message has been automatically generated in response to the creation of a trouble ticket regarding: "Infiniband support patch", a summary of which appears below. There is no need to reply to this message right now. Your ticket has been assigned an ID of [ISC-Bugs #44738]. Please include the string: [ISC-Bugs #44738] in the subject line of all future correspondence about this issue. To do so, you may reply to this message. Thank you, dhcp-bugs@isc.org ------------------------------------------------------------------------- Hi! I think/hope I've sent it to you already, but attached, an infiniband support patch. It's origins are at Mellanox (don't know the person). I've changed/improved it a lot to fit into dhcp-4.3.3 and address some places that weren't working well in the original. We will schedule to adopt and retest it for the upcoming 4.3.6 release ASAP and would be happy when it gets incorporated into 4.3.7 then :-) Gruesse / Regards, Marius Tomaschewski <mt@suse.de>, <mt@suse.com> -- SUSE LINUX GmbH, GF: Felix Imendörffer, Jane Smithard, Dilip Upmanyu, Graham Norton, HRB 21284 (AG Nürnberg), Maxfeldstraße 5, 90409 Nürnberg, Germany