Send dhcp-users mailing list submissions to
        dhcp-users@lists.isc.org

To subscribe or unsubscribe via the World Wide Web, visit
        https://lists.isc.org/mailman/listinfo/dhcp-users
or, via email, send a message with subject or body 'help' to
        dhcp-users-requ...@lists.isc.org

You can reach the person managing the list at
        dhcp-users-ow...@lists.isc.org

When replying, please edit your Subject line so it is more specific
than "Re: Contents of dhcp-users digest..."


Today's Topics:

   1. Re: dhclient binary size (Timothe Litt)
   2. Re: dhclient binary size (Victoria Risk)
   3. Re: A Question on Dynamic DHCP/DNS IP Lease Renew (Glenn Satchell)
   4. Re: A Question on Dynamic DHCP/DNS IP Lease Renew (Glenn Satchell)


----------------------------------------------------------------------

Message: 1
Date: Fri, 30 Sep 2016 10:47:50 -0400
From: Timothe Litt <l...@acm.org>
To: "dhcp-users@lists.isc.org" <dhcp-users@lists.isc.org>
Subject: Re: dhclient binary size
Message-ID: <7d566e7a-971c-c81a-6c1a-d99aa437d...@acm.org>
Content-Type: text/plain; charset="windows-1252"


On 30-Sep-16 10:07, dhcp-users-requ...@lists.isc.org wrote:
> Message: 1
> Date: Fri, 30 Sep 2016 09:38:25 -0400
> From: Thomas Markwalder <tm...@isc.org>
> To: dhcp-users@lists.isc.org
> Subject: Re: dhclient binary size
> Message-ID: <1a46fcfe-ffeb-be78-0be1-000c8c71f...@isc.org>
> Content-Type: text/plain; charset="windows-1252"
>
> Hi Matt:
>
> I suspect a fair portion of the size increase comes along with moving to
> the use of the Bind9 libraries, versus the code that resides in the
> "minires" directory.   We gained a good great deal by having common
> code,  with improved performance and robustness as well as in
> functionality, but apparently at a price of binary footprint.
>
> It may be possible to reduce the size of the Bind9 libraries based on
> capabilities excluded through compilation flags when building them but
> this would have to be researched.
>
> Regards,
>
> Thomas Markwalder
> ISC Software Engineering
>
>
Matt,

The other obvious approach is to make the libraries sharable, which will
help when more than one of dhclient, dhcpd and bind is present.  To be
effective, the libraries need to be modular enough that including one
doesn't drag in everything.  Optimizing code placement can also help;
many embedded environments have VMs, and can tolerate dead code storage
(but not activation).  FLASH is fairly cheap.  But if you're storage
limited and stuck with existing hardware, you're left with shrinking the
code.  And of course if you only need dhclient, sharable libraries are
counter-productive.  This is probably an exercise best left to ISC - or
at least, done once and contributed.

Finally, you can tweak your compiler flags to optimize for (small) code
size at the expense of speed.  The defaults will do unrolling & inlining
that you probably don't need.  dhcpd can be performance critical (though
rarely when it's embedded).  DHClient shouldn't care how slowly it
runs.  Perhaps the DHClient build flags should recognize this.

In any case it's nice to hear that there's someone else who care about
embedded systems.  I mention them to ISC fairly regularly, but it's not
obvious that we're heard.  Yet just about every NAS, router, and CPE
gateway that I look into uses dhcpd/dhclient. 

I run dhcpd and bind on old, slow XSCALE processors  (and RPi) - but I
have the advantage of USB disks so I just let VM negotiate a reasonable
working set.  I don't run DHclient.

Good luck.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4577 bytes
Desc: S/MIME Cryptographic Signature
URL: 
<https://lists.isc.org/pipermail/dhcp-users/attachments/20160930/86e9a42b/attachment-0001.bin>

------------------------------

Message: 2
Date: Fri, 30 Sep 2016 09:06:17 -0700
From: Victoria Risk <vi...@isc.org>
To: Users of ISC DHCP <dhcp-users@lists.isc.org>
Subject: Re: dhclient binary size
Message-ID: <95ad0efc-c211-4f77-aaa4-f17b5832e...@isc.org>
Content-Type: text/plain; charset="windows-1252"

The reason we kept DHCP 4.1.x around, after its planned EOL date had passed, 
was specifically to preserve an option for people who needed a smaller 
footprint.

All of the support we get for ISC DHCP today is for the server, so that is 
where we have been focusing our efforts. We do not have any financial support 
for maintaining or optimizing the client (despite the many devices apparently 
running the software).

If anyone on the list wants to discuss funding an engineer to work on reducing 
the client footprint, or funding a client based on Kea, you can contact me 
off-list at vi...@isc.org.  We are certainly open to it.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 842 bytes
Desc: Message signed with OpenPGP using GPGMail
URL: 
<https://lists.isc.org/pipermail/dhcp-users/attachments/20160930/d17200b8/attachment-0001.bin>

------------------------------

Message: 3
Date: Sat, 1 Oct 2016 14:10:22 +1000
From: "Glenn Satchell" <glenn.satch...@uniq.com.au>
To: "Users of ISC DHCP" <dhcp-users@lists.isc.org>
Subject: Re: A Question on Dynamic DHCP/DNS IP Lease Renew
Message-ID:
        <4fa3a50ef9281839f36940f2de8cf51c.squir...@mail.uniq.com.au>
Content-Type: text/plain;charset=iso-8859-1

On Fri, September 30, 2016 5:09 pm, Simon Hobson wrote:
> Glenn Satchell <glenn.satch...@uniq.com.au> wrote:
>
>> As an alternative you can just set a really long lease time, like a year
>> or two.
>
> But bear in mind that some systems explicitly release their leases on
> shutdown. So does that negate the long lease when it comes to the "least
> recently used" calculation ?
>
>
> As an aside ...
> I once had a printer (early generation of connected digital copiers) that
> just would not accept a lease via DHCP. I gave up in the end and manually
> configured it. Some time later I found that it had a requirement that any
> lease offered must be a minimum of 2 years. That was a WTF moment, and I
> had some questions about the parentage of the devs that came up with that
> requirement !

I recently resurrected an old HP UPS at work, and the management card
would only use bootp, so my normal range statement didn't help. In that
case host statement with fixed-address did the job.

regards,
-glenn




------------------------------

Message: 4
Date: Sat, 1 Oct 2016 14:29:40 +1000
From: "Glenn Satchell" <glenn.satch...@uniq.com.au>
To: "Users of ISC DHCP" <dhcp-users@lists.isc.org>
Subject: Re: A Question on Dynamic DHCP/DNS IP Lease Renew
Message-ID:
        <986fa560765620fa516c1e448fab94bc.squir...@mail.uniq.com.au>
Content-Type: text/plain;charset=iso-8859-1

Hi Simon

I think you will find it is the term used in the dhcpd.leases file, which
defines the client's last transaction time. If you look up dhcpd.leases
man page you'll find the acronym.

Looks like over-zealous pattern matching. It's been in their anti-spam
filters for many years. I even emailed the user once and he said it was
the corporate anti-spam filter and just too hard to get it changed.

regards,
-glenn

On Fri, September 30, 2016 7:07 pm, Simon Hobson wrote:
> Rather amusingly, I've had a notification that my last email was blocked
> to one person. Someone has some interesting ideas of what is pornography !
>
>> Bell Aliant Content Filtering Device:
>> ...
>> Content Rule: Policy Management (Inbound) : Block Pornographic Language
>> - BA
>
>
> _______________________________________________
> dhcp-users mailing list
> dhcp-users@lists.isc.org
> https://lists.isc.org/mailman/listinfo/dhcp-users
>




------------------------------

Subject: Digest Footer

_______________________________________________
dhcp-users mailing list
dhcp-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/dhcp-users

------------------------------

End of dhcp-users Digest, Vol 96, Issue 1
*****************************************

Reply via email to