netdev - Re: [PATCH 00/12] Ethernet: Add and use ether_<type>

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 05 Apr 2018 06:51:32 -0700
From:   Joe Perches <joe@...ches.com>
To:     Felix Fietkau <nbd@....name>, netdev@...r.kernel.org,
        linux-wireless@...r.kernel.org, b43-dev@...ts.infradead.org,
        bridge@...ts.linux-foundation.org, netfilter-devel@...r.kernel.org,
        coreteam@...filter.org
Cc:     linux-kernel@...r.kernel.org, brcm80211-dev-list.pdl@...adcom.com,
        brcm80211-dev-list@...ress.com
Subject: Re: [PATCH 00/12] Ethernet: Add and use ether_<type>_addr globals

On Thu, 2018-04-05 at 15:27 +0200, Felix Fietkau wrote:
> On 2018-03-31 09:05, Joe Perches wrote:
> > There are many local static and non-static arrays that are used for
> > Ethernet broadcast address output or comparison.
> > 
> > Centralize the array into a single separate file and remove the local
> > arrays.
> 
> I suspect that for many targets and configurations, the local arrays
> might actually be smaller than exporting a global.

I tried x86-64 allnoconfig and defconfig.
Those both did not increase vmlinux size.

The defconfig actually got smaller, but that might have been
some object alignment oddity.

>  You have to factor in
> not just the .text size, but the fact that referencing an exported
> symbol needs a .reloc entry as well, which also eats up some space (at
> least when the code is being built as module).

Thanks, the modules I built got smaller.

> In my opinion, your series probably causes more bloat in common
> configurations instead of reducing it.
> 
> You're also touching several places that could easily use
> eth_broadcast_addr and eth_zero_addr. I think making those changes would
> be more productive than what you did in this series.

Doubtful. AFAIK: possible unaligned addresses.