Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <4EB04B43.1020104@msgid.tls.msk.ru>
Date: Tue, 01 Nov 2011 23:40:51 +0400
From: Michael Tokarev <mjt@....msk.ru>
To: owl-dev@...ts.openwall.com
CC: Vasiliy Kulikov <segoon@...nwall.com>
Subject: Re: kernel size

On 24.10.2011 21:56, Vasiliy Kulikov wrote:
[]
>> I find it weird that the kernel became smaller for x86_64 (by approx. 1%),
>> but significantly larger for i686 (by approx. 7%).  This might indicate a
>> problem that we'd want to deal with irrespective of any size limits.
> 
> This might be kernel optimizations for specific gcc version (Something
> between 3.x and 4.x), which might be wrong for other versions.  I suppose
> if we try to compile Linux 2.6.32, we would get more adequate numbers
> (IOW, with the upgrade from 4.4 to 4.6).

For 32bit code, gcc4.6 produces significantly (sometimes >25%) larger
executables than 4.4 did.  I haven't investigated this further.

/mjt

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.