Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Tue, 11 Mar 2014 14:26:36 -0400
From: Rich Felker <dalias@...ifal.cx>
To: musl@...ts.openwall.com
Subject: Re: glibc vs musl sizeof types

On Tue, Mar 11, 2014 at 01:16:52PM +0100, Szabolcs Nagy wrote:
> i did sizeof comparisions of most structs on i386, x86_64, mips, arm, powerpc
> using embedian cross compilers vs musl-cross
> 
> (embedian toolchain is old though: gcc-4.4, glibc-2.13, binutils-2.20)
> 
> http://nsz.repo.hu/git/?p=musl-tables;a=blob;f=data/sizeof.arm.diff
> http://nsz.repo.hu/git/?p=musl-tables;a=blob;f=data/sizeof.i386.diff
> http://nsz.repo.hu/git/?p=musl-tables;a=blob;f=data/sizeof.mips.diff
> http://nsz.repo.hu/git/?p=musl-tables;a=blob;f=data/sizeof.powerpc.diff
> http://nsz.repo.hu/git/?p=musl-tables;a=blob;f=data/sizeof.x86_64.diff

Of these, mips has a number of mismatches in structs reflecting
register state, and powerpc has at least one of those too as well as
sysvipc mismatches which may be bugs. Any idea? If so now is the time
these should be fixed.

The differences to the others seem either mostly inconsequential or
due to bugs in glibc.

Rich

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.