dbus 1.2.16

[Thiago Macieira]

Lennart Poettering wrote:
>On Thu, 16.07.09 08:18, Schmottlach, Glenn (glenn.schmottlach at harman.com) 
wrote:
>> So this assumption is that the GCC intrinsic functions for atomic
>> operations are valid on all other platforms (e.g. ARM, SH4, PPC,
>> etc...) but not (apparently) for i386 class processors and
>> below. Therefore, has long as GCC generates good code for the
>> processors I'm interested in (ARM at the moment), all is good,
>> right?
>
>D-Bus didn't support atomic ops on ARM before, however with my __sync
>patch it now will if you have a recent enough gcc that can do it for
>ARM.

Not all ARM boards have atomic instructions. Some of them require 
mediation by the kernel. I have no idea if the GCC __sync stuff uses them.

ARMv6 is fine, though.

>So, in summary: this is actually a big improvement for ARM, not a
>regression.
>
>All what is lost is support for atomic ops on x86 with old gcc
>versions. Which I think is an Ok price to pay for the much cleaner
>code we now have.

And also anyone who forgot to pass -march=i486 or higher to GCC. There are 
no atomic fetch-and-add operations on i386.

-- 
  Thiago Macieira  -  thiago (AT) macieira.info - thiago (AT) kde.org
    PGP/GPG: 0x6EF45358; fingerprint:
    E067 918B B660 DBD1 105C  966C 33F5 F005 6EF4 5358
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://lists.freedesktop.org/archives/dbus/attachments/20090722/e5d2574f/attachment.pgp