[PATCH 1/2] [0.4.8] eloop extension for dbus control interface, v2

Jouni Malinen jkmaline at cc.hut.fi
Sun Apr 16 22:26:18 EDT 2006


On Thu, Apr 13, 2006 at 09:48:28AM -0400, Dan Williams wrote:

> The following patch adds the infrastructure necessary to support DBus
> bound to the eloop runloop.  Jouni, I think this takes care of the eloop
> API concerns you had at the wireless summit, let me know what you think.
> If the eloop interface looks good, feel free to commit and I'll keep
> working on the DBus control interface.  It should be fully contained and
> separate from the rest of the code, and needs to be enabled in the
> config file to be compiled in.

This looks quite good. However, type casting a function pointer to
different type and then calling read_sock handler function with one
extra parameter looks somewhat scary.. Is that guaranteed to work with
all commonly used ANSI C compilers and CPUs?

> Please review for commit to 0.4.8.  I'll whip up a 0.5.x patch too, if
> this one looks OK for 0.4.8.  Since most distros are shipping 0.4.8
> right now, and at least Fedora will be using 0.4.8+dbus in the near
> future, I'd rather do parallel development on 0.4.8 and 0.5.x rather
> than just targeting 0.5.x and making people who want to use dbus use a
> development version of wpa_supplicant.  (unless I misunderstand the
> structure of branches?)

All new features to wpa_supplicant need to be first added into the
current development branch (0.5.x) and once they have been shown to be
stable, they can be considered for merging into the latest stable branch
(0.4.x). The changes for 0.4.x can, of course, live outside my
repository if they are needed more quickly than the path through
development branch takes. I just want to be able to concentrate on a
single branch for new development and stable releases of wpa_supplicant
do indeed mean stable in the sense of not really changing much.

-- 
Jouni Malinen                                            PGP id EFC895FA



More information about the HostAP mailing list