This is the mail archive of the
ecos-discuss@sourceware.org
mailing list for the eCos project.
Ok, is this a race condition?
- From: Chuck McManis <ecos at mcmanis dot com>
- To: ECOS Discussion Group <ecos-discuss at sources dot redhat dot com>
- Date: Thu, 16 Feb 2006 22:04:41 -0800
- Subject: [ECOS] Ok, is this a race condition?
So I'm re-writing the RxEvent handler in my ethernet driver. The old rhine
code would find a descriptor that was owned by the host, call
(sc->funs->eth_drv->recv)(sc, len)
with the data and then free the packet by giving it back to the MAC.
Now, here is the question. How threaded is eCos? Or specifically, can I be
_sure_ that the call above will complete before I call free? Or does simply
queue a thread to come collect the data later, possibly suspending, and
getting back to my driver before it has run? Or is this going to
dereference a pointer to the vt8325_recv() function which is just below
this one? If its the latter then I can see its good to go this way.
I'm slowly getting a handle on how you guys (and gals?) have the state
machine set up and I wouldn't want my driver to return bogus data.
--Chuck
--
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss