From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ohad Ben-Cohen Subject: Re: [RFC 7/8] drivers: introduce rpmsg, a remote-processor messaging bus Date: Fri, 1 Jul 2011 18:13:54 +0300 Message-ID: References: <1308640714-17961-1-git-send-email-ohad@wizery.com> <1308640714-17961-8-git-send-email-ohad@wizery.com> <20110627222121.GD20865@ponder.secretlab.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Grant Likely Cc: linux-omap@vger.kernel.org, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, akpm@linux-foundation.org, Brian Swetland , Arnd Bergmann , davinci-linux-open-source , Rusty Russell List-Id: linux-omap@vger.kernel.org On Wed, Jun 29, 2011 at 6:43 PM, Grant Likely wrote: > So you are right that device_unregister drops the refcount to zero, > but the code still needs to be fixed to not call kfree() directly. Good point, thanks ! > It also looks like rpmsg_destroy_channel() needs to be fixed to remove > the kfree call Yes, I need to remove this direct kfree as well, and indeed just let rpmsg_release_device do its thing when the last reference is dropped. I should also remove the direct kfree when device_register fails: in that case, I need only to put_device and let the release handler do its thing too. > and an extra put_device() call. We need that extra put_device in rpmsg_destroy_channel because device_find_child() is doing get_device before returning it. Thanks, Grant! Ohad.