From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755670AbcH3Teu (ORCPT ); Tue, 30 Aug 2016 15:34:50 -0400 Received: from mx2.suse.de ([195.135.220.15]:54876 "EHLO mx2.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751372AbcH3Tet (ORCPT ); Tue, 30 Aug 2016 15:34:49 -0400 Date: Tue, 30 Aug 2016 21:34:46 +0200 From: "Luis R. Rodriguez" To: Daniel Wagner Cc: "Luis R. Rodriguez" , Daniel Wagner , linux-kernel@vger.kernel.org, Ming Lei , Greg Kroah-Hartman Subject: Re: [PATCH v3 1/3] firmware_class: encapsulate firmware loading status Message-ID: <20160830193446.GU3296@wotan.suse.de> References: <1472118723-22762-1-git-send-email-wagi@monom.org> <1472118723-22762-2-git-send-email-wagi@monom.org> <20160825175007.GA3296@wotan.suse.de> <151b26a2-6562-b37b-43e2-9040f0609a1c@bmw-carit.de> <1ffedc17-d2e8-7843-d1e5-5da20d4cae91@bmw-carit.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <1ffedc17-d2e8-7843-d1e5-5da20d4cae91@bmw-carit.de> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, Aug 29, 2016 at 04:18:33PM +0200, Daniel Wagner wrote: > On 08/29/2016 11:50 AM, Daniel Wagner wrote: > >On 08/25/2016 07:50 PM, Luis R. Rodriguez wrote: > >>>+#else /* CONFIG_FW_LOADER_USER_HELPER */ > >>>+ > >>>+static int loading_timeout = 60; > >>>+#define firmware_loading_timeout() (loading_timeout * HZ) > >>>+ > >>>+#define fw_status_wait_timeout(fw_st, long) 0 > >> > >>The timeout makes 0 sense for when !CONFIG_FW_LOADER_USER_HELPER so can > >>we do away with adding a silly 60 value to an int here and > >>the silly value of (loading_timeout * HZ) ? Its not used so its not > >>clear to me why this is here. > > > >So the main reason that silly timeout is needed is the usage of > >it in device_cache_fw_images(). I suggest we add a timeout > >argument to _request_firmware() and use the right timeout value > >at that level. > > > >That allows to move the loading_timeout into the > >CONFIG_FW_LOADER_USER_HELPER section. > > I forgot to answer your question. So we have the dependency to > loading_timeout/firmware_loading_timeout from the firmware caching > path. The patch added in the previous email removes that dependency. > > We still need the 60 second even in the > !CONFIG_FW_LOADER_USER_HELPER case. I think it would be a regression > if we change that value, no? Oh that might be the disconnect, see my series of pending patches, I did away with the cache stuff using the usermode helper, the cache stuff should not use the usermode helper as the cache stuff kills off the pending usermode helper requests right before suspend. Luis