From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH 6/6] bus: Add support for Tegra Generic Memory Interface To: Mirza Krak References: <1470512452-8322-1-git-send-email-mirza.krak@gmail.com> <1470512452-8322-7-git-send-email-mirza.krak@gmail.com> CC: Stephen Warren , Thierry Reding , Alexandre Courbot , , Prashant Gaikwad , , , , , Michael Turquette , , , , "Kumar Gala" , , , From: Jon Hunter Message-ID: Date: Tue, 9 Aug 2016 14:37:05 +0100 MIME-Version: 1.0 In-Reply-To: Return-Path: jonathanh@nvidia.com Content-Type: text/plain; charset="utf-8" List-ID: On 09/08/16 08:21, Mirza Krak wrote: > 2016-08-08 15:47 GMT+02:00 Jon Hunter : ... >>> + return of_platform_default_populate(of_node, NULL, dev); >> >> Seems odd to do this here. Why not as the last thing in probe once the >> GMI has been setup correctly? > > It is done here mostly to avoid clean up in probe. I could move it out > of tegra_gmi_parse_dt and put it right after that call but still > before clk enable? Is that less odd? I would have thought that clocks should be enabled and resets de-asserted before we even attempt to populate any children. Hence, I would do this last. Yes that will require some clean-up in probe if this does fail but seems safest. Cheers Jon -- nvpublic