From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from kirsty.vergenet.net ([202.4.237.240]:52072 "EHLO kirsty.vergenet.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751141AbcFFAjH (ORCPT ); Sun, 5 Jun 2016 20:39:07 -0400 Date: Mon, 6 Jun 2016 09:39:02 +0900 From: Simon Horman To: Sasha Levin Cc: stable@vger.kernel.org, stable-commits@vger.kernel.org, Greg Kroah-Hartman , linux-renesas-soc@vger.kernel.org, Yoshihiro Shimoda Subject: Re: [added to the 4.1 stable tree] usb: host: xhci-rcar: Use ARCH_RENESAS Message-ID: <20160606003858.GA17412@verge.net.au> References: <1464989831-16666-1-git-send-email-sasha.levin@oracle.com> <1464989831-16666-66-git-send-email-sasha.levin@oracle.com> <20160604004717.GB469@verge.net.au> <57522B69.3070105@oracle.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <57522B69.3070105@oracle.com> Sender: linux-renesas-soc-owner@vger.kernel.org List-ID: [Cc Shimoda-san] On Fri, Jun 03, 2016 at 09:14:17PM -0400, Sasha Levin wrote: > On 06/03/2016 08:47 PM, Simon Horman wrote: > > Hi, > > > > I do not believe this patch is suitable for v4.1 stable. > > > > ARCH_RENESAS is not otherwise present there and thus the driver in question > > will stop being compiled where it was previously compiled. > > > > Moreover, the transition to ARCH_RENESAS is (so far) being handled in such > > a way that there have not been any regressions that would necessitate a > > backport to -stable. > > Hey Simon, > > I've picked this commit because of what I saw in f879fc32a > ("usb: host: xhci-rcar: Avoid long wait in xhci_reset()"). > > >From my understanding of that change, we added a dependency on ARCH_RENESAS > for USB_XHCI_PLATFORM, but since USB_XHCI_RCAR depended on ARCH_SHMOBILE that > would cause silly errors for kconfig and would probably confuse users. > > Rather than changing it to a ARCH_SHMOBILE dependency, I've preferred > to pull in f879fc32a to keep the stable tree more aligned with upstream > and possibly prevent any future backporting errors for new commits that > add ARCH_RENESAS dependencies. > > If you don't agree with that approach I can change that to a ARCH_SHMOBILE > dependency and remove this commit, or take any other approach you think > is better? Thanks, this does seem to be a bit of a mess. Sorry about that. My feeling is that it would be better to change f879fc32a to use ARCH_SHMOBILE as ARCH_RENESAS does not exist in v4.1. One slight risk in this approach is that ARCH_SHMOBILE covers some (older) SH SoCs as well as (newer) ARM and arm64 based SoCs (Renesas arm64 based SoCs are not present in v4.1) whereas ARCH_RENESAS (in kernels that it exists) only covers ARM and arm64 based SoCs. So the coverage of ARCH_SHMOBILE is broader. But I don't think that should be a problem as that is already the dependency being used in v4.1. A quick compile test using ecovec24_defconfig + USB_XHCI_HCD + USB_XHCI_RCAR indicates that the driver compiles on that platform.