Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from aws-us-west-2-korg-lkml-1.web.codeaurora.org (localhost.localdomain [127.0.0.1]) by smtp.lore.kernel.org (Postfix) with ESMTP id 089E7C4332F for ; Tue, 20 Dec 2022 22:26:59 +0000 (UTC) Received: from mailout4.zoneedit.com (mailout4.zoneedit.com [64.68.198.64]) by mx.groups.io with SMTP id smtpd.web11.4867.1671575213861817562 for ; Tue, 20 Dec 2022 14:26:54 -0800 Authentication-Results: mx.groups.io; dkim=missing; spf=none, err=permanent DNS error (domain: denix.org, ip: 64.68.198.64, mailfrom: denis@denix.org) Received: from localhost (localhost [127.0.0.1]) by mailout4.zoneedit.com (Postfix) with ESMTP id E99B040C76; Tue, 20 Dec 2022 22:26:52 +0000 (UTC) Received: from mailout4.zoneedit.com ([127.0.0.1]) by localhost (zmo14-pco.easydns.vpn [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id YEEVVI4vEsFi; Tue, 20 Dec 2022 22:26:52 +0000 (UTC) Received: from mail.denix.org (pool-100-15-88-116.washdc.fios.verizon.net [100.15.88.116]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mailout4.zoneedit.com (Postfix) with ESMTPSA id B8FD340C5F; Tue, 20 Dec 2022 22:26:50 +0000 (UTC) Received: by mail.denix.org (Postfix, from userid 1000) id 84A3A1636B4; Tue, 20 Dec 2022 17:26:39 -0500 (EST) Date: Tue, 20 Dec 2022 17:26:39 -0500 From: Denys Dmytriyenko To: Ryan Eatmon Cc: Praneeth Bajjuri , meta-ti@lists.yoctoproject.org Subject: Re: [meta-ti][master/kirkstone][PATCH] conf: change from do_image_wic/tar to do_image Message-ID: <20221220222639.GZ22689@denix.org> References: <20221211190240.7246-1-reatmon@ti.com> <20221212181520.GE22689@denix.org> <17302955CCC60FDA.9432@lists.yoctoproject.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <17302955CCC60FDA.9432@lists.yoctoproject.org> User-Agent: Mutt/1.5.20 (2009-06-14) List-Id: X-Webhook-Received: from li982-79.members.linode.com [45.33.32.79] by aws-us-west-2-korg-lkml-1.web.codeaurora.org with HTTPS for ; Tue, 20 Dec 2022 22:26:59 -0000 X-Groupsio-URL: https://lists.yoctoproject.org/g/meta-ti/message/15496 On Mon, Dec 12, 2022 at 04:34:57PM -0500, Denys Dmytriyenko wrote: > On Mon, Dec 12, 2022 at 01:24:29PM -0600, Ryan Eatmon wrote: > > > > > > On 12/12/2022 12:15, Denys Dmytriyenko wrote: > > >On Sun, Dec 11, 2022 at 01:02:40PM -0600, Ryan Eatmon via lists.yoctoproject.org wrote: > > >>When trying to support OpenBMC we found that our use of do_image_wic > > >>and do_image_tar for adding depends and mcdepends was too limiting. > > >>Moving to do_image is a higher level in the tree that means for any > > >>image type the dependencies will be correct. > > > > > >There might be some unexpected side effects, since do_image is too broard. Not > > >every image has a boot partition and should depend on the corresponding boot > > >files being deployed. Things like ramdisks, initramfs and such have different > > >flows and own dependencies. Also images that include other images, like guests > > >for VMs or hypervisors or even bootstrapping. All those could potentially have > > >issues due to a global do_image dependency - I've seen circular dependency > > >breakages in the past in those areas. Not all use cases are being tested by > > >default, so not easy to tell if this change is compltetely harmless... > > > > I was worried something like that might be true. But how are we > > supposed to know which images (present and future) that we need to > > add dependencies for in this case? > > Yeah, that's a very good question. My thinking back then was that usually we > only want "wic" format for a bootable whole-system images, hence I limited > the dependency to do_image_wic (and the tar version was a by-product). Maybe > your change is not as dangerous as I was imagining :) - definitely needs more > thorough testing. Or we can come up with a completely different way of > specifying those dependencies? Ryan, After some digging and experimenting locally, I believe we can move towards using do_image_complete[mcdepends] for the required dependencies. It's further down the chain from do_image and before do_image_wic, so should be a bit safer in my opinion... I've done few builds with this change, but would you be able to run it through the full set of builds on the farm? Thanks. -- Denys