[coreboot] Chromium Upstreaming / Introduction
Hello all, I am currently working on helping the Chromium team get their coreboot patches upstreamed so I thought I should introduce myself to the community. My name is Isaac Christensen and I've been working for Sage Electronic Engineering since October. These will be my first pushes up to coreboot.org so if you have any comments on process or workflow feel free to let me know as I'm still learning. Thank you, Isaac -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
On Tue, Jul 29, 2014 at 9:55 AM, Isaac wrote: > Hello all, > > I am currently working on helping the Chromium team get their coreboot > patches upstreamed so I thought I should introduce myself to the community. > My name is Isaac Christensen and I've been working for Sage Electronic > Engineering since October. These will be my first pushes up to > coreboot.org > so if you have any comments on process or workflow feel free to let me know > as I'm still learning. > Great! We'll look forward to working with you on this. It might be beneficial to drop by #coreboot on irc.freenode.net to answer questions / feedback about patches in real-time. -- David Hendricks (dhendrix) Systems Software Engineer, Google Inc. -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
On 07/29/2014 09:20 PM, David Hendricks wrote: On Tue, Jul 29, 2014 at 9:55 AM, Isaac wrote: Hello all, I am currently working on helping the Chromium team get their coreboot patches upstreamed so I thought I should introduce myself to the community. My name is Isaac Christensen and I've been working for Sage Electronic Engineering since October. These will be my first pushes up to coreboot.org so if you have any comments on process or workflow feel free to let me know as I'm still learning. Great! We'll look forward to working with you on this. It might be beneficial to drop by #coreboot on irc.freenode.net to answer questions / feedback about patches in real-time. Excellent. A couple questions though: Do you plan to upstream all Chromebook coreboot and libpayload branches from Chromium git, or just the individual patches Sage finds useful and necessary for the boards You currently work on? Do we expect the original authors to review the rebased, possibly modified work, or is the plan these patches just get rubberstamped as +2 by a 3rd party once they build cleanly? I see some of you first patches had a couple commits from Chromium tree squashed into one. Why was the approach changed from how eg. Aaron and Stefan handled the process? This has the unpleasant effect that commit ownership (and eg. git blame) will no longer reflect the actual author of change and also git log --oneline no longer serves as a datapoint in attempt to compare which changes from Chromium branches have been upstreamed or not. Did you develop some nice method to keep track of which branches from Chromium we can consider as completely upstreamed? Do you have the facilities to do regular board-status script runs for recent Chromebooks? Kyösti -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Let me join in welcoming Isaac's efforts! Good to have you with the project, Isaac. Kyösti Mälkki wrote: > Excellent. A couple questions though: .. Thank you for asking these questions. I'm looking forward to good answers! This is important. //Peter -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
On Thu, Jul 31, 2014 at 12:00 PM, Kyösti Mälkki wrote: > I see some of you first patches had a couple commits from Chromium tree > squashed into one. Why was the approach changed from how eg. Aaron and > Stefan handled the process? > For the ARM platforms, there were a lot of files that were imported from other projects and were unnecessary and/or inconsistent with coreboot style. I suspect that squashing many of those commits helps to make it build cleanly against upstream without polluting the codebase with a lot of obsolete/unwanted stuff that got cleaned out anyway. > git log --oneline no longer serves as a datapoint in attempt to compare > which changes from Chromium branches have been upstreamed or not. > Perhaps, and hopefully the x86 stuff won't need as much squashing since there should be less churn in that part of the codebase. IMHO the most important reason to do this at all is to provide a good starting point to people who are interested in developing future products, not necessarily to bring in 100% of the patches from chromium for past products. Even if we miss a few patches or some history is lost -- and I obviously hope we don't lose anything important -- that's a small price to pay to make "upstream first" development easier for new products. > > Did you develop some nice method to keep track of which branches from > Chromium we can consider as completely upstreamed? > > Do you have the facilities to do regular board-status script runs for > recent Chromebooks? > > > Kyösti > > > -- > coreboot mailing list: coreboot@coreboot.org > http://www.coreboot.org/mailman/listinfo/coreboot > -- David Hendricks (dhendrix) Systems Software Engineer, Google Inc. -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Do you plan to upstream all Chromebook coreboot and libpayload branches from Chromium git, or just the individual patches Sage finds useful and necessary for the boards You currently work on? The goal of this effort is to minimize the differences between the Chromium chromeos-2013.04 branch and the coreboot master branch. As part of this any outstanding patches or differences between the branches will be addressed. Do we expect the original authors to review the rebased, possibly modified work, or is the plan these patches just get rubberstamped as +2 by a 3rd party once they build cleanly? Original authors are automatically added as reviewers if they have coreboot.org gerrit accounts. I try to get approval from the original authors, but they are not always available for comment. If they are not available and the patch is non-trivial, we look for someone else familiar with the area or the original author of the section for review. We would like to avoid too many rubber stamp reviews. Did you develop some nice method to keep track of which branches from Chromium we can consider as completely upstreamed? This hasn't been a problem yet since we are only focusing on the chromeos-2013.04 branch at this time. If you know of a good solution for this I would appreciate any information. Do you have the facilities to do regular board-status script runs for recent Chromebooks? Not at this time but we're coordinating with Google on this. Isaac -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Dear Isaac, Am Dienstag, den 29.07.2014, 10:55 -0600 schrieb Isaac: > I am currently working on helping the Chromium team get their coreboot > patches upstreamed so I thought I should introduce myself to the community. > My name is Isaac Christensen and I've been working for Sage Electronic > Engineering since October. welcome to coreboot and thank you very much for getting in contact with the coreboot community! > These will be my first pushes up to coreboot.org so if you have any > comments on process or workflow feel free to let me know as I'm still > learning. As already commented on your change sets, I’d prefer if you could avoid squashing commits. Kyösti made very good points, so I won’t repeat them. Also it’d be great to at least address such review comments and not ignore them and push the patches. I am looking forward to the time when the future Chromium OS branches can be rebased on the coreboot upstream master branches. ;-) Thanks, Paul signature.asc Description: This is a digitally signed message part -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
* Paul Menzel [140802 15:07]: > As already commented on your change sets, I’d prefer if you could avoid > squashing commits. Kyösti made very good points, so I won’t repeat them. One of the reasons Isaac has been squashing patches was that some in the community have been complaining about a vast amount of unsquashed patches last time, e.g.: https://www.mail-archive.com/coreboot@coreboot.org/msg40272.html http://www.coreboot.org/pipermail/coreboot/2013-January/073543.html As always, we have taken the community input under consideration, and tried acting upon it in the best possible way. Are you suggesting that the needs have changed here? Stefan -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
* Kyösti Mälkki [140731 21:00]: > Do you plan to upstream all Chromebook coreboot and libpayload > branches from Chromium git, or just the individual patches Sage > finds useful and necessary for the boards You currently work on? No, the plan for now is to only upstream the Chromium HEAD, since all Chrome OS product branches are derived from that. In theory all changes made to product branches should all go into Chromium HEAD as well. While this is not always true, the product branches are now all open on chromium.org. > Do we expect the original authors to review the rebased, possibly > modified work, or is the plan these patches just get rubberstamped > as +2 by a 3rd party once they build cleanly? Of course we value the feedback of the original authors. > I see some of you first patches had a couple commits from Chromium > tree squashed into one. Why was the approach changed from how eg. > Aaron and Stefan handled the process? It was changed due to community feedback. > This has the unpleasant effect that commit ownership (and eg. git > blame) will no longer reflect the actual author of change and also git > log --oneline no longer serves as a datapoint in attempt to compare > which changes from Chromium branches have been upstreamed or not. The idea is that we will switch to a new upstream coreboot version ASAP and get into a continuous upstreaming process where we don't accumulate such a large number of patches anymore but upstream at least every few weeks so we can stay closer to upstream coreboot in the future. > Did you develop some nice method to keep track of which branches > from Chromium we can consider as completely upstreamed? The method will be product firmware branches ==> Chrome OS HEAD ==> coreboot upstream > Do you have the facilities to do regular board-status script runs > for recent Chromebooks? Not at this time, but we are constantly working on improving our testing infrastructure to allow more flexible and broader testing. I hope at some point we can add that. However, if you rely on stability, you definitely want to run coreboot off the product firmware branches. Stefan -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Am 05.08.2014 um 20:36 schrieb Stefan Reinauer: > Are you suggesting that the needs have changed here? My only concern here would be to keep rebase/merge et al more functional, but it's probably already too late for that in the current state. I suppose, once the trees are somewhat synchronized, one could build a merge commit between both histories to create a new baseline - and redo that every now and then. Patrick signature.asc Description: OpenPGP digital signature -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Hello! Stefan, by what you posted there, (and correct me if I'm wrong) if I were to put together a system who would be running ChromeOS and of course using coreboot to bring it up, the OS would be constructed from the head of the entire Chromium set? Just checking. As of this moment I do not have these plans, but its always an interest to do so. - Gregg C Levine gregg.drw...@gmail.com "This signature fought the Time Wars, time and again." On Tue, Aug 5, 2014 at 2:59 PM, Stefan Reinauer wrote: > * Kyösti Mälkki [140731 21:00]: >> Do you plan to upstream all Chromebook coreboot and libpayload >> branches from Chromium git, or just the individual patches Sage >> finds useful and necessary for the boards You currently work on? > > No, the plan for now is to only upstream the Chromium HEAD, since all > Chrome OS product branches are derived from that. In theory all changes > made to product branches should all go into Chromium HEAD as well. > While this is not always true, the product branches are now all open on > chromium.org. > >> Do we expect the original authors to review the rebased, possibly >> modified work, or is the plan these patches just get rubberstamped >> as +2 by a 3rd party once they build cleanly? > > Of course we value the feedback of the original authors. > >> I see some of you first patches had a couple commits from Chromium >> tree squashed into one. Why was the approach changed from how eg. >> Aaron and Stefan handled the process? > > It was changed due to community feedback. > >> This has the unpleasant effect that commit ownership (and eg. git >> blame) will no longer reflect the actual author of change and also git >> log --oneline no longer serves as a datapoint in attempt to compare >> which changes from Chromium branches have been upstreamed or not. > > The idea is that we will switch to a new upstream coreboot version ASAP > and get into a continuous upstreaming process where we don't accumulate > such a large number of patches anymore but upstream at least every few > weeks so we can stay closer to upstream coreboot in the future. > >> Did you develop some nice method to keep track of which branches >> from Chromium we can consider as completely upstreamed? > > The method will be > > product firmware branches ==> Chrome OS HEAD ==> coreboot upstream > >> Do you have the facilities to do regular board-status script runs >> for recent Chromebooks? > > Not at this time, but we are constantly working on improving our testing > infrastructure to allow more flexible and broader testing. I hope at > some point we can add that. However, if you rely on stability, you > definitely want to run coreboot off the product firmware branches. > > Stefan > > > -- > coreboot mailing list: coreboot@coreboot.org > http://www.coreboot.org/mailman/listinfo/coreboot -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Dear Stefan, Am Dienstag, den 05.08.2014, 20:36 +0200 schrieb Stefan Reinauer: > * Paul Menzel [140802 15:07]: > > As already commented on your change sets, I’d prefer if you could avoid > > squashing commits. Kyösti made very good points, so I won’t repeat them. > > One of the reasons Isaac has been squashing patches was that some in the > community have been complaining about a vast amount of unsquashed > patches last time, e.g.: > > https://www.mail-archive.com/coreboot@coreboot.org/msg40272.html very interesting read. ;-) First of, I am sorry, that it looks like one time I write this and the other time that. But it is more my fault for not being clear. So let me clarify. By fix-ups I meant typos or wrong PCI IDs. Looking at the squashed commits in [1], from my point of view, these are no fix-ups and should not have been squashed. > http://www.coreboot.org/pipermail/coreboot/2013-January/073543.html Well, that message actually asks for more communication with the community, so is unrelated. > As always, we have taken the community input under consideration, and > tried acting upon it in the best possible way. Yes, thanks again for your answer. > Are you suggesting that the needs have changed here? The basic need has not changed and it looks like it was a misunderstanding in the past. Thanks, Paul [1] http://review.coreboot.org/6425 signature.asc Description: This is a digitally signed message part -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
* Patrick Georgi [140805 21:32]: > > > Am 05.08.2014 um 20:36 schrieb Stefan Reinauer: > > Are you suggesting that the needs have changed here? > > My only concern here would be to keep rebase/merge et al more > functional, but it's probably already too late for that in the current > state. Unfortunately, without significant amounts of testing, there is no guarantee that any given configuration in the Chromium HEAD or coreboot upstream HEAD will actually work on a given hardware. for every single build. If you want something that is tested for a given product, you will have to use a product firmware branch. If you want to make changes to a given configuration, you will have to have equipment and knowledge in place to test and fix up your configuration yourself. That's certainly not something you'd get for free in any open source project, and particularly not in a hardware related project. > I suppose, once the trees are somewhat synchronized, one could build a > merge commit between both histories to create a new baseline - and > redo that every now and then. The idea is indeed that the pain will go away if synchronization happens more regularly in both directions. Stefan -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
* Gregg Levine [140805 21:35]: > Hello! > Stefan, by what you posted there, (and correct me if I'm wrong) if I > were to put together a system who would be running ChromeOS and of > course using coreboot to bring it up, the OS would be constructed from > the head of the entire Chromium set? That is correct. We start out with chromium.org's HEAD and then move into a branch once development has stabilized. Stefan -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
Re: [coreboot] Chromium Upstreaming / Introduction
Hi Stefan, Good to know that Chromium HEAD is going upstream. Specially the starting branch chromeos-2013.04, which is really far ahead of what upstream has. Major one is ARMv8 support which is currently missing in upstream version. On Thu, Aug 7, 2014 at 2:08 AM, Stefan Reinauer wrote: > * Gregg Levine [140805 21:35]: >> Hello! >> Stefan, by what you posted there, (and correct me if I'm wrong) if I >> were to put together a system who would be running ChromeOS and of >> course using coreboot to bring it up, the OS would be constructed from >> the head of the entire Chromium set? > > That is correct. We start out with chromium.org's HEAD and then move > into a branch once development has stabilized. > > Stefan > > -- > coreboot mailing list: coreboot@coreboot.org > http://www.coreboot.org/mailman/listinfo/coreboot -- Regards, Alim -- coreboot mailing list: coreboot@coreboot.org http://www.coreboot.org/mailman/listinfo/coreboot
