On Fri, Apr 8, 2022 at 2:23 PM Robbie Harwood <rharwood@xxxxxxxxxx> wrote: > > Michel Alexandre Salim <salimma@xxxxxxxxxxxxxxxxx> writes: > > > - as I stated, there are offers to help with getting syslinux replaced > > with GRUB. what I've not stated originally is Chris Murphy brought up > > protective MBR and switching all new installs to inst.gpt, which let > > us future proof new installations for when we do kill off legacy BIOS. > > > > What people who want to help needs, though, is some sense that our > > contributions are welcome. > > (If you're looking for me to comment on proposals to change live media > generation and installs, I can't really do that - that's Brian's and > Jiri's areas of expertise.) > > > - Neal pointed out he has been working on addressing this for a while, > > and have not put up a Change Proposal because he thought it was > > premature > > While Neal's contributions are as welcome as anyone else's, Neal is not > a maintainer of (or regular contributor to) bootloader packages and it > would be very strange for Neal to propose such a change without > discussing it with us first. That said, if there was suggestion of such > a proposal, I have missed it (which is very possible given the size of > the thread at this point). > Alright, I'll bite. I am within my rights to propose any Change I want for Fedora Cloud, which I help steward with David Duncan. For Fedora Cloud, we've been discussing what we want to support going forward. I added hybrid boot support[1] in Fedora Linux 35 specifically so that we can start working on adding Fedora Cloud to Azure, who prefers UEFI for its Hyper-V VMs[2], but also supports legacy BIOS. We decided to not propose a Change to deprecate BIOS support yet specifically because we needed to have conversations with the various VPSes that people *actually use Fedora* on before we do that[3]. I was completely thrown by this Change proposal because it threatens to poison the conversations that Fedora Cloud wants to have with its actual stakeholders for a less acrimonious transition. As an aside, I examined the state of all release blocking Fedora deliverables, and something I noticed is that only the Workstation WG has Red Hatters actively engaged in it. That means that this Change comes with absolutely no understanding of the state of the world in Fedora across the various WGs and SIGs that deliver release-blocking artifacts. That in itself isn't necessarily a problem, but the fact that none of you are listening to us (David Duncan and myself for Cloud and KDE, Chris Murphy for Cloud and Workstation, and Peter Boy for Server) when we tell you this is too early is extremely tone-deaf. None of us want to keep supporting BIOS forever, but we all have *real-world experience* saying that we can't do this yet. We're trying to find a way to meet halfway to simplify legacy BIOS support, but you're not listening to us. We've also been trying to tell you that there are *real problems* with Fedora's UEFI support that need fixing before we can cut off BIOS support, but you're not listening to us. This thread has, at the time of my writing this post, has 269 posts across 62 individuals. It is the most active thread we've had since the switch to nano by default. However, unlike that change, almost every single respondent has brought up feedback in this discussion saying that we're not ready and providing examples of why we're not ready. However, you're *not listening*. I understand you want to drop BIOS support before Fedora Linux 40 is branched into RHEL 10. Obviously you could drop it in RHEL 10 even if Fedora doesn't, but it'd be better to drop it in Fedora first and make sure everything is flushed out. I would be fine with us doing that if there was some expectation that the UEFI experience in Fedora Linux was going to improve to resolve the issues people have *now* by the time we get there. And we can take incremental steps to get there, even now: 1. Switch Anaconda to default to GPT even on BIOS setups 2. Drop syslinux and use GRUB everywhere 3. Configure new installations to always do hybrid boot installations 4. Develop documentation and/or tooling to do MBR->GPT conversions and reconfigure for hybrid boot for existing systems These are all reasonably achievable things we can do. And that gives us time to work our relationships with our stakeholders to prepare them for the day legacy BIOS support is gone from the entire Red Hat family of distributions. It also gives room for improving the UEFI experience so it's *at least* as good as the BIOS one, if not better. Right now, it's not. And it needs to be in order to maintain the momentum we have now where Fedora Linux adoption is growing by leaps and bounds over the last couple of years. [1]: https://fedoraproject.org/wiki/Changes/FedoraCloudHybridBoot [2]: https://pagure.io/cloud-sig/issue/309 [3]: https://pagure.io/cloud-sig/issue/345 -- 真実はいつも一つ！/ Always, there's only one truth! _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx Do not reply to spam on the list, report it: https://pagure.io/fedora-infrastructure
