Date: Tue, 13 Dec 2022 12:23:53 -0700 From: Warner Losh <imp@bsdimp.com> To: "freebsd-arch@freebsd.org" <freebsd-arch@freebsd.org> Subject: Re: FreeBSD 14: Poll armv6 deprecated or removed Message-ID: <CANCZdfrZp2vNKfOqmLHKoMm8hSXRYGMLs_cR3okLtF6rJ_t7bA@mail.gmail.com> In-Reply-To: <CANCZdfquBUCQsxYL83cQoLm2s04qvO%2BOz-h88zYwT2DmXdKKBw@mail.gmail.com> References: <CANCZdfr_WsHbbeY6FyKxFdx7dOaFEhk%2BiJLYQ6F4rLOUuf-zRQ@mail.gmail.com> <CANCZdfquBUCQsxYL83cQoLm2s04qvO%2BOz-h88zYwT2DmXdKKBw@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--00000000000022ffc605efba90fe Content-Type: text/plain; charset="UTF-8" On Tue, Dec 13, 2022 at 10:48 AM Warner Losh <imp@bsdimp.com> wrote: > Sorry for the lag of a year, it's been a crazy time for me and this fell > through the cracks. > > On Thu, Oct 28, 2021 at 9:37 AM Warner Losh <imp@bsdimp.com> wrote: > >> Greetings, >> >> Given that the number of available and useful armv6 boards has fallen to >> almost zero, the time has come to look hard at armv6. >> >> There's a number of options. >> >> 1. Keep it as is. This will only happen if there's a lot more users than >> we think (and we think there's nearly zero users of FreeBSD 13 and newer >> that would want to run FreeBSD 14). >> >> 2. Stop building packages. Given it's small to non-existent user base, it >> makes no sense to provide a package building service for it. >> 2a. We should likely do this anyway for all stable branches since it's a >> net negative in terms of cost/benefit analysis: lots of effort to produce, >> very little use. >> >> 3. Disconnect it from universe: This will mean it will rot, though. It's >> a necessary step in removal. >> >> 4. Remove support for armv6 in base entirely. This will orphan any RPiB >> and RPi0 users out there. However, the RPiB hasn't been sold in a few >> years, and the RPI0's connectivity is severely lacking given no SDIO >> support. >> >> So, which of these steps do we do before FreeBSD 14 and which before >> FreeBSD 15? >> >> My vote would be to do 1-4 for 14 including 2a. >> > > After discussing this offline and distilling the responses here, on IRC, > etc, I'd like to propose we do the following: > > (1) Stop building packages for armv6 entirely, on all branches. While > there are some users, they can still use poudriere to build package sets > themselves. Usage data suggests there's not enough demand for these > packages. > (2) Move armv6 to an 'EXTRA' target in make universe (eg make universe > EXTRA_TARGETS=t). Powerpc does this today with powerpcspe since it's not > completely supported in base. People that care can continue to build it as > part of universe and we'll fix things that are broken, reported to us and > have patches that don't regress other platforms. > (3) After the stable/14 branch next May or June, we'll remove build > support from ports (very little) and src. We'll also start to tear down > armv6 support as we see it and are working in other areas, and expect that > work to be done before stable/15 is branched in a predicted 2025. > (4) Immediately stop including armv6 -current snapshots as generated by re@ > . > (5) re@ won't create armv6 release images or artifacts for stable/14, > 14.0, etc. Users wishing to build it can do so. > Oh, and #6: Demote armv6 (but not armv7) to tier 3. It's been kinda de-facto there for a while anyway, and this will make the signalling clear that the runway for armv6 is running out. > I've send email to re@ asking about #4 and #5. > > Warner > --00000000000022ffc605efba90fe Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"ltr"><div dir=3D"ltr"><br></div><br><div class=3D"gmail_quote">= <div dir=3D"ltr" class=3D"gmail_attr">On Tue, Dec 13, 2022 at 10:48 AM Warn= er Losh <<a href=3D"mailto:imp@bsdimp.com">imp@bsdimp.com</a>> wrote:= <br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8= ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"ltr= "><div>Sorry for the lag of a year, it's been a crazy time for me and t= his fell through the cracks.</div><br><div class=3D"gmail_quote"><div dir= =3D"ltr" class=3D"gmail_attr">On Thu, Oct 28, 2021 at 9:37 AM Warner Losh &= lt;<a href=3D"mailto:imp@bsdimp.com" target=3D"_blank">imp@bsdimp.com</a>&g= t; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0p= x 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div d= ir=3D"ltr">Greetings,<br><div><br></div><div>Given that the number of avail= able and useful armv6 boards has fallen to almost zero, the time has come t= o look hard at armv6.</div><div><br></div><div>There's a number of opti= ons.</div><div><br></div><div>1. Keep it as is. This will=C2=A0only happen = if there's a lot more users than we think (and we think=C2=A0there'= s nearly zero users of FreeBSD 13 and newer that would want to run FreeBSD = 14).</div><div><br></div><div>2. Stop building packages. Given it's sma= ll to non-existent user base, it makes no sense to provide a package buildi= ng service for it.</div><div>2a. We should likely do this anyway for all st= able branches since it's a net negative in terms of cost/benefit analys= is: lots of effort to produce, very little use.</div><div><br></div><div>3.= Disconnect it from universe: This will mean it will rot, though. It's = a necessary step in removal.</div><div><br></div><div>4. Remove support for= armv6 in base entirely. This will orphan any RPiB and RPi0 users out there= . However, the RPiB hasn't been sold in a few years, and the RPI0's= connectivity is severely=C2=A0lacking given no SDIO support.</div><div><br= ></div><div>So, which of these steps do we do before FreeBSD 14 and which b= efore FreeBSD 15?</div><div><br></div><div>My vote would be to do 1-4 for 1= 4 including 2a.</div></div></blockquote><div><br></div><div>After discussin= g this offline and distilling the responses here, on IRC, etc, I'd like= to propose we do the following:</div><div><br></div><div>(1) Stop building= packages for armv6 entirely, on all branches. While there are some users, = they can still use poudriere to build package sets themselves. Usage data s= uggests there's not enough demand for these packages.</div><div>(2) Mov= e armv6 to an 'EXTRA' target in make universe (eg make universe EXT= RA_TARGETS=3Dt). Powerpc does this today with powerpcspe since it's not= completely supported in base. People that care can continue to build it as= part of universe and we'll fix things that are broken, reported to us = and have patches that don't regress other platforms.</div><div>(3) Afte= r the stable/14 branch next May or June, we'll remove build support fro= m ports (very little) and src. We'll also start to tear down armv6 supp= ort as we see it and are working in other areas, and expect that work to be= done before stable/15 is branched in a predicted 2025.</div><div>(4) Immed= iately stop including armv6 -current snapshots as generated by re@.</div><d= iv>(5) re@ won't create armv6 release images or artifacts for stable/14= , 14.0, etc. Users wishing to build it can do so.</div></div></div></blockq= uote><div><br></div><div>Oh, and #6: Demote armv6 (but not armv7) to tier 3= . It's been kinda de-facto there for a while anyway, and this will make= the signalling clear that the runway for armv6 is running out.</div><div>= =C2=A0</div><blockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0= .8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir=3D"l= tr"><div class=3D"gmail_quote"><div>I've send email to re@ asking about= #4 and #5.</div><div><br></div><div>Warner</div></div></div> </blockquote></div></div> --00000000000022ffc605efba90fe--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfrZp2vNKfOqmLHKoMm8hSXRYGMLs_cR3okLtF6rJ_t7bA>