Date: Sat, 31 Aug 2019 01:43:00 +0000 From: bugzilla-noreply@freebsd.org To: toolchain@FreeBSD.org Subject: [Bug 230888] Missing 64 bit atomic functions for i386 (libatomic) Message-ID: <bug-230888-29464-EAv4Z6S4jO@https.bugs.freebsd.org/bugzilla/> In-Reply-To: <bug-230888-29464@https.bugs.freebsd.org/bugzilla/> References: <bug-230888-29464@https.bugs.freebsd.org/bugzilla/>
next in thread | previous in thread | raw e-mail | index | archive | help
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=3D230888 Warner Losh <imp@FreeBSD.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |imp@FreeBSD.org --- Comment #14 from Warner Losh <imp@FreeBSD.org> --- In the past we've kept 486 for two reasons. As a core technology, it was ar= ound in the embedded space well into the 686 era, so there were latter-day versi= ons of that technology well past the classic 486s that are being sneered at a b= it in this bug (though these too are now quite old). The Soekris box was one example. Now that it's become a burden, I think a good case could be made f= or its removal. We can make the default i686, say, and give people that are interested in 486/586 until just before the 13 branch to fix it or we remove it. That blu= nts the criticism somewhat, and make people put their money where their mouths are... And the 'let's remove i386' is an outlier position. There's strong support = for it at least being a userland ABI that we support as a tier 1 platform, with= the kernel dropping to tier 2 for 13. Now, this may change in 14, but that's 5 years off yet :). There's always radical positions within the project... Be= st not to take what any one person says seriously... But whatever you do, I'd strongly suggest talking about it in arch@. It hel= ps to have a firm plan and good justification for that plan. If I may be so bo= ld, I'd suggest removing 486 support in the kernel; support for generating new = 486 binaries and make the default i686, but allow i586 builds (unless there's a good technical reason for not doing that). I'd justify it with the amount of work to support the 486 has become burdensome and if we're going to change,= we might as well go to something a bit newer by default, but allow the folks t= hat need it to build binaries (or not, depending on the technical stuff). I sus= pect that this will be close enough to what most people want as to make it throu= gh an arch@ gauntlet and even though that might be a bit painful, it will get = us to buy in. My own experience is that 600MHz pentium III are still decent enough, though for a desktop with a modern web browser, you really need something quite a = bit more modern. I know people are still embedding 686 and maybe 586 boxes still with FreeBSD, though I know of no-one that still needs the 486 stuff. This = came up 6 months ago and that was the result of my survey then... --=20 You are receiving this mail because: You are the assignee for the bug.=
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?bug-230888-29464-EAv4Z6S4jO>