Date: Tue, 28 May 2013 15:28:44 -0600 From: Warner Losh <imp@bsdimp.com> To: David Chisnall <theraven@FreeBSD.org> Cc: "toolchain@freebsd.org" <toolchain@FreeBSD.org>, Pedro Giffuni <pfg@FreeBSD.org>, Rui Paulo <rpaulo@felyko.com> Subject: Re: [CFT] gcc: support for barcelona Message-ID: <75A860E0-B057-4F0C-84FF-AA7FF372E130@bsdimp.com> In-Reply-To: <3C29AD82-077D-4E6B-94C7-5D069A130348@FreeBSD.org> References: <51A38CBD.6000702@FreeBSD.org> <E9DC99EF-F2E9-4A5F-8370-36DA25DE2C89@felyko.com> <51A3B8AB.5080808@FreeBSD.org> <521EEFA1-E116-41F5-B618-238E7AA092A8@bsdimp.com> <3C29AD82-077D-4E6B-94C7-5D069A130348@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On May 28, 2013, at 12:10 PM, David Chisnall wrote: > On 28 May 2013, at 18:40, Warner Losh <imp@bsdimp.com> wrote: >=20 >> That's not going to happen soon. While it works OK for amd64, there's = still many bugs in its ARM support and even more in its MIPS support. = There's 0 chance it will be gone in 10... >=20 > I disagree. There is a significant chance that gcc in base will be = gone for all Tier 1 platforms in 10.0. There are still some reasons to = want gcc installed, but there are no compelling reasons to want an = ancient version of gcc installed on x86[-64] or ARM. For people who = need gcc, the ports collection provides a selection of recent versions. I think that's wildly optimistic. We have an integrated system, and = until clang makes it through a release, we need an easy to deploy backup = plan. While you can use clang in 9.x, it isn't default, so we're not = getting a lot of testing. While it is default in -current, there have = been many features that were default in current for years that didn't = find major, day-one problems when the release came. Today clang is useless for ARM kernels with WITNESS due to clang bugs. = While these bugs are being addressed (or maybe in the last week or two = have been addressed), there's be no stress testing of clang-built = systems to the level where we'd have high confidence that the result is = production ready. We haven't even begun to start to shake out all the = other ARM bugs that may be present. While it is true that Apple's #1 = target with clang is arm and amd64, it isn't for a system that's totally = identical to FreeBSD, so there are bound to be integration issues = lurking. And that doesn't even begin to cover ports, but at least there a = fallback to gcc strategy can be via the ports tree. History with the project follows the pattern of having wildly idealistic = goals, followed by realistic achievement in a time frame that was much = longer than was initially planned. Clang adaptation has followed this = pattern to date, and there's no reason to believe that it will be so = perfect that a fallback to gcc won't be needed for 10. Besides, it will still be needed for ia64, sparc64 and likely mips in = the 10 time frame, so it will still be in the system, and still = integrated into the system. That's really where my 0 chance comment came = from. The FreeBSD project is more than just tier 1 platforms. Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?75A860E0-B057-4F0C-84FF-AA7FF372E130>