From owner-freebsd-current@freebsd.org Mon Oct 5 17:41:01 2020 Return-Path: Delivered-To: freebsd-current@mailman.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mailman.nyi.freebsd.org (Postfix) with ESMTP id 2028342AE48; Mon, 5 Oct 2020 17:41:01 +0000 (UTC) (envelope-from adrian.chadd@gmail.com) Received: from mail-wm1-f66.google.com (mail-wm1-f66.google.com [209.85.128.66]) (using TLSv1.3 with cipher TLS_AES_128_GCM_SHA256 (128/128 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits) server-digest SHA256 client-signature RSA-PSS (2048 bits) client-digest SHA256) (Client CN "smtp.gmail.com", Issuer "GTS CA 1O1" (verified OK)) by mx1.freebsd.org (Postfix) with ESMTPS id 4C4nvJ5lfYz4GF1; Mon, 5 Oct 2020 17:41:00 +0000 (UTC) (envelope-from adrian.chadd@gmail.com) Received: by mail-wm1-f66.google.com with SMTP id l11so352390wmh.2; Mon, 05 Oct 2020 10:41:00 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to:cc; bh=nLO0M13GwLjijFJfvQFBQntYY7/T3Jbao2LdJr8AyGU=; b=LtdlrSIY3gvPY4BUd9zlSYQR/vKVCafTy3HZg5YZm/WCVRofcS6+qQM93Mr+l1TP59 hpm4UFfBjMdlgHIrek+pcA92q+6RK1E3Ji/eM6m1F+KidYJpGu3Bl3uXBuR326A48ZT4 Xn4FWekzpmjo6wiTeqLpYBSg8lLlfiG07pwjiTNR5GdcQuhGjNny3sZBC/DVwowyGA6U vShahB1jxWGQzIIliyZ4q2xzhxaDuBmJejjSDkrpolka6ljeVTkxBQKCDtWTvsu6Thvk OiCOTzmOXCXJsidlGlMQUO7fAHmI55Itm7TTs8nNyuPh0AQU2ximherEA3+wO/vRv1ns tOVg== X-Gm-Message-State: AOAM533WX3ojNaDUnJijnnufo+bbNqFcOMtZLNKs4kpW5pgXb292AYaK +F2mA6nC7Z/+aBdsP5zw7oWcV4zHmfjlJW622Z/DbPtl X-Google-Smtp-Source: ABdhPJwd9TGIayw9c3n3/iAMjpCJz0aDYvY4kf74y/r9GdBT/Vl4C28Fw2o1FjD1EfjKHnteifOwdAxCd6vMh/V7BOQ= X-Received: by 2002:a1c:bb84:: with SMTP id l126mr486876wmf.159.1601919658533; Mon, 05 Oct 2020 10:40:58 -0700 (PDT) MIME-Version: 1.0 References: <09E17145-6746-428D-87F2-65FBA934BCF5@FreeBSD.org> In-Reply-To: <09E17145-6746-428D-87F2-65FBA934BCF5@FreeBSD.org> From: Adrian Chadd Date: Mon, 5 Oct 2020 10:40:45 -0700 Message-ID: Subject: Re: mips32 + gcc9 -- still broken To: Dimitry Andric Cc: Alexander Richardson , freebsd-current , "freebsd-mips@freebsd.org" X-Rspamd-Queue-Id: 4C4nvJ5lfYz4GF1 X-Spamd-Bar: ---- X-Spamd-Result: default: False [-4.00 / 15.00]; ASN(0.00)[asn:15169, ipnet:209.85.128.0/17, country:US]; TAGGED_FROM(0.00)[]; REPLY(-4.00)[] Content-Type: text/plain; charset="UTF-8" X-Content-Filtered-By: Mailman/MimeDel 2.1.33 X-BeenThere: freebsd-current@freebsd.org X-Mailman-Version: 2.1.33 Precedence: list List-Id: Discussions about the use of FreeBSD-current List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Mon, 05 Oct 2020 17:41:01 -0000 I'm just testing the build out before I put up some patches. jhb mentioned we don't install the libatomic gcc bits right now? So we should likely poke at that somewhat too after this. I'll put up some patches to fix gcc builds and this c++ contention thing soon. -adrian On Sun, 4 Oct 2020 at 23:47, Dimitry Andric wrote: > Sorry, i hadn't picked up on this thread at all. What is the concrete > patch you are proposing? Is is in Phabricator somewhere? > > (Note that when building with gcc, we should really just automatically add > -latomic, to get rid of these errors.) > > -Dimitry > > > On 5 Oct 2020, at 05:40, Adrian Chadd wrote: > > > > ping! > > > > I've got the world building on gcc + mips32 just patching > cxx_contention_t to be 32 bits, but it looks like an ABI change. > > > > Would the better thing be to just make it 32 bits on FreeBSD + MIPS for > now? I don't think anyone is going to mind that changing at this point. > > > > > > > > -adrian > > > > > > On Sat, 3 Oct 2020 at 13:35, Adrian Chadd wrote: > > > > > > On Mon, 21 Sep 2020 at 00:43, Alexander Richardson < > arichardson@freebsd.org> wrote: > > > > On Mon, 21 Sep 2020, 07:38 Adrian Chadd, wrote: > > So, the big list of unknown symbols was my fault! Whoops. > > > > i've gotten further using gcc-6.4 by fixing some of the warnings/issues > > that have crept up. > > > > Here's a review for one of them: > > > > https://reviews.freebsd.org/D26504 > > > > However, now I've hit: > > > > /usr/local/bin/mips-unknown-freebsd13.0-ld: > > > /usr/home/adrian/work/freebsd/head-embedded/obj/mips_ap/usr/home/adrian/work/freebsd/head-embedded/src/mips.mips/tmp/usr/lib/libc++.so.1: > > undefined reference to `__atomic_fetch_sub_8' > > /usr/local/bin/mips-unknown-freebsd13.0-ld: > > > /usr/home/adrian/work/freebsd/head-embedded/obj/mips_ap/usr/home/adrian/work/freebsd/head-embedded/src/mips.mips/tmp/usr/lib/libc++.so.1: > > undefined reference to `__atomic_load_8' > > /usr/local/bin/mips-unknown-freebsd13.0-ld: > > > /usr/home/adrian/work/freebsd/head-embedded/obj/mips_ap/usr/home/adrian/work/freebsd/head-embedded/src/mips.mips/tmp/usr/lib/libc++.so.1: > > undefined reference to `__atomic_fetch_add_8' > > > > .. looks like we need some 64 bit atomics now in mips32 for libc++ / devd > > > > Those are now provided by compiler-rt when using clang. With GCC you'll > have to link libatomic. I had a quick look at the code in libc++ that uses > the 64-bit atomics a few weeks ago and I believe it's the futex fallback > code. The best solution would probably be to port it to use umtx but for > MIPS32 it might be fine to use a 32 bit atomic instead. > > > > It compiled fine with a 32 bit atomic. Why's it 64 bit for everything > but Linux? > > > > Can we make it a 32 bit atomic for all FreeBSD platforms? Are there ABI > concerns? > > > > > > > > -adrian > > > > > > Alex > >