From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 94517 invoked by alias); 6 Jun 2018 21:36:43 -0000 Mailing-List: contact gdb-patches-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: gdb-patches-owner@sourceware.org Received: (qmail 94500 invoked by uid 89); 6 Jun 2018 21:36:43 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=-23.5 required=5.0 tests=AWL,BAYES_00,GIT_PATCH_0,GIT_PATCH_1,GIT_PATCH_2,GIT_PATCH_3,KAM_LAZY_DOMAIN_SECURITY,SPF_HELO_PASS,TRACKER_ID autolearn=ham version=3.3.2 spammy= X-HELO: mx1.redhat.com Received: from mx3-rdu2.redhat.com (HELO mx1.redhat.com) (66.187.233.73) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Wed, 06 Jun 2018 21:36:41 +0000 Received: from smtp.corp.redhat.com (int-mx06.intmail.prod.int.rdu2.redhat.com [10.11.54.6]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id B64FF80125D6; Wed, 6 Jun 2018 21:36:39 +0000 (UTC) Received: from localhost (unused-10-15-17-196.yyz.redhat.com [10.15.17.196]) by smtp.corp.redhat.com (Postfix) with ESMTP id 8370D205D521; Wed, 6 Jun 2018 21:36:39 +0000 (UTC) From: Sergio Durigan Junior To: Alan Hayward Cc: GDB Patches , nd , Simon Marchi Subject: Re: [PATCH] Guard declarations of 'sve_*_from_*' macros on Aarch64 (and unbreak build) References: <87zi09vw7m.fsf@redhat.com> <20180605220556.7418-1-sergiodj@redhat.com> Date: Wed, 06 Jun 2018 21:36:00 -0000 In-Reply-To: (Alan Hayward's message of "Wed, 6 Jun 2018 07:34:20 +0000") Message-ID: <874lifwqag.fsf@redhat.com> User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/25.3 (gnu/linux) MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable X-IsSubscribed: yes X-SW-Source: 2018-06/txt/msg00161.txt.bz2 On Wednesday, June 06 2018, Alan Hayward wrote: >> On 5 Jun 2018, at 23:05, Sergio Durigan Junior wro= te: >>=20 >> Commit 122394f1476b1c925a281b15399119500c8231c1 ("Function for reading >> the Aarch64 SVE vector length") has added macros to manipulate SVE >> vector sizes based on Linux kernel sources, but did not guard them >> with #ifndef's, which breaks the build when the system headers already >> have these macros: >>=20 >> CXX aarch64-linux-nat.o >> In file included from ../../gdb/aarch64-tdep.h:25, >> from ../../gdb/aarch64-linux-nat.c:30: >> ../../gdb/arch/aarch64.h:79: error: "sve_vq_from_vl" redefined [-Werror] >> #define sve_vq_from_vl(vl) ((vl) / 0x10) >>=20 >> In file included from /usr/include/bits/sigcontext.h:30, >> from /usr/include/signal.h:291, >> from build-gnulib/import/signal.h:52, >> from ../../gdb/linux-nat.h:23, >> from ../../gdb/aarch64-linux-nat.c:26: >> /usr/include/asm/sigcontext.h:154: note: this is the location of the pr= evious definition >> #define sve_vq_from_vl(vl) ((vl) / SVE_VQ_BYTES) >>=20 >> In file included from ../../gdb/aarch64-tdep.h:25, >> from ../../gdb/aarch64-linux-nat.c:30: >> ../../gdb/arch/aarch64.h:80: error: "sve_vl_from_vq" redefined [-Werror] >> #define sve_vl_from_vq(vq) ((vq) * 0x10) >>=20 >> In file included from /usr/include/bits/sigcontext.h:30, >> from /usr/include/signal.h:291, >> from build-gnulib/import/signal.h:52, >> from ../../gdb/linux-nat.h:23, >> from ../../gdb/aarch64-linux-nat.c:26: >> /usr/include/asm/sigcontext.h:155: note: this is the location of the pr= evious definition >> #define sve_vl_from_vq(vq) ((vq) * SVE_VQ_BYTES) >>=20 >> In order to fix this breakage, this commit guards the declaration of >> the macros using #ifndef=E2=80=99s. > > Apologies for breaking this. > I originally had them in nat/aarch64-sve-linux-ptrace.h, within the > SVE_SIG_ZREGS_SIZE block, which does guard appropriately. > > Thanks for fixing so quickly. No problem. >> Tested by making sure GDB compiles fine again. Since the system I'm >> using doesn't have support for SVE, there's no way I can really test >> these changes. >>=20 > > Changes work fine for me on my SVE builds. Thanks for testing! >> gdb/ChangeLog: >> 2018-06-05 Sergio Durigan Junior >>=20 >> * arch/aarch64.h (sve_vg_from_vl): Guard with #ifndef. >> (sve_vl_from_vg): Likewise. >> (sve_vq_from_vl): Likewise. >> (sve_vl_from_vq): Likewise. >> (sve_vq_from_vg): Likewise. >> (sve_vg_from_vq): Likewise. >> --- >> gdb/arch/aarch64.h | 12 ++++++++++++ >> 1 file changed, 12 insertions(+) >>=20 >> diff --git a/gdb/arch/aarch64.h b/gdb/arch/aarch64.h >> index 9040d8d4c8..c378a4b239 100644 >> --- a/gdb/arch/aarch64.h >> +++ b/gdb/arch/aarch64.h >> @@ -74,12 +74,24 @@ enum aarch64_regnum >> VG : Vector Gradient. >> The number of 64bit chunks in an SVE Z register. */ >>=20 >> +#ifndef sve_vg_from_vl >> #define sve_vg_from_vl(vl) ((vl) / 8) >> +#endif >> +#ifndef sve_vl_from_vg >> #define sve_vl_from_vg(vg) ((vg) * 8) >> +#endif > > The guards around these first two aren=E2=80=99t needed. The kernel only > defines the VQ to/from VL macros - as those are the only values the > kernel cares about. Only GDB cares about VG because that is needed > for dwarf. Ah, fair enough. Removed. >> +#ifndef sve_vq_from_vl >> #define sve_vq_from_vl(vl) ((vl) / 0x10) >> +#endif >> +#ifndef sve_vl_from_vq >> #define sve_vl_from_vq(vq) ((vq) * 0x10) >> +#endif > > These two are fine! > >> +#ifndef sve_vq_from_vg >> #define sve_vq_from_vg(vg) (sve_vq_from_vl (sve_vl_from_vg (vg))) >> +#endif >> +#ifndef sve_vg_from_vq >> #define sve_vg_from_vq(vq) (sve_vg_from_vl (sve_vl_from_vq (vq))) >> +#endif > > Again these last two aren=E2=80=99t needed. Removed. On Wednesday, June 06 2018, Simon Marchi wrote: > On 2018-06-06 03:34 AM, Alan Hayward wrote: >> FYI, either today or tomorrow I=E2=80=99ll be posting a new set of SVE p= atches. >> As part of that series, due to review comments, I=E2=80=99ll be moving t= he all >> the linux kernel defines to two new files which contain only kernel >> defines (From ptrace.h and sigcontext.h). I=E2=80=99ll double check this= works >> with new header files. Regardless of that, I think your patch should >> still go in to unbreak the build until then. >>=20 >>=20 >> Changes are good to me if the VG guards are removed (but I=E2=80=99m not= a >> maintainer so cannot officially approve it). > > LGTM with Alan's proposed changes. Thanks, pushed: e5a77256e8961294b3ea7d483124834311ca363b --=20 Sergio GPG key ID: 237A 54B1 0287 28BF 00EF 31F4 D0EB 7628 65FC 5E36 Please send encrypted e-mail if possible http://sergiodj.net/