Date: Tue, 2 Jul 2024 11:58:16 -0600 From: Warner Losh <imp@bsdimp.com> To: John Baldwin <jhb@freebsd.org> Cc: Warner Losh <imp@freebsd.org>, src-committers <src-committers@freebsd.org>, dev-commits-src-all@freebsd.org, dev-commits-src-main@freebsd.org Subject: Re: git: 67d1a1cd9e77 - main - cdefs.h: Remove support for pre gcc 4.0 Message-ID: <CANCZdfoWdra0kQm1V-qa76GoCg77oWsbV60rMTCJyG9g0dWe2A@mail.gmail.com> In-Reply-To: <CANCZdfqRusqMiJtbh_iBr2cDwTccDyti0mJo36GeNee%2B7ns4dQ@mail.gmail.com> References: <202406210241.45L2fOkP057010@gitrepo.freebsd.org> <4ecc814d-5df1-4db8-b9e0-4919895c5e6a@FreeBSD.org> <CANCZdfrrLWW=JSZ_%2BJcZkbyXnz4Av6dbqn2RBZRyh=8=R6n4zQ@mail.gmail.com> <df672d3e-8b34-478a-b2a7-1ed7fece0b9b@FreeBSD.org> <CANCZdfqRusqMiJtbh_iBr2cDwTccDyti0mJo36GeNee%2B7ns4dQ@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--000000000000f04164061c477510 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Tue, Jul 2, 2024 at 10:30=E2=80=AFAM Warner Losh <imp@bsdimp.com> wrote: > Hey John, > > On Tue, Jul 2, 2024 at 9:44=E2=80=AFAM John Baldwin <jhb@freebsd.org> wro= te: > >> On 7/1/24 4:09 PM, Warner Losh wrote: >> > On Mon, Jul 1, 2024, 3:53=E2=80=AFPM John Baldwin <jhb@freebsd.org> wr= ote: >> > >> >> On 6/20/24 7:41 PM, Warner Losh wrote: >> >>> The branch main has been updated by imp: >> >>> >> >>> URL: >> >> >> https://cgit.FreeBSD.org/src/commit/?id=3D67d1a1cd9e772e2ef94003579f4fbc= 271d38be7d >> >>> >> >>> commit 67d1a1cd9e772e2ef94003579f4fbc271d38be7d >> >>> Author: Warner Losh <imp@FreeBSD.org> >> >>> AuthorDate: 2024-06-20 23:02:56 +0000 >> >>> Commit: Warner Losh <imp@FreeBSD.org> >> >>> CommitDate: 2024-06-21 02:41:08 +0000 >> >>> >> >>> cdefs.h: Remove support for pre gcc 4.0 >> >>> >> >>> All supported compilers support the gcc 3 attribute extensions= . >> >> Remove >> >>> the #else clauses for this. Also, latter-day pcc compilers als= o >> >> define >> >>> __GNUC__, so there's not need to also test for __PCC__. >> >>> >> >>> Reviewed by: brooks >> >>> Differential Revision: https://reviews.freebsd.org/D45654 >> >>> Sponsored by: Netflix >> >>> --- >> >>> sys/sys/cdefs.h | 42 ++++-------------------------------------- >> >>> 1 file changed, 4 insertions(+), 38 deletions(-) >> >>> >> >>> diff --git a/sys/sys/cdefs.h b/sys/sys/cdefs.h >> >>> index 88019819eb35..a6ecdca5d8b9 100644 >> >>> --- a/sys/sys/cdefs.h >> >>> +++ b/sys/sys/cdefs.h >> >>> @@ -408,15 +389,10 @@ >> >>> * assign pointer x to a local variable, to check that its type i= s >> >>> * compatible with member m. >> >>> */ >> >>> -#if __GNUC_PREREQ__(3, 1) >> >>> #define __containerof(x, s, m) ({ >> >> \ >> >>> const volatile __typeof(((s *)0)->m) *__x =3D (x); >> \ >> >>> __DEQUALIFY(s *, (const volatile char *)__x - __offsetof(s, >> m));\ >> >>> }) >> >>> -#else >> >>> -#define __containerof(x, s, m) >> >> \ >> >>> - __DEQUALIFY(s *, (const volatile char *)(x) - __offsetof(s, m)= ) >> >>> -#endif >> >>> >> >>> /* >> >>> * Compiler-dependent macros to declare that functions take >> printf-like >> >>> @@ -434,14 +410,8 @@ >> >>> #define __strftimelike(fmtarg, firstvararg) \ >> >>> __attribute__((__format__ (__strftime__, fmtarg, >> firstvararg))) >> >>> >> >>> -/* Compiler-dependent macros that rely on FreeBSD-specific >> extensions. >> >> */ >> >>> -#if defined(__FreeBSD_cc_version) && __FreeBSD_cc_version >=3D 3000= 01 >> && \ >> >>> - defined(__GNUC__) >> >>> #define __printf0like(fmtarg, firstvararg) \ >> >>> __attribute__((__format__ (__printf0__, fmtarg, >> firstvararg))) >> >>> -#else >> >>> -#define __printf0like(fmtarg, firstvararg) >> >>> -#endif >> >> >> >> Does this still work with external GCC? I didn't think printf0 was >> >> supported >> >> by external GCC (or maybe I had to readd it in the port and that's >> what I >> >> remember). Ah, yes, printf0 is a local patch in the devel/freebsd-gc= cX >> >> ports, but is not available in stock GCC (e.g. lang/gcc does not >> support >> >> it). >> >> >> > >> > Ah. That would explain why it just worked for me. That's what I tested >> > with. Clang also seemed happy with it. But that was the in tree clang. >> Is >> > there a similar issue? Gnuc is defined for both. >> >> So we don't support building the base system with lang/gcc, only >> devel/freebsd-gccX (which has a local patch to add printf0 support). >> The only question might be, do we support using __printf0like for things >> that aren't in the base system that could be built with lang/gcc. If so= , >> we might need to guard this somehow. I'm not sure though that we care >> about random software not in base using a FreeBSD-specific keyword from >> <sys/cdefs.h>. >> > > Yes. The question is "do we use __printf0like in our headers" since we > definitely > can't build FreeBSD itself w/o at least some of the extensions for other > things... > and the answer is "yes". err.h uses it, for example, as does setproctitle > in stdlib.h > > The interesting thing for me is that gcc13 will produce no warnings if I > include errr.h > because -Wsystem-header is off. With it on, warnings crop up too. This is > why my testing > didn't see it... > > It looks like clang has it as a builtin for all versions we care about, s= o > I'll create a phab to add some > of this back. Maybe we should upstream what we have, for this and > freebsd_printf since that's also > in clang and one of the small number of patches we have for the lang/gcc* > family? > https://reviews.freebsd.org/D45836 Warner --000000000000f04164061c477510 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, Jul 2, 2024 at 10:30=E2=80=AF= AM Warner 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 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><div dir= =3D"ltr"><div dir=3D"ltr">Hey John,<br></div><br><div class=3D"gmail_quote"= ><div dir=3D"ltr" class=3D"gmail_attr">On Tue, Jul 2, 2024 at 9:44=E2=80=AF= AM John Baldwin <<a href=3D"mailto:jhb@freebsd.org" target=3D"_blank">jh= b@freebsd.org</a>> wrote:<br></div><blockquote class=3D"gmail_quote" sty= le=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);paddi= ng-left:1ex">On 7/1/24 4:09 PM, Warner Losh wrote:<br> > On Mon, Jul 1, 2024, 3:53=E2=80=AFPM John Baldwin <<a href=3D"mailt= o:jhb@freebsd.org" target=3D"_blank">jhb@freebsd.org</a>> wrote:<br> > <br> >> On 6/20/24 7:41 PM, Warner Losh wrote:<br> >>> The branch main has been updated by imp:<br> >>><br> >>> URL:<br> >> <a href=3D"https://cgit.FreeBSD.org/src/commit/?id=3D67d1a1cd9e772= e2ef94003579f4fbc271d38be7d" rel=3D"noreferrer" target=3D"_blank">https://c= git.FreeBSD.org/src/commit/?id=3D67d1a1cd9e772e2ef94003579f4fbc271d38be7d</= a><br> >>><br> >>> commit 67d1a1cd9e772e2ef94003579f4fbc271d38be7d<br> >>> Author:=C2=A0 =C2=A0 =C2=A0Warner Losh <imp@FreeBSD.org>= <br> >>> AuthorDate: 2024-06-20 23:02:56 +0000<br> >>> Commit:=C2=A0 =C2=A0 =C2=A0Warner Losh <imp@FreeBSD.org>= <br> >>> CommitDate: 2024-06-21 02:41:08 +0000<br> >>><br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0cdefs.h: Remove support for pre gcc = 4.0<br> >>><br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0All supported compilers support the = gcc 3 attribute extensions.<br> >> Remove<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0the #else clauses for this. Also, la= tter-day pcc compilers also<br> >> define<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0__GNUC__, so there's not need to= also test for __PCC__.<br> >>><br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0Reviewed by:=C2=A0 =C2=A0 =C2=A0 =C2= =A0 =C2=A0 =C2=A0 brooks<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0Differential Revision:=C2=A0 <a href= =3D"https://reviews.freebsd.org/D45654" rel=3D"noreferrer" target=3D"_blank= ">https://reviews.freebsd.org/D45654</a><br>; >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0Sponsored by:=C2=A0 =C2=A0 =C2=A0 = =C2=A0 =C2=A0 =C2=A0Netflix<br> >>> ---<br> >>>=C2=A0 =C2=A0 sys/sys/cdefs.h | 42 ++++------------------------= --------------<br> >>>=C2=A0 =C2=A0 1 file changed, 4 insertions(+), 38 deletions(-)<= br> >>><br> >>> diff --git a/sys/sys/cdefs.h b/sys/sys/cdefs.h<br> >>> index 88019819eb35..a6ecdca5d8b9 100644<br> >>> --- a/sys/sys/cdefs.h<br> >>> +++ b/sys/sys/cdefs.h<br> >>> @@ -408,15 +389,10 @@<br> >>>=C2=A0 =C2=A0 =C2=A0* assign pointer x to a local variable, to = check that its type is<br> >>>=C2=A0 =C2=A0 =C2=A0* compatible with member m.<br> >>>=C2=A0 =C2=A0 =C2=A0*/<br> >>> -#if __GNUC_PREREQ__(3, 1)<br> >>>=C2=A0 =C2=A0 #define=C2=A0 =C2=A0 =C2=A0__containerof(x, s, m)= ({<br> >>=C2=A0 =C2=A0 =C2=A0 =C2=A0\<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0 const volatile __typeof(((s *)0)-&g= t;m) *__x =3D (x);=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 \= <br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0 __DEQUALIFY(s *, (const volatile ch= ar *)__x - __offsetof(s, m));\<br> >>>=C2=A0 =C2=A0 })<br> >>> -#else<br> >>> -#define=C2=A0 =C2=A0 =C2=A0 __containerof(x, s, m)<br> >>=C2=A0 =C2=A0 =C2=A0 =C2=A0 \<br> >>> -=C2=A0 =C2=A0 =C2=A0__DEQUALIFY(s *, (const volatile char *)(= x) - __offsetof(s, m))<br> >>> -#endif<br> >>><br> >>>=C2=A0 =C2=A0 /*<br> >>>=C2=A0 =C2=A0 =C2=A0* Compiler-dependent macros to declare that= functions take printf-like<br> >>> @@ -434,14 +410,8 @@<br> >>>=C2=A0 =C2=A0 #define=C2=A0 =C2=A0 =C2=A0__strftimelike(fmtarg,= firstvararg) \<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 __attribute__((__form= at__ (__strftime__, fmtarg, firstvararg)))<br> >>><br> >>> -/* Compiler-dependent macros that rely on FreeBSD-specific ex= tensions.<br> >> */<br> >>> -#if defined(__FreeBSD_cc_version) && __FreeBSD_cc_ver= sion >=3D 300001 && \<br> >>> -=C2=A0 =C2=A0 defined(__GNUC__)<br> >>>=C2=A0 =C2=A0 #define=C2=A0 =C2=A0 =C2=A0__printf0like(fmtarg, = firstvararg) \<br> >>>=C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 __attribute__((__form= at__ (__printf0__, fmtarg, firstvararg)))<br> >>> -#else<br> >>> -#define=C2=A0 =C2=A0 =C2=A0 __printf0like(fmtarg, firstvararg= )<br> >>> -#endif<br> >><br> >> Does this still work with external GCC?=C2=A0 I didn't think p= rintf0 was<br> >> supported<br> >> by external GCC (or maybe I had to readd it in the port and that&#= 39;s what I<br> >> remember).=C2=A0 Ah, yes, printf0 is a local patch in the devel/fr= eebsd-gccX<br> >> ports, but is not available in stock GCC (e.g. lang/gcc does not s= upport<br> >> it).<br> >><br> > <br> > Ah. That would explain why it just worked for me. That's what I te= sted<br> > with. Clang also seemed happy with it. But that was the in tree clang.= Is<br> > there a similar issue? Gnuc is defined for both.<br> <br> So we don't support building the base system with lang/gcc, only<br> devel/freebsd-gccX (which has a local patch to add printf0 support).<br> The only question might be, do we support using __printf0like for things<br= > that aren't in the base system that could be built with lang/gcc.=C2=A0= If so,<br> we might need to guard this somehow.=C2=A0 I'm not sure though that we = care<br> about random software not in base using a FreeBSD-specific keyword from<br> <sys/cdefs.h>.<br></blockquote><div><br></div><div>Yes. The question = is "do we use __printf0like=C2=A0 in our headers" since we defini= tely</div><div>can't build FreeBSD itself w/o at least some of the exte= nsions for other things...</div><div>and the answer is "yes". err= .h uses it, for example, as does setproctitle in stdlib.h</div><div><br></d= iv><div>The interesting thing for me is that gcc13 will produce no warnings= if I include errr.h=C2=A0</div><div>because -Wsystem-header=C2=A0is off. W= ith it on, warnings crop up too. This is why my testing</div><div>didn'= t see it...</div><div><br></div><div>It looks like clang has it as a builti= n for all versions we care about, so I'll create a phab to add some</di= v><div>of this back. Maybe we should upstream what we have, for this and fr= eebsd_printf since that's also</div><div>in clang and one of the small = number of patches we have for the lang/gcc* family?</div></div></div></bloc= kquote><div><br></div><div>=C2=A0<a href=3D"https://reviews.freebsd.org/D45= 836">https://reviews.freebsd.org/D45836</a></div><div><br></div><div>Warner= </div></div></div> --000000000000f04164061c477510--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANCZdfoWdra0kQm1V-qa76GoCg77oWsbV60rMTCJyG9g0dWe2A>