Skip site navigation (1)Skip section navigation (2)
Date:      Thu, 21 Apr 2022 18:31:43 -0700
From:      Mark Millard <marklmi@yahoo.com>
To:        Free BSD <freebsd-arm@freebsd.org>, freebsd-ports@freebsd.org
Cc:        Emmanuel Vadot <manu@freebsd.org>, toolchain@freebsd.org
Subject:   Re: It looks like devel/aarch64-none-elf-gcc and other aarch64 */gcc* 's suffer from https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101723
Message-ID:  <0F6C9A98-21DC-4AA4-9DDC-CAB0E0154E59@yahoo.com>
In-Reply-To: <DB8CAC13-CC31-452A-9926-30E9C1B498A9@yahoo.com>
References:  <C91744CC-ACB1-49C5-A4AF-9B71CD7BCB78@yahoo.com> <DB8CAC13-CC31-452A-9926-30E9C1B498A9@yahoo.com>

next in thread | previous in thread | raw e-mail | index | archive | help
On 2022-Apr-18, at 14:10, Mark Millard <marklmi@yahoo.com> wrote:

> On 2022-Apr-18, at 12:36, Mark Millard <marklmi@yahoo.com> wrote:
>=20
>> In doing some experiments, it looks like I've run into:
>>=20
>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D101723
>>=20
>> which had fixes applied to gcc9 , gcc10, and gcc11 in very
>> late 2021-July into Aug (all after the most recent tagged
>> release). gcc8 was already out of support or it would have
>> had a commit too. ( aarch64-none-elf-gcc is currently at
>> 8.4.0 .)
>>=20
>> The specific issue is that for compiles with -march=3Darmv8-a+crc
>> specified the result is as if the +crc had not been specified.
>> In my context:
>>=20
>> {standard input}: Assembler messages:
>> {standard input}:37: Error: selected processor does not support =
`crc32b w0,w0,w4'
>>=20
>> The description of the fixes starts with:
>>=20
>> QUOTE
>>   A change to the way gas interprets the .fpu directive in =
binutils-2.34
>>   means that issuing .fpu will clear any features set by =
.arch_extension
>>   that apply to the floating point or simd units.  This unfortunately
>>   causes problems for more recent versions of the architecture =
because
>>   we currently emit .arch, .arch_extension and .fpu directives at
>>   different times and try to suppress redundant changes.
>> . . .
>> END QUOTE
>>=20
>> So seeing the problem or not also depends on the vintage of gas
>> in use, possibly explaining this not having been noticed before.
>>=20
>> This was reported separately in:
>>=20
>> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D104439
>>=20
>> for 11.2.0 and more and specifically for crc32b messages, but was
>> classified as a duplicate of the earlier one listed above.
>>=20
>> As far as I can tell no tagged gcc release has the fix yet:
>>=20
>> releases/gcc-11.2.0 (2021-Jul-27, so close: 29th commit)
>> releases/gcc-9.4.0  (2021-Jun-01)
>> releases/gcc-10.3.0 (2021-Apr-08)
>>=20
>> This interfere with experimenting with an updated U-Boot for
>> aarch64.
>>=20
>=20
> I looked around and found:
>=20
> =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dcommit;h=3Dc1cdabe3aab817d95a8db0=
0a8b5e9f6bcdea936f
>=20
> arm: reorder assembler architecture directives [PR101723]
> author	Richard Earnshaw <rearnsha@arm.com>=09
> Thu, 29 Jul 2021 10:00:31 +0000 (11:00 +0100)
> committer	Richard Earnshaw <rearnsha@arm.com>=09
> Thu, 5 Aug 2021 11:51:14 +0000 (12:51 +0100)
> commit	c1cdabe3aab817d95a8db00a8b5e9f6bcdea936f
> tree	2e3b5461af89619f316ecefa9d46564aa7514d33	tree
> parent	6a37d0331c25f23628d4308e5a75624005c223b2	commit | =
diff
>=20
> As well as the refs/heads/releases/gcc-* related commits . . .
>=20
>=20
> =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dcommit;h=3Dc21ba5e57e49b870f16079=
44c0742e78feb7bc8d
> [Wed, 18 Aug 2021 15:22:38 +0000 (16:22 +0100)]
>=20
> via =
https://gcc.gnu.org/git?p=3Dgcc.git;a=3Dshortlog;h=3Drefs/heads/releases/g=
cc-11
>=20
>=20
> =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dcommit;h=3D02d5a1988247207da46f25=
ce8b58515e25c1f250
> [Mon, 23 Aug 2021 14:31:16 +0000 (15:31 +0100)]
>=20
> via =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dshortlog;h=3Drefs/heads/releases/=
gcc-10
>=20
>=20
> =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dcommit;h=3D04c568961e793a1d7ad862=
48b4ca929fc84acf8d
> [Mon, 23 Aug 2021 14:39:20 +0000 (15:39 +0100)]
>=20
> via =
https://gcc.gnu.org/git/?p=3Dgcc.git;a=3Dshortlog;h=3Drefs/heads/releases/=
gcc-9
>=20
>=20
> So possibly anything based on more recent snapshots than those
> refs/heads/releases/gcc-* related commits have the commit in
> place already.
>=20

I looked and gcc 11.3.0 has source code from the fix. So, for
lang/gcc* (non -devel),

git: 06b81ba6804c - main - lang/gcc11: upgrade to 11.3.0 Piotr Kubaj

looks to bee the first (and so-far only) lang/gcc* (non -devel)
to contain the fix. The content of the other versions predate
the fix (so far).

A simple bugzilla search of what had been fixed in 11.3 did not list
PR 101723 but looking at the likes of

=
https://github.com/gcc-mirror/gcc/blob/releases/gcc-11.3.0/gcc/config/arm/=
arm.c

showed a history entry for the fix and code from the fix.

=3D=3D=3D
Mark Millard
marklmi at yahoo.com




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?0F6C9A98-21DC-4AA4-9DDC-CAB0E0154E59>