View previous topic :: View next topic |
Author |
Message |
Perfect Gentleman Veteran
Joined: 18 May 2014 Posts: 1245
|
Posted: Thu May 03, 2018 3:58 am Post subject: Gcc 8.1 |
|
|
Did anyone already update to it? |
|
Back to top |
|
|
fedeliallalinea Administrator
Joined: 08 Mar 2003 Posts: 30837 Location: here
|
Posted: Thu May 03, 2018 5:54 am Post subject: |
|
|
I doubt that someone have migrated, this version is added only a few hours ago _________________ Questions are guaranteed in life; Answers aren't. |
|
Back to top |
|
|
Perfect Gentleman Veteran
Joined: 18 May 2014 Posts: 1245
|
Posted: Thu May 03, 2018 6:01 am Post subject: |
|
|
anyway,
Are you going to? |
|
Back to top |
|
|
fedeliallalinea Administrator
Joined: 08 Mar 2003 Posts: 30837 Location: here
|
Posted: Thu May 03, 2018 6:09 am Post subject: |
|
|
Perfect Gentleman wrote: | anyway,
Are you going to? |
No also because isn't yet keyworded. Usually I expect that the package to be at least unstable form my arch _________________ Questions are guaranteed in life; Answers aren't. |
|
Back to top |
|
|
Perfect Gentleman Veteran
Joined: 18 May 2014 Posts: 1245
|
Posted: Thu May 03, 2018 6:18 am Post subject: |
|
|
I updated Glbc to 2.27 long time ago. No problems.
I'll try GCC 8.1 when it's in main Arch repos/ |
|
Back to top |
|
|
fturco Veteran
Joined: 08 Dec 2010 Posts: 1181 Location: Italy
|
Posted: Thu May 03, 2018 1:23 pm Post subject: |
|
|
I'm also going to try GCC 8.1 after it will be added to ~amd64. |
|
Back to top |
|
|
freke l33t
Joined: 23 Jan 2003 Posts: 962 Location: Somewhere in Denmark
|
Posted: Thu May 03, 2018 4:39 pm Post subject: |
|
|
Currently awaiting toolchain update after emerging glibc 2.27.
But GCC8.1 is in portage now - so it'll be next |
|
Back to top |
|
|
Tyler_Durden Apprentice
Joined: 27 Jul 2004 Posts: 189 Location: Germany
|
Posted: Thu May 03, 2018 6:05 pm Post subject: |
|
|
for the kernel you need >=4.16.7 or just the patched version of /usr/src/linux/tools/lib/str_error_r.c _________________ Gentoo 17.1 x86_64
Intel Core i9-9900K
Asus MAXIMUS XI HERO
AMD Radeon 6800XT
64GB DDR4
Samsung SSD 970 EVO Plus 1TB
8x Seagate Archive (SATA-RAID 64TB)
Digital Devices Cine S2 V6.5 DVB Adapter |
|
Back to top |
|
|
thumper Guru
Joined: 06 Dec 2002 Posts: 550 Location: Venice FL
|
Posted: Thu May 03, 2018 7:59 pm Post subject: |
|
|
I'm waiting for the GCC 8.1 blocker bug to see just what the shape of things are.
Then I decide if I want to wait for ~arch or not.
George |
|
Back to top |
|
|
xaviermiller Bodhisattva
Joined: 23 Jul 2004 Posts: 8706 Location: ~Brussels - Belgique
|
Posted: Thu May 03, 2018 8:12 pm Post subject: |
|
|
One machine switched to gcc 8.1 without any problem (except nasm, bug already referenced) _________________ Kind regards,
Xavier Miller |
|
Back to top |
|
|
thumper Guru
Joined: 06 Dec 2002 Posts: 550 Location: Venice FL
|
Posted: Fri May 04, 2018 1:43 am Post subject: |
|
|
xaviermiller wrote: | One machine switched to gcc 8.1 without any problem (except nasm, bug already referenced) |
And already fixed, very cool.
https://bugs.gentoo.org/654716
George |
|
Back to top |
|
|
xaviermiller Bodhisattva
Joined: 23 Jul 2004 Posts: 8706 Location: ~Brussels - Belgique
|
Posted: Fri May 04, 2018 5:58 am Post subject: |
|
|
Yes
3 Gentoo systems are now live, next victim will be a raspberry pi. _________________ Kind regards,
Xavier Miller |
|
Back to top |
|
|
freke l33t
Joined: 23 Jan 2003 Posts: 962 Location: Somewhere in Denmark
|
Posted: Fri May 04, 2018 7:31 pm Post subject: |
|
|
3x PC Engines APU2 boards switched to 8.1.0-r1
Toolchain re-compiled fine - just waiting for kernel compile (4.16.7) and a reboot now - then it's time for emptytree and testing what I still have left in no-lto.conf |
|
Back to top |
|
|
Chiitoo Administrator
Joined: 28 Feb 2010 Posts: 2551 Location: Here and Away Again
|
Posted: Fri May 04, 2018 7:49 pm Post subject: ><)))°€ |
|
|
Built @system with GCC 8.1.0 just fine, as well as a bunch of 'dev-qt/*-5.11.9999' packages.
So far I'm only seeing issues with 'dev-qt/qtwebengine'. _________________ Kindest of regardses. |
|
Back to top |
|
|
pjp Administrator
Joined: 16 Apr 2002 Posts: 20053
|
Posted: Fri May 04, 2018 8:05 pm Post subject: |
|
|
Tyler_Durden wrote: | for the kernel you need >=4.16.7 or just the patched version of /usr/src/linux/tools/lib/str_error_r.c | Anyone know if this is historically something likely to be backported? 4.14 seems to remain the latest "longterm maintenance" release. _________________ Quis separabit? Quo animo? |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Fri May 04, 2018 9:46 pm Post subject: |
|
|
Currently, ~800 of ~1500 packages compiled, even with the new flags -fcf-protection=full -fstack-clash-protection.
Except for the necessity to filter some flags for some packages which previously had compiled with these flags, there are no issues so far.
freke wrote: | and testing what I still have left in no-lto.conf |
The list increased with every new compiler release instead of shrinking. With gcc-8 again some packages needed to be added (I do no longer test for removal since this never happened). |
|
Back to top |
|
|
Hu Moderator
Joined: 06 Mar 2007 Posts: 21489
|
Posted: Sat May 05, 2018 12:53 am Post subject: |
|
|
pjp wrote: | Tyler_Durden wrote: | for the kernel you need >=4.16.7 or just the patched version of /usr/src/linux/tools/lib/str_error_r.c | Anyone know if this is historically something likely to be backported? 4.14 seems to remain the latest "longterm maintenance" release. | This is already backported. v4.16: git log v4.16..v4.16.7 -- tools/lib/str_error_r.c: | commit 4b0110291ae1d618384cab105b1e2d208a31a6e3
Author: Josh Poimboeuf <jpoimboe@redhat.com>
Date: Thu Mar 15 22:11:54 2018 -0500
objtool, perf: Fix GCC 8 -Wrestrict error
commit 854e55ad289ef8888e7991f0ada85d5846f5afb9 upstream.
Starting with recent GCC 8 builds, objtool and perf fail to build with
the following error:
../str_error_r.c: In function ‘str_error_r’:
../str_error_r.c:25:3: error: passing argument 1 to restrict-qualified parameter aliases with argument 5 [-Werror=restrict]
snprintf(buf, buflen, "INTERNAL ERROR: strerror_r(%d, %p, %zd)=%d", errnum, buf, buflen, err);
The code seems harmless, but there's probably no benefit in printing the
'buf' pointer in this situation anyway, so just remove it to make GCC
happy.
Reported-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/20180316031154.juk2uncs7baffctp@treble
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Fredrik Schön <fredrikschon@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
| v4.14: git log v4.14..v4.14.39 -- tools/lib/str_error_r.c: | commit d1f1f7771a6a5f81047ecf948c5a580c916f6c3d
Author: Josh Poimboeuf <jpoimboe@redhat.com>
Date: Thu Mar 15 22:11:54 2018 -0500
objtool, perf: Fix GCC 8 -Wrestrict error
commit 854e55ad289ef8888e7991f0ada85d5846f5afb9 upstream.
Starting with recent GCC 8 builds, objtool and perf fail to build with
the following error:
../str_error_r.c: In function ‘str_error_r’:
../str_error_r.c:25:3: error: passing argument 1 to restrict-qualified parameter aliases with argument 5 [-Werror=restrict]
snprintf(buf, buflen, "INTERNAL ERROR: strerror_r(%d, %p, %zd)=%d", errnum, buf, buflen, err);
The code seems harmless, but there's probably no benefit in printing the
'buf' pointer in this situation anyway, so just remove it to make GCC
happy.
Reported-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Tested-by: Laura Abbott <labbott@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/20180316031154.juk2uncs7baffctp@treble
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Fredrik Schön <fredrikschon@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
|
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Sat May 05, 2018 2:20 am Post subject: |
|
|
Couldn't they have made it: Code: |
void *ptr=buf;
snprintf(buf, buflen, "INTERNAL ERROR: strerror_r(%d, %p, %zd)=%d", errnum, ptr, buflen, err); | and avoided the error without losing information? |
|
Back to top |
|
|
pjp Administrator
Joined: 16 Apr 2002 Posts: 20053
|
Posted: Sat May 05, 2018 4:57 am Post subject: |
|
|
Excellent, thanks!
Updates seem pretty frequent, so I've been lagging behind a bit to avoid chasing the kernel dragon. I'll have to upgrade soonish. _________________ Quis separabit? Quo animo? |
|
Back to top |
|
|
mv Watchman
Joined: 20 Apr 2005 Posts: 6747
|
Posted: Sun May 06, 2018 1:42 pm Post subject: |
|
|
world is re-emerged
Very few packages had failed, and mplayer/mpv/... segfault if ffmpeg is emerged with -fno-common. (Can somebody explain to me how -fno-common might cause a segfault?)
Meanwhile there is a gcc-8 tracker bug; I guess it is meanwhile complete, and practically all reported packages have patches: There is no reason to keep any older gcc than 8.1 on the system, anymore! |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Sun May 06, 2018 2:29 pm Post subject: |
|
|
Tony0945 wrote: | Couldn't they have made it: Code: |
void *ptr=buf;
snprintf(buf, buflen, "INTERNAL ERROR: strerror_r(%d, %p, %zd)=%d", errnum, ptr, buflen, err); | and avoided the error without losing information? | Sure, though we'd probably prefer a: const void *p; (or w/e.)
If it's just an unexpected error from strerror_r, though, or an unknown errno (neither of which should happen at kernel-level), there really is no utility in knowing the buffer address that happened to be in use at the callsite. (It was always excess verbiage due to irrelevance.) |
|
Back to top |
|
|
Tony0945 Watchman
Joined: 25 Jul 2006 Posts: 5127 Location: Illinois, USA
|
Posted: Sun May 06, 2018 4:37 pm Post subject: |
|
|
steveL wrote: | If it's just an unexpected error from strerror_r, though, or an unknown errno (neither of which should happen at kernel-level), there really is no utility in knowing the buffer address that happened to be in use at the callsite. (It was always excess verbiage due to irrelevance.) | True. I just instinctively think of working my way around the compiler when it stops me from doing what I want to do. Maybe my old boss was right and I am still an assembly programmer at heart. |
|
Back to top |
|
|
steveL Watchman
Joined: 13 Sep 2006 Posts: 5153 Location: The Peanut Gallery
|
Posted: Mon May 07, 2018 12:37 am Post subject: |
|
|
Tony0945 wrote: | Maybe my old boss was right and I am still an assembly programmer at heart. | Once an asm-coder, always an asm-coder. ;-)
The perspective never leaves you.
And that's a Good thing. |
|
Back to top |
|
|
wrc1944 Advocate
Joined: 15 Aug 2002 Posts: 3432 Location: Gainesville, Florida
|
|
Back to top |
|
|
adamk90 n00b
Joined: 20 Dec 2017 Posts: 14
|
Posted: Mon May 28, 2018 1:15 pm Post subject: |
|
|
I have settled to gcc 8.1 too.. after emerge -e @world (1208 packages) only 12 failed then after a reboot, 10 of these 12 succeeded, but LLVM 5.0.1 does not compile it says ninja error (LLVM 6.0.0 works) and qtwebengine-5.9.4 also fails with ninja subcommand error. Guess time will cure these. |
|
Back to top |
|
|
|