Re: GCC 7.1 Released
On May 12, 2017 3:42 AM, "Chris Johns"wrote: On 12/05/2017 18:39, Sebastian Huber wrote: > On 12/05/17 10:35, Chris Johns wrote: > >> On 12/05/2017 15:56, Sebastian Huber wrote: >> >>> On 11/05/17 17:29, Joel Sherrill wrote: >>> >> I talked to Joel late his time yesterday and he is reporting increased >> code size and as a result some tests are not fitting into the target. >> Rather than head on into the change we think it is better to step back >> and visit it again after we have a release. What do you think? >> > > A code size increase would be not an issue for me if its in the range of > a small percent range. However, we should identify this and create a GCC > PR. > > Gaisler suggested to use GCC 7.1. Did Gaisler observe some significant > code size changes on SPARC? > > Sorry I do not have any detail. I am sure Joel can provide some when he is online and has some time. I think all he did was build all the BSPs with tests and saw some new failures. 21 BSPs failed to link some.test due to a memory overflow. Mix or at least arm and 68k 4am here. I will post when i am really awake.b Chris ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 12/05/2017 18:39, Sebastian Huber wrote: On 12/05/17 10:35, Chris Johns wrote: On 12/05/2017 15:56, Sebastian Huber wrote: On 11/05/17 17:29, Joel Sherrill wrote: I talked to Joel late his time yesterday and he is reporting increased code size and as a result some tests are not fitting into the target. Rather than head on into the change we think it is better to step back and visit it again after we have a release. What do you think? A code size increase would be not an issue for me if its in the range of a small percent range. However, we should identify this and create a GCC PR. Gaisler suggested to use GCC 7.1. Did Gaisler observe some significant code size changes on SPARC? Sorry I do not have any detail. I am sure Joel can provide some when he is online and has some time. I think all he did was build all the BSPs with tests and saw some new failures. Chris ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 12/05/2017 15:56, Sebastian Huber wrote: On 11/05/17 17:29, Joel Sherrill wrote: Ok to On May 11, 2017 5:09 AM, "Sebastian Huber"> wrote: Hello, if someone wants to use GCC 7.1 for RTEMS 4.12, then the RSB should be updated accordingly as soon as possible. Chris posted a patch to bump gcc and binutils. He just needs to push it. I didn't see the patch. According to I have not posted anything. contrib/download_prerequisites we should use gmp='gmp-6.1.0.tar.bz2' mpfr='mpfr-3.1.4.tar.bz2' mpc='mpc-1.0.3.tar.gz' isl='isl-0.16.1.tar.bz2' for GCC 7. I talked to Joel late his time yesterday and he is reporting increased code size and as a result some tests are not fitting into the target. Rather than head on into the change we think it is better to step back and visit it again after we have a release. What do you think? Chris ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 11/05/17 17:29, Joel Sherrill wrote: Ok to On May 11, 2017 5:09 AM, "Sebastian Huber"> wrote: Hello, if someone wants to use GCC 7.1 for RTEMS 4.12, then the RSB should be updated accordingly as soon as possible. Chris posted a patch to bump gcc and binutils. He just needs to push it. I didn't see the patch. According to contrib/download_prerequisites we should use gmp='gmp-6.1.0.tar.bz2' mpfr='mpfr-3.1.4.tar.bz2' mpc='mpc-1.0.3.tar.gz' isl='isl-0.16.1.tar.bz2' for GCC 7. -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.hu...@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
Ok to On May 11, 2017 5:09 AM, "Sebastian Huber" < sebastian.hu...@embedded-brains.de> wrote: Hello, if someone wants to use GCC 7.1 for RTEMS 4.12, then the RSB should be updated accordingly as soon as possible. Chris posted a patch to bump gcc and binutils. He just needs to push it. I had a local patch to just do gcc but thanks to timezone differences, he posted first. :) -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.hu...@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
Hello, if someone wants to use GCC 7.1 for RTEMS 4.12, then the RSB should be updated accordingly as soon as possible. -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.hu...@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 09/05/17 14:54, Daniel Hellstrom wrote: On 2017-05-03 07:11, Sebastian Huber wrote: On 03/05/17 03:33, Chris Johns wrote: On 3/5/17 3:04 am, Joel Sherrill wrote: > >Thoughts on upgrading the 4.12 tools? > Looks good with updated dependent packages for gcc and with binutils-2.28. There are no build failures for just the tests. Attached is the warnings report which looks good if the networking warnings are ignored. I would rather use the GCC 6 which received some bug fix releases in the mean time and was used most of the time of the RTEMS 4.12 development. For RTEMS 5 I would use GCC 8. For the gaisler toolchain we are currently planning on using GCC 7.1. If you tested this version thoroughly than this is fine for me. Did you build all RTEMS targets which currently use GCC 6 with this version and can you provide the RSB patches? If yes, then please use the versions specified by GCC contrib/download_prerequisites for the support libraries. A new GCC version probably means new warnings to fix. We should manage the GCC selection better in the future. GCC 6 received considerable more testing and regression fixes than GCC 7.1 which is only two weeks old. -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.hu...@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 2017-05-03 07:11, Sebastian Huber wrote: On 03/05/17 03:33, Chris Johns wrote: On 3/5/17 3:04 am, Joel Sherrill wrote: > >Thoughts on upgrading the 4.12 tools? > Looks good with updated dependent packages for gcc and with binutils-2.28. There are no build failures for just the tests. Attached is the warnings report which looks good if the networking warnings are ignored. I would rather use the GCC 6 which received some bug fix releases in the mean time and was used most of the time of the RTEMS 4.12 development. For RTEMS 5 I would use GCC 8. For the gaisler toolchain we are currently planning on using GCC 7.1. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 03/05/17 03:33, Chris Johns wrote: On 3/5/17 3:04 am, Joel Sherrill wrote: > >Thoughts on upgrading the 4.12 tools? > Looks good with updated dependent packages for gcc and with binutils-2.28. There are no build failures for just the tests. Attached is the warnings report which looks good if the networking warnings are ignored. I would rather use the GCC 6 which received some bug fix releases in the mean time and was used most of the time of the RTEMS 4.12 development. For RTEMS 5 I would use GCC 8. -- Sebastian Huber, embedded brains GmbH Address : Dornierstr. 4, D-82178 Puchheim, Germany Phone : +49 89 189 47 41-16 Fax : +49 89 189 47 41-09 E-Mail : sebastian.hu...@embedded-brains.de PGP : Public key available on request. Diese Nachricht ist keine geschäftliche Mitteilung im Sinne des EHUG. ___ devel mailing list devel@rtems.org http://lists.rtems.org/mailman/listinfo/devel
Re: GCC 7.1 Released
On 3/5/17 3:04 am, Joel Sherrill wrote: > > Thoughts on upgrading the 4.12 tools? > Looks good with updated dependent packages for gcc and with binutils-2.28. There are no build failures for just the tests. Attached is the warnings report which looks good if the networking warnings are ignored. The test results for erc32 are: Passed:548 Failed: 2 User Input: 4 Expected Fail: 0 Indeterminate: 0 Benchmark: 3 Timeout: 1 Invalid: 1 -- Total: 559 Failures: tar01.exe spcontext01.exe User Input: fileio.exe top.exe termios.exe monitor.exe Benchmark: dhrystone.exe whetstone.exe linpack.exe Timeouts: capture.exe Invalid: minimum.exe I think capture needs to be tagged as user input. Chris RTEMS Tools Project - RTEMS Kernel BSP Builder, 4.12.not_released Date: Wed May 3 21:04:05 2017 command: /opt/work/rtems/4.12/bin/rtems-bsp-builder --rtems-\ tools=/build/rtems/tools/4.12\ --rtems=/opt/work/chris/rtems/kernel/rtems.git --build=tests\ --log=x --warnings-report=warnings-tier-1-tests.txt +=+ | By Architecture (total : 288) | +-+ | common : 33 warning(s) | +-+++-+ | Shared :0 | BSP :0 | Network : 136 |Tests :0 | | LibCPU :0 | CPU Kit :0 || | +---+-+++-+ | 4 | cpukit/librpc/src/rpc/get_myaddress.c:130:5 | | 2 | cpukit/libnetworking/libc/getservbyport.c:52:4 | | 2 | cpukit/libnetworking/netinet/tcp_usrreq.c:106:16| | 2 | cpukit/librpc/src/rpc/rpc_dtablesize.c:55:1 | | 2 | cpukit/libnetworking/libc/getservbyname.c:52:4 | | 2 | cpukit/librpc/src/rpc/rpc_prot.c:215:2 | | 2 | cpukit/librpc/src/rpc/get_myaddress.c:80:1 | | 2 | cpukit/librpc/src/rpc/rpc_prot.c:255:2 | | 2 | cpukit/librpc/src/rpc/rpcdname.c:78:1 | | 2 | cpukit/libnetworking/libc/res_update.c:90:12| | 2 | cpukit/libnetworking/netinet/tcp_output.c:408:4 | | 2 | cpukit/libnetworking/netinet/tcp_output.c:422:5 | | 2 | cpukit/librpc/src/rpc/getrpcport.c:52:1 | | 2 | cpukit/libnetworking/netinet/tcp_usrreq.c:497:16| | 2 | cpukit/libnetworking/lib/getprotoby.c:37:1 | | 2 | cpukit/librpc/src/rpc/rpc_prot.c:115:2 | | 2 | cpukit/librpc/src/rpc/auth_unix.c:102:1 | | 2 | cpukit/libnetworking/netinet/tcp_output.c:393:4 | | 2 | cpukit/libnetworking/lib/getprotoby.c:22:1 | | 2 | cpukit/librpc/src/rpc/svc_simple.c:71:1 | | 2 | cpukit/libnetworking/netinet/tcp_output.c:429:5 | | 2 | cpukit/libnetworking/netinet/tcp_usrreq.c:182:16| | 2 | cpukit/librpc/src/rpc/clnt_simple.c:65:1| | 2 | cpukit/librpc/src/rpc/rpc_prot.c:138:1 | | 2 | cpukit/librpc/src/rpc/rpc_prot.c:105:1 | | 2 | cpukit/librpc/src/rpc/svc_udp.c:346:5 | | 2 | cpukit/librpc/src/rpc/svc_auth_unix.c:147:1 | | 2 | cpukit/librpc/src/rpc/svc_auth_unix.c:59:1 | | 2 | cpukit/libnetworking/netinet/tcp_usrreq.c:292:16| | 2 | cpukit/libnetworking/libc/gethostnamadr.c:169:4 | | 2 | cpukit/librpc/src/rpc/rtime.c:75:1 | | 2 | cpukit/libnetworking/netinet/tcp_usrreq.c:484:16| | 2 | cpukit/librpc/src/rpc/rpc_prot.c:146:2 | +---+-+ | arm : 1 warning(s) | +-+++-+ | Shared :0 | BSP :0 | Network : 175 |Tests :0 | | LibCPU :0 | CPU Kit :0 || | +---+-+++-+ | 5 | cpukit/libnetworking/netinet/in_cksum.c:84:1| +---+-+ | i386 : 1 warning(s) | +-+++-+ | Shared :0 | BSP :0 | Network : 35 |Tests :0 | | LibCPU :0 | CPU Kit :0 || |