Re: Host CPUTYPE?=bdver2 unable to build release target for CPUTYPE?=slm

From: Alastair Hogge <agh_at_fastmail.fm>
Date: Fri, 27 Oct 2017 13:24:13 +0800
Source tree is at r325004

On Fri, 27 Oct 2017-10-27 12:59:30 Alastair Hogge wrote:
> Hi,
> 
> I am attempting to build a release ${SRC}/release/release.sh -c
> ${custom_release.conf}, however, the build fails with:
> 
> make[1]: "/scratch/fafnir/usr/src/Makefile.inc1" line 162: SYSTEM_COMPILER:
> Determined that CC=cc matches the source tree.  Not bootstrapping a cross-
> compiler.
> --------------------------------------------------------------
> 
> >>> World build started on Fri Oct 27 00:41:56 UTC 2017
> 
> --------------------------------------------------------------
> 
> --------------------------------------------------------------
> 
> >>> Rebuilding the temporary build tree
> 
> --------------------------------------------------------------
> 
> --------------------------------------------------------------
> 
> >>> stage 1.1: legacy release compatibility shims
> 
> --------------------------------------------------------------
> ===> tools/build (obj,includes,all,install)
> 
> --------------------------------------------------------------
> 
> >>> stage 1.2: bootstrap tools
> 
> --------------------------------------------------------------
> ===> lib/clang/libllvmminimal (obj,all,install)
> ===> usr.bin/clang/llvm-tblgen (obj,all,install)
> ===> usr.bin/clang/clang-tblgen (obj,all,install)
> ===> kerberos5/tools/make-roken (obj,all,install)
> ===> kerberos5/lib/libroken (obj,all,install)
> ===> kerberos5/lib/libvers (obj,all,install)
> ===> kerberos5/tools/asn1_compile (obj,all,install)
> yacc: 4 shift/reduce conflicts.
> ===> kerberos5/tools/slc (obj,all,install)
> ===> usr.bin/compile_et (obj,all,install)
> ===> usr.bin/fortune/strfile (obj,all,install)
> ===> usr.bin/dtc (obj,all,install)
> ===> usr.bin/lorder (obj,all,install)
> ===> lib/libopenbsd (obj,all,install)
> ===> usr.bin/mandoc (obj,all,install)
> ===> usr.bin/rpcgen (obj,all,install)
> ===> lib/liby (obj,all,install)
> ===> usr.bin/yacc (obj,all,install)
> ===> usr.bin/xinstall (obj,all,install)
> ===> usr.sbin/config (obj,all,install)
> ===> usr.bin/vtfontcvt (obj,all,install)
> ===> usr.bin/localedef (obj,all,install)
> In file included from
> /scratch/fafnir/usr/src/usr.bin/localedef/collate.c:50: In file included
> from /scratch/fafnir/usr/src/lib/libc/locale/collate.h:42:
> /scratch/fafnir/usr/src/lib/libc/locale/xlocale_private.h:168:18: warning:
> passing 'long *' to parameter of type 'volatile u_long *' (aka 'volatile
> unsigned long *') converts between pointers to integer types with different
> sign [-Wpointer-sign]
>         atomic_add_long(&(obj->retain_count), 1);
> 
> [log trimmed]
> 
> ===> usr.sbin/crunch (installconfig)
> ===> usr.sbin/crunch/crunchgen (installconfig)
> ===> usr.sbin/crunch/crunchide (installconfig)
> ===> usr.sbin/unbound (installconfig)
> ===> usr.sbin/unbound/daemon (installconfig)
> ===> usr.sbin/unbound/anchor (installconfig)
> ===> usr.sbin/unbound/checkconf (installconfig)
> ===> usr.sbin/unbound/control (installconfig)
> ===> usr.sbin/unbound/local-setup (installconfig)
> ===> usr.sbin/uathload (installconfig)
> ===> usr.sbin/uhsoctl (installconfig)
> ===> usr.sbin/usbconfig (installconfig)
> ===> usr.sbin/usbdump (installconfig)
> ===> usr.sbin/ac (installconfig)
> ===> usr.sbin/lastlogin (installconfig)
> ===> usr.sbin/utx (installconfig)
> ===> etc (installconfig)
> ELF ldconfig path: /lib /usr/lib /usr/lib/compat
> 32-bit compatibility ldconfig path: /usr/lib32
> make: "/usr/src/Makefile" line 313: warning: "LC_ALL=C date" exited on a
> signal make[1]: "/usr/src/Makefile.inc1" line 162: SYSTEM_COMPILER:
> Determined that CC=cc matches the source tree.  Not bootstrapping a
> cross-compiler. make[2]: "/usr/src/release/Makefile.ec2" line 9: warning:
> "date +-%Y-%m-%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.azure" line 20: warning: "date
> +-%Y-%m-%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.gce" line 22: warning: "date +-%Y-%m-%d"
> exited on a signal
> make[2]: "/usr/src/release/Makefile.vagrant" line 32: warning: "date
> +-%Y%m%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.vagrant" line 35: warning: "date
> +%Y.%m.%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.ec2" line 9: warning: "date +-%Y-%m-%d"
> exited on a signal
> make[2]: "/usr/src/release/Makefile.azure" line 20: warning: "date
> +-%Y-%m-%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.gce" line 22: warning: "date +-%Y-%m-%d"
> exited on a signal
> make[2]: "/usr/src/release/Makefile.vagrant" line 32: warning: "date
> +-%Y%m%d" exited on a signal
> make[2]: "/usr/src/release/Makefile.vagrant" line 35: warning: "date
> +%Y.%m.%d" exited on a signal
> --------------------------------------------------------------
> 
> >>> World build started on
> 
> --------------------------------------------------------------
> 
> --------------------------------------------------------------
> 
> >>> Rebuilding the temporary build tree
> 
> --------------------------------------------------------------
> Illegal instruction (core dumped)
> *** [_worldtmp] Error code 132
> 
> make[1]: stopped in /usr/src
> 1 error
> 
> make[1]: stopped in /usr/src
> *** [buildworld] Error code 2
> 
> make: stopped in /usr/src
> 1 error
> 
> make: stopped in /usr/src
> 
> Command exit status: 2
> 
> I have three other hosts that I build for, the only major difference between
> each hosts is the collections of ports, wireless/wired options in the
> kernel, and CPUTYPE. The build host is built with CPUTYPE?=bdver2, the
> others hosts are btver1, corei7, corei7-avx, and slm. There used to a core2
> host, but I retired it, however, I mention it because it built
> successfully.
> 
> The failed build log above is from the bdver2 host building a release for
> the slm host, all other builds complete without error.
> 
> If I use the slm config and make build{kernel|world} from ${SRC}, the build
> completes, if I populate a jail with stock or my bdver2 distribution set,
> the build also succeeds, the only time the slm build fails is during
> ${SRC}/release/release.sh on the bdver2 host.
> 
> Is it possible to build a x86_64 Silvermont release from an AMD Bulldozer?
> Why does the regular build from the top of the source tree complete and the
> one from the release sub directory fail? I normally export the custom
> builds via NFS, tho, I still need release images and distribution sets.
> 
> The release build with customisations works on the slm host itself, it just
> takes too many hours.
> 
> $ cat /etc/slm-make.conf:
> ALWAYS_CHECK_MAKE=      YES
> CPUTYPE?=               slm
> DEFAULT_VERSIONS=       bdb=5 linux=c7_64 ssl=openssl
> LICENSES_ACCEPTED+=     MSPAT
> KERNCONF=FAFNIR
> MALLOC_PRODUCTION=
> MODULES_OVERRIDE=       vmm
> WITH_PKG=               devel
> WITH_SSP_PORTS=
> 
> $ cat /etc/bdver2-make.conf:
> ALWAYS_CHECK_MAKE=      YES
> CPUTYPE?=               bdver2
> DEFAULT_VERSIONS=       bdb=5 linux=c7_64 ssl=openssl
> LOADER_FIREWIRE_SUPPORT=
> LICENSES_ACCEPTED+=     MSPAT
> KERNCONF=               DIREWOLF
> MALLOC_PRODUCTION=
> MODULES_OVERRIDE=       linux_common linuxkpi linprocfs linsysfs linux64 \
> linux vmm
> QT4_OPTIONS=            CUPS
> TEX_DEFAULT=            texlive
> WINE_CROSS_BUILD=
> WITH_NVIDIA_GL=
> WITH_PKG=               devel
> WITH_SSP_PORTS=
> 
> Thanks,
> Alastair

-- 
Reisner's Rule of Conceptual Inertia:
	If you think big enough, you'll never have to do it.

Received on Fri Oct 27 2017 - 03:24:32 UTC

This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:13 UTC