>From dmesg radeondrm0: 1024x768, 32bpp wsdisplay0 at radeondrm0 mux 1: console (std, vt100 emulation), using wskbd0 wsdisplay0: screen 1-5 added (std, vt100 emulation) trap [machdep-ml32.exe]43360/78357 type 6: sp fffe0a80 not inside 7f7ffdfe2000-7f7ffffe2000 map stack 0xe7b9ee9000-0xe7b9eef000 of map 0xffffff01b70e9110 failed: bad protection map stack for pid 4268 failed map stack 0x599a8e30000-0x599a8e36000 of map 0xffffff01b70e9b60 failed: bad protection map stack for pid 20069 failed map stack 0x15df534ca000-0x15df534d0000 of map 0xffffff01b70e9110 failed: bad protection map stack for pid 8896 failed map stack 0xb213ba0f000-0xb213ba15000 of map 0xffffff01b70e9218 failed: bad protection map stack for pid 87617 failed map stack 0x11c91a952000-0x11c91a958000 of map 0xffffff01b70e9b60 failed: bad protection map stack for pid 47459 failed map stack 0x1d4d83c51000-0x1d4d83c57000 of map 0xffffff01b70e9110 failed: bad protection map stack for pid 98623 failed map stack 0x18ddb3921000-0x18ddb3927000 of map 0xffffff0120511320 failed: bad protection map stack for pid 36066 failed map stack 0xaa0c5d28000-0xaa0c5d2e000 of map 0xffffff01403edc78 failed: bad protection map stack for pid 69906 failed map stack 0x3c2c0518000-0x3c2c051e000 of map 0xffffff01403edc78 failed: bad protection map stack for pid 93650 failed
Looking at machdep-ml $ pkglocate machdep-ml cil-1.5.1p6:devel/cil:/usr/local/libexec/cil/machdep-ml.exe Search through dpb logs $ find /usr/ports/logs/amd64/packages/ /usr/ports/logs/amd64/paths/ -type f -name "*.log" | xargs grep machdep-ml 2>/dev/null /usr/ports/logs/amd64/paths/devel/cil.log:config.status: creating src/machdep-ml.c /usr/ports/logs/amd64/paths/devel/cil.log:echo "(* This module was generated automatically by code in Makefile and machdep-ml.c *)" >obj/x86_OPENBSD/machdep.ml /usr/ports/logs/amd64/paths/devel/cil.log:if cc -D_GNUCC src/machdep-ml.c -o obj/x86_OPENBSD/machdep-ml.exe ;then \ /usr/ports/logs/amd64/paths/devel/cil.log: echo "machdep-ml.exe created succesfully." \ /usr/ports/logs/amd64/paths/devel/cil.log:machdep-ml.exe created succesfully. /usr/ports/logs/amd64/paths/devel/cil.log:obj/x86_OPENBSD/machdep-ml.exe >>obj/x86_OPENBSD/machdep.ml /usr/ports/logs/amd64/paths/devel/cil.log:if cl /D_MSVC src/machdep-ml.c /Feobj/x86_OPENBSD/machdep-ml.exe /Foobj/x86_OPENBSD/machdep-ml.obj ;then \ /usr/ports/logs/amd64/paths/devel/cil.log: obj/x86_OPENBSD/machdep-ml.exe >>obj/x86_OPENBSD/machdep.ml ;\ /usr/ports/logs/amd64/paths/devel/cil.log:if cc -D_GNUCC -m32 src/machdep-ml.c -o obj/x86_OPENBSD/machdep-ml32.exe ;then \ /usr/ports/logs/amd64/paths/devel/cil.log: sed -e "s|nogcc32model|` obj/x86_OPENBSD/machdep-ml32.exe --env`|" lib/Cilly.pm > lib/Cilly.pm.tmp; \ /usr/ports/logs/amd64/paths/devel/cil.log:/usr/bin/ld: warning: i386 architecture of input file `/tmp/machdep-ml-5e93ab.o' is incompatible with i386:x86-64 output /usr/ports/logs/amd64/paths/devel/cil.log:if cc -D_GNUCC -m64 src/machdep-ml.c -o obj/x86_OPENBSD/machdep-ml64.exe ;then \ /usr/ports/logs/amd64/paths/devel/cil.log: sed -e "s|nogcc64model|` obj/x86_OPENBSD/machdep-ml64.exe --env`|" lib/Cilly.pm > lib/Cilly.pm.tmp; \ /usr/ports/logs/amd64/paths/devel/cil.log:/usr/ports/pobj/cil-1.5.1/bin/install -c -s -m 755 /usr/ports/pobj/cil-1.5.1/cil-1.5.1/obj/*_OPENBSD/machdep-ml.exe /usr/ports/pobj/cil-1.5.1/fake-amd64/usr/local/libexec/cil $ >From dpb log if cc -D_GNUCC -m32 src/machdep-ml.c -o obj/x86_OPENBSD/machdep-ml32.exe ;then \ sed -e "s|nogcc32model|` obj/x86_OPENBSD/machdep-ml32.exe --env`|" lib/Cilly.pm > lib/Cilly.pm.tmp; \ mv lib/Cilly.pm.tmp lib/Cilly.pm; \ fi /usr/bin/ld: warning: i386 architecture of input file `/tmp/machdep-ml-5e93ab.o' is incompatible with i386:x86-64 output Segmentation fault (core dumped) devel/cil - appears to be the issue with the trap, not sure why machdep-ml32.exe is being executed .exe suggests it's for Windows. Check rebuilding devel/cil and is the cause of trap but none of the map stack. While not nice, can live with the issue for now. That deals with trap, the "map stack", can't tell if from one port or multiple ports, as only building a selected 319 ports at the time, that narrows down the ports, some fail with clang6. Rebuilt release so fully up to date, and started building the 319 ports again one by one, with a dmesg | wc before and after, extra lines in dmesg could be a map stack. lang/gforth dmesg line count increased from 191 to 349 which is 158 lines matching the 158 found before. The one by one build completed, and building lang/gforth was the cause of all the "map stack" messages.
