buildkernel with -j4 gives "No entry found [dwarf_next_cu_header_c(61)]" for hpt27xx, hptmv, hptnr, hptrr. Using -B fixes the issue. ========================== Working Copy Root Path: /usr/src URL: https://svn.freebsd.org/base/stable/11 Relative URL: ^/stable/11 Repository Root: https://svn.freebsd.org/base Repository UUID: ccf9f872-aa2e-dd11-9fc8-001c23d0bc1f Revision: 328591 ========================== --- hpt27xx_lib.o --- uudecode < /usr/src/sys/dev/hpt27xx/amd64-elf.hpt27xx_lib.o.uu ctfconvert -L VERSION -g hpt27xx_lib.o ERROR: ctfconvert: rc = -1 No entry found [dwarf_next_cu_header_c(61)] --- hptmvraid.o --- uudecode < /usr/src/sys/dev/hptmv/amd64-elf.raid.o.uu ctfconvert -L VERSION -g hptmvraid.o ERROR: ctfconvert: rc = -1 No entry found [dwarf_next_cu_header_c(61)] --- hptnr_lib.o --- uudecode < /usr/src/sys/dev/hptnr/amd64-elf.hptnr_lib.o.uu ctfconvert -L VERSION -g hptnr_lib.o ERROR: ctfconvert: rc = -1 No entry found [dwarf_next_cu_header_c(61)] --- hptrr_lib.o --- uudecode < /usr/src/sys/dev/hptrr/amd64-elf.hptrr_lib.o.uu ctfconvert -L VERSION -g hptrr_lib.o ERROR: ctfconvert: rc = -1 No entry found [dwarf_next_cu_header_c(61)] ===========================
Can you expand on "Using -B fixes issue" - what's the exact command you issued? This error is really a false positive. It is effectively ctfconvert reporting that there is no DWARF debug info in the .o file. We should fix up the build system to avoid emitting spurious false positive errors, but it doesn't indicate a real problem with the kernel you've built.
I get the error when doing "make -j4 buildkernel KERNCONF=VIMAGE". When I add -B "make -B -j4 buildkernel KERNCONF=VIMAGE" there are no errors. So -B uses backwards compatiblty by executing a single shell per command. Does this looks like a timing issue? ======== cat /etc/src-env.conf WITH_META_MODE = YES # WITH_DIRDEPS_BUILD = yes cat /etc/src.conf WITH_CCACHE_BUILD = yes WITH_FAST_DEPEND = yes ========= Thank you for the help. Robert
This is a harmless build issue; the .o files in question don't contain debug info, but make's pattern matching rules somehow cause us to run ctfconvert.
^Triage: close as OBE. I'm sorry that this PR did not get addressed in a timely fashion. Please let us know if this still occurs on a supported OSVERSION.