On Sun, Dec 24, 2017 at 07:37:23PM +0000, Ivan Kozik wrote: > On Fri, Dec 22, 2017 at 8:44 AM, Greg Kroah-Hartman > <gre...@linuxfoundation.org> wrote: > > This is the start of the stable review cycle for the 4.14.9 release. > > There are 159 patches in this series, all will be posted as a response > > to this one. If anyone has any issues with these being applied, please > > let me know. > > > > Responses should be made by Sun Dec 24 08:45:36 UTC 2017. > > Anything received after that time might be too late. > > > > > Josh Poimboeuf <jpoim...@redhat.com> > > x86/unwind: Make CONFIG_UNWINDER_ORC=y the default in kconfig for 64-bit > > This is uncovering a very difficult-to-debug build failure with NVIDIA DKMS: > with CONFIG_UNWINDER_ORC=y, out-of-tree modules hit this rule in > scripts/Makefile.build: > > $(obj)/%.o: $(src)/%.c $(recordmcount_source) $(objtool_dep) FORCE > > and fail (here, at least) to build tools/objtool/objtool (note: I do have > libelf-dev installed)
Is this problem also in Linus's tree? > After editing dkms.conf to do `make --debug=a -j1`, I see make output: > > Considering target file > '/var/lib/dkms/nvidia-current/387.34/build/nvidia/nv-gpu-numa.o'. > File '/var/lib/dkms/nvidia-current/387.34/build/nvidia/nv-gpu-numa.o' > does not exist. > Looking for an implicit rule for > '/var/lib/dkms/nvidia-current/387.34/build/nvidia/nv-gpu-numa.o'. > [...] > Trying rule prerequisite 'tools/objtool/objtool'. > Looking for a rule with intermediate file 'tools/objtool/objtool'. > Avoiding implicit rule recursion. > > then silently fail to build objtool, silently fail to build all the .o files, > then continue until ld finally errors out trying to link nonexistent object > files. Is this just a bug in the nvidia Makefile somehow? > If things look alright to you, and objtool is known to work with out-of-tree > modules, and the Debian packaging just needs to be adjusted, please ignore; > I figured I'd send this anyway because it was such a pain to debug. objtool should work with out-of-tree modules, again, try Linus's tree to verify this. thanks, greg k-h