Re: Weird bug (maybe in clang - inetd.c does not compile without INET6)

On 17 Aug 2026, at 10:21, Milan Obuch <[email protected]> wrote:
> 
> I am building my systems from source and usually with WITHOUT_INET6
> knob among others. Partly inertia, partly minimalisation, partly
> speedup. At the moment, my boxes are at 15-STABLE, last working build
> is from May 30, 2026, git hash cfbe04310e254890c471f97f03dc7c3c5e9af01c
> - that's with compiler
> 
> # cc -v
> FreeBSD clang version 19.1.7 (https://github.com/llvm/llvm-project.git llvmorg-19.1.7-0-gcd708029e0b2)
> 
> Then new LLVM is being imported, builds do not work with error
> 
> ---- 8< ------------------------------------------------------------
> 
> ===> usr.bin/clang/llvm-min-tblgen (obj,all,install)
> make[3]: don't know how to make TableGen.cpp. Stop
> 
> make[3]: stopped making "all" in /usr/src/usr.bin/clang/llvm-min-tblgen
> *** Error code 2
> 
> Stop.
> make[2]: stopped making "bootstrap-tools" in /usr/src
>        0.40 real         0.16 user         0.04 sys
> *** Error code 1
> 
> Stop.
> make[1]: stopped making "buildworld" in /usr/src
> *** Error code 1
> 
> Stop.
> make: stopped making "buildworld" in /usr/src
> 
> ---- 8< ------------------------------------------------------------
> 
> for commits 0008bdb8bd28d1c15670084ca63015b551a33f76,
> c76ec29427955975909939e41052e19d000cf008,
> 85298a854720b7e712d4500b8d6ca340f6c39dd8,
> 4575fcccafed03e93f306ddd287229c4568e7280 - import was divided into more
> commits.
> 
> Next commits does not build as well, with different error:
> 
> ---- 8< ------------------------------------------------------------
> 
> In file included from
> /usr/src/contrib/llvm-project/libcxx/src/algorithm.cpp:9: In file
> included from
> /usr/src/contrib/llvm-project/libcxx/include/algorithm:1865: In file
> included from
> /usr/src/contrib/llvm-project/libcxx/include/__algorithm/inplace_merge.h:27:
> In file included from
> /usr/src/contrib/llvm-project/libcxx/include/__memory/unique_ptr.h:48:
> /usr/src/contrib/llvm-project/libcxx/include/__type_traits/remove_extent.h:9:2:
> error: unterminated conditional directive 9 | #ifndef
> _LIBCPP___TYPE_TRAITS_REMOVE_EXTENT_H |  ^ 3 errors generated. ***
> Error code 1
> 
> Stop.
> make[4]: stopped making "all" in /usr/src/lib/libc++
> *** Error code 1
> 
> Stop.
> make[3]: stopped making "_prebuild_libs" in /usr/src
> *** Error code 1
> 
> Stop.
> make[2]: stopped making "libraries" in /usr/src
>        8.67 real         6.56 user         0.65 sys
> *** Error code 1
> 
> Stop.
> make[1]: stopped making "buildworld" in /usr/src
> *** Error code 1
> 
> Stop.
> make: stopped making "buildworld" in /usr/src
> 
> ---- 8< ------------------------------------------------------------
> 
> Commits were fe97aeb6c4fb57f9cd3ae834e66d85bcb36b91ee,
> cf7d88e89f1dcc1b66c3e8f327e16122d42d6a6c,
> d08fb64ae2157d3bc65777589728e880e6409d72,
> c1db7ad65d704e4ae81bbfe7bdaf049120d324a2 - I think with those import of
> new LLVM is being finished.

Yeah, don't try to build the intermediate commits. I was not able to test all possible combinations, while wrestling with git.


> Now, one could expect things going back to normal, but for me, with
> WITHOUT_INET6 knob in action, did not work, the error is
> 
> ---- 8< ------------------------------------------------------------
> 
> ===> usr.sbin/inetd (all)
> cc -target x86_64-unknown-freebsd15.1
> --sysroot=/usr/obj/usr/src/amd64.amd64/tmp
> -B/usr/obj/usr/src/amd64.amd64/tmp/usr/bin -fpie -DPIC   -O2 -pipe
> -fno-common   -DLOGIN_CAP -DLIBWRAP -DIPSEC -fPIE
> -ffile-prefix-map=/usr/src=/usr/src
> -ffile-prefix-map=/usr/obj/usr/src=/usr/obj -MD  -MF.depend.inetd.pieo
> -MTinetd.pieo -std=gnu17 -Wno-format-zero-length
> -fstack-protector-strong -Wsystem-headers -Werror -Wall -Wno-format-y2k
> -W -Wno-unused-parameter -Wstrict-prototypes -Wmissing-prototypes
> -Wpointer-arith -Wreturn-type -Wcast-qual -Wwrite-strings -Wswitch
> -Wshadow -Wunused-parameter -Wcast-align -Wchar-subscripts
> -Wnested-externs -Wold-style-definition -Wno-pointer-sign -Wdate-time
> -Wmissing-variable-declarations -Wthread-safety -Wno-empty-body
> -Wno-string-plus-int -Wno-unused-const-variable
> -Wno-error=unused-but-set-parameter
> -Wno-error=cast-function-type-mismatch  -Qunused-arguments      -c
> /usr/src/usr.sbin/inetd/inetd.c -o inetd.pieo
> /usr/src/usr.sbin/inetd/inetd.c:737:26: warning: cast from 'struct
> servtab *(*)(int)' to 'bi_fn_t *' (aka 'void (*)(int, struct servtab
> *)') converts to incompatible function type
> [-Wcast-function-type-mismatch] 737 |
> sep->se_bi->bi_fn == (bi_fn_t *) tcpmux) { |
>                          ^~~~~~~~~~~~~~~~~~
> /usr/src/usr.sbin/inetd/inetd.c:1365:44: error: variable 'nbuf2' is
> uninitialized when passed as a const pointer argument here
> [-Werror,-Wuninitialized-const-pointer] 1365 |
>       rpcb_set(sep->se_rpc_prog, i, netid2, &nbuf2); |
>                                                       ^~~~~ 1 warning

This just looks like a warning which is only produced when INET6 is off. I don't know much about inetd, but with #ifdef mazes something always goes wrong. :)

-Dimitry