Re: [PATCH bpf-next v7 2/9] selftests/bpf: drop stale lines, restore two header dependencies

[Eduard Zingerman <eddyz87@gmail.com>]

On Sun, 2026-08-23 at 12:42 -0700, Mykola Lysenko wrote:
> Four target-specific lines name objects nothing builds. Three name
> the BPF objects by their pre-rename names: commit afef88e65554
> ("selftests/bpf: Store BPF object files with .bpf.o extension") left
> them without a target; the flow_dissector_load.o dependency names an
> intermediate the generic one-step compile+link rule does not produce.
> 
> Three linked-skeleton dependency map entries (xsk_xdp_progs,
> xdp_hw_metadata, xdp_features) were dead on arrival: commit
> f0a249df1b07 ("selftests/xsk: get rid of built-in XDP program"),
> commit 297a3f124155 ("selftests/bpf: Simple program to dump XDP RX
> metadata") and commit 4dba3e7852b7 ("selftests/bpf: introduce XDP
> compliance test tool") each added one, mimicking the neighboring
> entries without adding the skeleton to LINKED_SKELS, the only place
> the -deps map is read from; these three skeletons are generated by
> the regular pattern rule, which never consults it.
> 
> The '-fno-inline' pair (test_l4lb_noinline, test_xdp_noinline) had
> stopped taking effect even earlier, with commit 74b5a5968fe8
> ("selftests/bpf: Replace test_progs and test_maps w/ general rule"):
> since then the BPF compile recipe expands TRUNNER_BPF_CFLAGS, a
> simply-expanded copy of BPF_CFLAGS taken when the runner rules are
> instantiated, which a target-specific 'BPF_CFLAGS +=' cannot reach.
> Both programs have compiled without the flag since, and nothing was
> lost: every function they define
> carries a noinline annotation, except the SEC() entry points, which
> nothing in the file calls, and the single __always_inline helper in
> each, a deliberate exception the flag never overrode; and
> compiling with -fno-inline restored yields byte-identical objects.
> 
> The two header dependencies with remaining value are restored in
> working form. flow_dissector_load.h moves to the binary itself, which
> is linked straight from its .c by the generic '$(OUTPUT)/%:%.c' rule
> - editing the header now rebuilds it - on the line that already lists
> the binary's helper object. cgroup_getset_retval_hooks.h -
> added by commit e7215f574079 ("selftests/bpf: Make sure
> bpf_{g,s}et_retval is exposed everywhere") days before the rename
> orphaned it - lived in the top directory, outside the progs/*.h
> blanket prerequisite of the BPF object rules, so editing it never
> rebuilt the BPF object. Move it under progs/, next to its only BPF
> consumer, where the blanket prerequisite covers it - the arrangement
> several other headers shared between progs/ and prog_tests/ already
> use. Its userspace consumer is tracked exactly by the
> compiler-generated dependency files.
> 
> 'CURDIR := $(abspath .)' redefines make's builtin to the value it
> already has, and 'OBJCOPY ?= $(CROSS_COMPILE)objcopy' defines a
> variable nothing in the selftests build or the included kselftest
> infrastructure ever reads; drop both.
> 
> Signed-off-by: Mykola Lysenko <nickolay.lysenko@gmail.com>
> ---

Acked-by: Eduard Zingerman <eddyz87@gmail.com>

...