[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAP-5=fVHFMJvaY_UE4QdV-PW+gy1EuyiHDXqWJmHVxS9Mr3XEQ@mail.gmail.com>
Date: Wed, 1 Feb 2023 10:57:40 -0800
From: Ian Rogers <irogers@...gle.com>
To: Jiri Olsa <olsajiri@...il.com>
Cc: Alexei Starovoitov <ast@...nel.org>,
Daniel Borkmann <daniel@...earbox.net>,
Andrii Nakryiko <andrii@...nel.org>,
Martin KaFai Lau <martin.lau@...ux.dev>,
Song Liu <song@...nel.org>, Yonghong Song <yhs@...com>,
John Fastabend <john.fastabend@...il.com>,
KP Singh <kpsingh@...nel.org>,
Stanislav Fomichev <sdf@...gle.com>,
Hao Luo <haoluo@...gle.com>,
Connor OBrien <connoro@...gle.com>,
Nathan Chancellor <nathan@...nel.org>,
Kumar Kartikeya Dwivedi <memxor@...il.com>,
bpf@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v1] tools/resolve_btfids: Tidy host CFLAGS forcing
On Wed, Feb 1, 2023 at 2:43 AM Jiri Olsa <olsajiri@...il.com> wrote:
>
> On Wed, Feb 01, 2023 at 11:00:02AM +0100, Jiri Olsa wrote:
> > On Tue, Jan 31, 2023 at 05:50:15PM -0800, Ian Rogers wrote:
> > > Avoid passing CROSS_COMPILE to submakes and ensure CFLAGS is forced to
> > > HOSTCFLAGS for submake builds. This fixes problems with cross
> > > compilation.
> > >
> > > Tidy to not unnecessarily modify/export CFLAGS, make the override for
> > > prepare and build clearer.
> > >
> > > Fixes: 13e07691a16f ("tools/resolve_btfids: Alter how HOSTCC is forced")
> > > Reported-by: Nathan Chancellor <nathan@...nel.org>
> > > Signed-off-by: Ian Rogers <irogers@...gle.com>
> >
> > hum, that seems to build just the fixdep and skip the resolve_btfids binary
> >
> > make[1]: Entering directory '/home/jolsa/kernel/linux-qemu/build'
> > GEN Makefile
> > CALL ../scripts/checksyscalls.sh
> > DESCEND bpf/resolve_btfids
> > HOSTCC /home/jolsa/kernel/linux-qemu/build/tools/bpf/resolve_btfids/fixdep.o
> > HOSTLD /home/jolsa/kernel/linux-qemu/build/tools/bpf/resolve_btfids/fixdep-in.o
> > LINK /home/jolsa/kernel/linux-qemu/build/tools/bpf/resolve_btfids/fixdep
> > UPD include/generated/utsversion.h
> > CC init/version-timestamp.o
> > LD .tmp_vmlinux.btf
> > BTF .btf.vmlinux.bin.o
> > die__process_unit: DW_TAG_label (0xa) @ <0x4f0d4> not handled!
> >
> > ...
> >
> > die__process_unit: DW_TAG_label (0xa) @ <0xaf91cc3> not handled!
> > die__process_unit: DW_TAG_label (0xa) @ <0xb032fa7> not handled!
> > LD .tmp_vmlinux.kallsyms1
> > NM .tmp_vmlinux.kallsyms1.syms
> > KSYMS .tmp_vmlinux.kallsyms1.S
> > AS .tmp_vmlinux.kallsyms1.S
> > LD .tmp_vmlinux.kallsyms2
> > NM .tmp_vmlinux.kallsyms2.syms
> > KSYMS .tmp_vmlinux.kallsyms2.S
> > AS .tmp_vmlinux.kallsyms2.S
> > LD .tmp_vmlinux.kallsyms3
> > NM .tmp_vmlinux.kallsyms3.syms
> > KSYMS .tmp_vmlinux.kallsyms3.S
> > AS .tmp_vmlinux.kallsyms3.S
> > LD vmlinux
> > BTFIDS vmlinux
> > ../scripts/link-vmlinux.sh: line 277: ./tools/bpf/resolve_btfids/resolve_btfids: No such file or directory
> > make[2]: *** [../scripts/Makefile.vmlinux:35: vmlinux] Error 127
> > make[2]: *** Deleting file 'vmlinux'
> > make[1]: *** [/home/jolsa/kernel/linux-qemu/Makefile:1264: vmlinux] Error 2
> > make[1]: Leaving directory '/home/jolsa/kernel/linux-qemu/build'
> > make: *** [Makefile:242: __sub-make] Error 2
> >
> > we actually have the hostprogs support in tools/build and we use it for
> > fixdep, I think we should use it also here, I'll check
>
> it doesn't look that bad.. the change below fixes the build for me,
> perhaps we should do that for all the host tools
>
> jirka
I don't mind this. The fixdep vs all thing is just cause by the
ordering in the Makefile, you can fix by specifying the target or add
this patch:
```
--- a/tools/bpf/resolve_btfids/Makefile
+++ b/tools/bpf/resolve_btfids/Makefile
@@ -58,11 +58,11 @@ HOST_OVERRIDES_BUILD := $(HOST_OVERRIDES_PREPARE) \
LIBS = $(LIBELF_LIBS) -lz
+all: $(BINARY)
+
export srctree OUTPUT Q
include $(srctree)/tools/build/Makefile.include
-all: $(BINARY)
-
prepare: $(BPFOBJ) $(SUBCMDOBJ)
$(OUTPUT) $(OUTPUT)/libsubcmd $(LIBBPF_OUT):
```
Should we do this and the hostprogs migration as a follow up? There
isn't that much use of hostprogs in tools, but I like that your change
will show HOSTCC rather than CC during compilation. If we use
hostprogs can we just avoid the overrides altogether?
Thanks,
Ian
>
> ---
> diff --git a/tools/bpf/resolve_btfids/Build b/tools/bpf/resolve_btfids/Build
> index ae82da03f9bf..077de3829c72 100644
> --- a/tools/bpf/resolve_btfids/Build
> +++ b/tools/bpf/resolve_btfids/Build
> @@ -1,3 +1,5 @@
> +hostprogs := resolve_btfids
> +
> resolve_btfids-y += main.o
> resolve_btfids-y += rbtree.o
> resolve_btfids-y += zalloc.o
> @@ -7,4 +9,4 @@ resolve_btfids-y += str_error_r.o
>
> $(OUTPUT)%.o: ../../lib/%.c FORCE
> $(call rule_mkdir)
> - $(call if_changed_dep,cc_o_c)
> + $(call if_changed_dep,host_cc_o_c)
> diff --git a/tools/bpf/resolve_btfids/Makefile b/tools/bpf/resolve_btfids/Makefile
> index daed388aa5d7..de513fd08535 100644
> --- a/tools/bpf/resolve_btfids/Makefile
> +++ b/tools/bpf/resolve_btfids/Makefile
> @@ -22,6 +22,9 @@ HOST_OVERRIDES := AR="$(HOSTAR)" CC="$(HOSTCC)" LD="$(HOSTLD)" ARCH="$(HOSTARCH)
> EXTRA_CFLAGS="$(HOSTCFLAGS) $(KBUILD_HOSTCFLAGS)"
>
> RM ?= rm
> +HOSTCC ?= gcc
> +HOSTLD ?= ld
> +HOSTAR ?= ar
> CROSS_COMPILE =
>
> OUTPUT ?= $(srctree)/tools/bpf/resolve_btfids/
> @@ -64,7 +67,7 @@ $(BPFOBJ): $(wildcard $(LIBBPF_SRC)/*.[ch] $(LIBBPF_SRC)/Makefile) | $(LIBBPF_OU
> LIBELF_FLAGS := $(shell $(HOSTPKG_CONFIG) libelf --cflags 2>/dev/null)
> LIBELF_LIBS := $(shell $(HOSTPKG_CONFIG) libelf --libs 2>/dev/null || echo -lelf)
>
> -CFLAGS += -g \
> +HOSTCFLAGS += -g \
> -I$(srctree)/tools/include \
> -I$(srctree)/tools/include/uapi \
> -I$(LIBBPF_INCLUDE) \
> @@ -73,7 +76,7 @@ CFLAGS += -g \
>
> LIBS = $(LIBELF_LIBS) -lz
>
> -export srctree OUTPUT CFLAGS Q
> +export srctree OUTPUT HOSTCFLAGS Q HOSTCC HOSTLD HOSTAR
> include $(srctree)/tools/build/Makefile.include
>
> $(BINARY_IN): fixdep FORCE prepare | $(OUTPUT)
Powered by blists - more mailing lists