From: Mark Rutland <mark.rutland@arm.com>
To: linux-kernel@vger.kernel.org
Cc: linux-tip-commits@vger.kernel.org,
Ian Rogers <irogers@google.com>,
Josh Poimboeuf <jpoimboe@kernel.org>,
x86@kernel.org
Subject: Re: [tip: objtool/core] objtool: Fix HOSTCC flag usage
Date: Wed, 1 Feb 2023 17:02:16 +0000 [thread overview]
Message-ID: <Y9qbGHDBFtGoqnKK@FVFF77S0Q05N> (raw)
In-Reply-To: <167526879495.4906.2898311831401901292.tip-bot2@tip-bot2>
Hi,
I just spotted this breaks cross-compiling; details below.
On Wed, Feb 01, 2023 at 04:26:34PM -0000, tip-bot2 for Ian Rogers wrote:
> The following commit has been merged into the objtool/core branch of tip:
>
> Commit-ID: 93eacc15687a491a9cf829f94b6891bf544084f3
> Gitweb: https://git.kernel.org/tip/93eacc15687a491a9cf829f94b6891bf544084f3
> Author: Ian Rogers <irogers@google.com>
> AuthorDate: Thu, 26 Jan 2023 11:06:06 -08:00
> Committer: Josh Poimboeuf <jpoimboe@kernel.org>
> CommitterDate: Mon, 30 Jan 2023 16:28:18 -08:00
>
> objtool: Fix HOSTCC flag usage
>
> HOSTCC is always wanted when building objtool. Setting CC to HOSTCC
> happens after tools/scripts/Makefile.include is included, meaning
> flags (like CFLAGS) are set assuming say CC is gcc, but then it can be
> later set to HOSTCC which may be clang. tools/scripts/Makefile.include
> is needed for host set up and common macros in objtool's
> Makefile. Rather than override the CC variable to HOSTCC, just pass CC
> as HOSTCC to the sub-makes of Makefile.build, the libsubcmd builds and
> also to the linkage step.
>
> Signed-off-by: Ian Rogers <irogers@google.com>
> Link: https://lore.kernel.org/r/20230126190606.40739-4-irogers@google.com
> Signed-off-by: Josh Poimboeuf <jpoimboe@kernel.org>
> ---
> tools/objtool/Makefile | 25 ++++++++++++++-----------
> 1 file changed, 14 insertions(+), 11 deletions(-)
>
> diff --git a/tools/objtool/Makefile b/tools/objtool/Makefile
> index d54b669..29a8cd7 100644
> --- a/tools/objtool/Makefile
> +++ b/tools/objtool/Makefile
> @@ -2,11 +2,6 @@
> include ../scripts/Makefile.include
> include ../scripts/Makefile.arch
>
> -# always use the host compiler
> -AR = $(HOSTAR)
> -CC = $(HOSTCC)
> -LD = $(HOSTLD)
So as of this change, CC is now the cross-compiler CC ....
> -
> ifeq ($(srctree),)
> srctree := $(patsubst %/,%,$(dir $(CURDIR)))
> srctree := $(patsubst %/,%,$(dir $(srctree)))
> @@ -34,13 +29,18 @@ INCLUDES := -I$(srctree)/tools/include \
> -I$(srctree)/tools/objtool/include \
> -I$(srctree)/tools/objtool/arch/$(SRCARCH)/include \
> -I$(LIBSUBCMD_OUTPUT)/include
> +# Note, EXTRA_WARNINGS here was determined for CC and not HOSTCC, it
> +# is passed here to match a legacy behavior.
> WARNINGS := $(EXTRA_WARNINGS) -Wno-switch-default -Wno-switch-enum -Wno-packed -Wno-nested-externs
> -CFLAGS := -Werror $(WARNINGS) $(KBUILD_HOSTCFLAGS) -g $(INCLUDES) $(LIBELF_FLAGS)
> -LDFLAGS += $(LIBELF_LIBS) $(LIBSUBCMD) $(KBUILD_HOSTLDFLAGS)
> +OBJTOOL_CFLAGS := -Werror $(WARNINGS) $(KBUILD_HOSTCFLAGS) -g $(INCLUDES) $(LIBELF_FLAGS)
> +OBJTOOL_LDFLAGS := $(LIBELF_LIBS) $(LIBSUBCMD) $(KBUILD_HOSTLDFLAGS)
>
> # Allow old libelf to be used:
> elfshdr := $(shell echo '$(pound)include <libelf.h>' | $(CC) $(CFLAGS) -x c -E - | grep elf_getshdr)
... and so as of this patch, this check uses the cross-compiler CC rather than
the HOSTCC (and IIUC the wrong CFLAGS too).
So that probably wants to be `$(HOSTCC) $(OBJTOOL_CFLAGS)` ?
The cross CC might not even have libelf in its header path ...
> -CFLAGS += $(if $(elfshdr),,-DLIBELF_USE_DEPRECATED)
> +OBJTOOL_CFLAGS += $(if $(elfshdr),,-DLIBELF_USE_DEPRECATED)
... which means that we might end up setting DLIBELF_USE_DEPRECATED when we
didn't mean to.
I saw this blowing up in some experiments with an arm64 objtool built on
x86_64, and I believe the reverse can also occur (i.e. building for x86_64 on
arm64).
As of this patch, I get a splat:
| [mark@lakrids:~/src/linux]% usekorg 12.1.0 make ARCH=arm64 CROSS_COMPILE=aarch64-linux- tools/objtool/elf.o
| CALL scripts/checksyscalls.sh
| DESCEND objtool
| <stdin>:1:10: fatal error: libelf.h: No such file or directory
| compilation terminated.
| INSTALL libsubcmd_headers
| CC /home/mark/src/linux/tools/objtool/elf.o
| elf.c: In function ‘read_sections’:
| elf.c:273:2: error: ‘elf_getshnum’ is deprecated [-Werror=deprecated-declarations]
| 273 | if (elf_getshdrnum(elf->elf, §ions_nr)) {
| | ^~
| In file included from /usr/include/gelf.h:32,
| from /home/mark/src/linux/tools/objtool/include/objtool/elf.h:10,
| from elf.c:22:
| /usr/include/libelf.h:310:12: note: declared here
| 310 | extern int elf_getshnum (Elf *__elf, size_t *__dst)
| | ^~~~~~~~~~~~
| elf.c:278:2: error: ‘elf_getshstrndx’ is deprecated [-Werror=deprecated-declarations]
| 278 | if (elf_getshdrstrndx(elf->elf, &shstrndx)) {
| | ^~
| In file included from /usr/include/gelf.h:32,
| from /home/mark/src/linux/tools/objtool/include/objtool/elf.h:10,
| from elf.c:22:
| /usr/include/libelf.h:322:12: note: declared here
| 322 | extern int elf_getshstrndx (Elf *__elf, size_t *__dst)
| | ^~~~~~~~~~~~~~~
| cc1: all warnings being treated as errors
| make[3]: *** [/home/mark/src/linux/tools/build/Makefile.build:97: /home/mark/src/linux/tools/objtool/elf.o] Error 1
| make[2]: *** [Makefile:66: /home/mark/src/linux/tools/objtool/objtool-in.o] Error 2
| make[1]: *** [Makefile:73: objtool] Error 2
| make: *** [Makefile:1439: tools/objtool] Error 2
Thanks,
Mark.
> +
> +# Always want host compilation.
> +HOST_OVERRIDES := CC="$(HOSTCC)" LD="$(HOSTLD)" AR="$(HOSTAR)"
>
> AWK = awk
> MKDIR = mkdir
> @@ -61,12 +61,14 @@ export BUILD_ORC
> export srctree OUTPUT CFLAGS SRCARCH AWK
> include $(srctree)/tools/build/Makefile.include
>
> -$(OBJTOOL_IN): fixdep FORCE
> +$(OBJTOOL_IN): fixdep $(LIBSUBCMD) FORCE
> $(Q)$(CONFIG_SHELL) ./sync-check.sh
> - $(Q)$(MAKE) $(build)=objtool
> + $(Q)$(MAKE) $(build)=objtool $(HOST_OVERRIDES) CFLAGS="$(OBJTOOL_CFLAGS)" \
> + LDFLAGS="$(OBJTOOL_LDFLAGS)"
> +
>
> $(OBJTOOL): $(LIBSUBCMD) $(OBJTOOL_IN)
> - $(QUIET_LINK)$(CC) $(OBJTOOL_IN) $(LDFLAGS) -o $@
> + $(QUIET_LINK)$(HOSTCC) $(OBJTOOL_IN) $(OBJTOOL_LDFLAGS) -o $@
>
>
> $(LIBSUBCMD_OUTPUT):
> @@ -75,6 +77,7 @@ $(LIBSUBCMD_OUTPUT):
> $(LIBSUBCMD): fixdep $(LIBSUBCMD_OUTPUT) FORCE
> $(Q)$(MAKE) -C $(LIBSUBCMD_DIR) O=$(LIBSUBCMD_OUTPUT) \
> DESTDIR=$(LIBSUBCMD_OUTPUT) prefix= subdir= \
> + $(HOST_OVERRIDES) EXTRA_CFLAGS="$(OBJTOOL_CFLAGS)" \
> $@ install_headers
>
> $(LIBSUBCMD)-clean:
next prev parent reply other threads:[~2023-02-01 17:02 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-01-26 19:06 [PATCH v4 0/3] objtool build improvements Ian Rogers
2023-01-26 19:06 ` [PATCH v4 1/3] objtool: Install libsubcmd in build Ian Rogers
2023-02-01 16:26 ` [tip: objtool/core] " tip-bot2 for Ian Rogers
2023-01-26 19:06 ` [PATCH v4 2/3] objtool: Properly support make V=1 Ian Rogers
2023-02-01 16:26 ` [tip: objtool/core] " tip-bot2 for Ian Rogers
2023-01-26 19:06 ` [PATCH v4 3/3] objtool: Alter how HOSTCC is forced Ian Rogers
2023-02-01 16:26 ` [tip: objtool/core] objtool: Fix HOSTCC flag usage tip-bot2 for Ian Rogers
2023-02-01 17:02 ` Mark Rutland [this message]
2023-02-01 17:36 ` Josh Poimboeuf
2023-02-03 18:25 ` Vladimir Oltean
2023-02-04 17:05 ` Josh Poimboeuf
2023-02-04 21:21 ` Ian Rogers
2023-02-04 10:17 ` tip-bot2 for Ian Rogers
2024-04-20 22:43 ` [PATCH v4 3/3] objtool: Alter how HOSTCC is forced Tan Nayir
2023-01-26 19:34 ` [PATCH v4 0/3] objtool build improvements Josh Poimboeuf
2023-01-31 0:25 ` Josh Poimboeuf
2023-01-31 0:31 ` Ian Rogers
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Y9qbGHDBFtGoqnKK@FVFF77S0Q05N \
--to=mark.rutland@arm.com \
--cc=irogers@google.com \
--cc=jpoimboe@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-tip-commits@vger.kernel.org \
--cc=x86@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).