From: Kees Cook <keescook@chromium.org>
To: Masahiro Yamada <masahiroy@kernel.org>
Cc: linux-kbuild@vger.kernel.org,
Sami Tolvanen <samitolvanen@google.com>,
linux-kernel@vger.kernel.org,
Michal Marek <michal.lkml@markovi.net>,
Nick Desaulniers <ndesaulniers@google.com>
Subject: Re: [PATCH 03/13] kbuild: detect objtool changes correctly and remove .SECONDEXPANSION
Date: Wed, 18 Aug 2021 19:55:37 -0700 [thread overview]
Message-ID: <202108181952.14AEEED@keescook> (raw)
In-Reply-To: <20210819005744.644908-4-masahiroy@kernel.org>
On Thu, Aug 19, 2021 at 09:57:34AM +0900, Masahiro Yamada wrote:
> This reverts commit 8852c5524029 ("kbuild: Fix objtool dependency for
> 'OBJECT_FILES_NON_STANDARD_<obj> := n'"), and fix the dependency in a
> cleaner way.
>
> Using .SECONDEXPANSION is expensive since Makefile.build is parsed
> twice every time, and the escaping dollars makes the code unreadable.
>
> Adding include/config/* as dependency is not maintainable either because
> objtool_args is dependent on more CONFIG options.
>
> A better fix is to include the objtool command in *.cmd files so any
> command change is naturally detected by if_change.
>
> Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
> ---
>
> scripts/Makefile.build | 28 ++++++++++------------------
> 1 file changed, 10 insertions(+), 18 deletions(-)
>
> diff --git a/scripts/Makefile.build b/scripts/Makefile.build
> index 31154e44c251..3e4cd1439cd4 100644
> --- a/scripts/Makefile.build
> +++ b/scripts/Makefile.build
> @@ -155,7 +155,7 @@ $(obj)/%.ll: $(src)/%.c FORCE
> # (See cmd_cc_o_c + relevant part of rule_cc_o_c)
>
> quiet_cmd_cc_o_c = CC $(quiet_modtag) $@
> - cmd_cc_o_c = $(CC) $(c_flags) -c -o $@ $<
> + cmd_cc_o_c = $(CC) $(c_flags) -c -o $@ $< $(cmd_objtool)
>
> ifdef CONFIG_MODVERSIONS
> # When module versioning is enabled the following steps are executed:
> @@ -223,6 +223,8 @@ endif # CONFIG_FTRACE_MCOUNT_USE_RECORDMCOUNT
>
> ifdef CONFIG_STACK_VALIDATION
>
> +objtool := $(objtree)/tools/objtool/objtool
> +
> # Objtool arguments are also needed for modfinal with LTO, so we define
> # then here to avoid duplication.
> objtool_args = \
> @@ -237,26 +239,19 @@ objtool_args = \
>
> ifndef CONFIG_LTO_CLANG
>
> -__objtool_obj := $(objtree)/tools/objtool/objtool
> -
> # 'OBJECT_FILES_NON_STANDARD := y': skip objtool checking for a directory
> # 'OBJECT_FILES_NON_STANDARD_foo.o := 'y': skip objtool checking for a file
> # 'OBJECT_FILES_NON_STANDARD_foo.o := 'n': override directory skip for a file
> cmd_objtool = $(if $(patsubst y%,, \
> $(OBJECT_FILES_NON_STANDARD_$(basetarget).o)$(OBJECT_FILES_NON_STANDARD)n), \
> - $(__objtool_obj) $(objtool_args) $@)
> -objtool_obj = $(if $(patsubst y%,, \
> - $(OBJECT_FILES_NON_STANDARD_$(basetarget).o)$(OBJECT_FILES_NON_STANDARD)n), \
> - $(__objtool_obj))
> + ; $(objtool) $(objtool_args) $@)
This is extremely clever -- pasting commands together. :)
Does this correctly propagate failures in the first half of the command?
For example, now we'd have:
gcc flags..... -c -o out in ; objtool...
But I think objtool will run even on gcc failure, and "make" won't see
the failure from gcc? I need to go test this.
-Kees
> +
> +# Rebuild all objects when objtool is updated
> +objtool_dep = $(objtool)
>
> endif # CONFIG_LTO_CLANG
> endif # CONFIG_STACK_VALIDATION
>
> -# Rebuild all objects when objtool changes, or is enabled/disabled.
> -objtool_dep = $(objtool_obj) \
> - $(wildcard include/config/ORC_UNWINDER \
> - include/config/STACK_VALIDATION)
> -
> ifdef CONFIG_TRIM_UNUSED_KSYMS
> cmd_gen_ksymdeps = \
> $(CONFIG_SHELL) $(srctree)/scripts/gen_ksymdeps.sh $@ >> $(dot-target).cmd
> @@ -270,7 +265,6 @@ define rule_cc_o_c
> $(call cmd,gen_ksymdeps)
> $(call cmd,checksrc)
> $(call cmd,checkdoc)
> - $(call cmd,objtool)
> $(call cmd,modversions_c)
> $(call cmd,record_mcount)
> endef
> @@ -278,13 +272,11 @@ endef
> define rule_as_o_S
> $(call cmd_and_fixdep,as_o_S)
> $(call cmd,gen_ksymdeps)
> - $(call cmd,objtool)
> $(call cmd,modversions_S)
> endef
>
> # Built-in and composite module parts
> -.SECONDEXPANSION:
> -$(obj)/%.o: $(src)/%.c $(recordmcount_source) $$(objtool_dep) FORCE
> +$(obj)/%.o: $(src)/%.c $(recordmcount_source) $(objtool_dep) FORCE
> $(call if_changed_rule,cc_o_c)
> $(call cmd,force_checksrc)
>
> @@ -367,7 +359,7 @@ $(obj)/%.s: $(src)/%.S FORCE
> $(call if_changed_dep,cpp_s_S)
>
> quiet_cmd_as_o_S = AS $(quiet_modtag) $@
> - cmd_as_o_S = $(CC) $(a_flags) -c -o $@ $<
> + cmd_as_o_S = $(CC) $(a_flags) -c -o $@ $< $(cmd_objtool)
>
> ifdef CONFIG_ASM_MODVERSIONS
>
> @@ -386,7 +378,7 @@ cmd_modversions_S = \
> fi
> endif
>
> -$(obj)/%.o: $(src)/%.S $$(objtool_dep) FORCE
> +$(obj)/%.o: $(src)/%.S $(objtool_dep) FORCE
> $(call if_changed_rule,as_o_S)
>
> targets += $(filter-out $(subdir-builtin), $(real-obj-y))
> --
> 2.30.2
>
--
Kees Cook
next prev parent reply other threads:[~2021-08-19 2:55 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-08-19 0:57 [PATCH 00/13] kbuild: refactoring after Clang LTO Masahiro Yamada
2021-08-19 0:57 ` [PATCH 01/13] kbuild: move objtool_args back to scripts/Makefile.build Masahiro Yamada
2021-08-19 2:45 ` Kees Cook
2021-08-19 0:57 ` [PATCH 02/13] gen_compile_commands: extract compiler command from a series of commands Masahiro Yamada
2021-08-19 6:48 ` Kees Cook
2021-08-19 0:57 ` [PATCH 03/13] kbuild: detect objtool changes correctly and remove .SECONDEXPANSION Masahiro Yamada
2021-08-19 2:55 ` Kees Cook [this message]
2021-08-19 3:18 ` Masahiro Yamada
2021-08-19 0:57 ` [PATCH 04/13] kbuild: remove unused quiet_cmd_update_lto_symversions Masahiro Yamada
2021-08-19 6:19 ` Kees Cook
2021-08-19 0:57 ` [PATCH 05/13] kbuild: remove stale *.symversions Masahiro Yamada
2021-08-19 6:29 ` Kees Cook
2021-08-19 0:57 ` [PATCH 06/13] kbuild: merge vmlinux_link() between the ordinary link and Clang LTO Masahiro Yamada
2021-08-19 2:42 ` Kees Cook
2021-08-19 0:57 ` [PATCH 07/13] kbuild: do not remove 'linux' link in scripts/link-vmlinux.sh Masahiro Yamada
2021-08-19 6:32 ` Kees Cook
2021-08-19 0:57 ` [PATCH 08/13] kbuild: merge vmlinux_link() between ARCH=um and other architectures Masahiro Yamada
2021-08-19 2:43 ` Kees Cook
2021-08-19 0:57 ` [PATCH 09/13] kbuild: do not create built-in.a.symversions or lib.a.symversions Masahiro Yamada
2021-08-19 6:41 ` Kees Cook
2021-08-28 11:57 ` Masahiro Yamada
2021-08-19 0:57 ` [PATCH 10/13] kbuild: build modules in the same way with/without Clang LTO Masahiro Yamada
2021-08-19 6:59 ` Kees Cook
2021-08-19 8:11 ` Masahiro Yamada
2021-08-19 0:57 ` [PATCH 11/13] kbuild: always postpone CRC links for module versioning Masahiro Yamada
2021-08-19 6:43 ` Kees Cook
2021-08-19 0:57 ` [PATCH 12/13] kbuild: merge cmd_modversions_c and cmd_modversions_S Masahiro Yamada
2021-08-19 3:06 ` Kees Cook
2021-08-19 0:57 ` [PATCH 13/13] kbuild: merge cmd_ar_builtin and cmd_ar_module Masahiro Yamada
2021-08-19 3:01 ` Kees Cook
2021-08-25 4:57 ` [PATCH 00/13] kbuild: refactoring after Clang LTO Masahiro Yamada
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=202108181952.14AEEED@keescook \
--to=keescook@chromium.org \
--cc=linux-kbuild@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=masahiroy@kernel.org \
--cc=michal.lkml@markovi.net \
--cc=ndesaulniers@google.com \
--cc=samitolvanen@google.com \
/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).