public inbox for llvm@lists.linux.dev
 help / color / mirror / Atom feed
From: Masahiro Yamada <masahiroy@kernel.org>
To: linux-kbuild@vger.kernel.org
Cc: linux-kernel@vger.kernel.org,
	Masahiro Yamada <masahiroy@kernel.org>,
	Nicolas Schier <nicolas@fjasle.eu>,
	Albert Ou <aou@eecs.berkeley.edu>, Borislav Petkov <bp@alien8.de>,
	Dave Hansen <dave.hansen@linux.intel.com>,
	"H. Peter Anvin" <hpa@zytor.com>, Ingo Molnar <mingo@redhat.com>,
	Nathan Chancellor <nathan@kernel.org>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Palmer Dabbelt <palmer@dabbelt.com>,
	Paul Walmsley <paul.walmsley@sifive.com>,
	Thomas Gleixner <tglx@linutronix.de>, Tom Rix <trix@redhat.com>,
	linux-riscv@lists.infradead.org, llvm@lists.linux.dev,
	x86@kernel.org
Subject: [PATCH v3 1/5] kbuild: add test-{le,ge,lt,gt} macros
Date: Sun, 27 Nov 2022 07:56:20 +0900	[thread overview]
Message-ID: <20221126225624.751661-1-masahiroy@kernel.org> (raw)

Because GNU Make is handle strings, it is very hard to perform math in
Makefiles.

When we compare two integers, we invokes shell. One example is in the
top Makefile:

  ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 130000; echo $$?),0)

This is more expensive than using built-in functions since it forks a
process.

If we know the two have the same number of digits, we can do better.

This commit adds four macros, test-le, test-ge, test-lt, test-gt.

$(call test-lt, A, B) is evaluated to 'y' if A is less than B, or
empty otherwise. This will replace $(call shell test A -lt B).

Again, the limitation is that A and B must have the same number of
digits because these macros are based on $(sort ) function.

  $(call test-lt, 1, 9)    -->  y        (Works!)
  $(call test-lt, 10, 9)   -->  y        (Not work...)

To make the latter work, you need to add '0' prefix to align the number
of digits:

  $(call test-lt, 10, 09)  -->  empty    (Works!)

Actually, we can live with this limitation in many places. As for the
example above, we know $(CONFIG_LLD_VERSION) is 6-digits because the
minimal supported version of LLVM is 11.0.0 (or CONFIG_LLD_VERSION=0
if the linker is not LLD, this case also works.)

So, the shell invocation can be replaced with more efficient code:

  ifeq ($(call test-lt, $(CONFIG_LLD_VERSION), 130000),y)

Of course, this assumption will break when LLVM 100 is released, but it
will be far in the future.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
---

Changes in v3:
  - Ensure the given parameters are not empty

 Makefile               |  2 +-
 arch/riscv/Makefile    |  2 +-
 arch/x86/Makefile      |  2 +-
 scripts/Kbuild.include | 10 ++++++++++
 4 files changed, 13 insertions(+), 3 deletions(-)

diff --git a/Makefile b/Makefile
index 6f846b1f2618..eb80332f7b51 100644
--- a/Makefile
+++ b/Makefile
@@ -986,7 +986,7 @@ KBUILD_LDFLAGS += -mllvm -import-instr-limit=5
 # Check for frame size exceeding threshold during prolog/epilog insertion
 # when using lld < 13.0.0.
 ifneq ($(CONFIG_FRAME_WARN),0)
-ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 130000; echo $$?),0)
+ifeq ($(call test-lt, $(CONFIG_LLD_VERSION), 130000),y)
 KBUILD_LDFLAGS	+= -plugin-opt=-warn-stack-size=$(CONFIG_FRAME_WARN)
 endif
 endif
diff --git a/arch/riscv/Makefile b/arch/riscv/Makefile
index 0d13b597cb55..faf2c2177094 100644
--- a/arch/riscv/Makefile
+++ b/arch/riscv/Makefile
@@ -37,7 +37,7 @@ else
 endif
 
 ifeq ($(CONFIG_LD_IS_LLD),y)
-ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 150000; echo $$?),0)
+ifeq ($(call test-lt, $(CONFIG_LLD_VERSION), 150000),y)
 	KBUILD_CFLAGS += -mno-relax
 	KBUILD_AFLAGS += -mno-relax
 ifndef CONFIG_AS_IS_LLVM
diff --git a/arch/x86/Makefile b/arch/x86/Makefile
index 415a5d138de4..e72c7a49cd59 100644
--- a/arch/x86/Makefile
+++ b/arch/x86/Makefile
@@ -211,7 +211,7 @@ endif
 KBUILD_LDFLAGS += -m elf_$(UTS_MACHINE)
 
 ifdef CONFIG_LTO_CLANG
-ifeq ($(shell test $(CONFIG_LLD_VERSION) -lt 130000; echo $$?),0)
+ifeq ($(call test-lt, $(CONFIG_LLD_VERSION), 130000),y)
 KBUILD_LDFLAGS	+= -plugin-opt=-stack-alignment=$(if $(CONFIG_X86_32),4,8)
 endif
 endif
diff --git a/scripts/Kbuild.include b/scripts/Kbuild.include
index cbe28744637b..4b8cf464b53b 100644
--- a/scripts/Kbuild.include
+++ b/scripts/Kbuild.include
@@ -11,6 +11,16 @@ space   := $(empty) $(empty)
 space_escape := _-_SPACE_-_
 pound := \#
 
+###
+# Comparison macros.
+# Usage: $(call test-le, A, B)
+# works like shell's "test A -le B", use with care bacause it is ASCII sort.
+# If A and B have the same number of digits, it works as expected.
+test-le = $(and $(strip $1),$(strip $2),$(filter $1, $(firstword $(sort $1 $2))),y)
+test-ge = $(call test-le, $2, $1)
+test-lt = $(and $(strip $1),$(strip $2),$(filter-out $2, $(firstword $(sort $1 $2))),y)
+test-gt = $(call test-lt, $2, $1)
+
 ###
 # Name of target with a '.' as filename prefix. foo/bar.o => foo/.bar.o
 dot-target = $(dir $@).$(notdir $@)
-- 
2.34.1


             reply	other threads:[~2022-11-26 22:56 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-26 22:56 Masahiro Yamada [this message]
2022-11-26 22:56 ` [PATCH v3 2/5] kbuild: implement {gcc,clang}-min-version only with built-in functions 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=20221126225624.751661-1-masahiroy@kernel.org \
    --to=masahiroy@kernel.org \
    --cc=aou@eecs.berkeley.edu \
    --cc=bp@alien8.de \
    --cc=dave.hansen@linux.intel.com \
    --cc=hpa@zytor.com \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-riscv@lists.infradead.org \
    --cc=llvm@lists.linux.dev \
    --cc=mingo@redhat.com \
    --cc=nathan@kernel.org \
    --cc=ndesaulniers@google.com \
    --cc=nicolas@fjasle.eu \
    --cc=palmer@dabbelt.com \
    --cc=paul.walmsley@sifive.com \
    --cc=tglx@linutronix.de \
    --cc=trix@redhat.com \
    --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