linux-kernel - Re: [PATCH 5/7] selftests/x86: fix cross build logic

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <552DB1C5.7020704@amacapital.net>
Date:	Tue, 14 Apr 2015 17:33:09 -0700
From:	Andy Lutomirski <luto@...capital.net>
To:	Tyler Baker <tyler.baker@...aro.org>,
	Shuah Khan <shuahkh@....samsung.com>
CC:	Kevin Hilman <khilman@...nel.org>,
	John Stultz <john.stultz@...aro.org>,
	Darren Hart <dvhart@...radead.org>,
	David Herrmann <dh.herrmann@...il.com>,
	Michael Ellerman <mpe@...erman.id.au>,
	linux-kernel@...r.kernel.org, linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH 5/7] selftests/x86: fix cross build logic

On 04/14/2015 03:52 PM, Tyler Baker wrote:
> x86 tests should not be built when ARCH != x86. Reused the logic from
> breakpoints to determine when it's appropriate to build.

In the future, please cc the author of recently-written code that you're 
fixing :)

This patch is really weird.  You're generating ARCH, then you're 
modifying it part way through, and you're changing the rule for all_32 
and all_64 to print an error (and succeed!) if run on the wrong arch.

I don't see the point of the latter at all.  Just let all have no 
dependencies if you're on the wrong ARCH.

As for the former, can you do this more cleanly?  This is really ugly. 
It looks to me (on naive reading) like 64-bit hosts won't do the right 
thing.  (The right thing is to build both bitnesses of tests.)

--Andy

>
> Signed-off-by: Tyler Baker <tyler.baker@...aro.org>
> ---
>   tools/testing/selftests/x86/Makefile | 35 +++++++++++++++++++++++------------
>   1 file changed, 23 insertions(+), 12 deletions(-)
>
> diff --git a/tools/testing/selftests/x86/Makefile b/tools/testing/selftests/x86/Makefile
> index f0a7918..7be67a0 100644
> --- a/tools/testing/selftests/x86/Makefile
> +++ b/tools/testing/selftests/x86/Makefile
> @@ -7,31 +7,36 @@ BINARIES_64 := $(TARGETS_C_BOTHBITS:%=%_64)
>
>   CFLAGS := -O2 -g -std=gnu99 -pthread -Wall
>
> -UNAME_P := $(shell uname -p)
> -
> -# Always build 32-bit tests
> +# Taken from perf makefile
> +uname_M := $(shell uname -m 2>/dev/null || echo not)
> +ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/)
> +ifeq ($(ARCH),i386)
> +ARCH := x86
>   all: all_32
> -
> +endif
> +ifeq ($(ARCH),x86_64)
> +ARCH := x86
>   # If we're on a 64-bit host, build 64-bit tests as well
> -ifeq ($(shell uname -p),x86_64)
> -all: all_64
> +all: all_32 all_64
>   endif
>
>   all_32: check_build32 $(BINARIES_32)
>
>   all_64: $(BINARIES_64)
>
> -clean:
> -	$(RM) $(BINARIES_32) $(BINARIES_64)
> -
> -run_tests:
> -	./run_x86_tests.sh
> -
>   $(TARGETS_C_BOTHBITS:%=%_32): %_32: %.c
> +ifeq ($(ARCH),x86)
>   	$(CC) -m32 -o $@ $(CFLAGS) $(EXTRA_CFLAGS) $^ -lrt -ldl
> +else
> +	echo "Not an x86 target, can't build x86 tests"
> +endif
>
>   $(TARGETS_C_BOTHBITS:%=%_64): %_64: %.c
> +ifeq ($(ARCH),x86)
>   	$(CC) -m64 -o $@ $(CFLAGS) $(EXTRA_CFLAGS) $^ -lrt -ldl
> +else
> +	echo "Not an x86 target, can't build x86 tests"
> +endif
>
>   check_build32:
>   	@if ! $(CC) -m32 -o /dev/null trivial_32bit_program.c; then	\
> @@ -46,3 +51,9 @@ check_build32:
>   	  echo "  yum install glibc-devel.*i686";			\
>   	  exit 1;							\
>   	fi
> +
> +run_tests:
> +	./run_x86_tests.sh
> +
> +clean:
> +	$(RM) $(BINARIES_32) $(BINARIES_64)
>

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/