On Thu, Feb 24, 2022 at 2:58 PM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
>
> Add x86-64 target for Clang+um and update user-offsets.c to use
> Clang-friendly assembly, similar to the fix from commit cf0c3e68aa81
> ("kbuild: fix asm-offset generation to work with clang").
>
> This lets me run KUnit tests with Clang:
>
> $ ./tools/testing/kunit/kunit.py config --make_options LLVM=1
> ...
> $ ./tools/testing/kunit/kunit.py run --make_options LLVM=1
> ...
>
> Cc: Jeff Dike <jdike@xxxxxxxxxxx>
> Cc: Richard Weinberger <richard@xxxxxx>
> Cc: Anton Ivanov <anton.ivanov@xxxxxxxxxxxxxxxxxx>
> Cc: Masahiro Yamada <masahiroy@xxxxxxxxxx>
> Cc: Nick Desaulniers <ndesaulniers@xxxxxxxxxx>
> Cc: Nathan Chancellor <nathan@xxxxxxxxxx>
> Cc: David Gow <davidgow@xxxxxxxxxx>
> Cc: linux-um@xxxxxxxxxxxxxxxxxxx
> Cc: linux-kbuild@xxxxxxxxxxxxxxx
> Cc: linux-kselftest@xxxxxxxxxxxxxxx
> Cc: kunit-dev@xxxxxxxxxxxxxxxx
> Cc: llvm@xxxxxxxxxxxxxxx
> Reviewed-by: Nathan Chancellor <nathan@xxxxxxxxxx>
> Link: https://lore.kernel.org/lkml/Yg2YubZxvYvx7%2Fnm@dev-arch.archlinux-ax161/
> Tested-by: David Gow <davidgow@xxxxxxxxxx>
> Link: https://lore.kernel.org/lkml/CABVgOSk=oFxsbSbQE-v65VwR2+mXeGXDDjzq8t7FShwjJ3+kUg@xxxxxxxxxxxxxx/
> Signed-off-by: Kees Cook <keescook@xxxxxxxxxxxx>
> ---
> v1: https://lore.kernel.org/lkml/20220217002843.2312603-1-keescook@xxxxxxxxxxxx/
> v2:
> - tweak commit log phrasing and alphabetize targets (nathan)
> - fix a missing implicit fallthrough under 32-bit builds
> - add review tags
> ---
> arch/um/os-Linux/execvp.c | 1 +
> arch/x86/um/user-offsets.c | 4 ++--
> scripts/Makefile.clang | 1 +
> 3 files changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/arch/um/os-Linux/execvp.c b/arch/um/os-Linux/execvp.c
> index 84a0777c2a45..c09a5fd5e225 100644
> --- a/arch/um/os-Linux/execvp.c
> +++ b/arch/um/os-Linux/execvp.c
> @@ -93,6 +93,7 @@ int execvp_noalloc(char *buf, const char *file, char *const argv[])
> up finding no executable we can use, we want to diagnose
> that we did find one but were denied access. */
> got_eacces = 1;
> + break;
> case ENOENT:
> case ESTALE:
> case ENOTDIR:
> diff --git a/arch/x86/um/user-offsets.c b/arch/x86/um/user-offsets.c
> index bae61554abcc..d9071827b515 100644
> --- a/arch/x86/um/user-offsets.c
> +++ b/arch/x86/um/user-offsets.c
> @@ -10,10 +10,10 @@
> #include <asm/types.h>
>
> #define DEFINE(sym, val) \
> - asm volatile("\n->" #sym " %0 " #val : : "i" (val))
> + asm volatile("\n.ascii \"->" #sym " %0 " #val "\"": : "i" (val))
Another way might be #include <linux/kbuild.h> and delete this macro
definition.
> #define DEFINE_LONGS(sym, val) \
> - asm volatile("\n->" #sym " %0 " #val : : "i" (val/sizeof(unsigned long)))
> + asm volatile("\n.ascii \"->" #sym " %0 " #val "\"": : "i" (val/sizeof(unsigned long)))
This generates wrong comments.
In include/generated/user_constants.h,
I see this:
#define HOST_BX 5 /* RBX */
(Here, the value of RBX is 40.)
#define DEFINE_LONGS(sym, val) asm volatile("\n.ascii \"->" #sym
" %0 " #val "/sizeof(unsigned long)\"": : "i" (val/sizeof(unsigned
long)))
creates valid comments:
#define HOST_BX 5 /* RBX/sizeof(unsigned long) */
Another way might be to do this indirectly.
#define DEFINE_LONGS(sym, val) DEFINE(sym, val / sizeof(unsigned long))
The comments in include/generated/user_constans.h do not retain
the original macro names, though...
> void foo(void)
> {
> diff --git a/scripts/Makefile.clang b/scripts/Makefile.clang
> index 51fc23e2e9e5..6e49344c6db2 100644
> --- a/scripts/Makefile.clang
> +++ b/scripts/Makefile.clang
> @@ -9,6 +9,7 @@ CLANG_TARGET_FLAGS_mips := mipsel-linux-gnu
> CLANG_TARGET_FLAGS_powerpc := powerpc64le-linux-gnu
> CLANG_TARGET_FLAGS_riscv := riscv64-linux-gnu
> CLANG_TARGET_FLAGS_s390 := s390x-linux-gnu
> +CLANG_TARGET_FLAGS_um := x86_64-linux-gnu
Personally, I like Nathan's idea, but we can live with the hard-coding
since we see no efforts for UML on other host arch.
> CLANG_TARGET_FLAGS_x86 := x86_64-linux-gnu
> CLANG_TARGET_FLAGS := $(CLANG_TARGET_FLAGS_$(SRCARCH))
>
> --
> 2.30.2
>
--
Best Regards
Masahiro Yamada
