The Makefile open-codes compiler invocations that ../lib.mk already
provides.

Avoid this by using a Make feature that allows setting per-target
variables, which in this case are: CFLAGS and LDFLAGS. This approach
generates the exact same compiler invocations as before, but removes all
of the code duplication, along with the quirky mangled variable names.
So now the Makefile is smaller, less unusual, and easier to read.

The new dependencies are listed after including lib.mk, in order to
let lib.mk provide the first target ("all:"), and are grouped together
with their respective source file dependencies, for visual clarity.

Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

There were a couple of errors here:

1. TEST_GEN_PROGS was incorrectly prepending $(OUTPUT) to each program
to be built. However, lib.mk already does that because it assumes "bare"
program names are passed in, so this ended up creating
$(OUTPUT)/$(OUTPUT)/file.c, which of course won't work as intended.

2. lib.mk was included before TEST_GEN_PROGS was set, which led to
lib.mk's "all:" target not seeing anything to rebuild.

So nothing worked, which caused the author to force things by creating
an "all:" target locally--while still including ../lib.mk.

Fix all of this by including ../lib.mk at the right place, and removing
the $(OUTPUT) prefix to the programs to be built, and removing the
duplicate "all:" target.

Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

When building with clang, via:

    make LLVM=1 -C tools/testing/selftests

...there are several warnings, and an error. This fixes all of those and
allows these tests to run and pass.

1. Fix linker error (undefined reference to memcpy) by providing a local
   version of memcpy.

2. clang complains about using this form:

    if (g = h & 0xf0000000)

...so factor out the assignment into a separate step.

3. The code is passing a signed const char* to elf_hash(), which expects
   a const unsigned char *. There are several callers, so fix this at
   the source by allowing the function to accept a signed argument, and
   then converting to unsigned operations, once inside the function.

4. clang doesn't have __attribute__((externally_visible)) and generates
   a warning to that effect. Fortunately, gcc 12 and gcc 13 do not seem
   to require that attribute in order to build, run and pass tests here,
   so remove it.

Reviewed-by: Carlos Llamas <cmllamas@google.com>
Reviewed-by: Edward Liaw <edliaw@google.com>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

It could not find __vdso_getcpu and __vdso_gettimeofday when test getcpu
and gettimeofday on LoongArch.

  # make headers && cd tools/testing/selftests/vDSO && make
  # ./vdso_test_getcpu
  Could not find __vdso_getcpu
  # ./vdso_test_gettimeofday
  Could not find __vdso_gettimeofday

One simple way is to add LoongArch case to define version and name, just
like commit d942f231afc0 ("selftests/vDSO: Add riscv getcpu & gettimeofday
test"), but it is not the best way.

Since each architecture has already defined names and versions in
vdso_config.h, it is proper to include vdso_config.h to get version and
name for all archs.

Link: https://lkml.kernel.org/r/20240428030530.24399-3-yangtiezhu@loongson.cn
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Tested-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: Shuah Khan <skhan@linuxfoundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

Patch series "selftests/vDSO: Fix errors on LoongArch", v4.


This patch (of 2):

There exist the following errors when build vDSO selftests on LoongArch:

  # make headers && cd tools/testing/selftests/vDSO && make
  ...
  error: 'VDSO_VERSION' undeclared (first use in this function)
  ...
  error: 'VDSO_NAMES' undeclared (first use in this function)

We can see the following code in arch/loongarch/vdso/vdso.lds.S:

VERSION
{
        LINUX_5.10 {
        global:
                __vdso_getcpu;
                __vdso_clock_getres;
                __vdso_clock_gettime;
                __vdso_gettimeofday;
                __vdso_rt_sigreturn;
        local: *;
        };
}

so VDSO_VERSION should be 6 and VDSO_NAMES should be 1 for LoongArch,
add them to fix the building errors on LoongArch.

Link: https://lkml.kernel.org/r/20240428030530.24399-1-yangtiezhu@loongson.cn
Link: https://lkml.kernel.org/r/20240428030530.24399-2-yangtiezhu@loongson.cn
Signed-off-by: Tiezhu Yang <yangtiezhu@loongson.cn>
Reviewed-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Kees Cook <keescook@chromium.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: Shuah Khan <skhan@linuxfoundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

There are a couple of raw printf() calls in vdso_test_abi which result in
non KTAP conforment output such as

   [vDSO kselftest] VDSO_VERSION: LINUX_2.6

Convert them to use ksft_print_msg() so that they don't cause confusion for
parsers.

Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

When logging the ID of the currently tested clock vdso_test_clock() puts a
spurious newline at the start of the format string resulting in output
such as

  #
  clock_id: CLOCK_BOOTTIME

which is a valid but empty KTAP informational message followed by a non
conferment output line. Remove the initial newline to create a more KTAP
friendly

  # clock_id: CLOCK_BOOTTIME

Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

The test results from vdso_abi_test are all formatted using a series of
macros:

  #define VDSO_TEST_PASS_MSG()	"\n%s(): PASS\n", __func__
  #define VDSO_TEST_FAIL_MSG(x)	"\n%s(): %s FAIL\n", __func__, x
  #define VDSO_TEST_SKIP_MSG(x)	"\n%s(): SKIP: Could not find %s\n", __func__, x

which don't play nicely with automated KTAP parsers since the actual KTAP
lines are in the form

   ok 1

with no test name and we get an additional log line such as

  vdso_test_gettimeofday(): PASS

with no preceeding # as KTAP requires. The lack of a test name means that
many automation systems will have a hard time distinguishing between the
different tests or correlating results between runs, the lack of # is less
severe but could potentially cause confusion.

Fix these issues by rewriting all the result reporting to include both the
vDSO function name being tested and (where there is one) the name of the
clock being tested in the main KTAP line. Since we have tests both with and
without a specific clock we abandon the helper macros and just put the
format strings used directly in the ksft_ API calls. When we fail to look
up the relevant vDSO symbol we add a separate print statement explaining
why the skip is being done. This gives output such as:

  ok 1 __vdso_gettimeofday
  # clock_id: CLOCK_REALTIME
  # The time is 1700673118.58091596
  ok 2 __vdso_clock_gettime CLOCK_REALTIME

which is much easier for test automation to work with.

Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

In the unlikely case that CLOCK_REALTIME is not defined, variable ret is
not initialized and further accumulation of return values to ret can leave
ret in an undefined state. Fix this by initialized ret to zero and changing
the assignment of ret to an accumulation for the CLOCK_REALTIME case.

Fixes: 03f55c7952c9 ("kselftest: Extend vDSO selftest to clock_getres")
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>

Enable vDSO getcpu & gettimeofday test for riscv. But only riscv64
supports __vdso_gettimeofday and riscv32 is under development.

VERSION
{
        LINUX_4.15 {
        global:
                __vdso_rt_sigreturn;
                __vdso_gettimeofday;
                __vdso_clock_gettime;
                __vdso_clock_getres;
                __vdso_getcpu;
                __vdso_flush_icache;
        local: *;
        };
}

Co-developed-by: haocheng.zy <haocheng.zy@linux.alibaba.com>
Signed-off-by: haocheng.zy <haocheng.zy@linux.alibaba.com>
Suggested-by: Mao Han <han_mao@linux.alibaba.com>
Reviewed-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Guo Ren <guoren@linux.alibaba.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Elliott Hughes <enh@google.com>
Signed-off-by: Shuah Khan <skhan@linuxfoundation.org>