Re: [PATCH v6 5/5] selftests/mm: add UFFDIO_MOVE ioctl test

[Suren Baghdasaryan]

On Sun, Dec 10, 2023 at 5:01 PM Suren Baghdasaryan <surenb@xxxxxxxxxx> wrote:
>
> On Sun, Dec 10, 2023 at 6:26 AM Mark Brown <broonie@xxxxxxxxxx> wrote:
> >
> > On Wed, Dec 06, 2023 at 02:36:59AM -0800, Suren Baghdasaryan wrote:
> > > Add tests for new UFFDIO_MOVE ioctl which uses uffd to move source
> > > into destination buffer while checking the contents of both after
> > > the move. After the operation the content of the destination buffer
> > > should match the original source buffer's content while the source
> > > buffer should be zeroed. Separate tests are designed for PMD aligned and
> > > unaligned cases because they utilize different code paths in the kernel.
> > >
> > > Signed-off-by: Suren Baghdasaryan <surenb@xxxxxxxxxx>
> > > ---
> > >  tools/testing/selftests/mm/uffd-common.c     |  24 +++
> > >  tools/testing/selftests/mm/uffd-common.h     |   1 +
> > >  tools/testing/selftests/mm/uffd-unit-tests.c | 189 +++++++++++++++++++
> > >  3 files changed, 214 insertions(+)
> >
> > This breaks the build in at least some configurations with separate
> > output directories like those used by KernelCI:
> >
> > make KBUILD_BUILD_USER=KernelCI FORMAT=.xz ARCH=arm64 HOSTCC=gcc CROSS_COMPILE=aarch64-linux-gnu- CROSS_COMPILE_COMPAT=arm-linux-gnueabihf- CC="ccache aarch64-linux-gnu-gcc" O=/tmp/kci/linux/build -C/tmp/kci/linux -j10 kselftest-gen_tar
> >
> > (full logs for both arm64 and x86_64 at):
> >
> >    https://storage.kernelci.org/next/master/next-20231208/arm64/defconfig/gcc-10/logs/kselftest.log
> >    https://storage.kernelci.org/next/master/next-20231208/x86_64/x86_64_defconfig/clang-17/logs/kselftest.log
> >
> > or tuxmake:
> >
> > make --silent --keep-going --jobs=16 O=/home/broonie/.cache/tuxmake/builds/25/build INSTALL_PATH=/home/broonie/.cache/tuxmake/builds/25/build/kselftest_install ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu- CROSS_COMPILE_COMPAT=arm-linux-gnueabihf- kselftest-install
> >
> > The specific failure:
> >
> > aarch64-linux-gnu-gcc -Wall -I /tmp/kci/linux/tools/testing/selftests/../../..  -isystem /tmp/kci/linux/build/usr/include     uffd-stress.c vm_util.c uffd-common.c -lrt -lpthread -lm -o /tmp/kci/linux/build/kselftest/mm/uffd-stress
> > uffd-common.c: In function ‘move_page’:
> > uffd-common.c:636:21: error: storage size of ‘uffdio_move’ isn’t known
> >   636 |  struct uffdio_move uffdio_move;
> >       |                     ^~~~~~~~~~~
> > uffd-common.c:643:21: error: ‘UFFDIO_MOVE_MODE_ALLOW_SRC_HOLES’ undeclared (first use in this function)
> >   643 |  uffdio_move.mode = UFFDIO_MOVE_MODE_ALLOW_SRC_HOLES;
> >       |                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > uffd-common.c:643:21: note: each undeclared identifier is reported only once for each function it appears in
> > uffd-common.c:645:17: error: ‘UFFDIO_MOVE’ undeclared (first use in this function); did you mean ‘UFFDIO_COPY’?
> >   645 |  if (ioctl(ufd, UFFDIO_MOVE, &uffdio_move)) {
> >       |                 ^~~~~~~~~~~
> >       |                 UFFDIO_COPY
> > uffd-common.c:636:21: warning: unused variable ‘uffdio_move’ [-Wunused-variable]
> >   636 |  struct uffdio_move uffdio_move;
> >       |                     ^~~~~~~~~~~
>
> Thanks for reporting! I'll try that later today.
> Just to clarify, are you using mm-unstable and if so, has it been
> rebased since Friday? There was an update to this patchset in
> mm-unstable which Andrew merged on Friday and the failure does look
> like something that would happen with the previous version.

I tried reproducing the issue but so far unsuccessfully. Could you
please confirm that on the latest mm-unstable branch it's still
reproducible and if so, please provide detailed instructions on how
you reproduce it.
Thanks,
Suren.