On Wed, Aug 28, 2024 at 02:15:30PM -0400, Brian Foster wrote: > Hi all, > > Here's v2 of the patches for improved test coverage (via fsx) for EOF > related zeroing. The most notable change is that the discussion on v1 > uncovered that some of the existing fsx behavior was wrong wrt to > simulated ops, so patch 1 is factored out as a standalone bug fix to > address that. > > Brian > > v2: > - Factored out patch 1 to fix simulation mode. > - Use MAP_FAILED, don't inject data for simulated ops. > - Rebase to latest master and renumber test. > - Use run_fsx and -S 0 by default (timestamp seed). > v1: https://lore.kernel.org/fstests/20240822144422.188462-1-bfoster@xxxxxxxxxx/ > > Brian Foster (4): > fsx: don't skip file size and buf updates on simulated ops > fsx: factor out a file size update helper > fsx: support eof page pollution for eof zeroing test coverage > generic: test to run fsx eof pollution Thanks for updating the fsx.c, fsstress and fsx are important test program in fstests, it's always warm welcome improving them :) I'll merge this patchset after a basic regression test. Thanks, Zorro > > ltp/fsx.c | 134 ++++++++++++++++++++++++++++++++---------- > tests/generic/363 | 23 ++++++++ > tests/generic/363.out | 2 + > 3 files changed, 127 insertions(+), 32 deletions(-) > create mode 100755 tests/generic/363 > create mode 100644 tests/generic/363.out > > -- > 2.45.0 > >
