Dave Chinner wrote:
> On Sat, Aug 13, 2016 at 07:04:18PM +0200, Felix Janda wrote:
> > This patch series does several things related to large file support.
> >
> > Patches 1-3 enable transparent LFS in the build system and make it
> > mandatory.
> >
> > Patches 4-9 and 12 replace *64 function and structure aliases.
> >
> > Patches 10 and 11 disable fsr on Mac OS X and do cleanup to enable
> > Patch 12. Further cleanup of the portability code is possible later.
> >
> > Patch 13 makes tranparent LFS also mandatory for all users of libxfs.
> >
> > Patch 14 finally replaces off64_t by off_t.
> >
> >
> > In comparison to v1:
> >
> > Patches 1, 3 and 14 are identical to previous patches. Patches 4-8 are
> > identical to previous patches, except that some of them are merged.
> > Patch 9 was previously send separately from the patch series. Patch
> > 13 is identical to a previous patch except for the commit message. The
> > other patches are new, grown out of review by Christoph Hellwig.
> >
> > Felix Janda (14):
> > configure: use AC_SYS_LARGEFILE
> > configure: error out when LFS does not work
> > remove unecessary definitions of _FILE_OFFSET_BITS
> > replace [fl]stat64 by equivalent [fl]stat
> > replace ftruncate64 by equivalent ftruncate
> > replace lseek64 by equivalent lseek
> > replace pread64/pwrite64 by equivalent pread/pwrite
> > replace sendfile64 by equivalent sendfile
> > fadvise.c: replace posix_fadvise64 by equivalent posix_fadvise
> > Makefile: disable fsr for Mac OS X
> > fsr: remove workaround for statvfs on Mac OS X
> > replace statvfs64 by equivalent statvfs
> > xfs.h: require transparent LFS for all users
> > platform: remove use of off64_t
>
> So, the patches are fine, and everything works. Problem is, it
> screws up xfstests because it changes all the error messages
> that are output to stderr and captured by the test harness.
> There are quite a few tests that this causes failures for,
> and because it's stderr, it's not as simple as just adding a new
> filter to do 'sed -e "s/^\(.*\)64\(: .*$\)/\1\2/"' on stderr.
Thanks for testing!
I can rework the patches to leave stderr unchanged. I guess that this
is preferable as opposed to updating the output expected by xfstests
since xfstests should be usable with both old and new xfsprogs.
> Further, errors returned change, which further screws up xfstests.
> e.g. old code gives EFBIG when we try to use offsets beyond the
> supported range, this patchset returns EINVAL. That further
> complicates any sort of error filtering we'll need to do.
I am very surprised that something apart from the error messages has
changed. I would be interested to know on what architecture and for
which test(s) (where) this happened, if you still remember.
> I don't have the time or patience to fix up xfstests for every
> change that people want to make and this series is a non-critical
> cleanup, so I'm dropping this until the fixups for xfstests are
> worked out. I'm not going to get to this for weeks at the current
> rate patches are being thrown at me for inclusion, so I'm not
> breaking xfstests for everyone while I'm bottlenecked on other,
> higher priority changes.
That make sense.
--Felix
|