tree a02ce92211f2dd75fbe0e92b9a1b751c3595d785
parent 1a48049adb98a20fad05ed86bc00c2f9120a006e
author Andreas Gruenbacher <agruenba@redhat.com> 1567777898 +0100
committer Andreas Gruenbacher <agruenba@redhat.com> 1573156864 +0100

gfs2: Improve mmap write vs. punch_hole consistency

When punching a hole in a file, use filemap_write_and_wait_range to
write back any dirty pages in the range of the hole.  As a side effect,
if the hole isn't page aligned, this marks unaligned pages at the
beginning and the end of the hole read-only.  This is required when the
block size is smaller than the page size: when those pages are written
to again after the hole punching, we must make sure that page_mkwrite is
called for those pages so that the page will be fully allocated and any
blocks turned into holes from the hole punching will be reallocated.
(If a page is writably mapped, page_mkwrite won't be called.)

Fixes xfstest generic/567.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>