Hi,
Andreas had some suggestions for v4, so here is v5.
Bob
---
gfs2: clean_journal improperly set sd_log_flush_head
This patch fixes regressions in 588bff95c94efc05f9e1a0b19015c9408ed7c0ef.
Due to that patch, function clean_journal was setting the value of
sd_log_flush_head, but that's only
Hi,
On 28 March, I posted a v3 patch for this problem, but that had problems too,
which Andreas found using xfstests/034 test.
This v4 version fixes those problems and, as Steve Whitehouse suggested,
uses the faster iomap interface to map blocks from other journals, rather
than the now obsolete