Currently, if we do get_node_of_data before f2fs_lock_op, there may be dead
lock
as follows, where process A would be in infinite loop, and B will NOT be awaked.
Process A(cp): Process B:
f2fs_lock_all(sbi)
get_dnode_of_data <---- lock dn.node_page
flush_nodes f2fs_lock_op
However, in do_write_data_page path, we had better do get_node_of_data first,
since
we can check IPU in prior and do IPU without f2fs_lock_op. It make senses for
our
current FSYNC IPU to avoid being blocked by cp.
What's more, in patch 3), we do IPU if sync io is blocked by cp.
Hou Pengyang (3):
f2fs: do get_dnode_of_data before f2fs_lock_op
f2fs: check IPU after reading dnode from disk
f2fs: try IPU if sync io is blocked by cp
fs/f2fs/data.c | 77 ++++++++++++++++++++++++++++++++++++++++++++++++----------
fs/f2fs/f2fs.h | 6 +++++
fs/f2fs/node.c | 1 +
3 files changed, 71 insertions(+), 13 deletions(-)
--
2.10.1
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________
Linux-f2fs-devel mailing list
Linux-f2fs-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
