What follows are four series of patches that lead up to making
it clear exactly which data in an osd operation is outgoing
(being sent to an osd) and which is incoming (the result of
a read or object method call operation). They build on each
other, with the results of one making the subsequent ones
possible (or at least easier to understand and verify).
The four series are:
[PATCH 0/3] libceph: a few cleanups
[PATCH 1/3] libceph: use (void *) for untyped data in osd ops
[PATCH 2/3] libceph: kill ceph_msg->pagelist_count
[PATCH 3/3] libceph: rename ceph_calc_object_layout()
These are sort of miscellaneous cleanup, not directly
related but fixed while I was working with nearby code.
[PATCH 0/3] libceph: simplify incoming message allocation
[PATCH 1/3] libceph: drop mutex while allocating a message
[PATCH 2/3] libceph: define mds_alloc_msg() method
[PATCH 3/3] libceph: no need for alignment for mds message
This simplifies and localizes the code that allocates
messages for incoming data.
[PATCH 0/3] ceph: assign message data fields consistently
[PATCH 1/3] ceph: use calc_pages_for() in start_read()
[PATCH 2/3] ceph: simplify ceph_sync_write() page_align calculation
[PATCH 3/3] libceph: don't assign page info in ceph_osdc_new_request()
This allows the assignment of everything about the pages
for holding data in an osd request to happen in consecutive
lines of code, and such that it's done the same way throughout
the ceph code.
[PATCH 0/3] libceph: distinguish osd request read and write data
[PATCH 1/3] libceph: separate osd request data info
[PATCH 2/3] libceph: distinguish page and bio requests
[PATCH 3/3] libceph: separate read and write data
This encapsulates information about osd request data into a
separate data structure, then defines separate instances
of that structure for incoming and outgoing data.
These patches are available in branch "review/wip-rwsep" on
the ceph-client git repository. Included before them are two
others I posted recently:
[PATCH] libceph: fix wrong opcode use in osd_req_encode_op()
[PATCH] libceph: complete lingering requests only once
-Alex
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
