linux-stable.git/net/ceph, branch linux-4.5.y

libceph: don't spam dmesg with stray reply warnings

2016-02-24T19:28:51+00:00

Commit d15f9d694b77 ("libceph: check data_len in ->alloc_msg()")
mistakenly bumped the log level on the "tid %llu unknown, skipping"
message.  Turn it back into a dout() - stray replies are perfectly
normal when OSDs flap, crash, get killed for testing purposes, etc.

Cc: stable@vger.kernel.org # 4.3+
Signed-off-by: Ilya Dryomov 
Reviewed-by: Alex Elder

libceph: use the right footer size when skipping a message

2016-02-24T19:28:40+00:00

ceph_msg_footer is 21 bytes long, while ceph_msg_footer_old is only 13.
Don't skip too much when CEPH_FEATURE_MSG_AUTH isn't negotiated.

Cc: stable@vger.kernel.org # 3.19+
Signed-off-by: Ilya Dryomov 
Reviewed-by: Alex Elder

libceph: don't bail early from try_read() when skipping a message

2016-02-24T19:28:31+00:00

The contract between try_read() and try_write() is that when called
each processes as much data as possible.  When instructed by osd_client
to skip a message, try_read() is violating this contract by returning
after receiving and discarding a single message instead of checking for
more.  try_write() then gets a chance to write out more requests,
generating more replies/skips for try_read() to handle, forcing the
messenger into a starvation loop.

Cc: stable@vger.kernel.org # 3.10+
Reported-by: Varada Kari 
Signed-off-by: Ilya Dryomov 
Tested-by: Varada Kari 
Reviewed-by: Alex Elder

libceph: MOSDOpReply v7 encoding

2016-02-04T17:26:08+00:00

Empty request_redirect_t (struct ceph_request_redirect in the kernel
client) is now encoded with a bool.  NEW_OSDOPREPLY_ENCODING feature
bit overlaps with already supported CRUSH_TUNABLES5.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

crush: decode and initialize chooseleaf_stable

2016-02-04T17:25:58+00:00

Also add missing \n while at it.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

crush: add chooseleaf_stable tunable

2016-02-04T17:25:55+00:00

Add a tunable to fix the bug that chooseleaf may cause unnecessary pg
migrations when some device fails.

Reflects ceph.git commit fdb3f664448e80d984470f32f04e2e6f03ab52ec.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

crush: ensure take bucket value is valid

2016-02-04T17:25:50+00:00

Ensure that the take argument is a valid bucket ID before indexing the
buckets array.

Reflects ceph.git commit 93ec538e8a667699876b72459b8ad78966d89c61.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

crush: ensure bucket id is valid before indexing buckets array

2016-02-04T17:25:23+00:00

We were indexing the buckets array without verifying the index was
within the [0,max_buckets) range.  This could happen because
a multistep rule does not have enough buckets and has CRUSH_ITEM_NONE
for an intermediate result, which would feed in CRUSH_ITEM_NONE and
make us crash.

Reflects ceph.git commit 976a24a326da8931e689ee22fce35feab5b67b76.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil

libceph: remove outdated comment

2016-01-21T18:36:09+00:00

MClientMount{,Ack} are long gone.  The receipt of bare monmap doesn't
actually indicate a mount success as we are yet to authenticate at that
point in time.

Signed-off-by: Ilya Dryomov

libceph: kill off ceph_x_ticket_handler::validity

2016-01-21T18:36:09+00:00

With it gone, no need to preserve ceph_timespec in process_one_ticket()
either.

Signed-off-by: Ilya Dryomov 
Reviewed-by: Sage Weil