Define a function ceph_msg_data_set_pages(), which more clearly
abstracts the assignment page-related fields for data in a ceph
message structure.  Use this new function in the osd client and mds
client.

Ideally, these fields would never be set more than once (with
BUG_ON() calls to guarantee that).  At the moment though the osd
client sets these every time it receives a message, and in the event
of a communication problem this can happen more than once.  (This
will be resolved shortly, but setting up these helpers first makes
it all a bit easier to work with.)

Rearrange the field order in a ceph_msg structure to group those
that are used to define the possible data payloads.

This partially resolves:
    http://tracker.ceph.com/issues/4263Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Rather than explicitly initializing many fields to 0, NULL, or false
in a newly-allocated message, just use kzalloc() for allocating new
messages.  This will become a much more convenient way of doing
things anyway for upcoming patches that abstract the data field.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

While processing an outgoing pagelist (either the data pagelist or
trail) in a ceph message, the messenger cycles through each of the
pages on the list.  This is accomplished in out_msg_pos_next(), if
the end of the first page on the list is reached, the first page is
moved to the end of the list.

There is a list operation, list_rotate_left(), which performs
exactly this operation, and by using it, what's really going on
becomes more obvious.

So replace these two list_move_tail() calls with list_rotate_left().
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Define a new function in_msg_pos_next() to match out_msg_pos_next(),
and use it in place of code at the end of read_partial_message_pages()
and read_partial_message_bio().

Note that the page number is incremented and offset reset under
slightly different conditions from before.  The result is
equivalent, however, as explained below.

Each time an incoming message is going to arrive, we find out how
much room is left--not surpassing the current page--and provide that
as the number of bytes to receive.  So the amount we'll use is the
lesser of:  all that's left of the entire request; and all that's
left in the current page.

If we received exactly how many were requested, we either reached
the end of the request or the end of the page.  In the first case,
we're done, in the second, we move onto the next page in the array.

In all cases but (possibly) on the last page, after adding the
number of bytes received, page_pos == PAGE_SIZE.  On the last page,
it doesn't really matter whether we increment the page number and
reset the page position, because we're done and we won't come back
here again.  The code previously skipped over that last case,
basically.  The new code handles that case the same as the others,
incrementing and resetting.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

There is only one caller for read_partial_message_bio(), and it
always passes &msg->bio_iter and &bio_seg as the second and third
arguments.  Furthermore, the message in question is always the
connection's in_msg, and we can get that inside the called function.

So drop those two parameters and use their derived equivalents.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Change the type of the "more" parameter from int to bool.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Some values printed are not (necessarily) in CPU order.  We already
have a copy of the converted versions, so use them.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

This is probably unnecessary but the code read as if it were wrong
in read_partial_message().
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

In ceph_con_in_msg_alloc() it is possible for a connection's
alloc_msg method to indicate an incoming message should be skipped.
By default, read_partial_message() initializes the skip variable
to 0 before it gets provided to ceph_con_in_msg_alloc().

The osd client, mon client, and mds client each supply an alloc_msg
method.  The mds client always assigns skip to be 0.

The other two leave the skip value of as-is, or assigns it to zero,
except:
    - if no (osd or mon) request having the given tid is found, in
      which case skip is set to 1 and NULL is returned; or
    - in the osd client, if the data of the reply message is not
      adequate to hold the message to be read, it assigns skip
      value 1 and returns NULL.
So the returned message pointer will always be NULL if skip is ever
non-zero.

Clean up the logic a bit in ceph_con_in_msg_alloc() to make this
state of affairs more obvious.  Add a comment explaining how a null
message pointer can mean either a message that should be skipped or
a problem allocating a message.

This resolves:
    http://tracker.ceph.com/issues/4324Reported-by: NGreg Farnum <greg@inktank.com>
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NGreg Farnum <greg@inktank.com>

The only user of the ceph messenger that doesn't define an alloc_msg
method is the mds client.  Define one, such that it works just like
it did before, and simplify ceph_con_in_msg_alloc() by assuming the
alloc_msg method is always present.

This and the next patch resolve:
    http://tracker.ceph.com/issues/4322Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NGreg Farnum <greg@inktank.com>

In ceph_con_in_msg_alloc(), if no alloc_msg method is defined for a
connection a new message is allocated with ceph_msg_new().

Drop the mutex before making this call, and make sure we're still
connected when we get it back again.

This is preparing for the next patch, which ensures all connections
define an alloc_msg method, and then handles them all the same way.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NGreg Farnum <greg@inktank.com>

The pagelist_count field is never actually used, so get rid of it.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Use distinct fields for tracking the number of pages in a message's
page array and in a message's page list.  Currently only one or the
other is used at a time, but that will be changing soon.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

The bio_seg field is used by the ceph messenger in iterating through
a bio.  It should never have a negative value, so make it an
unsigned.  (I contemplated making it unsigned short to match the
struct bio definition, but it offered no benefit.)

Change variables used to hold bio_seg values to all be unsigned as
well.  Change two variable names in init_bio_iter() to match the
convention used everywhere else.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

This just converts a manually-implemented loop into a do..while loop
in con_work().  It also moves handling of EAGAIN inside the blocks
where it's already been determined an error code was returned.

Also update a few dout() calls near the affected code for
consistency.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

This just rearranges the logic in con_work() a little bit so that a
flag is used to indicate a fault has occurred.  This allows both the
fault and non-fault case to be handled the same way and avoids a
couple of nearly consecutive gotos.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

An error occurring on a ceph connection is treated as a fault,
causing the connection to be reset.  The initial part of this fault
handling has to be done while holding the connection mutex, but
it must then be dropped for the last part.

Separate the part of this fault handling that executes without the
lock into its own function, con_fault_finish().  Move the call to
this new function, as well as call that drops the connection mutex,
into ceph_fault().  Rename that function con_fault() to reflect that
it's only handling the connection part of the fault handling.

The motivation for this was a warning from sparse about the locking
being done here.  Rearranging things this way keeps all the mutex
manipulation within ceph_fault(), and this stops sparse from
complaining.

This partially resolves:
    http://tracker.ceph.com/issues/4184Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Collect the code that tests for and implements a backoff delay for a
ceph connection into a new function, ceph_backoff().

Make the debug output messages in that part of the code report
things consistently by reporting a message in the socket closed
case, and by making the one for PREOPEN state report the connection
pointer like the rest.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Eliminate most of the problems in the libceph code that cause sparse
to issue warnings.
    - Convert functions that are never referenced externally to have
      static scope.
    - Pass NULL rather than 0 for a pointer argument in one spot in
      ceph_monc_delete_snapid()

This partially resolves:
    http://tracker.ceph.com/issues/4184Reported-by: NFengguang Wu <fengguang.wu@intel.com>
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

Define and use functions that encapsulate operations performed on
a connection's flags.

This resolves:
    http://tracker.ceph.com/issues/4234Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

The ceph messenger has a few spots that are only used when
bio messages are supported, and that's only when CONFIG_BLOCK
is defined.  This surrounds a couple of spots with #ifdef's
that would cause a problem if CONFIG_BLOCK were not present
in the kernel configuration.

This resolves:
    http://tracker.ceph.com/issues/3976Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NJosh Durgin <josh.durgin@inktank.com>

We should not set con->state to CLOSED here; that happens in
ceph_fault() in the caller, where it first asserts that the state
is not yet CLOSED.  Avoids a BUG when the features don't match.

Since the fail_protocol() has become a trivial wrapper, replace
calls to it with direct calls to reset_connection().
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

A number of assertions in the ceph messenger are implemented with
BUG_ON(), killing the system if connection's state doesn't match
what's expected.  At this point our state model is (evidently) not
well understood enough for these assertions to trigger a BUG().
Convert all BUG_ON(con->state...) calls to be WARN_ON(con->state...)
so we learn about these issues without killing the machine.

We now recognize that a connection fault can occur due to a socket
closure at any time, regardless of the state of the connection.  So
there is really nothing we can assert about the state of the
connection at that point so eliminate that assertion.
Reported-by: NUgis <ugis22@gmail.com>
Tested-by: NUgis <ugis22@gmail.com>
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

When a connection's socket disconnects, or if there's a protocol
error of some kind on the connection, a fault is signaled and
the connection is reset (closed and reopened, basically).  We
currently get an error message on the log whenever this occurs.

A ceph connection will attempt to reestablish a socket connection
repeatedly if a fault occurs.  This means that these error messages
will get repeatedly added to the log, which is undesirable.

Change the error message to be a warning, so they don't get
logged by default.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

A connection's socket can close for any reason, independent of the
state of the connection (and without irrespective of the connection
mutex).  As a result, the connectino can be in pretty much any state
at the time its socket is closed.

Handle those other cases at the top of con_work().  Pull this whole
block of code into a separate function to reduce the clutter.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

The ceph_on_in_msg_alloc() method calls the ->alloc_msg() helper which
may return NULL.  It also drops con->mutex while it allocates a message,
which means that the connection state may change (e.g., get closed).  If
that happens, we clean up and bail out.  Avoid calling ceph_msg_put() on
a NULL return value and triggering a crash.

This was observed when an ->alloc_msg() call races with a timeout that
resends a zillion messages and resets the connection, and ->alloc_msg()
returns NULL (because the request was resent to another target).

Fixes http://tracker.newdream.net/issues/3342Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

The ceph_on_in_msg_alloc() method drops con->mutex while it allocates a
message.  If that races with a timeout that resends a zillion messages and
resets the connection, and the ->alloc_msg() method returns a NULL message,
it will call ceph_msg_put(NULL) and BUG.

Fix by only calling put if msg is non-NULL.

Fixes http://tracker.newdream.net/issues/3142Signed-off-by: NSage Weil <sage@inktank.com>

This patch defines a single function, queue_con_delay() to call
queue_delayed_work() for a connection.  It basically generalizes
what was previously queue_con() by adding the delay argument.
queue_con() is now a simple helper that passes 0 for its delay.
queue_con_delay() returns 0 if it queued work or an errno if it
did not for some reason.

If con_work() finds the BACKOFF flag set for a connection, it now
calls queue_con_delay() to handle arranging to start again after a
delay.

Note about connection reference counts:  con_work() only ever gets
called as a work item function.  At the time that work is scheduled,
a reference to the connection is acquired, and the corresponding
con_work() call is then responsible for dropping that reference
before it returns.

Previously, the backoff handling inside con_work() silently handed
off its reference to delayed work it scheduled.  Now that
queue_con_delay() is used, a new reference is acquired for the
newly-scheduled work, and the original reference is dropped by the
con->ops->put() call at the end of the function.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

Both ceph_fault() and con_work() include handling for imposing a
delay before doing further processing on a faulted connection.
The latter is used only if ceph_fault() is unable to.

Instead, just let con_work() always be responsible for implementing
the delay.  After setting up the delay value, set the BACKOFF flag
on the connection unconditionally and call queue_con() to ensure
con_work() will get called to handle it.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

If ceph_fault() is unable to queue work after a delay, it sets the
BACKOFF connection flag so con_work() will attempt to do so.

In con_work(), when BACKOFF is set, if queue_delayed_work() doesn't
result in newly-queued work, it simply ignores this condition and
proceeds as if no backoff delay were desired.  There are two
problems with this--one of which is a bug.

The first problem is simply that the intended behavior is to back
off, and if we aren't able queue the work item to run after a delay
we're not doing that.

The only reason queue_delayed_work() won't queue work is if the
provided work item is already queued.  In the messenger, this
means that con_work() is already scheduled to be run again.  So
if we simply set the BACKOFF flag again when this occurs, we know
the next con_work() call will again attempt to hold off activity
on the connection until after the delay.

The second problem--the bug--is a leak of a reference count.  If
queue_delayed_work() returns 0 in con_work(), con->ops->put() drops
the connection reference held on entry to con_work().  However,
processing is (was) allowed to continue, and at the end of the
function a second con->ops->put() is called.

This patch fixes both problems.
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

In write_partial_msg_pages(), pages need to be kmapped in order to
perform a CRC-32c calculation on them.  As an artifact of the way
this code used to be structured, the kunmap() call was separated
from the kmap() call and both were done conditionally.  But the
conditions under which the kmap() and kunmap() calls were made
differed, so there was a chance a kunmap() call would be done on a
page that had not been mapped.

The symptom of this was tripping a BUG() in kunmap_high() when
pkmap_count[nr] became 0.
Reported-by: NBryan K. Wright <bryan@virginia.edu>
Signed-off-by: NAlex Elder <elder@inktank.com>
Reviewed-by: NSage Weil <sage@inktank.com>

Because the Ceph client messenger uses a non-blocking connect, it is
possible for the sending of the client banner to race with the
arrival of the banner sent by the peer.

When ceph_sock_state_change() notices the connect has completed, it
schedules work to process the socket via con_work().  During this
time the peer is writing its banner, and arrival of the peer banner
races with con_work().

If con_work() calls try_read() before the peer banner arrives, there
is nothing for it to do, after which con_work() calls try_write() to
send the client's banner.  In this case Ceph's protocol negotiation
can complete succesfully.

The server-side messenger immediately sends its banner and addresses
after accepting a connect request, *before* actually attempting to
read or verify the banner from the client.  As a result, it is
possible for the banner from the server to arrive before con_work()
calls try_read().  If that happens, try_read() will read the banner
and prepare protocol negotiation info via prepare_write_connect().
prepare_write_connect() calls con_out_kvec_reset(), which discards
the as-yet-unsent client banner.  Next, con_work() calls
try_write(), which sends the protocol negotiation info rather than
the banner that the peer is expecting.

The result is that the peer sees an invalid banner, and the client
reports "negotiation failed".

Fix this by moving con_out_kvec_reset() out of
prepare_write_connect() to its callers at all locations except the
one where the banner might still need to be sent.

[elder@inktak.com: added note about server-side behavior]
Signed-off-by: NJim Schutt <jaschut@sandia.gov>
Reviewed-by: NAlex Elder <elder@inktank.com>

We drop the lock when calling the ->alloc_msg() con op, which means
we need to (a) not clobber con->in_msg without the mutex held, and (b)
we need to verify that we are still in the OPEN state when we retake
it to avoid causing any mayhem.  If the state does change, -EAGAIN
will get us back to con_work() and loop.
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

This function's calling convention is very limiting.  In particular,
we can't return any error other than ENOMEM (and only implicitly),
which is a problem (see next patch).

Instead, return an normal 0 or error code, and make the skip a pointer
output parameter.  Drop the useless in_hdr argument (we have the con
pointer).
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

The ceph_fault() function takes the con mutex, so we should avoid
dropping it before calling it.  This fixes a potential race with
another thread calling ceph_con_close(), or _open(), or similar (we
don't reverify con->state after retaking the lock).

Add annotation so that lockdep realizes we will drop the mutex before
returning.
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

We drop the con mutex when delivering a message.  When we retake the
lock, we need to verify we are still in the OPEN state before
preparing to read the next tag, or else we risk stepping on a
connection that has been closed.
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

If the connect() call immediately fails such that sock == NULL, we
still need con_close_socket() to reset our socket state to CLOSED.
Signed-off-by: NSage Weil <sage@inktank.com>
Reviewed-by: NAlex Elder <elder@inktank.com>

Signed-off-by: NSage Weil <sage@inktank.com>

Rename flags with CON_FLAG prefix, move the definitions into the c file,
and (better) document their meaning.
Signed-off-by: NSage Weil <sage@inktank.com>

Use a simple set of 6 enumerated values for the socket states (CON_STATE_*)
and use those instead of the state bits.  All of the con->state checks are
now under the protection of the con mutex, so this is safe.  It also
simplifies many of the state checks because we can check for anything other
than the expected state instead of various bits for races we can think of.

This appears to hold up well to stress testing both with and without socket
failure injection on the server side.
Signed-off-by: NSage Weil <sage@inktank.com>