Re: [10.0_STABLE] Hard lock

[mlelstv%serpens.de@localhost (Michael van Elst)]

joel.bertrand%systella.fr@localhost (=?ISO-8859-1?Q?BERTRAND_Jo=EBl?=) writes:

>>bouyer@antioche=2Eeu=2Eorg (Manuel Bouyer) writes:
>>>This clearly means there is a locking problem here=2E And it may explain=
> the
>>>hard lock

So far I have never seen a LOCKDEBUG error, but there are at least two
locking problems.

It's possible that a request completes before a timeout for it is even
started. That caused the ccb_timeout without a (related) connection.

It's possible that a PDU is attempted to be freed twice (as part of
sending out, and as part of timeout handling). The second free is
ignored (causing the message about an unused PDU being freed),
unless that PDU is already reused with possibly more severe
consequences.