Slaves can fail to correctly read their SII information, leading to
zero or corrupted SII information. This leads to slaves that remain
in PREOP or SAFEOP.
This is caused by the cached external (slave) datagrams being cycled
through too quickly due to being marked as consumed when they are not.
Datagrams can then be lost or incorrectly shared between multiple slaves
as the ring returns a datagram that is still in use by a read SII call.
This may also occur if multiple services are contending for a slaves
mailbox.
I have not added checks into ec_master_get_external_datagram() to
check that a returned datagram is not currently in use.
Sdo directory now only fetched on request.
The time-consuming SDO directory fetch during slave scan can now be
skipped by setting the EC_SKIP_SDO_DICT in globals.h. The directory
will now instead be fethed the first time an ethercat sdos command
is executed with the ethercat tool.
Skip output statistics during re-scan.
No reason to write output statistics in syslog when issuing a slave
scanning where UNMATCHED datagrams are expected behavior.
When rescanning a slave (and thus discarding any prior pending FSMs),
explicitly clear the slave mailbox, to avoid getting confused by a
stale response.
Introduce a new lock for synchronization of the master->domains list, and use
that instead of the big master_sem lock, improving real-time performance of
fx. ecrt_domain_process() when EoE thread is being shut down (or other
use-cases where master_sem is held for extended periods).
The following locking order is used:
* master->master_sem
* master->domains_lock
* master->io_sem
* domain->datagram_pairs_lock
Meaning that it is allowed to take domains_lock while holding master_sem, and
it is allowed to hold domains_lock while locking io_sem and
datagram_pairs_lock.
[RV: For now, the lockdep_assert_held() annotations are commented out,
since they don't compile with CONFIG_PROVE_LOCKING and (the default)
--disable-rtmutex.]
With this, we should have in place so that master->io_sem is always held when
accessing master->datagram_queue.
Note, the use of master->master_sem in ec_ioctl_send() and
ec_ioctl_receive() were wrong (introduced in commit cef602586c). The
only thing needing to be synchronized there is the access to
master->datagram_queue, which is protected by master->io_sem.
This corresponds roughly to the fixes for stable-1.5 that are seen in
https://gitlab.com/etherlab.org/ethercat/-/merge_requests/36 .
Graeme Foot
Knud Baastrup
Florian Pose
Esben Haabendal
Nicola Fontana
Dave Page
Gavin Lambert
Andreas Stewering-Bone