this is tested with read hardware, the tests obviously don't work,
we have to grandfather them.

very old terminals do not have SDO definitions, but have their
PDO mapping defined in the eeprom. Namely the EK1814. This makes
this particular terminal work, let's hope it also works for similar
terminals.

if a device is only reading from terminals, it can be added to
several sync groups. Start to build this concept and improve
tests.

the use of hashmaps is just hopelessly outdated.

See merge request teichman/ebpfcat!16

with the PacketVar descriptor, one can simply address data in packets.

this also adds support for network byte order, and minor improvements.

See merge request teichman/ebpfcat!15

this also improves readabiliy

especially the reverse divisions were incorrect. Now the code
is shorter as well.

this modifies the tests to do some multiplication not long -
but it actually seems legit.

this makes a difference with signed extension to 64 bit

this is not actually related to the ebpfcat PacketVar, but kinda
similar.

See merge request teichman/ebpfcat!14

this improves ethercat compatibility by actually setting the correct PDO size and other minor details.

See merge request teichman/ebpfcat!12

we still have to artificially tell that this is a float
value, although that's actually in the SDO.

we used to initialize the sync manager to what was
in the eeprom, which is big enough normally, but not always.

Now we read out the pdo assignment first, and set the sync
manager size accordingly.

See merge request teichman/ebpfcat!13

for most operations, pre-operational is good enough a state.
By thinko we went to safe operational, which is not needed.