Pluto-P

cnc-club.ru for English speaking users

Pluto-P

Сообщение mimi » 27 авг 2013, 19:30

Hi!

I've seen that there are some pluto-p users.
I've found some race condition bugs in pluto-p drivers (in verilog code).
Currently the changes are only for the pluto-servo code due to the lack of stepper equipment (its on the way).

https://github.com/tinkercnc/linuxcnc-m ... o_firmware
https://github.com/tinkercnc/linuxcnc-m ... 014bcd3cb4

Copying the pluto_servo.rbf in an existing src tree and compiling lasts out because the changes in the driver dir are not necessary.

The stepper fix will be given later. :)

---------
Matsche
Аватара пользователя
mimi
Новичок
 
Сообщения: 2
Зарегистрирован: 27 авг 2013, 14:51
Откуда: Austria
Репутация: 0

Re: Pluto-P

Сообщение Impartial » 13 сен 2013, 20:32

mimi писал(а):I've found some race condition bugs in pluto-p drivers (in verilog code).

Could you show where were errors please.
Impartial
Мастер
 
Сообщения: 896
Зарегистрирован: 23 фев 2011, 01:50
Репутация: 16

Re: Pluto-P

Сообщение mimi » 25 сен 2013, 01:15

yes, look at the comments in
https://github.com/tinkercnc/linuxcnc-m ... re/servo.v
for instance:
// the first part of the "right way" (see below @ addr_reg test)
...
// the second part of the "right way"
...
// yes, it was the wrong way.
...

here is the diff:
https://github.com/tinkercnc/linuxcnc-m ... xes#diff-6

the problem of the bug is, it stochastically muddles the EPP bytes .
and it depends on several environmental factors and is not easy to detect.
I've found it because I've installed leds on the output pins for debugging. :)

the Open-Drain/Collector buffer is for my optocouplers and does not belong to the bugfix.

poka
Аватара пользователя
mimi
Новичок
 
Сообщения: 2
Зарегистрирован: 27 авг 2013, 14:51
Откуда: Austria
Репутация: 0


Вернуться в English forum

Кто сейчас на конференции

Зарегистрированные пользователи: anthro, Bing [Bot], FedorovDrugoy, Google [Bot], Heify, sas_75, Yandex [bot], Фрост

Поделиться

Reputation System ©'