- Apr 19, 2016
-
-
Lucas Russo authored
This was include in commits from lnls-bpm/bpm-gw@09f35bd8f7 to lnls-bpm/bpm-gw@3c537b769f
-
Lucas Russo authored
-
- Apr 18, 2016
-
-
Lucas Russo authored
-
- Apr 15, 2016
-
-
Lucas Russo authored
The FPGA gateware was changed to uniformize the WB register map for the FMC ACTIVE part.
-
- Apr 14, 2016
-
-
Lucas Russo authored
-
Lucas Russo authored
Now, if EPICS IOCs are started with systemd they have a dependency with bpm-sw@.service
-
Lucas Russo authored
Removing the socket handler on *_cfg handle was wrong as the devio_destroy_smio () already calls devio_destroy_actor (), which, in turn, removes the socket from the reader. Also, as we expect only a single message on the handler, it`s safer and cleaner to not poll the socket until there is no message.
-
- Apr 11, 2016
-
-
Lucas Russo authored
If an SMIO fails for some reason, continue to spawn the other ones.
-
- Apr 04, 2016
-
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
This was changed on commit 5a769af3
-
- Apr 01, 2016
-
-
Lucas Russo authored
As an optimization, we registered the SMIO message and SMIO config handlers before creating the actor. For the first case this is OK, as as register the handler to an existing socket creted by zsys_create_pipe (). For the second case the registered socket does not yet exist, as it will be creted by zactor_new ().
-
Lucas Russo authored
-
Lucas Russo authored
This fixes #114 github issue
-
- Mar 31, 2016
-
-
Lucas Russo authored
It's easier to standardize the initial EPICS BPM ID from 1 to 24 instead of from 0 to 23. In this way, the reported ID to clients stau consistent with the FPGA driver, as well.
-
Lucas Russo authored
-
Lucas Russo authored
In this way, we avoid creating an SMIO instance without a listener on the other side.
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
This allows us to clean all of the generated APPS objects
-
Lucas Russo authored
-
Lucas Russo authored
This would help solving a long-standing issue of not having the DEVIO looping through messages before spawing all of the SMIO modules, which used to happen before the main loop
-
Lucas Russo authored
We are now using a reactor API, zloop, which is better coded and more standard than a custom poller loop.
-
Lucas Russo authored
We are now using a reactor API, zloop, which is better coded and more standard than a custom poller loop.
-
Lucas Russo authored
Instead of using the index of the smio_mod_dispatch table, which is weak and error prone, we use a handler pointer instead.
-
- Mar 24, 2016
-
-
Lucas Russo authored
-
Lucas Russo authored
The system integration will take care of daemonizing the whole set of process if needed
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
The specified address is a one-hot encoded. So, for each ADC, only 1 bit must be set in the order required by that specific ADC.
-
Lucas Russo authored
The MSB of the ISLA216P transaction must be R/W bit followed by Length specification, Register address and data (for write transactions)
-
- Mar 23, 2016
-
-
Lucas Russo authored
On BIDIR mode, the received data is on base address register RX0 and not RX0_SINGLE
-
Lucas Russo authored
If we are using SPI in bidir mode, the character length should be copied to the 7 LSB of "SPI_PROTO_REG_CFG_BIDIR" and not "SPI_PROTO_REG_CTRL"
-
Lucas Russo authored
It should be comparing with the number of bytes, not with SUCCESS conditions.
-