- 15 Oct, 2015 10 commits
-
-
Lucas Russo authored
We now only pass the cfg_file as argument and not the broker endpoint. The later is read from the configuration file, among other options.
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
This allow systemd to take care of spawning different dev_io services independently whenever the corresponding device is added/removed from the system bus.
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
This is needed as we use a custom mapping between board/bpm and the bpm_id used to spawn EPICS IOC processes.
-
Lucas Russo authored
Now, we don't map our hash with <pattern>/afe or <pattern>/dbe, just <pattern> is enough and we split different configuration options in hash structure fields.
-
Lucas Russo authored
-
- 09 Oct, 2015 8 commits
-
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
Now, we don't specify the configuration filename through the Makefile, but in the command-line of dev_mngr.
-
Lucas Russo authored
This option is only valid if daemonize parameter is "yes".
-
Lucas Russo authored
-
Lucas Russo authored
-
- 07 Oct, 2015 6 commits
-
-
Lucas Russo authored
This is not needed anymore as we use system services
-
Lucas Russo authored
-
Lucas Russo authored
This allows us to selective spawn the broker. Useful in cases that the broker is an independent system service, as Malamute.
-
Lucas Russo authored
-
Lucas Russo authored
This is related to commits a0f3f992 and 9d80fdf5
-
Lucas Russo authored
We were not multiplying the sample size by the number of samples to give the number of total bytes per block.
-
- 03 Oct, 2015 4 commits
-
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
blink_leds and ad9510_cfg_defaults were not using our simple messaging protocol, which causes some confusion for the client programmer, as well as for libbpmclient.
-
Lucas Russo authored
The timeout variable must support negative numbers, -1, meaning an infinite wait.
-
- 30 Sep, 2015 2 commits
-
-
Lucas Russo authored
-
Lucas Russo authored
Now, we can stop the acquisition FSM. This is useful if we are acquiring data in triggered mode and we want to abort the acquisition.
-
- 25 Sep, 2015 3 commits
-
-
Lucas Russo authored
For now, we reset the FPGA on PCIe open.
-
Lucas Russo authored
-
Lucas Russo authored
-
- 24 Sep, 2015 1 commit
-
-
Lucas Russo authored
-
- 22 Sep, 2015 4 commits
-
-
Lucas Russo authored
For now, enabling synthetic data on the FMC130 module, we can acquire consecutive data and check if the output file is correct with these functions.
-
Lucas Russo authored
-
Lucas Russo authored
-
Lucas Russo authored
-
- 21 Sep, 2015 2 commits
-
-
Lucas Russo authored
As we were not acquiring from non-paged aligned address (only skip trigger acquisition was used), we didn't see the possible errors coming from this. The first one is the start address. On the first read, we need to check for page size overflow. If this occurs, the remaining bytes need to be read from the next page. Secondly, the output data pointer was not being updated to account for split (from 2 consecutive pages) reads. Without this, we were copying data to the wrong destination address.
-
Lucas Russo authored
-