Commits · 716b7a6109e3efc583f196ba5e823883a32b919f · Projects / FMC TDC 1ns 5cha

20 Aug, 2018 8 commits

lib: fix to_ns convertion · 716b7a61
Federico Vaga authored Aug 17, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
716b7a61
lib: convert all aquired timestamps in a dedicated func · a869ec6b
Federico Vaga authored Aug 17, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
a869ec6b

*: set timeout coalescing dynamically · 013a37f9

Federico Vaga authored Aug 16, 2018

The driver exports a sysfs attrubte to change the coalescing timeout. A new
library funciton can be used to set/get this value; and the test program allows
the configuration before an acquisition.

Keep in mind that, for FIFOs the configuration is global (users are warned on
dmesg) while for DMA it is per-channel. I though it was better to have a common
interface for both. Users are warnd in the documentation as well. In the long
run, perhaps also the FIFO can have its own register.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

013a37f9

drv: bugfix, do not touch burst size dynamically · 8f34868a

Federico Vaga authored Aug 17, 2018

It is not an IRQ coalescing size, but the size that triggers the dump
from the FIFO to the DDR. so, it is internal stuff that does not need
to get exposed to the user
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

8f34868a

drv: WARN if spurious IRQ · 19d22fd4
Federico Vaga authored Aug 20, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
19d22fd4
drv: continue acquisition on ZIO failure · 4280de0d
Federico Vaga authored Aug 20, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
4280de0d
drv: add bitfield shift value for channels · 1627ac98
Federico Vaga authored Aug 20, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
1627ac98

drv: bugfix ZIO names are 12bytes, make this 11 · cc39d6ac

Federico Vaga authored Aug 17, 2018

This way the name is properly terminated. Perhaps there is a bug in ZIO
that expect a \0 and it does not check any upper limit.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>

cc39d6ac

16 Aug, 2018 4 commits
- tools:tstamp: init pollfd array · a91b9e35
  Federico Vaga authored Aug 16, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  a91b9e35
- tools:tstamp: add testing mode · 5a8e8f55
  Federico Vaga authored Aug 16, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  5a8e8f55
- drv: fix fmc-bus type · 34941a92
  Federico Vaga authored Aug 16, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  34941a92
- drv: print only on DMA error · 3219d260
  Federico Vaga authored Aug 16, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  3219d260
15 Aug, 2018 28 commits
- drv: add helper to configure gennum · 5dc76e3e
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  5dc76e3e
- drv: add some ZIO helpers · d7bb4319
  Federico Vaga authored Aug 15, 2018
```
They should be implemented in ZIO
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  d7bb4319
- drv: move out HW_BUSY flag set|get and use ZIO API · e9567174
  Federico Vaga authored Aug 15, 2018
```
the HW_BUSY flag must be cleared before doing data_done in order to hold
only during DMA.
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  e9567174
- drv: relax CPU on active wait · 31d98b13
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  31d98b13
- drv: put basic gennum DMA ops in header · c33b9ad3
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  c33b9ad3
- drv: change misleading name in gennum header · 1d9a21b3
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  1d9a21b3
- drv: re-organize gennum header file · 75779622
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  75779622
- drv: bugfix terminate string · 0f9d5767
  Federico Vaga authored Aug 15, 2018
```
it happened that ZIO was taking some extra characters after the name.
to play safe I put a terminator there
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  0f9d5767
- drv: bugfix on DMA IRQ clear · 98b8165e
  Federico Vaga authored Aug 15, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  98b8165e
- drv: remove old comment · c3111ade
  Federico Vaga authored Aug 14, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  c3111ade
- drv: remove unused variable · d0169fc1
  Federico Vaga authored Aug 14, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  d0169fc1
- drv: gennum LAST item meaning is reversed · 6391f63c
  Federico Vaga authored Aug 14, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  6391f63c
- drv: fix gennum DMA done bit · 0ffdccec
  Federico Vaga authored Aug 14, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  0ffdccec
- drv: release HW_BUSY flag on data_done · 3bdb77a5
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  3bdb77a5
- tools: use option -D to set device ID · 2727a10a
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  2727a10a
- tools: add Makefile dependency with the tdc library · de3abb08
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  de3abb08
- drv: apply offset and update control on data_done · 9ccbb839
  Federico Vaga authored Aug 13, 2018
```
This also fix a bug. The bug was that I was coping data
at the wrong offset withing the ZIO block. It starts always
at offset 0, but I was shifing by "start" bytes
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  9ccbb839
- drv: re-organize DMA readout · 1ae714cc
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  1ae714cc
- drv: add dedicated ZIO trigger · b19881b9
  Federico Vaga authored Aug 09, 2018
```
For the time being is used only to change the BURST size using POST-SAMPLEs

This solves also a bug with the ZIO trigger "user" which does arm for us, but
this is not exactly what we want because we need to arm only when we know the
number of samples to acquire
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  b19881b9
- drv: configure FIFO IRQ coalescing only for FIFO · ab7367ed
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  ab7367ed
- drv: split gn4124_dma_read in smaller functions · 1b3e6b2e
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  1b3e6b2e
- drv: access DMA register as inline function · 72405bbe
  Federico Vaga authored Aug 13, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  72405bbe
- drv: bugfix mark channel-set busy during acquisition · 64a89493
  Federico Vaga authored Aug 10, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  64a89493
- drv: enable and ack DMA interrupts · 53e9d78b
  Federico Vaga authored Aug 10, 2018
```
Not reallyused at this point
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  53e9d78b
- tools: read more samples when ask to skip · ea5be5bb
  Federico Vaga authored Aug 10, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  ea5be5bb
- drv: do a pre-allocation to store all timestamps · 113ff999
  Federico Vaga authored Aug 10, 2018
```
This allows bigger read from the driver, which reduces the number of syscall
and make the system faster
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  113ff999
- drv: store all samples directly into the ZIO buffer · fe5b9727
  Federico Vaga authored Aug 09, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  fe5b9727
- drv: remove debug print · 180fce07
  Federico Vaga authored Aug 09, 2018
```
Signed-off-by: Federico Vaga <federico.vaga@cern.ch>
```
  180fce07