FreeCalypso > hg > fc-magnetite
comparison src/cs/layer1/tpu_drivers/source0/tpudrv12.h @ 600:8f50b202e81f
board preprocessor conditionals: prep for more FC hw in the future
This change eliminates the CONFIG_TARGET_FCDEV3B preprocessor symbol and
all preprocessor conditionals throughout the code base that tested for it,
replacing them with CONFIG_TARGET_FCFAM or CONFIG_TARGET_FCMODEM. These
new symbols are specified as follows:
CONFIG_TARGET_FCFAM is intended to cover all hardware designs created by
Mother Mychaela under the FreeCalypso trademark. This family will include
modem products (repackagings of the FCDEV3B, possibly with RFFE or even
RF transceiver changes), and also my desired FreeCalypso handset product.
CONFIG_TARGET_FCMODEM is intended to cover all FreeCalypso modem products
(which will be firmware-compatible with the FCDEV3B if they use TI Rita
transceiver, or will require a different fw build if we switch to one of
Silabs Aero transceivers), but not the handset product. Right now this
CONFIG_TARGET_FCMODEM preprocessor symbol is used to conditionalize
everything dealing with MCSI.
At the present moment the future of FC hardware evolution is still unknown:
it is not known whether we will ever have any beyond-FCDEV3B hardware at all
(contingent on uncertain funding), and if we do produce further FC hardware
designs, it is not known whether they will retain the same FIC modem core
(triband), if we are going to have a quadband design that still retains the
classic Rita transceiver, or if we are going to switch to Silabs Aero II
or some other transceiver. If we produce a quadband modem that still uses
Rita, it will run exactly the same fw as the FCDEV3B thanks to the way we
define TSPACT signals for the RF_FAM=12 && CONFIG_TARGET_FCFAM combination,
and the current fcdev3b build target will be renamed to fcmodem. OTOH, if
that putative quadband modem will be Aero-based, then it will require a
different fw build target, the fcdev3b target will stay as it is, and the
two targets will both define CONFIG_TARGET_FCFAM and CONFIG_TARGET_FCMODEM,
but will have different RF_FAM numbers. But no matter which way we are
going to evolve, it is not right to have conditionals on CONFIG_TARGET_FCDEV3B
in places like ACI, and the present change clears the way for future
evolution.
author | Mychaela Falconia <falcon@freecalypso.org> |
---|---|
date | Mon, 01 Apr 2019 01:05:24 +0000 |
parents | d380b62e1019 |
children | 92dbfa906f66 |
comparison
equal
deleted
inserted
replaced
599:4626d7a955de | 600:8f50b202e81f |
---|---|
272 #define RU_1900 ( PA_OFF | FEM_PINS ^ 0 ) | 272 #define RU_1900 ( PA_OFF | FEM_PINS ^ 0 ) |
273 #define RD_1900 ( PA_OFF | FEM_PINS ^ 0 ) | 273 #define RD_1900 ( PA_OFF | FEM_PINS ^ 0 ) |
274 #define TU_1900 ( PA_HI_BAND | FEM_PINS ^ FEM_8 ) | 274 #define TU_1900 ( PA_HI_BAND | FEM_PINS ^ FEM_8 ) |
275 #define TD_1900 ( PA_OFF | FEM_PINS ^ 0 ) | 275 #define TD_1900 ( PA_OFF | FEM_PINS ^ 0 ) |
276 | 276 |
277 #elif defined(CONFIG_TARGET_GTAMODEM) || defined(CONFIG_TARGET_FCDEV3B) | 277 #elif defined(CONFIG_TARGET_GTAMODEM) |
278 | 278 |
279 // Openmoko's triband configuration is a bastardized version | 279 // Openmoko's triband configuration is a bastardized version |
280 // of TI's quadband one from Leonardo/E-Sample | 280 // of TI's quadband one from Leonardo/E-Sample |
281 | 281 |
282 #define FEM_7 BIT_2 // act2 | 282 #define FEM_7 BIT_2 // act2 |
320 | 320 |
321 /* | 321 /* |
322 * In our FreeCalypso hw family, we would like to be able to use | 322 * In our FreeCalypso hw family, we would like to be able to use |
323 * both triband and quadband RFFEs. Our current FCDEV3B is triband, | 323 * both triband and quadband RFFEs. Our current FCDEV3B is triband, |
324 * copied from Openmoko, and the same is expected to be the case for | 324 * copied from Openmoko, and the same is expected to be the case for |
325 * future low-cost designs, but if someone pays for a new RF layout, | 325 * future low-budget designs, but if someone pays for a new RF layout, |
326 * we can use a quadband RFFE instead. If we ever have two different | 326 * we can use a quadband RFFE instead. If we ever have two different |
327 * hw platforms or variants that differ in the RFFE but are otherwise | 327 * hw platforms or variants that differ in the RFFE but are otherwise |
328 * firmware-compatible, we would like to have the same fw build | 328 * firmware-compatible, we would like to have the same fw build |
329 * work with both triband and quadband RFFEs. How is it possible? | 329 * work with both triband and quadband RFFEs. How is it possible? |
330 * The trick is that we define our set of TSPACT RFFE control signals | 330 * The trick is that we define our set of TSPACT RFFE control signals |