PSOC6: Update CSP to latest #11324
PSOC6: Update CSP to latest #11324
Conversation
|
@kyle-cypress, thank you for your changes. |
|
CI started |
Test run: FAILEDSummary: 3 of 4 test jobs failed Failed test jobs:
|
kyle-cypress
force-pushed
the
pr/psoc6-csp
branch
from
2873555
to
977eb83
Compare
|
@ARMmbed/mbed-os-maintainers This one should go first (see how many PRs depend on this one , above referenced) |
|
CI restarted |
Test run: FAILEDSummary: 3 of 4 test jobs failed Failed test jobs:
|
Update HAL adapter for interface changes Misc minor fixes to HAL adapter
kyle-cypress
force-pushed
the
pr/psoc6-csp
branch
from
141742e
to
8ad377a
Compare
|
The CI failures for CY8CKIT_062_WIFI_BT and CYW943012P6EVB_01 are due to a psoc6csp change which is dependent on additional functions being provided by the BSPs. To avoid creating a circular dependency, I have revised the commit history to remove those changes from this PR and move them to #11323 |
|
CI restarted |
Test run: SUCCESSSummary: 11 of 11 test jobs passed |
|
@kyle-cypress Can you update dependencies now and let us know which one is readY? |
There was a problem hiding this comment.
Transferring some thoughts from #11378.
Also:
The 12->16-bit scaling in cyhal_adc_read_u16 is a little bit off - it ends up just multiplying by 16, so you get an answer between 0x0000 and 0xFFF0. You could never get 1.0 from the float version.
An better and still-efficient way to increase number of bits is to do this, replicating the top nibble:
uint16_t scaled_result = (result << 4) | (result >> 8);
So 0xABC becomes 0xABCA. (And 0x000->0x0000, 0xFFF->0xFFFF).
| @@ -51,7 +51,7 @@ void analogout_write_u16(dac_t *obj, uint16_t value) | |||
|
|
|||
| float analogout_read(dac_t *obj) | |||
| { | |||
| return 100.0f / UINT16_MAX * analogout_read_u16(obj); | |||
| return analogout_read_u16(obj) / UINT16_MAX; | |||
There was a problem hiding this comment.
This almost always returns 0, as it's integer division. Should be
return analogout_read_u16(obj) * (1.0f / UINT64_MAX);
(Multiplication by reciprocal will be more efficient than actual FP division).
Same maths should be used in analogin_read.
Replace with multiplication by the floating point reciprocal, to produce a floating point result while being more efficient than floating point division. Addresses the issue raised by kjbracey-arm in ARMmbed#11324 (review)
Replace with multiplication by the floating point reciprocal, to produce a floating point result while being more efficient than floating point division. Addresses the issue raised by kjbracey-arm in ARMmbed#11324 (review)
Replace with multiplication by the floating point reciprocal, to produce a floating point result while being more efficient than floating point division. Addresses the issue raised by kjbracey-arm in ARMmbed#11324 (review)
Replace with multiplication by the floating point reciprocal, to produce a floating point result while being more efficient than floating point division. Addresses the issue raised by kjbracey-arm in #11324 (review)
Replace with multiplication by the floating point reciprocal, to produce a floating point result while being more efficient than floating point division. Addresses the issue raised by kjbracey-arm in #11324 (review)
Description
Update PSoC 6 Chip Support Package and HAL to latest versions.
Update HAL adapter for interface changes
Misc minor fixes to HAL adapter
Pull request type
Reviewers
@ARMmbed/team-cypress
Release Notes