applications: sdp: mspi: RX High frequency fix #21112
Conversation
github-actions
bot
added
the
changelog-entry-required
CI InformationTo view the history of this post, clich the 'edited' button above Inputs:Sources:sdk-nrf: PR head: 804cb68e80f9a59afca0ae25f0a411b255d4d3db more detailssdk-nrf:
Github labels
List of changed files detected by CI (2)Outputs:ToolchainVersion: 4ffa2202d5 Test Spec & Results: ✅ Success; ❌ Failure; 🟠 Queued; 🟡 Progress; ◻️ Skipped;
|
|
You can find the documentation preview for this PR here. |
Modified hrt rx to work in quad mode in 21.333333MHz. Signed-off-by: Michal Frankiewicz <michal.frankiewicz@nordicsemi.no>
mif1-nordic
force-pushed
the
rx_21MHz_quad
branch
from
2e9fb59 to
804cb68
Compare
| while (nrf_vpr_csr_vio_shift_cnt_in_get() > 0) { | ||
| } | ||
| /* Wait until timer 0 stops. | ||
| * counter stops 1 clock to early. |
There was a problem hiding this comment.
| * counter stops 1 clock to early. | |
| * counter stops 1 clock too early. |
As it was before.
| @@ -41,6 +41,34 @@ static nrf_vpr_csr_vio_shift_ctrl_t xfer_shift_ctrl = { | |||
| .in_mode = NRF_VPR_CSR_VIO_MODE_IN_CONTINUOUS, | |||
| }; | |||
|
|
|||
| NRF_STATIC_INLINE bool is_counter_running(uint8_t counter) | |||
There was a problem hiding this comment.
Nit, to be more readable:
| NRF_STATIC_INLINE bool is_counter_running(uint8_t counter) | |
| NRF_STATIC_INLINE bool is_counter_running(uint8_t cnt_idx) |
| return cnt != nrf_vpr_csr_vtim_simple_counter_get(counter); | ||
| } | ||
|
|
||
| NRF_STATIC_INLINE uint32_t get_shift_count(volatile hrt_xfer_t *hrt_xfer_params, |
There was a problem hiding this comment.
Nit: I would change name of this function to something different, "get" may suggest it is read from a register. But I am not sure what, "calculate" does not fit either. Maybe something like get_next_shift_count?
| switch (word_count - *index) { | ||
| case 1: /* Last transfer */ | ||
| (*index)++; | ||
| return SHIFTCNTB_VALUE(hrt_xfer_params->xfer_data[HRT_FE_DATA].last_word_clocks); | ||
| case 2: /* Last but one transfer */ | ||
| (*index)++; | ||
| return SHIFTCNTB_VALUE( | ||
| hrt_xfer_params->xfer_data[HRT_FE_DATA].penultimate_word_clocks); | ||
| default: | ||
| (*index)++; | ||
| return SHIFTCNTB_VALUE(BITS_IN_WORD / hrt_xfer_params->bus_widths.data); | ||
| } |
There was a problem hiding this comment.
If index is incremented for every case, then why is it changed in this function? I think it would be better to do it outside and pass index directly, not by a pointer.
| * nrf_vpr_csr_vio_shift_cnt_in_get function inside while cannot be used because in | ||
| * higher frequencies shifting stops before it is called 1-st time. |
There was a problem hiding this comment.
English has a strict order of words in a sentence for a reason ;)
| * nrf_vpr_csr_vio_shift_cnt_in_get function inside while cannot be used because in | |
| * higher frequencies shifting stops before it is called 1-st time. | |
| * nrf_vpr_csr_vio_shift_cnt_in_get function cannot be used inside while loop because at | |
| * higher frequencies shifting stops before it is called the first time. |
| nrf_vpr_csr_vio_in_buffered_reversed_byte_get(); | ||
| nrf_vpr_csr_vio_shift_ctrl_buffered_set(&rx_shift_ctrl); | ||
|
|
||
| for (uint32_t i = 0; i < word_count - 1; i++) { |
There was a problem hiding this comment.
You could start the for loop from i = 2, then index variable is not needed.
Modified hrt rx to work in quad mode up to 21.333333MHz.