Skip to content

iio: trx-rf: ad9088: Support IC13 - #3639

Merged
gastmaier merged 7 commits into
mainfrom
staging/ad9084/ic13
Nov 16, 2026
Merged

gastmaier merged 7 commits into
mainfrom
staging/ad9084/ic13

Conversation

@gastmaier

Copy link
Copy Markdown
Collaborator

PR Description

Resolve quite a few quirks to support IC13 4G IBW; including distinct lane rates on side A and B (known 'fixme'), clock scale for the lane rate itself, and a mapping fix for case 1xNx, where the n-th FDUC wouldn't sum into any CDUC, then silently fallback to CDUC0, which made thing worse, since the NCO wouldn't be reached at all.

PR Type

  • Bug fix (a change that fixes an issue)
  • New feature (a change that adds new functionality)
  • Breaking change (a change that affects other repos or cause CIs to fail)

PR Checklist

  • I have conducted a self-review of my own code changes
  • I have compiled my changes, including the documentation
  • I have tested the changes on the relevant hardware
  • I have updated the documentation outside this repo accordingly
  • I have provided links for the relevant upstream lore

@gastmaier
gastmaier added this pull request to stack #3638 October 30, 2026 18:02
@gastmaier gastmaier added the llm review Request a review from a LLM Reviewer label Oct 30, 2026
@github-actions

This comment was marked as resolved.

@gastmaier
gastmaier force-pushed the staging/ad9084/ic13 branch 2 times, most recently from 0b9b1e5 to a011d85 Compare October 30, 2026 18:38
@gastmaier
gastmaier force-pushed the staging/ad9084/fnco_ffh_profile branch from ffad15a to 5365496 Compare October 30, 2026 18:42
@gastmaier
gastmaier force-pushed the staging/ad9084/ic13 branch from a011d85 to d1e1bb5 Compare October 30, 2026 18:42
@nunojsa
nunojsa force-pushed the staging/ad9084/ic13 branch from d1e1bb5 to 932d997 Compare November 2, 2026 16:58
@gastmaier
gastmaier marked this pull request as ready for review November 3, 2026 10:33
@nunojsa

nunojsa commented Nov 3, 2026

Copy link
Copy Markdown
Collaborator

@gastmaier any special reason to merge this into a staging branch?!

@gastmaier
gastmaier removed this pull request from stack #3638 November 3, 2026 15:19
@gastmaier
gastmaier changed the base branch from staging/ad9084/fnco_ffh_profile to main November 3, 2026 15:19
@gastmaier

gastmaier commented Nov 3, 2026 •

Copy link
Copy Markdown
Collaborator Author

@gastmaier any special reason to merge this into a staging branch?!

nothing in particular, just wanted to use the new stack pr feature 👉👈

@nunojsa nunojsa left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just two minor nits

Comment thread drivers/iio/trx-rf/ad9088/ad9088.c Resolved
dmas = <&rx_dma_b 0>;
dma-names = "rx";
clocks = <&trx0_ad9084 0>; /* FIXME */
clocks = <&trx0_ad9084 2>;

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Could be a different patch given it's a different arch. But no strong feelings

clock-cells takes <&phandle N>, so set from 0 to 1. Also change
dev_clk-clock-scales to "-clock-scales$" pattern to encompass also the
sample clock scales.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
The TX channel map took JESD pair k as FDUC k and looked it up in the
FDUC->CDUC summer, falling back to CDUC0 when no CDUC summed it;
problematic with FDUC interpolation 1 (1xNx).

In 1xNx pair k feeds the k-th FDUC summed into a CDUC, in CDUC order,
not FDUC k. It would also fallback to CDUC0, then labelled it
FDUC1->CDUC0->DAC0, and its CNCO/FNCO attributes programmed CDUC0
(channel 0's) and FDUC1 (unused), so channel 1's NCOs could not be
reached.

In 1xNx on 4T4R, take the k-th summed FDUC. When a channel's FDUC is
still not summed into any CDUC, label it Side-X:FDUCn->unrouted and
return no channel map.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Register rx_sampl_clk_link2 and tx_sampl_clk_link2 (outputs 2 and 3),
with rates from side B's profile and their own clock scales, as ad9081
does for its second link. Allows asymmetric at different rates and
resolves 'fixme' were side A clock rates were used for side B.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Add 204C_8T8R_M8_L2_NP16_8p0_4x8.bin and 204C_M4_L12_NP12_20p0_4x1.bin
to CONFIG_EXTRA_FIRMWARE.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Point the side B TPL cores (rx_b, tx_b) to the side B sample clock
outputs of the converter, 2 and 3, instead of side A's.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Add the VCU118 devicetree for the AD9088 default 8T8R use case
(204C_8T8R_M8_L2_NP16_8p0_4x8: M=8, L=2, NP=16 per side at 16.5 Gbps),
based on the AD9084 one. Needs the asymmetric HDL build, 16.5 Gbps uses
the QPLL1 in full-rate mode (VCO at 8.25 GHz from the 250 MHz
FPGA_REFCLK). The QPLL0 (VCO 9.8 to 16.375 GHz) can't generate it.

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
Add the 204C_M4_L12_NP12_20p0_4x1 use case (IC13: M=4, L=12, NP=12,
20.625 Gbps, DAC 20 GS/s, 4x1), asymmetric.

hdl/projects/ad9084_ebz/vcu118$ make \
  JESD_MODE=64B66B ASYMMETRIC_A_B_MODE=1 \
  RX_LANE_RATE=20.625 TX_LANE_RATE=20.625 \
  RX_B_LANE_RATE=20.625 TX_B_LANE_RATE=20.625 \
  RX_JESD_M=4 TX_JESD_M=4 RX_JESD_L=12 TX_JESD_L=12 \
  RX_JESD_S=2 TX_JESD_S=2 RX_JESD_NP=12 TX_JESD_NP=12 \
  RX_B_JESD_M=4 TX_B_JESD_M=4 RX_B_JESD_L=12 TX_B_JESD_L=12 \
  RX_B_JESD_S=2 TX_B_JESD_S=2 RX_B_JESD_NP=12 TX_B_JESD_NP=12

Signed-off-by: Jorge Marques <jorge.marques@analog.com>
@gastmaier
gastmaier merged commit d292d40 into main Nov 16, 2026
21 checks passed
@gastmaier
gastmaier deleted the staging/ad9084/ic13 branch November 16, 2026 16:48
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

llm review Request a review from a LLM Reviewer

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants