Staging/koror support by stefpopa · Pull Request #3094 · analogdevicesinc/linux

stefpopa · 2026-01-27T13:06:36Z

PR Description

This PR adds initial Linux kernel driver support for the Analog Devices ADRV904X (Koror) family of wideband transceivers.

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 · 2026-01-27T13:33:13Z

Can drivers/iio/adc/adrv904x/initdata.c and drivers/iio/adc/adrv904x/initdata.h are API files right?
can they placed under any of the rules below?

local api_match="(public/include|public/src|private/include|private/src|navassa/common|adrv902x/common|adrv903x/common|adrv903x/platforms)/"

the rules are at https://github.com/analogdevicesinc/linux/blob/ci/ci/build.sh#L35
feel free to adjust if needed, but focus on 'containerizing' the API code.

The license.pdf is a no no, please remove

All files should have unix endings, even api ones, csv, etc, use dos2unix to convert

I will now take a look on the build_gcc_arm

gastmaier · 2026-01-27T14:51:30Z

The api is generating calls to __aeabi_uldivmod(), only compatible with arm64 due to u64 operations.
please check if you can do it in 32 bits, divide first, find gcd (lib/math/gcd.c), etc.
and provide back to the api unit
If not possible, please isolate the 64bit bits into defines, so we don't have to blacklist the whole driver as 'arm64-compatible only'

thanks!

gastmaier · 2026-01-27T15:20:30Z

0001-API-dirty-adrv904x-use-math64-and-gcd.patch
Attached should do the trick, please read commit message body.

stefpopa · 2026-01-28T14:32:03Z

@gastmaier thank you for your review. See the changelog below:

Moved initdata.c/initdata.h to devices/adrv904x/private/src/ per API directory structure requirements
Updated Makefile with new paths and include directory
Updated adrv904x.c include path (now resolved via Makefile)
Removed License.pdf
Fixed CRLF line endings in initdata.c
added your arm32 build patch which to which I added a platform abstraction layer in adi_library_types.h, as per a request made by Sean Harte
Added kernel math64/gcd includes
Created ADI_LIBRARY_DIV_* and ADI_LIBRARY_GCD wrapper macros

gastmaier · 2026-01-28T17:17:19Z

Hi,

is adrv904x/platforms API right? so I added to the
https://github.com/analogdevicesinc/linux/blob/ci/ci/build.sh#L35
as a broad

"(public/include|public/src|private/include|private/src|iio/adc/.*/common|iio/adc/.*/platforms)/"

rule

gastmaier · 2026-01-28T17:24:53Z

Adding new comment to fixup my last comment, you did fix the line endings at
fa6d679

but checkpatch checks per commit

What you can do, if you don't want to force push, is to create a fixup commit with

git commit -s --fixup 7c27e7ba145863d1d069fd9bf8b1d34c5d890511 -m "optional extra msg"

Those commits are intentionally auto squashed during the ci
https://git-scm.com/docs/git-rebase#Documentation/git-rebase.txt---autosquash

stefpopa · 2026-01-29T08:35:35Z

Hi,

is adrv904x/platforms API right? so I added to the https://github.com/analogdevicesinc/linux/blob/ci/ci/build.sh#L35 as a broad
"(public/include|public/src|private/include|private/src|iio/adc/.*/common|iio/adc/.*/platforms)/"
rule

Yes, adrv904x/platforms contains the platform abstraction layer (adi_library_types.h, adi_library.c). Thanks for adding the board rule!

stefpopa · 2026-02-03T09:20:31Z

@gastmaier Thank you again for the review! Could you take another look and let me know if there are any remaining topics?
@nunojsa I would love to have your opinion on this PR.

stefpopa · 2026-02-10T11:53:34Z

Friendly reminder: please don't forget to review this PR.
Thank you!

gastmaier

Do we have documentation on the dt bindings .yaml? How will you document how to use?

Please resolve the clang/gcc analyzer caught issues, they are not static analyzer and are mostly always correct, either directly correct, or surface deeper structure issues.

The issues are, for convenience, shown as annotations, in the changed files; you can also now filter the commits shows in the gui, so de-select the api and fw commits, and focus on your files.

Thanks!

gastmaier · 2026-02-12T11:55:27Z

drivers/iio/adc/adrv904x/adrv904x.c

+					break;
+				}
+
+				ret = adrv904x_gainindex_to_gain(phy, chan->channel,


this ret value is never read.
should you error check? or discard explicitly

Same in all locations

gastmaier · 2026-02-12T11:57:29Z

drivers/iio/adc/adrv904x/adrv904x.c

+	adrv904x_add_debugfs_entry(phy, "bist_tone", DBGFS_BIST_TONE);
+
+	for (i = 0; i < phy->adrv904x_debugfs_entry_index; i++)
+		d = debugfs_create_file(phy->debugfs_entry[i].propname, 0644,


also d is never read.

gastmaier · 2026-02-12T11:58:20Z

drivers/iio/adc/adrv904x/adrv904x.c

+	case OBS_SAMPL_CLK:
+		phy->orx_iqRate_kHz = phy->kororDevice->initExtract.jesdSetting.framerSetting[1].iqRate_kHz;
+		init.ops = &bb_clk_ops;
+		clk_priv->rate = phy->orx_iqRate_kHz;;


clean-up double column

gastmaier · 2026-02-12T12:03:04Z

drivers/iio/adc/adrv904x/adrv904x.c

+		adrv904x_TxLinkSamplingRateFind(phy->kororDevice,
+						ADI_ADRV904X_DEFRAMER_0,
+						&rate);
+		init.ops = &bb_clk_ops;
+		clk_priv->rate = rate;


Error checking before storing value? Does rate really get a value on all branches?
If deframerSel selects more than one, it never writes iqRate_kHz.
But in general just always error check.
do you have a wrapper to convert recoveryAction like device->common.errPtr->errDebugInfo.highestPriorityAction, in a more palatable format?

stefpopa · 2026-02-12T15:44:36Z

changelog v3:
Resolved clang/gcc analyzer issues in adrv904x.c:

Fixed unused/overwritten values:

Changed ret = -EINVAL; to return -EINVAL; (value was never returned)
Added error check for adi_adrv904x_HwReset() return value before overwriting with PreMcsInit() call
Removed unused return value assignment from adrv904x_gainindex_to_gain()
Fixed uninitialized rate variable by adding error check for adrv904x_TxLinkSamplingRateFind()

Fixed unused variables:

Removed unused variable d from debugfs_create_file() call
Added __maybe_unused to clk_priv (used only in dev_dbg)
Added __maybe_unused to dev variables (used only in dev_dbg calls)

Fixed syntax issues:

Removed double semicolon in orx_iqRate_kHz assignment

Added dt-bindings documentation:

Created Documentation/devicetree/bindings/iio/adc/adi,adrv9040.yaml

gastmaier · 2026-02-13T13:04:32Z

For the

gcc_fanalayzer: the frame size of 1088 bytes is larger than 1024 bytes

warnings,
could the api providers kindly allocate the buffers instead of putting in the stack?

Add Koror API 2.15.0.5. Signed-off-by: George Mois <george.mois@analog.com>

Add binaries to be used with the ADRV904X driver. New adrv904x firmware to reflect the two available use cases, without ORX (default profile) and with ORX (NLS profile) Signed-off-by: George Mois <george.mois@analog.com> Signed-off-by: Andrei Dragomir <andrei.dragomir@analog.com> Signed-off-by: Stefan Popa <stefan.popa@analog.com>

stefpopa · 2026-02-23T16:38:18Z

v4

Fix MODULE_LICENSE string: "GPL v2" → "GPL"
Fix spurious trailing semicolons after closing braces in four JESD204 state handlers
Replace pr_err with dev_err/dev_dbg in JESD204 device init and link init callbacks
Replace dev_err with dev_dbg for per-link status logging in clks_enable
Replace bare -1 return codes with proper errno values (-ENOMEM, -EIO)
Fix devm_clk_get error path: propagate PTR_ERR instead of silently ignoring it
Check return value of adrv904x_gainindex_to_gain() in read_raw
Remove variable shadowing of ret inside TX_SAMPL_CLK case block
Migrate of_property_read_string to device_property_read_string for firmware-agnostic property access
Add struct adrv904x_chip_info with per-compatible .data in of_device_id table; use device_get_match_data() instead of spi_get_device_id()->driver_data
Migrate clock provider: devm_clk_register → devm_clk_hw_register, of_clk_add_provider → devm_of_clk_add_hw_provider with clk_hw_onecell_data; remove manual of_clk_del_provider calls from error paths and remove()
Remove redundant indio_dev->dev.parent assignment and simplify indio_dev->name
Remove switch on device ID — only one device variant exists
Fix IS_ERR_OR_NULL checks and EXPORT_SYMBOL placement in adrv904x_conv.c
Rename indexed chip info table to single named adrv9040_axiadc_chip_info in adrv904x_conv.c
Replace struct clk_onecell_data with struct clk_hw_onecell_data * in adrv904x_rf_phy; replace int spi_device_id with const struct adrv904x_chip_info *chip_info

Add the ADRV9040 initial driver including: - Core driver implementation with JESD204 support - IIO channel definitions and read/write operations - Clock framework integration - Linux platform abstraction layer Add support for observation receiver (ORX) channels including: - Add OBS_SAMPL_CLK clock for ORX sampling rate - Differentiate between RX and OBS_RX channels in read_raw - Add axi-adrv904x-obs compatible string support Signed-off-by: George Mois <george.mois@analog.com> Signed-off-by: Stefan Popa <stefan.popa@analog.com>

Add dts file for adrv904x (Koror). Signed-off-by: George Mois <george.mois@analog.com>

New adrv904x dts to implement the use case with ORX in NLS mode Signed-off-by: AndrDragomir <andrei.dragomir@analog.com>

Direct 64-bit division on ARM32 causes undefined references to __aeabi_uldivmod and __aeabi_ldivmod. Add platform abstraction wrappers in adi_library_types.h for division operations and use them throughout the driver. Use GCD reduction before squaring in scale_with_squared_ratio() to prevent overflow in power threshold calculations. Co-authored-by: Jorge Marques <jorge.marques@analog.com> Signed-off-by: Stefan Popa <stefan.popa@analog.com>

Signed-off-by: Stefan Popa <stefan.popa@analog.com>

Add device tree bindings documentation for the Analog Devices ADRV904X RF transceiver family. The ADRV904X is a highly integrated RF transceiver supporting 8 transmitter and 8 receiver channels with observation receivers, designed for cellular infrastructure applications. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

Disable VLA and frame-size warnings for ADI vendor API files that trigger build failures when compiled with -Werror. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

…e_dpd The prototype for adi_adrv904x_WBBufSegConfigGet() in adi_adrv904x_dfe_dpd.h is conditionally compiled under in adi_adrv904x_dfe_dpd.c lacks the same guard. When ADI_LIBRARY_RM_FLOATS is defined the compiler sees no prior declaration for the function and emits -Wmissing-prototypes, which is promoted to an error by -Werror, breaking the build. Wrap the function definition with the same guard as its prototype. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

Fix variable length array compilation errors in ADI vendor API code that trigger build failures with -Werror=vla. The VLAs were caused by using const local variables as array dimensions in RxGainTableLoad and RxGainTableChecksumRead, which the compiler cannot evaluate as constant expressions in C90 mode. Replace the local const variables LINE_BUFFER_SIZE and HEADER_BUFFER_SIZE with file-scope #define macros so the array sizes are true compile-time constants. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

adrv904x_dfe_vswr.c defines adrv904x_VswrPlaybackDatNumOfSamplesInit but does not include its own header adrv904x_dfe_vswr.h, where the function prototype is declared. This causes a -Werror=missing-prototypes build failure since the compiler sees the definition before any declaration. Add the missing self-include to fix the build error. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

stefpopa requested review from CiprianRegus, amiclaus, btogorean, buha, chegbeli, danmois, dbogdan, machschmitt, mhennerich, nunojsa and rbolboac as code owners January 27, 2026 13:06

stefpopa force-pushed the staging/koror_support branch 3 times, most recently from 3c1012d to a37e455 Compare January 28, 2026 13:49

stefpopa force-pushed the staging/koror_support branch from 5fb6eed to fa6d679 Compare January 28, 2026 15:26

stefpopa force-pushed the staging/koror_support branch from fa6d679 to 1405cc2 Compare January 29, 2026 08:29

gastmaier requested changes Feb 12, 2026

View reviewed changes

stefpopa force-pushed the staging/koror_support branch from 1405cc2 to db3c42e Compare February 12, 2026 15:41

stefpopa force-pushed the staging/koror_support branch 2 times, most recently from ef097a1 to 3e64a1d Compare February 13, 2026 11:02

stefpopa marked this pull request as draft February 23, 2026 08:19

danmois added 2 commits February 23, 2026 14:29

iio: adc: adrv904x: Add Koror API

7a4b9e5

Add Koror API 2.15.0.5. Signed-off-by: George Mois <george.mois@analog.com>

stefpopa force-pushed the staging/koror_support branch 2 times, most recently from d5f3f58 to d1c095e Compare February 23, 2026 16:37

stefpopa force-pushed the staging/koror_support branch 4 times, most recently from 545c194 to 2e6cda5 Compare February 25, 2026 10:05

danmois and others added 10 commits February 25, 2026 12:50

arm64: dts: Add zynqmp-zcu102-rev10-adrv904x.dts

bbca5a1

Add dts file for adrv904x (Koror). Signed-off-by: George Mois <george.mois@analog.com>

arm64: dts: Add zynqmp-zcu102-rev10-adrv904x-nls.dts

3100b5d

New adrv904x dts to implement the use case with ORX in NLS mode Signed-off-by: AndrDragomir <andrei.dragomir@analog.com>

fixup! firmware: Add ADRV9040 firmware

b1480fb

Signed-off-by: Stefan Popa <stefan.popa@analog.com>

iio: adc: adrv904x: Suppress build warnings for vendor API files

01b2a55

Disable VLA and frame-size warnings for ADI vendor API files that trigger build failures when compiled with -Werror. Signed-off-by: Stefan Popa <stefan.popa@analog.com>

stefpopa force-pushed the staging/koror_support branch from 2e6cda5 to 8f0c2f0 Compare February 25, 2026 10:51

Conversation

stefpopa commented Jan 27, 2026

PR Description

PR Type

PR Checklist

Uh oh!

gastmaier commented Jan 27, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gastmaier commented Jan 27, 2026

Uh oh!

gastmaier commented Jan 27, 2026

Uh oh!

stefpopa commented Jan 28, 2026

Uh oh!

gastmaier commented Jan 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

gastmaier commented Jan 28, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

stefpopa commented Jan 29, 2026

Uh oh!

stefpopa commented Feb 3, 2026

Uh oh!

stefpopa commented Feb 10, 2026

Uh oh!

gastmaier left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gastmaier Feb 12, 2026

Choose a reason for hiding this comment

Uh oh!

gastmaier Feb 12, 2026

Choose a reason for hiding this comment

Uh oh!

gastmaier Feb 12, 2026

Choose a reason for hiding this comment

Uh oh!

gastmaier Feb 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

stefpopa commented Feb 12, 2026

Uh oh!

gastmaier commented Feb 13, 2026

Uh oh!

stefpopa commented Feb 23, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

gastmaier commented Jan 27, 2026 •

edited

Loading

gastmaier commented Jan 28, 2026 •

edited

Loading

gastmaier commented Jan 28, 2026 •

edited

Loading

gastmaier left a comment •

edited

Loading

gastmaier Feb 12, 2026 •

edited

Loading