Skip to content

hwmon: (pmbus/ltc2978) Add support for LT7175, LT7176P, LTM4681, LTM4682 - #3596

Open
cdeguzm4 wants to merge 6 commits into
analogdevicesinc:mainfrom
cdeguzm4:pmbus-lt7175-ltm4681
Open

cdeguzm4 wants to merge 6 commits into
analogdevicesinc:mainfrom
cdeguzm4:pmbus-lt7175-ltm4681

Conversation

@cdeguzm4

@cdeguzm4 cdeguzm4 commented Oct 9, 2026 •

Copy link
Copy Markdown

PR Description

Adds support for the following chips to the existing ltc2978 PMBus driver:

  • LT7175 / LT7175-1 (part of LT7170 family, MFR_SPECIAL_ID 0x1C1D masked to 0x1C10)
  • LT7176P / LT7176P-1 (same family, same MFR_SPECIAL_ID)
  • LTM4681 — quad 31.25A / single 125A µModule step-down regulator (commits inherited from reviewed PR Dev/ltm4681 add LTM4681 and LTM4682 #3085, now owned by this PR)
  • LTM4682 — same architecture as LTM4681 (commits inherited from reviewed PR Dev/ltm4681 add LTM4681 and LTM4682 #3085, now owned by this PR)

Changes include:

  • DT bindings YAML updates (Documentation/devicetree/bindings/hwmon/lltc,ltc2978.yaml)
  • Driver code additions in drivers/hwmon/pmbus/ltc2978.c (chip IDs, ltc2978_get_id() string matches, probe switch cases, i2c_id/of_match tables)
  • Driver documentation updates (Documentation/hwmon/ltc2978.rst)
  • Kconfig help text update

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

@github-actions

This comment was marked as low quality.

@github-actions

This comment was marked as outdated.

Comment thread drivers/hwmon/pmbus/ltc2978.c Resolved
@cdeguzm4
cdeguzm4 force-pushed the pmbus-lt7175-ltm4681 branch from 536bd67 to a13f21a Compare October 15, 2026 05:08
@cdeguzm4

cdeguzm4 commented Oct 15, 2026 •

Copy link
Copy Markdown
Author

Changelog:

  • Fixed the LTM4681_ID, from 0x4140 to 0x4150
  • Removed wrong comment in LTC3880_MFR_TEMPERATURE2_PEAK line
  • Added missing case ltm4681: in probe switch
  • Corrected typo "TM4686" and added LT7175/LT7176P help text in Kconfig
  • Updated commit authorship

@cdeguzm4
cdeguzm4 force-pushed the pmbus-lt7175-ltm4681 branch from a13f21a to 3548c88 Compare October 15, 2026 08:08
@cdeguzm4

Copy link
Copy Markdown
Author

Hello @gastmaier, I would appreciate if you could take a look at changes again. Thank you for the review.

gastmaier

This comment was marked as outdated.

@cdeguzm4
cdeguzm4 force-pushed the pmbus-lt7175-ltm4681 branch from 3548c88 to 6f1f438 Compare October 16, 2026 10:57
@cdeguzm4

Copy link
Copy Markdown
Author

Simply noting that this code is tested in the actual hardware (EVAL-LT7175-AZ):

  • I²C address 0x4F detected on Pi 3B i2c-1 bus
  • Driver probes successfully, chip identified as lt7175 via PMBUS_IC_DEVICE_ID string match
  • hwmon2 registered with name "lt7175"
  • Telemetry readings (all via IEEE754):
  • VIN: 3292 mV (PSU: 3.3V)
  • VOUT: 499 mV (configured default)
  • IOUT: -53 mA (no load, within offset tolerance)
  • temp1: 23593 mdegreeC (room temperature)

@gastmaier gastmaier removed the llm review Request a review from a LLM Reviewer label Oct 21, 2026
@github-actions

Copy link
Copy Markdown
Contributor

LLM review

This series adds support for LTM4681/LTM4682 and LT7175/LT7176P to hwmon/pmbus/ltc2978. Most of the findings from the previous round are fixed. One functional issue remains.

Session: 39574341045

396a602facb0 - Add support for LTM4681 and LTM4682

LTM4681 decoded as LTM4680: ltc2978_get_id() has chip_id == LTM4680_ID || chip_id == LTM4681_ID returning ltm4680. As a result:

  • Every lltc,ltm4681 device logs Device mismatch: Configured ltm4681 ..., detected ....
  • The new case ltm4681: is only reachable through the MFR_MODEL fallback.

It should return ltm4681. The two extended register comments also go past 80 columns.

The Kconfig TM4686 typo is introduced here and only fixed in 54457c633040. There's no patch for it because a fixup would conflict with the later commit's hunk.

015cab93a4d5 - Document LTM4681 and LTM4682 support

The wrapped temperature. line is indented with 8 spaces instead of tabs. Docutils then parses it as a new row of the attribute table, which splits the temp[N]_input entry. Also, LTM4682, LTM4681 in in[N]_label is out of order.

54457c633040 - Add support for LT7175 and LT7176P

The case lt7175: case lt7176p: probe block is the same as the lt7170/lt7171 block. Add the new IDs as case labels on the existing block instead of copying it. The enum chips line is now 94 columns.

aed355b7f16d - dt-bindings

The binding itself (vout0 - vout1) is correct: each I2C address exposes two PAGEs. The commit message says "outputs four channels", which needs that clarification. "Unlike other LTM46xx" is also inaccurate, since the LTM4644 is quad too.

CI warnings

None that matter. The annotations are informational only, and checkpatch --strict reports nothing for any commit.

Verification data

ltm4680.md, ltm4681.md, ltm4682.md and lt7176_lt7176-1.md came from the docling corpus. I used them to check:

  • MFR_SPECIAL_ID values: 0x414X, 0x415X, 0x418X and 0x1C1D.
  • Two pages per address.
  • That the LTM4680 feature set covers the new parts (READ_IIN, READ_PIN, 0xE1, 0xE3, 0xF4, L16 VOUT).

No LT7175/LT7176P datasheet is in the corpus. For those parts I relied on the author's preliminary datasheets and the hardware test results posted in this PR.

Validation steps:

  • Build: x86 with adi_ci_defconfig + auto_set_kconfig + regulator support, W=1. It builds cleanly at the head and after autosquashing the fixups, for every commit.
  • Bindings: dt_binding_check passes on the binding.
  • rst: the table split was confirmed by rendering the file with docutils.

Suggested patches

  • 0001-fixup-hwmon-pmbus-ltc2978-Add-support-for-LTM4681-an.patch (396a602facb0): returns ltm4681 for LTM4681_ID and wraps the comments.
  • 0002-fixup-hwmon-ltc2978-Document-LTM4681-and-LTM4682-sup.patch (015cab93a4d5): fixes the rst indentation and the list order.
  • 0003-fixup-hwmon-pmbus-ltc2978-Add-support-for-LT7175-and.patch (54457c633040): merges the duplicate probe case and rewraps the enum.

Apply with:

cd path/to/repository
export GITHUB_TOKEN=ghp_***
apply-patches --repo=analogdevicesinc/linux 39574341045
Install instructions

The following one-liner installs the script if not present already:

curl -fSsL "https://raw.githubusercontent.com/analogdevicesinc/doctools/refs/heads/main/ci/scripts/apply-patches.sh"      -o ~/.local/bin/apply-patches.sh &&   grep -q "/apply-patches.sh" ~/.bashrc || echo "source ~/.local/bin/apply-patches.sh" >> $_ ; . $_

More information at AI Usage.

@gastmaier

Copy link
Copy Markdown
Collaborator

The Kconfig TM4686 typo is introduced here and only fixed in 54457c633040.

please rewrite history fixing the type at the commit it added.

the proposed patches seems pertinent

Comment thread Documentation/hwmon/ltc2978.rst Resolved
Comment thread Documentation/hwmon/ltc2978.rst Resolved
@cdeguzm4

Copy link
Copy Markdown
Author

Add LTM4681 and LTM4682 to the supported devices of LTC2978. Both are
quad-output µModule regulators. Each I2C address exposes two PMBus
pages, so the regulators are vout0 - vout1.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
Add support for LTM4681 and LTM4682. The LTM4681 and LTM4682 are quad
31.25A and single 125A, switching mode DC/DC step-down µModule regulators
integrated with 4-channel power system manager.

This adds the chip IDs, the checks for the manufacturer special ID, the
switch case for pmbus_driver_info population, and the relevant attributes.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
Document support for the LTM4681 and LTM4682 PMBus µModule regulators
added in the previous commit.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
Document the compatible strings for the LT7175/LT7175-1 and
LT7176P/LT7176P-1 PMBus buck regulators newly supported by the
ltc2978 driver.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
Add support for LT7175/LT7175-1 and LT7176P/LT7176P-1 PMBus buck
regulators. These parts share the LT7170 family MFR_SPECIAL_ID
(0x1C1D masked to 0x1C10) and use the same PMBus register layout,
so they reuse the existing lt7170/lt7171 probe path. Variants are
distinguished by reading the PMBUS_IC_DEVICE_ID ASCII string.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
Document support for the LT7175/LT7175-1 and LT7176P/LT7176P-1 PMBus
buck regulators added in the previous commit.

Signed-off-by: Christopher de Guzman <christopher.deguzman@analog.com>
@cdeguzm4
cdeguzm4 force-pushed the pmbus-lt7175-ltm4681 branch from 6f1f438 to 50c237d Compare November 16, 2026 16:58
@cdeguzm4

Copy link
Copy Markdown
Author

Changelog (rebased on current main):

@gastmaier's review

  • Fixed the rst indentation at the lines you pointed out. The LT7170, LT7171 lines under in[N]_label, temp[N]_input and curr[N]_label had spaces before the tabs. This is fixed in "Document LTM4681 and LTM4682 support", so "Document LT7175 and LT7176P support" now only adds the new names.

LLM review (run 39574341045)

  • LTM4681 is no longer detected as LTM4680: LTM4681_ID now returns ltm4681.
  • Wrapped the two register comments that went past 80 columns.
  • The TM4686 typo is now fixed in the commit that introduced it, so no later commit has to correct it.
  • Fixed the wrapped temperature. line in the rst, which was indented with spaces. temp[N]_input renders as one table row again (checked with docutils). I also fixed the LTM4681, LTM4682 ordering under in[N]_label.
  • Merged case lt7175:/case lt7176p: into the existing lt7170/lt7171 probe block, and rewrapped the enum chips line to 80 columns.
  • Reworded the LTM4681/LTM4682 dt-binding commit message: both are quad-output parts, and each I2C address exposes two PMBus pages (vout0 - vout1).

Checks

  • Every commit builds with W=1, with no warnings.
  • checkpatch --strict is clean on all 6 commits.
  • dt_binding_check passes for lltc,ltc2978.yaml.

@gastmaier, could you please take another look? Thank you!

This branch has not been deployed

No deployments
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants