[PATCH AUTOSEL 6.18-6.12] platform/x86: thinkpad_acpi: Fix fan speed reporting on Edge E330
From: Sasha Levin
Date: Mon Oct 19 2026 - 10:31:19 EST
From: Andrew Onyshchuk <andryk.rv@xxxxxxxxx>
[ Upstream commit 8d840dad456f88374d8a3b03851452a280d6ee50 ]
The ThinkPad Edge E330 with H3 firmware uses the non-standard EC fan
register block. Without a matching quirk, thinkpad_acpi reads the legacy
tachometer registers and reports 0 RPM.
Add the H3 BIOS family to TPACPI_FAN_NS so the driver reads the fan
period from EC register 0x95 using the existing non-standard reporting
path.
Tested on a ThinkPad Edge E330 with BIOS H3ET77WW and EC H3EC35WW.
Signed-off-by: Andrew Onyshchuk <andryk.rv@xxxxxxxxx>
Reviewed-by: Mark Pearson <mpearson-lenovo@xxxxxxxxx>
Link: https://patch.msgid.link/20260806154417.618575-1-andryk.rv@xxxxxxxxx
Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
LLM Generated explanations, may be completely bogus:
## Local tree
This checkout is **pending-6.12** (`p-6.12-3-gc504cf0baa3d8`), kernel
**6.12.111**. The verdict is only for this tree.
The driver is still `drivers/platform/x86/thinkpad_acpi.c` (not yet
moved to `drivers/platform/x86/lenovo/`). The H3 `TPACPI_FAN_NS` quirk
is **not** present. The non-standard fan path that the quirk selects
**is** present.
---
## PHASE 1: COMMIT MESSAGE FORENSICS
**Step 1.1: PARSE THE SUBJECT LINE**
Record: **[platform/x86 thinkpad_acpi] [Fix] [fan speed reporting on
ThinkPad Edge E330]**
**Step 1.2: PARSE ALL COMMIT MESSAGE TAGS**
Record:
- `Reviewed-by: Mark Pearson <mpearson-lenovo@xxxxxxxxx>` (Lenovo
thinkpad_acpi reviewer)
- `Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>`
(platform/x86 maintainer)
- `Signed-off-by: Andrew Onyshchuk <andryk.rv@xxxxxxxxx>` (author)
- `Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>`
(maintainer)
- `Link:
https://patch.msgid.link/20260806154417.618575-1-andryk.rv@xxxxxxxxx`
- No `Fixes:` tag (expected)
- No `Reported-by:` / `Tested-by:` tags (testing is described in the
body instead)
- Ignore pipeline `Signed-off-by: Sasha Levin` / `[Upstream commit ...]`
if present
**Step 1.3: ANALYZE THE COMMIT BODY TEXT**
Record:
- **Bug:** Edge E330 with H3 firmware uses the non-standard EC fan
register block.
- **Symptom:** without a quirk, the driver reads the legacy tachometer
registers and reports **0 RPM**.
- **Root cause:** missing BIOS-family match in `fan_quirk_table`.
- **Fix:** add `TPACPI_Q_LNV('H', '3', TPACPI_FAN_NS)` so the existing
NS path reads period from EC **0x95**.
- **Version info:** tested on BIOS **H3ET77WW**, EC **H3EC35WW**. No
kernel-version range given.
- **Failure mode:** wrong hwmon/procfs fan speed (always 0), not a
crash.
**Step 1.4: DETECT HIDDEN BUG FIXES**
Record: Not a disguised cleanup. It is an explicit hardware-quirk bug
fix.
---
## PHASE 2: DIFF ANALYSIS
**Step 2.1: INVENTORY THE CHANGES**
Record:
- One file: `drivers/platform/x86/lenovo/thinkpad_acpi.c` in the
submitted patch; local equivalent is
`drivers/platform/x86/thinkpad_acpi.c`.
- **+1 line**, 0 removed.
- Function/data: `fan_quirk_table[]`.
- Scope: **single-file surgical hardware quirk**.
**Step 2.2: UNDERSTAND THE CODE FLOW CHANGE**
Record:
- **Before:** `tpacpi_check_quirks(fan_quirk_table)` does not match BIOS
family `H3`. `fan_init()` uses `TPACPI_FAN_RD_TPEC` and reads
`fan_rpm_offset` **0x84/0x85**. On this EC those registers are unused,
so speed is 0.
- **After:** BIOS `H3` matches `TPACPI_FAN_NS`. `fan_init()` sets
`fan_with_ns_addr`, selects `TPACPI_FAN_RD_TPEC_NS`, and
`fan_get_speed()` reads `fan_rpm_status_ns` (**0x95**) and converts
period to RPM via `FAN_RPM_CAL_CONST / lo`.
Verified in this tree:
```8156:8162:drivers/platform/x86/thinkpad_acpi.c
case TPACPI_FAN_RD_TPEC_NS:
if (!acpi_ec_read(fan_rpm_status_ns, &lo))
return -EIO;
if (speed)
*speed = lo ? FAN_RPM_CAL_CONST / lo : 0;
break;
```
**Step 2.3: IDENTIFY THE BUG MECHANISM**
Record: **(h) Hardware workaround** — quirk-table entry enabling an
existing EC register path. Same class as the X120e TPR and L13/X13 Yoga
NS fan quirks already in this file.
**Step 2.4: ASSESS THE FIX QUALITY**
Record:
- Obviously correct: one table entry, existing matcher, existing NS
reader.
- BIOS matching is right: `H3ET77WW` is the 8-char `xxyTkkWW` form, so
`tpacpi_parse_fw_id()` stores `TPID('H','3')`, which is exactly what
`TPACPI_Q_LNV('H','3', ...)` matches (`ec = TPACPI_MATCH_ANY`).
- Minimal; no API/data-structure change.
- Regression risk is limited to H3-family machines. FAN_NS also marks
host fan control as undefined (`fan_ctrl_status_undef = 1`) and uses
NS status register 0x93. That is the same behavior already used for
L13/X13 Yoga, L380/L390, 11e Gen5. Mark Pearson noted the platform is
~14 years old and he could not independently confirm; the author
tested it. Overmatch of unrelated H3 BIOS families is possible in
theory but H3ET is the E330/E335 firmware family.
---
## PHASE 3: GIT HISTORY INVESTIGATION
**Step 3.1: BLAME THE CHANGED LINES**
Record: The insertion sits among NS quirks added by:
- `66e92e23a7276` (2023-12-14) Vishnu Sankar — introduced
`TPACPI_FAN_NS` (L13 Yoga Gen 2, X13 Yoga Gen 2)
- `a6dcd3f1725e0` (2024-02-29) Vishnu Sankar — more NS models (L380,
L390, 11e Gen5, …)
Both **are ancestors of this HEAD**. The E330 bug is the missing table
row, not a newly introduced regression. This machine has been mis-
reported for as long as the default 0x84/0x85 path has been used.
**Step 3.2: FOLLOW THE FIXES: TAG**
Record: No `Fixes:` tag. Not applicable.
**Step 3.3: CHECK FILE HISTORY FOR RELATED CHANGES**
Record: Related fan quirks already in this 6.12 tree:
- `f92324135f09e` — X120e TPR fan-speed quirk (stable-style backport
already here)
- `1be765b292577` — L480 `TPACPI_FAN_DECRPM` incorrect RPM
- `a6dcd3f1725e0` / `66e92e23a7276` — NS fan register path
- `48792ab445529` — T495/E560 ACPI fan quirk
Standalone one-line table add; not patch X/Y.
**Step 3.4: CHECK THE AUTHOR'S OTHER COMMITS**
Record: `git log --author='Onyshchuk'` on `drivers/platform/x86/` in
this tree is empty. Author is an outside contributor with Lenovo +
maintainer review, not the subsystem maintainer.
**Step 3.5: CHECK FOR DEPENDENT/PREREQUISITE COMMITS**
Record: Depends only on `TPACPI_FAN_NS` / `fan_rpm_status_ns` /
`TPACPI_Q_LNV`, all present here. Can apply standalone (path tweak
only).
---
## PHASE 4: MAILING LIST AND EXTERNAL RESEARCH
**Step 4.1: FIND THE ORIGINAL PATCH DISCUSSION**
Record:
- `b4 dig -c 051292bf7697e72a760401b6883c55a98beecc5b` failed: SHA is
not in this tree (`Not a valid commit name`).
- `b4 mbox` for msgid `20260806154417.618575-1-andryk.rv@xxxxxxxxx`
returned 404.
- lore.kernel.org fetches returned 403.
- Recovered the thread from Patchew and LKML mirrors:
- https://patchew.org/linux/20260806154417.618575-1-
andryk.rv@xxxxxxxxx/
- https://lkml.iu.edu/2608.0/11369.html
- Single `[PATCH]`, not a series. Applied by Ilpo as
`051292bf7697e72a760401b6883c55a98beecc5b`.
- No NAK. Ilpo noted Linus’s policy that even fixes usually go via for-
next unless they are current-cycle regressions — that is merge-window
routing, not a stable NAK.
- No reviewer asked for `Cc: stable`.
**Step 4.2: CHECK WHO REVIEWED THE PATCH**
Record: Mark Pearson (Lenovo) Reviewed-by; Ilpo Järvinen applied it.
Recipients from Openwall copy: platform-driver-x86, ibm-acpi-devel,
linux-kernel, plus Pearson / Clark / Holschuh / Hans.
**Step 4.3: SEARCH FOR THE BUG REPORT**
Record: No syzbot/bugzilla. Evidence is the author’s hardware test
(H3ET77WW / H3EC35WW). Mark Pearson: platform is ~14 years old; he could
not confirm but was glad the author figured it out.
**Step 4.4: CHECK FOR RELATED PATCHES AND SERIES**
Record: Standalone. Same pattern as earlier NS fan quirk additions.
E330/E335 share the H3ET BIOS family; the 2-char `H3` match is the
driver’s normal family match.
**Step 4.5: CHECK STABLE MAILING LIST HISTORY**
Record: lore.kernel.org/stable search returned 403. No evidence this was
already requested or rejected on stable. UNVERIFIED beyond that.
---
## PHASE 5: CODE SEMANTIC ANALYSIS
semcode MCP is not available; used grep/read.
**Step 5.1: IDENTIFY KEY FUNCTIONS**
Record: No function body change. Touches `fan_quirk_table`, consumed by
`fan_init()`. Downstream: `fan_get_speed()`, `fan_get_status()`,
`fan_read()`, sysfs `fan1_input`.
**Step 5.2: TRACE CALLERS**
Record:
- `fan_get_speed()` → `fan_fan1_input_show()` (hwmon `fan1_input`) and
`fan_read()` (`/proc/acpi/ibm/fan`).
- Userspace: lm-sensors, thinkfan, TLP, etc., on ThinkPads with
`CONFIG_THINKPAD_ACPI`.
- Init-time quirk match only; no IRQ path.
**Step 5.3: TRACE CALLEES**
Record: Existing `acpi_ec_read(0x95)` and RPM conversion. No new
allocation/locking.
**Step 5.4: FOLLOW THE CALL CHAIN**
Record: Reachable from userspace sysfs/proc reads on E330 with
thinkpad_acpi loaded. Unprivileged read of hwmon is typical. Not a
syscall crash path.
**Step 5.5: SEARCH FOR SIMILAR PATTERNS**
Record: Nine existing `TPACPI_FAN_NS` rows in the same table; this is
the tenth. Thermal NS table for the Yoga/11e set is **not** extended
here (E330 was not claimed to need thermal NS).
---
## PHASE 6: CROSS-REFERENCING AGAINST THE LOCAL TREE
**Step 6.1: DOES THE BUGGY CODE EXIST IN THIS TREE?**
Record: Yes. `fan_quirk_table` in `drivers/platform/x86/thinkpad_acpi.c`
(lines 8805–8837) has NS entries through 11e Gen5 and then jumps to
`TPACPI_FAN_NOFAN`. No `'H','3'` row. Default TPEC RPM path is used → 0
RPM on E330. Infrastructure needed for the fix is here since
`66e92e23a7276`.
**Step 6.2: CHECK FOR BACKPORT COMPLICATIONS**
Record: **Minor path adjustment only.** Surrounding context in this tree
matches the hunk exactly (same 11e Gen5 / X1 Tablet / L480 / x120e
neighbors). File move to `lenovo/` happened after 6.12. Expected apply:
change path, otherwise clean.
**Step 6.3: CHECK IF RELATED FIXES ARE ALREADY HERE**
Record: NS path and other model quirks are here. This specific H3/E330
line is not. `git merge-base --is-ancestor 051292bf7697… HEAD` fails
because that SHA is not in this object store.
---
## PHASE 7: SUBSYSTEM AND MAINTAINER CONTEXT
**Step 7.1: IDENTIFY THE SUBSYSTEM AND ITS CRITICALITY**
Record: **platform/x86 thinkpad_acpi** — **PERIPHERAL/IMPORTANT for
ThinkPad users**. Fan sysfs is widely used; not core mm/fs/net.
**Step 7.2: ASSESS SUBSYSTEM ACTIVITY**
Record: Actively maintained; recent 6.12 backports already include other
thinkpad_acpi fan quirks (X120e TPR, T495 NOACPI). Mature driver, long-
lived quirk table.
---
## PHASE 8: IMPACT AND RISK ASSESSMENT
**Step 8.1: DETERMINE WHO IS AFFECTED**
Record: **Driver-specific / DMI-specific.** Only Lenovo machines whose
BIOS parses as family `H3` (E330 H3ET; likely E335 siblings). Everyone
else is unchanged.
**Step 8.2: DETERMINE THE TRIGGER CONDITIONS**
Record: Every read of `fan1_input` or `/proc/acpi/ibm/fan` on those
machines. Common for anyone monitoring fans. Unprivileged sysfs read.
Not a race.
**Step 8.3: DETERMINE THE FAILURE MODE SEVERITY**
Record: Incorrect fan RPM (always 0). **MEDIUM** for affected users
(broken monitoring; thinkfan/sensors may treat the fan as stopped). Not
panic, UAF, or corruption. Still a real hardware bug with a standard
quirk fix.
**Step 8.4: CALCULATE RISK-BENEFIT RATIO**
Record:
- **Benefit:** Makes fan tachometer work on E330 (and H3-family
siblings) on 6.12 LTS. Same class of fix already accepted in this tree
for X120e/L480/Yoga.
- **Risk:** Very low — one line, model-gated, existing code path.
- **Ratio:** Favorable.
---
## PHASE 9: FINAL SYNTHESIS
**Step 9.1: COMPILE THE EVIDENCE**
FOR:
- Real, tested hardware bug: E330 reports 0 RPM.
- Classic stable-eligible **hardware quirk** (one `TPACPI_Q_LNV` row).
- Existing NS path already in 6.12; no new feature/API.
- 1-line, obviously correct; Lenovo + pdx86 review.
- Analogous fan quirks already live in this tree.
- Applies here with a trivial path change.
AGAINST:
- Not crash/security/corruption.
- Old niche laptop; small population.
- Maintainer sent it via for-next, not as a current-cycle regression.
- Author-only test; Pearson could not independently verify.
UNRESOLVED:
- Stable-list discussion (lore 403).
- Whether every H3 BIOS SKU (e.g. E335) truly uses 0x95; family-level
match is the driver’s normal pattern.
**Step 9.2: APPLY THE STABLE RULES CHECKLIST**
1. Obviously correct and tested? **PASS** — 1-line quirk; tested on E330
H3ET77WW; two Reviewed-by.
2. Fixes a real bug that affects users? **PASS** — fan speed stuck at 0.
3. Important issue? **PASS** under the hardware-quirk exception;
severity MEDIUM, not CRITICAL.
4. Small and contained? **PASS** — +1 line, one table.
5. No new features or APIs? **PASS** — enables existing NS path for one
BIOS family.
6. Can apply to the local tree? **PASS** — NS path present; path rename
only.
**Step 9.3: CHECK FOR EXCEPTION CATEGORIES**
Record: **Hardware quirk/workaround** — qualifies even though it adds a
table row.
**Step 9.4: DECISION**
This is the same kind of thinkpad_acpi fan quirk this 6.12 tree already
carries. The bug exists here, the fix is surgical, and the risk is
confined to H3-family ThinkPads.
---
### What the commit solves for this tree
On ThinkPad Edge E330 (BIOS family H3), thinkpad_acpi reads the standard
EC tachometer at 0x84/0x85 and always reports 0 RPM. Adding
`TPACPI_FAN_NS` for `H3` switches that model onto the existing 0x95
period reader so hwmon/`/proc/acpi/ibm/fan` show a real speed. 6.12
already has that reader; it is only missing this DMI row.
### Risk vs benefit
Benefit is correct fan monitoring on those laptops. Risk is a one-line,
model-gated change. Backport needs only retargeting the path from
`lenovo/thinkpad_acpi.c` to `drivers/platform/x86/thinkpad_acpi.c`.
---
### Verification
- [Phase 1] Parsed subject/tags/body from the provided commit message;
Link msgid `20260806154417.618575-1-andryk.rv@xxxxxxxxx`.
- [Phase 2] Diff is +1 `TPACPI_Q_LNV('H','3', TPACPI_FAN_NS)` in
`fan_quirk_table`.
- [Phase 2] Read `fan_get_speed()` NS case (0x95) and `fan_init()` NS
setup in this tree.
- [Phase 2] Read `tpacpi_parse_fw_id()`: `H3ET77WW` → `TPID('H','3')`,
matches `TPACPI_Q_LNV`.
- [Phase 3] `git blame -L 8822,8837`: NS rows from `66e92e23a7276` and
`a6dcd3f1725e0`.
- [Phase 3] `git merge-base --is-ancestor 66e92e23a7276 HEAD` and
`a6dcd3f1725e0 HEAD`: both in tree.
- [Phase 3] `git log --author=Onyshchuk` on platform/x86: no local
commits.
- [Phase 4] `b4 dig`/`b4 mbox`/lore.kernel.org failed (missing SHA / 404
/ 403). Thread recovered from Patchew + lkml.iu.edu + Openwall:
Pearson Reviewed-by, Ilpo applied as `051292bf7697`.
- [Phase 4] Web search: H3ET is E330/E335 BIOS family; author tested
H3ET77WW.
- [Phase 5] semcode unavailable; `grep fan_get_speed` → sysfs
`fan1_input` and `fan_read()`.
- [Phase 6] `git describe HEAD` → `p-6.12-3-gc504cf0baa3d8`; `make
kernelversion` → 6.12.111.
- [Phase 6] Grep: no `'H', '3'` / Edge E330 in local file; NS table ends
at 11e Gen5.
- [Phase 6] File exists only at `drivers/platform/x86/thinkpad_acpi.c`.
- [Phase 6] `051292bf7697` is not a valid object in this repo.
- [Phase 7] `CONFIG_THINKPAD_ACPI` in `drivers/platform/x86/Kconfig`.
- [Phase 8] Failure mode: 0 RPM reporting, not oops.
- UNVERIFIED: stable@ lore discussion (HTTP 403); independent
confirmation on E335.
**YES**
drivers/platform/x86/lenovo/thinkpad_acpi.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/platform/x86/lenovo/thinkpad_acpi.c b/drivers/platform/x86/lenovo/thinkpad_acpi.c
index 075543cd0e77e..f3b2e3c794571 100644
--- a/drivers/platform/x86/lenovo/thinkpad_acpi.c
+++ b/drivers/platform/x86/lenovo/thinkpad_acpi.c
@@ -8854,6 +8854,7 @@ static const struct tpacpi_quirk fan_quirk_table[] __initconst = {
TPACPI_Q_LNV3('R', '0', 'T', TPACPI_FAN_NS), /* 11e Gen5 GL */
TPACPI_Q_LNV3('R', '1', 'D', TPACPI_FAN_NS), /* 11e Gen5 GL-R */
TPACPI_Q_LNV3('R', '0', 'V', TPACPI_FAN_NS), /* 11e Gen5 KL-Y */
+ TPACPI_Q_LNV('H', '3', TPACPI_FAN_NS), /* Edge E330 */
TPACPI_Q_LNV3('N', '1', 'O', TPACPI_FAN_NOFAN), /* X1 Tablet (2nd gen) */
TPACPI_Q_LNV3('R', '0', 'Q', TPACPI_FAN_DECRPM),/* L480 */
TPACPI_Q_LNV('8', 'F', TPACPI_FAN_TPR), /* ThinkPad x120e */
--
2.53.0