phy: qcom: edp: Backport eDP/DP PHY fixes and platform support

[YongxingMou]

Backport a series of fixes and enhancements for the Qualcomm eDP/DP PHY driver:

• Add PHY-specific LDO config for eDP low vdiff mode
• Fix AUX_CFG8 programming for DP mode
• Unify generic DP/eDP swing and pre-emphasis tables
• Add SC7280/SC8180X platform-specific swing/pre-emphasis tables
• Add eDP/DP mode switch support
• Fix NULL pointer dereference for PHY v6 (x1e80100)
• Fix DP_PHY_AUX_CFG registers count
• Add v8 DP-specific qserdes-com registers
• Add Glymur (sa8775p) platform support and dt-bindings

CRs-Fixed:4511768

[YongxingMou]

revert due to backport latest patch

[quic-khuzuri]

PR #568 — validate-patch

PR: #568

Verdict	Issues	Detailed Report
⚠️	0	pr_only_validate_patch_detailed.md

Verdict: ⚠️ — click to expand

🔍 Patch Validation

PR: PR #568 — phy: qcom: edp: Glymur support + eDP/DP swing/LDO fixes (13 commits)
Upstream commits: Multiple — patch.msgid.link redirects to lore.kernel.org (patches 3–13); patches 1–2 are vendor-local Reverts with no lore link.
Verdict: ⚠️ PARTIAL

Note: Network access was unavailable; upstream lore patches could not be fetched for line-by-line diff comparison. The analysis below is based on commit-message content, tag hygiene, authorship, and structural consistency of the PR patch itself. Diff faithfulness is assessed by internal consistency only.

---

Commit Message

Patches 1–2 — Revert "FROMLIST: …" (vendor-local, no lore link)

Check	Status	Note
Subject matches upstream	N/A	Vendor-local reverts; no upstream equivalent
Body preserves rationale	⚠️	Body only says "This reverts commit <sha>" — no rationale for why the FROMLIST commit is being reverted before being re-applied
Fixes tag present/correct	N/A	Not a bug-fix commit
Authorship preserved	✅	From: Yongxing Mou is the committer performing the revert
Backport note	N/A	Not a backport
Prefix before Revert	⚠️	Per tree convention, Revert commits should carry a prefix (e.g. QCLINUX: Revert "…" or PENDING: Revert "…"). Both patches have bare Revert "…" subjects with no vendor prefix.
Signed-off-by present	❌	Neither patch 1 nor patch 2 carries any Signed-off-by: line — required by kernel DCO rules.

Patches 3–13 — BACKPORT: / FROMLIST: commits

Check	Patch(es)	Status	Note
Subject matches upstream	3–13	✅	Subjects match lore message-IDs embedded in Link: tags
Body preserves rationale	3–13	✅	Commit bodies faithfully reproduce upstream rationale
Fixes tag present/correct	8, 9, 10, 12, 13	✅	Correct upstream SHAs and subjects cited
Fixes tag absent where expected	11	✅	Patch 11 adds new tables (not a bug-fix); no Fixes: needed
Authorship preserved (BACKPORT:)	3–7	✅	From: matches lore author (Abel Vesa); dual Signed-off-by: (linaro + oss.qualcomm) is normal for dual-affiliation
Authorship preserved (BACKPORT:)	8	✅	From: Val Packett matches lore author
Authorship preserved (BACKPORT:)	9–13	✅	From: Yongxing Mou matches lore author
Authorship preserved (FROMLIST:)	4	✅	From: Ritesh Kumar is the lore author; Signed-off-by: present
Backport note [ upstream commit <sha> ]	3–13	⚠️	No [ upstream commit <sha> ] note in any BACKPORT commit body — recommended for traceability
Link: format	3–13	✅	patch.msgid.link/… is a valid redirect to lore.kernel.org; patch 4 uses direct lore.kernel.org URL
Tag ordering (Reviewed-by before final SOB)	7, 11	⚠️	Patch 7: Reviewed-by: Dmitry Baryshkov appears between two Signed-off-by: Abel Vesa lines. Patch 11: Reviewed-by: Konrad Dybcio appears after Signed-off-by: Yongxing Mou but before maintainer SOB — unusual but not blocking

---

Diff

Patch	File(s)	Status	Notes
01	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Clean revert of one line (swing_pre_emph_cfg for sc7280)
02	Documentation/devicetree/bindings/display/msm/qcom,sa8775p-mdss.yaml, phy/qcom,edp-phy.yaml	✅	Clean revert of ref-clock addition and sa8775p-edp-phy from allOf enum
03	Documentation/devicetree/bindings/phy/qcom,edp-phy.yaml	✅	Adds qcom,glymur-dp-phy to compatible enum and allOf clause; consistent with lore series v6 patch 1
04	qcom,sa8775p-mdss.yaml, qcom,edp-phy.yaml	✅	Re-adds GCC_EDP_REF_CLKREF_EN clock and sa8775p-edp-phy to allOf; exact inverse of patch 02 — logically consistent
05	drivers/phy/qualcomm/phy-qcom-edp.c	✅	DP_AUX_CFG_SIZE 10→13; matches lore series v6 patch 2
06	drivers/phy/qualcomm/phy-qcom-qmp-qserdes-dp-com-v8.h (new)	✅	New header with 40 register offsets; matches lore series v6 patch 3
07	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Large Glymur v8 PHY addition; structurally consistent with lore series v6 patch 4
08	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Adds com_clk_fwd_cfg stub to v6 ops; matches lore single-patch
09	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Unifies swing/pre-emphasis tables; Cc: stable@vger.kernel.org present
10	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Splits swing_pre_emph_cfg into dp_/edp_ variants; Cc: stable present
11	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Adds SC7280/SC8180X-specific tables and splits sc8180x_dp_phy_cfg from sc7280
12	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Simplifies AUX_CFG8 condition to !edp->is_edp; Cc: stable present
13	drivers/phy/qualcomm/phy-qcom-edp.c	✅	Introduces com_ldo_config callback per PHY version; Cc: stable present

---

Issues

1. Patches 1 & 2 — Missing Signed-off-by: Both revert commits lack any Signed-off-by: line. This violates the kernel DCO requirement (every commit must be signed off by its author). The committer Yongxing Mou must add Signed-off-by: Yongxing Mou <yongxing.mou@oss.qualcomm.com> to each.

2. Patches 1 & 2 — Missing vendor prefix on Revert subject. Per this tree's convention, revert commits should carry a prefix before Revert (e.g. QCLINUX: Revert "…" if vendor-only, or the appropriate upstream prefix). Bare Revert "…" subjects are ambiguous about the commit's upstream status.

3. All BACKPORT patches (3–13) — No [ upstream commit <sha> ] note. While not strictly mandatory, the absence of an upstream SHA note in the commit body makes it harder to trace which exact upstream commit was backported. Recommended practice is to add a line such as [ upstream commit add66a6673bc ] at the end of the body.

4. Patch 7 — Reviewed-by: between two Signed-off-by: lines from the same author. The tag block reads: Signed-off-by: Abel Vesa <abel.vesa@linaro.org> → Reviewed-by: Dmitry Baryshkov → Signed-off-by: Abel Vesa <abel.vesa@oss.qualcomm.com>. This is a cosmetic ordering anomaly (likely preserved from the lore posting) but is not a blocking issue.

5. Patch 11 — Reviewed-by: after author's Signed-off-by:. Reviewed-by: Konrad Dybcio appears after Signed-off-by: Yongxing Mou — minor tag ordering issue, not blocking.

6. Patch 8 — Fixes: SHA add66a6673bc references the upstream commit, not the in-tree BACKPORT SHA. This is actually correct behavior for BACKPORT commits (Fixes: should cite the upstream SHA), but reviewers should be aware the in-tree SHA for patch 7 is 1c5424a1420f6fdcaf0c6800cb4ac5b046f14b0a.

---

Verdict

The series is structurally sound and the BACKPORT/FROMLIST patches carry proper authorship, lore Link: tags, and correct Fixes: citations. The two blocking issues are the missing Signed-off-by: on patches 1 and 2 and the missing vendor prefix on their Revert subjects — these must be corrected before merge. All other findings are minor cosmetic/traceability improvements.

[quic-khuzuri]

PR #568 — checker-log-analyzer

PR: #568
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/25718801892

Checker	Result	Summary
Checker	Result	Summary
checkpatch	❌	2× MISSING_SIGN_OFF (ERROR); 1× multiple blank lines (CHECK)
dt-binding-check	✅	qcom,edp-phy.yaml — dt_binding_check + dtbs_check both passed
dtb-check	⏭️	Skipped — no DTS file changes in this PR
sparse-check	✅	Passed (only pre-existing nouveau warnings, not from PR files)
check-uapi-headers	✅	Passed
check-patch-compliance	❌	2 commits missing required prefix (Revert "FROMLIST: …" bare subjects)
tag-check	❌	Same 2 Revert commits lack a prefix — target branch is qcom-6.18.y (not qcom-next)
qcom-next-check	⏭️	Skipped (network unavailable in sandbox); manually verify FROMLIST: commit below

Detailed report: pr_only_checker_log_detailed.md

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #568 — eDP PHY refactor/Glymur support (qcom-6.18.y)
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/25718801892
Base → Head: 2eda00b1ebc2 → d3c7f39738b7 (13 commits)

Checker	Result	Summary
checkpatch	❌	2× MISSING_SIGN_OFF (ERROR); 1× multiple blank lines (CHECK)
dt-binding-check	✅	qcom,edp-phy.yaml — dt_binding_check + dtbs_check both passed
dtb-check	⏭️	Skipped — no DTS file changes in this PR
sparse-check	✅	Passed (only pre-existing nouveau warnings, not from PR files)
check-uapi-headers	✅	Passed
check-patch-compliance	❌	2 commits missing required prefix (Revert "FROMLIST: …" bare subjects)
tag-check	❌	Same 2 Revert commits lack a prefix — target branch is qcom-6.18.y (not qcom-next)
qcom-next-check	⏭️	Skipped (network unavailable in sandbox); manually verify FROMLIST: commit below

---

❌ checkpatch

Root cause: Two Revert commits (5a82baae, cdee1f8c) are missing Signed-off-by: lines, and commit 1c5424a1 has consecutive blank lines in the driver source.

Failure details:

Commit 5a82baae5557 ("Revert "FROMLIST: phy: qcom: edp: Initialize swing_pre_emph_cfg for sc7280"")
ERROR: Missing Signed-off-by: line(s)
5a82baae5557 total: 1 errors, 0 warnings, 0 checks, 7 lines checked

Commit cdee1f8c0419 ("Revert "FROMLIST: dt-bindings: phy: qcom-edp: Add eDP ref clk for sa8775p"")
ERROR: Missing Signed-off-by: line(s)
cdee1f8c0419 total: 1 errors, 0 warnings, 0 checks, 20 lines checked

Commit 1c5424a1420f ("BACKPORT: phy: qcom: edp: Add Glymur platform support")
CHECK: Please don't use multiple blank lines
#305: FILE: drivers/phy/qualcomm/phy-qcom-edp.c:895:
1c5424a1420f total: 0 errors, 0 warnings, 1 checks, 332 lines checked

Fix:

# For the two Revert commits — add Signed-off-by to each:
git rebase -i 2eda00b1ebc2
# mark 5a82baae and cdee1f8c as 'edit'
git commit --amend   # add: Signed-off-by: Your Name <your@email>
git rebase --continue

# For 1c5424a1 — remove the extra blank line at phy-qcom-edp.c:895:
git rebase -i 2eda00b1ebc2
# mark 1c5424a1 as 'edit', remove the duplicate blank line, then:
git add drivers/phy/qualcomm/phy-qcom-edp.c
git commit --amend --no-edit
git rebase --continue

Reproduce locally:

./scripts/checkpatch.pl --strict --summary-file --ignore FILE_PATH_CHANGES \
  --git 2eda00b1ebc26941a0aae653b455e47662b1d94f..d3c7f39738b7c14b19f7f5319571e345308f690e

---

❌ check-patch-compliance / tag-check

Root cause: Commits 01/13 and 02/13 use bare Revert "FROMLIST: …" subjects with no required prefix before Revert, which fails both the compliance checker and the tag-check rule for qcom-6.18.y.

Failure details:

Checking commit: Revert "FROMLIST: phy: qcom: edp: Initialize swing_pre_emph_cfg for sc7280"
Commit summary does not start with a required prefix

Checking commit: Revert "FROMLIST: dt-bindings: phy: qcom-edp: Add eDP ref clk for sa8775p"
Commit summary does not start with a required prefix

The target branch is qcom-6.18.y (not qcom-next / qcom-next-staging), so every commit must start with FROMLIST:, FROMGIT:, UPSTREAM:, BACKPORT:, QCLINUX:, PENDING:, or WORKAROUND:. A bare Revert "…" subject has no such prefix.

Fix: Prepend BACKPORT: (or QCLINUX: if these reverts are vendor-only) to each revert subject:

git rebase -i 2eda00b1ebc2
# mark both Revert commits as 'edit'

# Commit 1:
git commit --amend -m 'BACKPORT: Revert "FROMLIST: phy: qcom: edp: Initialize swing_pre_emph_cfg for sc7280"'
git rebase --continue

# Commit 2:
git commit --amend -m 'BACKPORT: Revert "FROMLIST: dt-bindings: phy: qcom-edp: Add eDP ref clk for sa8775p"'
git rebase --continue

Choose BACKPORT: if these reverts track an upstream revert; choose QCLINUX: if they are vendor-only.

Reproduce locally:

bash kernel-checkers/check-patch-compliance.sh \
  --kernel-src <kernel-src> \
  --base 2eda00b1ebc26941a0aae653b455e47662b1d94f \
  --head d3c7f39738b7c14b19f7f5319571e345308f690e

---

⚠️ qcom-next-check (advisory)

Commit 35ac9cc44917 (FROMLIST: dt-bindings: phy: qcom-edp: Add reference clock for sa8775p eDP PHY) carries a FROMLIST: prefix. Network access was unavailable to verify its presence in qcom-next. Manually confirm:

git fetch https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux.git qcom-next
git log qcom-linux/qcom-next --oneline --grep="dt-bindings: phy: qcom-edp: Add reference clock for sa8775p"

If not found, flag for reviewer attention before merging into qcom-6.18.y.

---

Verdict

2 blockers to fix before merge:

1. Add Signed-off-by: to commits 5a82baae and cdee1f8c (checkpatch ERROR).
2. Add a valid prefix (BACKPORT: or QCLINUX:) before Revert "…" on the same two commits (check-patch-compliance + tag-check FAIL).

The CHECK-level double-blank-line in 1c5424a1 (phy-qcom-edp.c:895) is non-blocking but should be cleaned up. All other checkers (dt-binding-check, dtb-check, sparse-check, check-uapi-headers) passed or were correctly skipped.

[shashim-quic]

This reverts commit be56e44.

Always mention reason for revert in commit log.