PCI: qcom: Add register dump support for PCIe Link Down events

[ziyuezhang-123]

When the PCIe link goes down unexpectedly, being able to inspect the
state of key controller registers at the time of failure is valuable
for root-causing the issue.

If a storage endpoint is present downstream, the dump is printed
directly via dev_err() so it is visible in dmesg immediately, since a
devcoredump read from userspace could otherwise race with a storage
failure. Otherwise, the buffer is handed to the devcoredump framework
so it can be collected from /sys/class/devcoredump/ for offline
analysis.

To keep the dump readable given the number of registers involved,
entries are packed as "offset: value" pairs, QCOM_PCIE_DUMP_REGS_PER_LINE
(4) per dmesg line, with a section header ("DBI registers:", "PARF
registers:", "MHI registers:") identifying each register group.

Also expose the same dump on demand via a new "regdump" debugfs file
under the existing per-controller debugfs directory, so the register
state can be inspected without needing an actual Link Down event.

Link: https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com

[qlijarvis]

🔨 Build Failure Analysis — PR #1677

PR: #1677
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/31572676600

#	Error	File:Line	PR-introduced?	Root Cause
1	PARF_INT_ALL_STATUS undeclared	drivers/pci/controller/dwc/pcie-qcom.c:201	Yes	Macro used in qcom_pcie_parf_dump_regs[] array but never defined. Patch defines PARF_INT_ALL_2_STATUS but not PARF_INT_ALL_STATUS.
2	PARF_STATUS undeclared	drivers/pci/controller/dwc/pcie-qcom.c:215	Yes	Macro used in qcom_pcie_ext_parf_dump_regs[] array but never defined in the patch.

Verdict

Both compilation errors are introduced by this PR due to missing macro definitions for registers referenced in newly added register dump arrays.

📎 Detailed analysis: Full report

[qlijarvis]

🔨 Build Failure Analysis — PR #1677

PR: #1677
Build run: https://github.com/qualcomm-linux/kernel-config/actions/runs/31572676600

#	Error	File:Line	PR-introduced?	Root Cause
1	PARF_INT_ALL_STATUS undeclared	drivers/pci/controller/dwc/pcie-qcom.c:201	Yes	Macro used in qcom_pcie_parf_dump_regs[] array but never defined
2	PARF_INT_ALL_STATUS undeclared	drivers/pci/controller/dwc/pcie-qcom.c:208	Yes	Macro used in qcom_pcie_ext_parf_dump_regs[] array but never defined
3	PARF_STATUS undeclared	drivers/pci/controller/dwc/pcie-qcom.c:215	Yes	Macro used in qcom_pcie_ext_parf_dump_regs[] array but never defined
4	qcom_pcie_dump_regs unused	drivers/pci/controller/dwc/pcie-qcom.c:2023	Yes	Function defined but never called (warning treated as error)

Verdict

All 4 errors are introduced by this PR. The patch adds register dump functionality but references undefined register macros and defines an unused function.

📎 Detailed analysis: Full report

[qlijarvis]

PR #1677 — validate-patch

PR: #1677

Verdict	Issues	Detailed Report
⚠️	2	Full report

Final Summary

1. Lore link present: Yes — https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
2. Lore link matches PR commits: Partial — diff content faithful to lore with 2 extra register definitions required for older tree baseline (PARF_INT_ALL_STATUS, PARF_STATUS)
3. Upstream patch status: ⏳ Decision Pending — posted Aug 11, 2026 (1 day ago); no maintainer review signals yet (no Reviewed-by, Acked-by, applied, or NAK messages in thread)
4. PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence

Verdict: ⚠️ — click to expand

🔍 Patch Validation

PR: #1677 - "FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events"
Upstream commit: https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
Verdict: ⚠️ PARTIAL

Commit Message

Check	Status	Note
Subject matches upstream	✅	Subject identical except for FROMLIST: prefix (expected)
Body preserves rationale	✅	Commit message body is identical to lore patch
Fixes tag present/correct	N/A	No Fixes tag in either patch (new feature, not a fix)
Authorship preserved	✅	Original author Krishna Chaitanya Chundru preserved in From: and first Signed-off-by:
Backport note (if applicable)	N/A	FROMLIST: commit, not a backport
Co-developed-by usage	✅	Not used; submitter Ziyue Zhang correctly added own Signed-off-by:

Diff

File	Status	Notes
drivers/pci/controller/dwc/pcie-designware.h	✅	Identical - adds SD-EQ register definitions
drivers/pci/controller/dwc/pcie-qcom.c	⚠️	Adds 2 extra register definitions not in lore patch (see below)
include/uapi/linux/pci_regs.h	✅	Identical - adds PCI_PL_16GT_STATUS and PCI_PL_32GT_STATUS

Issues

Extra definitions in PR (not in lore patch):
The PR adds two register definitions that are not present in the upstream lore patch:

• +#define PARF_INT_ALL_STATUS 0x228 (line 78 in pr.patch)
• +#define PARF_STATUS 0x230 (line 79 in pr.patch)

Root cause: The lore patch is based on a newer upstream tree where PARF_INT_ALL_STATUS and PARF_STATUS already exist. The PR's target tree (kernel-topics) does not have these definitions yet, so they must be added for the patch to compile.

Analysis: This is a legitimate adaptation for applying a FROMLIST: patch to an older tree. The lore patch uses PARF_INT_ALL_STATUS in the register dump arrays (lines 252, 259 in lore mbox) but doesn't define it because it already exists upstream. The PR correctly adds the missing definitions.

Verification: The PR has 338 insertions vs 336 in the lore patch - exactly 2 extra lines for the missing register definitions.

Verdict

⚠️ PARTIAL - The patch faithfully represents the lore upstream with necessary adaptations for the target tree's older baseline. The two extra register definitions are required for compilation and do not alter the patch's functionality.

Recommendation: Merge as-is. The extra definitions are a correct adaptation for the kernel-topics tree baseline. Consider adding a note in the commit message explaining the extra definitions if this pattern is common.

Final Summary

1. Lore link present: Yes — https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
2. Lore link matches PR commits: Partial — diff content faithful to lore with 2 extra register definitions required for older tree baseline (PARF_INT_ALL_STATUS, PARF_STATUS)
3. Upstream patch status: ⏳ Decision Pending — posted Aug 11, 2026 (1 day ago); no maintainer review signals yet (no Reviewed-by, Acked-by, applied, or NAK messages in thread)
4. PR present in qcom-next/topics: Partial — integration_presence_report.md shows "partial - subject or partial tree evidence found, but full change was not verified" (1/1 commits partial, 0/1 present, 0/1 missing)

Deterministic Integration Presence

Integration Presence Report

This report is generated by Jarvis before validate-patch runs.
It is the authoritative source for whether PR changes are already present
in qcom-next or in the kernel topic branches.

Kernel repo: /local/mnt/workspace/sgaud/Qgenie/image_pipeline/kernel
qcom-next ref: 49dbe0dae5cfb7a1eb3434cde6fc7ba37924fe94
topics remote: topics -> https://github.com/qualcomm-linux/kernel-topics
topics fetch: fetched

Commit	Subject	qcom-next	topics	Final
1/1	[PATCH] FROMLIST: PCI: qcom: Add register dump support for PCIe Link	partial - subject or partial tree evidence found, but full change was not verified	missing - no subject, patch-id, or full tree-content match found	partial

Final Status

overall_status: PARTIAL
present_commits: 0/1
partial_commits: 1/1
missing_commits: 0/1
topics_checked_for_commits: 1/1
final_summary: PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence

[qlijarvis]

PR #1677 — checker-log-analyzer

PR: #1677
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/31572674648

Checker	Result	Summary
Checker	Result	Summary
checkpatch	⚠️	1 CHECK: multiple blank lines
dt-binding-check	⏭️	No DT binding changes
dtb-check	⏭️	No DTS changes
sparse-check	✅	Passed
check-uapi-headers	✅	Passed
check-patch-compliance	❌	Content mismatch with upstream link
tag-check	✅	Subject has valid FROMLIST: prefix

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1677 - FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/31572674648
Target branch: tech/bus/pci/all (topic branch, not qcom-next/qcom-next-staging)

Checker	Result	Summary
checkpatch	⚠️	1 CHECK: multiple blank lines
dt-binding-check	⏭️	No DT binding changes
dtb-check	⏭️	No DTS changes
sparse-check	✅	Passed
check-uapi-headers	✅	Passed
check-patch-compliance	❌	Content mismatch with upstream link
tag-check	✅	Subject has valid FROMLIST: prefix

---

⚠️ checkpatch

Root cause: Multiple consecutive blank lines in drivers/pci/controller/dwc/pcie-qcom.c:2062

Failure details:

CHECK: Please don't use multiple blank lines
#473: FILE: drivers/pci/controller/dwc/pcie-qcom.c:2062:
+
+

2ab8551988b38f009279efec45ac2f913711a9d2 total: 0 errors, 0 warnings, 1 checks, 444 lines checked

Fix: Remove one of the duplicate blank lines at line 2062 in drivers/pci/controller/dwc/pcie-qcom.c

git rebase -i <base_sha>   # mark commit as 'edit'
# Edit drivers/pci/controller/dwc/pcie-qcom.c and remove the extra blank line
git add drivers/pci/controller/dwc/pcie-qcom.c
git commit --amend --no-edit
git rebase --continue

Reproduce locally:

./scripts/checkpatch.pl --strict --ignore FILE_PATH_CHANGES --git f04463e51d69987f9ed441a0b69e6d104ca0656b..26e9acfd49432b7845b4b47f910d429dee84d5dd

Severity: CHECK — style issue, not a blocker, but should be fixed for consistency with kernel coding style.

---

❌ check-patch-compliance

Root cause: The patch content in the PR differs from the upstream patch at the provided lore.kernel.org link.

Failure details:

Checking commit: FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events
Change is different from the one mentioned in Link

Link: https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com

Analysis: The checker detected that the patch content in commit 2ab8551988b3 does not match the upstream patch referenced in the Link: trailer. This could be due to:

1. Context-only differences — line numbers shifted but actual changes are identical (false positive)
2. Legitimate adaptations — changes needed for the target tree (should be documented in commit message)
3. Missing hunks — some changes from upstream were omitted
4. Extra hunks — additional changes not in the upstream patch

Fix: Verify the difference manually:

# Fetch the upstream patch
b4 am --single-message -C -l -3 https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com -o /tmp/out

# Compare the +/- lines between the PR commit and upstream
git format-patch -1 2ab8551988b3 --stdout | awk '/^diff/,/^--$/' | grep -E '^[+-][^+-]' > /tmp/pr-changes.txt
awk '/^diff/,/^--$/' /tmp/out/*.mbx | grep -E '^[+-][^+-]' > /tmp/upstream-changes.txt
diff /tmp/pr-changes.txt /tmp/upstream-changes.txt

Action required:

• If the differences are context-only (line number shifts), this is a false positive — no action needed.
• If there are legitimate adaptations for the target tree, document them in the commit message body (e.g., "Adapted for qcom-next: ...").
• If there are missing or extra hunks, either add the missing changes or split the extra changes into a separate commit with appropriate attribution.

---

Verdict

2 issues to address before merge:

1. checkpatch (CHECK) — Remove duplicate blank line at drivers/pci/controller/dwc/pcie-qcom.c:2062 (trivial fix)
2. check-patch-compliance (FAIL) — Verify and document the content difference from the upstream link (requires manual review)

The check-patch-compliance failure is the blocker — the content mismatch must be investigated and either:

• Confirmed as a false positive (context-only difference), or
• Documented in the commit message if legitimate adaptations were made, or
• Fixed by aligning the patch with the upstream version

All other checkers passed or were skipped (no relevant changes).

[qlijarvis]

PR #1677 — validate-patch

PR: #1677

Verdict	Issues	Detailed Report
⚠️	2	Full report

Final Summary

1. Lore link present: Yes — https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
2. Lore link matches PR commits: Partial — Core functionality is identical, but PR adds one extra register definition (PARF_INT_ALL_STATUS) not in the lore patch. This is a legitimate tree adaptation, not a semantic change.
3. Upstream patch status: ⏳ Decision Pending — Patch posted Aug 11, 2026 (1 day ago). No maintainer response, review comments, or merge signals found in the lore thread. This is expected for a patch less than 48 hours old.
4. PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence

Verdict: ⚠️ — click to expand

🔍 Patch Validation

PR: #1677 - FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events
Upstream commit: https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
Verdict: ⚠️ PARTIAL

Commit Message

Check	Status	Note
Subject matches upstream	✅	Subject correctly prefixed with FROMLIST:
Body preserves rationale	✅	Commit message body is identical to lore patch
Fixes tag present/correct	✅	No Fixes tag required (new feature)
Authorship preserved	✅	Original author Krishna Chaitanya Chundru present in Signed-off-by: chain; submitter Ziyue Zhang added own Signed-off-by: (correct for FROMLIST:)
Backport note (if applicable)	N/A	Not a backport; FROMLIST: prefix is correct

Diff

File	Status	Notes
drivers/pci/controller/dwc/pcie-designware.h	✅	Identical - adds SD-EQ register definitions
drivers/pci/controller/dwc/pcie-qcom.c	⚠️	Extra define added: PR adds #define PARF_INT_ALL_STATUS 0x228 which is NOT in the lore patch. The lore patch uses PARF_INT_ALL_STATUS in the register dump arrays but does not define it, suggesting it expects this define to already exist in the base tree. The PR also adds #define PARF_STATUS 0x230 which already exists as a context line in the lore patch's base.
include/uapi/linux/pci_regs.h	✅	Identical - adds PCI_EXP_LNKSTA2_FLIT register definitions

Issues

1. Extra register definition not in upstream patch

The PR adds this line which is NOT present in the lore patch:

+#define PARF_INT_ALL_STATUS			0x228

The lore patch uses PARF_INT_ALL_STATUS in two places in the register dump arrays:

• Line ~400: PARF_INT_ALL_STATUS, (in qcom_pcie_parf_dump_regs[])
• Line ~420: PARF_INT_ALL_STATUS, (in another register array)

However, the lore patch does NOT define this constant, suggesting the upstream author expects it to already exist in the mainline tree at the time of posting.

Analysis: The PR is based on a different tree version than the lore patch. The lore patch (posted Aug 11, 2026) targets a mainline tree where PARF_INT_ALL_STATUS is already defined. The PR targets a Qualcomm kernel tree where this define does not yet exist, so the submitter correctly added it to make the patch compile.

2. PARF_STATUS definition context mismatch

The PR adds:

+#define PARF_STATUS				0x230

But in the lore patch, this line appears as a context line (without +), meaning it already exists in the base tree. This confirms the PR is based on an older tree version.

Impact: These are legitimate adaptations for the target tree. The extra define is necessary for the code to compile and does not change the semantic behavior. However, this should be documented in the commit message or as a separate preparatory commit.

Verdict

The PR faithfully represents the lore patch's intent and functionality. The extra PARF_INT_ALL_STATUS definition is a necessary adaptation because the PR targets a tree where this constant does not yet exist, while the upstream patch assumes it's already present in mainline.

Recommendation:

Option 1 (Preferred): Split into two commits:

1. QCLINUX: PCI: qcom: Add PARF_INT_ALL_STATUS register definition - preparatory commit adding the missing define
2. FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events - the actual lore patch without the extra define

Option 2: Merge as-is but add a note to the commit message:

[ Qualcomm tree adaptation: Added PARF_INT_ALL_STATUS definition which
  is not yet present in this tree but exists in mainline. ]

Final Summary

1. Lore link present: Yes — https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com
2. Lore link matches PR commits: Partial — Core functionality is identical, but PR adds one extra register definition (PARF_INT_ALL_STATUS) not in the lore patch. This is a legitimate tree adaptation, not a semantic change.
3. Upstream patch status: ⏳ Decision Pending — Patch posted Aug 11, 2026 (1 day ago). No maintainer response, review comments, or merge signals found in the lore thread. This is expected for a patch less than 48 hours old.
4. PR present in qcom-next/topics: Partial — Integration presence report shows "partial - subject or partial tree evidence found, but full change was not verified" in qcom-next, and "missing" in topics. This suggests related changes may exist but this exact patch has not been integrated yet.

Deterministic Integration Presence

Integration Presence Report

This report is generated by Jarvis before validate-patch runs.
It is the authoritative source for whether PR changes are already present
in qcom-next or in the kernel topic branches.

Kernel repo: /local/mnt/workspace/sgaud/Qgenie/image_pipeline/kernel
qcom-next ref: 49dbe0dae5cfb7a1eb3434cde6fc7ba37924fe94
topics remote: topics -> https://github.com/qualcomm-linux/kernel-topics
topics fetch: fetched

Commit	Subject	qcom-next	topics	Final
1/1	[PATCH] FROMLIST: PCI: qcom: Add register dump support for PCIe Link	partial - subject or partial tree evidence found, but full change was not verified	missing - no subject, patch-id, or full tree-content match found	partial

Final Status

overall_status: PARTIAL
present_commits: 0/1
partial_commits: 1/1
missing_commits: 0/1
topics_checked_for_commits: 1/1
final_summary: PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence

[qlijarvis]

PR #1677 — checker-log-analyzer

PR: #1677
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/31586676231

Checker	Result	Summary
Checker	Result	Summary
checkpatch	❌	1 CHECK: multiple blank lines
dt-binding-check	⏭️	No DT binding changes
dtb-check	⏭️	No DTS changes
sparse-check	✅	Passed
check-uapi-headers	✅	Passed
check-patch-compliance	❌	Content mismatch with upstream link
tag-check	✅	Subject has valid FROMLIST: prefix

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1677 - FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/31586676231
Target branch: tech/bus/pci/all (not qcom-next/qcom-next-staging)

Checker	Result	Summary
checkpatch	❌	1 CHECK: multiple blank lines
dt-binding-check	⏭️	No DT binding changes
dtb-check	⏭️	No DTS changes
sparse-check	✅	Passed
check-uapi-headers	✅	Passed
check-patch-compliance	❌	Content mismatch with upstream link
tag-check	✅	Subject has valid FROMLIST: prefix

---

❌ checkpatch

Root cause: Multiple consecutive blank lines in drivers/pci/controller/dwc/pcie-qcom.c:2064

Failure details:

CHECK: Please don't use multiple blank lines
#477: FILE: drivers/pci/controller/dwc/pcie-qcom.c:2064:
+
+

Fix: Remove one of the two consecutive blank lines at line 2064 in pcie-qcom.c:

git rebase -i e97ed4115c2fb66d752be5ed8d52f95589922c16   # mark commit 935b283 as 'edit'
# Edit drivers/pci/controller/dwc/pcie-qcom.c and remove the extra blank line at line 2064
git add drivers/pci/controller/dwc/pcie-qcom.c
git commit --amend --no-edit
git rebase --continue

Reproduce locally:

./scripts/checkpatch.pl --strict --ignore FILE_PATH_CHANGES --git e97ed4115c2fb66d752be5ed8d52f95589922c16..a492bc0e71ea0013cf9a456ebd1c726d5e4cfb4a

---

❌ check-patch-compliance

Root cause: The patch content in the PR differs from the upstream patch at the provided lore.kernel.org link.

Failure details:

Checking commit: FROMLIST: PCI: qcom: Add register dump support for PCIe Link Down events
Change is different from the one mentioned in Link

Link provided: https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com

Fix: Verify the content difference between the PR patch and the upstream patch:

# Fetch the upstream patch
b4 am --single-message -C -l -3 https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com -o /tmp/out

# Compare the diff content (ignoring context lines)
diff <(git format-patch -1 935b2836886f --stdout | awk '/^diff/,/^--$/' | grep -E '^[+-][^+-]') \
     <(awk '/^diff/,/^--$/' /tmp/out/*.mbx | grep -E '^[+-][^+-]')

Possible causes:

• Context-only shift (not a real difference) — acceptable
• Legitimate adaptation for the target tree — document in commit message
• Missing hunk from upstream — add it
• Extra hunk not in upstream — remove or attribute separately

If the difference is legitimate (e.g., adaptation for the target tree), add a note in the commit message explaining the deviation. If it's an error, align the patch with the upstream version.

Reproduce locally:

cd kernel
git checkout 935b2836886f
b4 am --single-message -C -l -3 https://lore.kernel.org/r/20260811-regdump-v1-1-f22db3fb37c2@oss.qualcomm.com

---

Verdict

2 issues to fix before merge:

1. checkpatch (minor): Remove the extra blank line at pcie-qcom.c:2064 — trivial style fix.
2. check-patch-compliance (blocker): Investigate and resolve the content mismatch with the upstream patch. If the difference is intentional, document it in the commit message. If it's an error, align with upstream.

The FROMLIST: prefix is correctly present, and all other checkers passed. Once the content mismatch is resolved and the blank line is removed, the PR should be ready to merge.