Skip to content

FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range - #1365

Open
akakum-qualcomm wants to merge 1 commit into
qualcomm-linux:qcom-6.18.yfrom
akakum-qualcomm:akash/dwc3-qcom-nominal-votes
Open

akakum-qualcomm wants to merge 1 commit into
qualcomm-linux:qcom-6.18.yfrom
akakum-qualcomm:akash/dwc3-qcom-nominal-votes

Conversation

@akakum-qualcomm

@akakum-qualcomm akakum-qualcomm commented Oct 29, 2026 •

Copy link
Copy Markdown

The USB memory peak vote of 5500 MB/s, together with the Apps USB bandwidth votes, can raise the interconnect request into the turbo range during USB activity. This causes DDR to operate beyond the nominal range and can result in unnecessary frequency changes and power consumption.

Reduce the USB memory peak vote to 4500 MB/s, clear the Apps USB average vote, and reduce the Apps USB peak vote from 80 MB/s to 40 MB/s. These changes prevent the Apps USB vote from unnecessarily triggering turbo levels and keep the interconnect bandwidth within the nominal range.

Maintaining the bandwidth votes within the nominal range avoids DDR undervolting and improves stability during simultaneous USB audio and high-throughput data transfers.

Link: https://lore.kernel.org/all/20261105100626.2003771-1-akash.kumar@oss.qualcomm.com/

Crs-Fixed: 4700998

@qswat-orbit-external

Copy link
Copy Markdown

Merge Check Failed: No Change Task Found

No associated change tasks found for CR 4700998 on any of the following entities:

Entities:

  • kernel.qli.2.0

CR: 4700998

Please ensure the CR has a change task associated with at least one of the entities for this branch.

@akakum-qualcomm
akakum-qualcomm force-pushed the akash/dwc3-qcom-nominal-votes branch from 022aacd to 727cd6d Compare October 29, 2026 14:48
@qswat-orbit-external

Copy link
Copy Markdown

Merge Check Failed: No Change Task Found

No associated change tasks found for CR 4700998 on any of the following entities:

Entities:

  • kernel.qli.2.0

CR: 4700998

Please ensure the CR has a change task associated with at least one of the entities for this branch.

@qlijarvis

Copy link
Copy Markdown

PR #1365 — validate-patch

PR: #1365

Verdict Issues Detailed Report
❌ 0 Full report

Final Summary

  1. Lore link present: Yes — https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/

  2. Lore link matches PR commits: ❌ No — The PR implements the exact opposite change: lore increases bandwidth (2500→5500, 0→40, 40→80), PR decreases bandwidth (5500→4500, 40→0, 80→40). All three macro changes are semantically inverted.

  3. Upstream patch status: ⏳ Decision Pending — Posted June 11, 2026; received review feedback from Dmitry Baryshkov on June 12, 2026 with unresolved concerns (missing macro, unclear target scope, insufficient justification); no acceptance signals; thread stalled with no author follow-up; contains identified bugs.

  4. PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence

Verdict: ❌ — click to expand

🔍 Patch Validation Report

PR: #1365 - usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Upstream: https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/
Verdict: ❌ FAIL


Critical Issue: PR Implements the OPPOSITE Change

The PR patch and the linked lore patch make completely opposite changes to the same code:

Macro Lore (upstream) PR Direction
USB_MEMORY_PEAK_SS_BW 2500 → 5500 5500 → 4500 ⬆️ INCREASE vs ⬇️ DECREASE
APPS_USB_AVG_BW 0 → 40 40 → 0 ⬆️ INCREASE vs ⬇️ DECREASE
APPS_USB_PEAK_BW 40 → 80 80 → 40 ⬆️ INCREASE vs ⬇️ DECREASE

Lore patch rationale: "Increase nominal max votes to improve DDR performance and USB audio use case handling" — increases bandwidth to fix audio glitches.

PR patch rationale: "Keep USB bandwidth votes in nominal range" — decreases bandwidth to avoid turbo mode and reduce power consumption.

These are contradictory goals with opposite implementations.


Commit Message

Check Status Note
Subject matches upstream ❌ Lore: "Update nominal max votes for qcom usb" vs PR: "Keep USB bandwidth votes in nominal range" — completely different intent
Body preserves rationale ❌ PR describes reducing bandwidth to avoid turbo; lore describes increasing bandwidth to fix glitches — opposite rationales
Fixes tag present/correct N/A Neither patch has a Fixes tag
Authorship preserved ⚠️ Same author (Akash Kumar) but different email addresses: lore uses akash.kumar@oss.qualcomm.com, PR uses akakum@qti.qualcomm.com
Backport note (if applicable) N/A Not a backport
Link tag ✅ PR correctly links to the lore thread

Diff Comparison

File Status Notes
drivers/usb/dwc3/dwc3-qcom.c ❌ All three hunks are inverted — PR decreases values that lore increases, and vice versa

Detailed hunk analysis:

  1. USB_MEMORY_PEAK_SS_BW:

    • Lore: MBps_to_icc(2500) → MBps_to_icc(5500) (+3000 MB/s)
    • PR: MBps_to_icc(5500) → MBps_to_icc(4500) (-1000 MB/s)
    • Verdict: Complete semantic reversal
  2. APPS_USB_AVG_BW:

    • Lore: 0 → 40 (enables average bandwidth vote)
    • PR: 40 → 0 (disables average bandwidth vote)
    • Verdict: Complete semantic reversal
    • Additional issue: Lore patch is missing MBps_to_icc() macro (bug caught by Dmitry Baryshkov in review)
  3. APPS_USB_PEAK_BW:

    • Lore: MBps_to_icc(40) → MBps_to_icc(80) (+40 MB/s)
    • PR: MBps_to_icc(80) → MBps_to_icc(40) (-40 MB/s)
    • Verdict: Complete semantic reversal

Upstream Patch Status

Community verdict: ⏳ Decision Pending — under review with unresolved concerns

Evidence:

  • Posted: June 11, 2026
  • Last activity: June 12, 2026 (Dmitry Baryshkov review)
  • No acceptance signals (applied, queued, Reviewed-by, Acked-by) found
  • No explicit NAK, but Dmitry raised multiple concerns:
    • Questioned the math behind DDR frequency claims
    • Asked which targets are affected (Agatti? SDX55? IPQ8064?)
    • Identified missing MBps_to_icc() macro on APPS_USB_AVG_BW line
    • Requested explanation for why APPS_USB_*_BW changes are needed
  • Thread appears stalled with no follow-up from author addressing concerns
  • Sashiko AI bot flagged the missing MBps_to_icc() macro as a High severity issue

Status: The lore patch has not been accepted upstream and has unresolved review feedback.


Integration Presence (qcom-next/topics)

Per integration_presence_report.md:

Commit qcom-next topics Final
1/1 partial missing partial

Interpretation: Some related change may exist in qcom-next (possibly the lore version or an earlier iteration), but the exact PR change (the reversal) is not present in either qcom-next or kernel-topics.

Status: ⚠️ Partial presence — 0/1 commits fully present, 1/1 commits have partial evidence


Issues Found

  1. ❌ CRITICAL: Semantic inversion — The PR implements the exact opposite of the linked lore patch. All three macro changes are reversed in direction.

  2. ❌ CRITICAL: Contradictory rationale — The PR commit message describes a completely different goal (reduce bandwidth to avoid turbo) than the lore patch (increase bandwidth to fix audio glitches).

  3. ❌ Subject line mismatch — PR subject "Keep USB bandwidth votes in nominal range" does not match lore subject "Update nominal max votes for qcom usb" and describes opposite intent.

  4. ❌ Upstream patch not accepted — The lore patch has unresolved review concerns and has not been merged into any maintainer tree or mainline.

  5. ⚠️ Authorship inconsistency — Email addresses differ between lore (akash.kumar@oss.qualcomm.com) and PR (akakum@qti.qualcomm.com), though both are Akash Kumar.

  6. ⚠️ Upstream patch has bugs — Dmitry Baryshkov identified that the lore patch is missing MBps_to_icc() on the APPS_USB_AVG_BW line, which would cause incorrect bandwidth voting.


Recommendation

❌ DO NOT MERGE — This PR must be rejected for the following reasons:

  1. The PR does not match the linked lore patch — it implements the opposite change. The Link tag is misleading.

  2. The upstream patch is not accepted — it has unresolved review concerns and contains bugs identified by maintainers.

  3. Unclear intent — It's unclear whether:

    • The PR author intended to revert the lore patch (in which case the commit message should say "Revert" and explain why)
    • The PR author made an error and applied the wrong diff
    • The PR represents a different solution to a different problem (in which case it should not link to the lore patch)

Required actions before this can be reconsidered:

  1. Clarify intent: Is this a revert, a mistake, or a different patch?
  2. Remove or correct the Link tag: The current lore link is incorrect for this change.
  3. If this is a revert: Use proper revert commit format and explain why the original change is being reverted.
  4. If this is a new patch: Post to lore.kernel.org for upstream review first, address maintainer feedback, and only then submit to the vendor tree with the correct lore link.
  5. Address the base state confusion: The PR assumes a starting state (5500/40/80) that matches the lore patch's end state, suggesting the lore patch may have already been applied somewhere. Clarify the tree state.

Final Summary

  1. Lore link present: Yes — https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/

  2. Lore link matches PR commits: ❌ No — The PR implements the exact opposite change: lore increases bandwidth (2500→5500, 0→40, 40→80), PR decreases bandwidth (5500→4500, 40→0, 80→40). All three macro changes are semantically inverted.

  3. Upstream patch status: ⏳ Decision Pending — Posted June 11, 2026; received review feedback from Dmitry Baryshkov on June 12, 2026 with unresolved concerns (missing macro, unclear target scope, insufficient justification); no acceptance signals; thread stalled with no author follow-up; contains identified bugs.

  4. PR present in qcom-next/topics: ⚠️ Partial — 0/1 commits fully present; 1/1 commits have partial evidence (possibly the lore version or related change exists, but not this exact reversal).

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: c7896917e523946138ce654d4b6c1e33d5a8f4e5
topics remote: topics -> https://github.com/qualcomm-linux/kernel-topics
topics fetch: fetched

Commit Subject qcom-next topics Final
1/1 [PATCH] usb: dwc3: qcom: Keep USB bandwidth votes in nominal range 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

Copy link
Copy Markdown

PR #1365 — checker-log-analyzer

PR: #1365
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/40895328880

Checker Result Summary
Checker Result Summary
checkpatch ✅ No style issues
dt-binding-check ⏭️ No binding changes
dtb-check ⏭️ No DTS changes
sparse-check ✅ No new warnings
check-uapi-headers ✅ No UAPI changes
check-patch-compliance ❌ Missing subject prefix
tag-check ❌ Missing subject prefix (mandatory for qcom-6.18.y)

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1365 - usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/40895328880
Target branch: qcom-6.18.y

Checker Result Summary
checkpatch ✅ No style issues
dt-binding-check ⏭️ No binding changes
dtb-check ⏭️ No DTS changes
sparse-check ✅ No new warnings
check-uapi-headers ✅ No UAPI changes
check-patch-compliance ❌ Missing subject prefix
tag-check ❌ Missing subject prefix (mandatory for qcom-6.18.y)

❌ check-patch-compliance

Root cause: Commit subject does not start with a required prefix tag.

Failure details:

Checking commit: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Commit summary does not start with a required prefix

The commit has a Link: tag pointing to lore.kernel.org, indicating this patch was posted upstream, but the subject line is missing the required FROMLIST: prefix.

Fix:

Since the patch has a lore link (https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/), it should use the FROMLIST: prefix:

git rebase -i bf8dffb05bda45b8490ab4ee04998dfa9b8be292
# mark commit 727cd6d8857b as 'edit'
git commit --amend -m "FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range

The USB memory peak vote of 5500 MB/s, together with the Apps USB
bandwidth votes, can raise the interconnect request into the turbo
range during USB activity. This causes DDR to operate beyond the
nominal range and can result in unnecessary frequency changes and power
consumption.

Reduce the USB memory peak vote to 4500 MB/s, clear the Apps USB average
vote, and reduce the Apps USB peak vote from 80 MB/s to 40 MB/s. These
changes prevent the Apps USB vote from unnecessarily triggering turbo
levels and keep the interconnect bandwidth within the nominal range.

Maintaining the bandwidth votes within the nominal range avoids DDR
undervolting and improves stability during simultaneous USB audio and
high-throughput data transfers.

Link: https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/

Signed-off-by: Akash Kumar <akakum@qti.qualcomm.com>"
git rebase --continue

Reproduce locally:

cd /path/to/kernel
./scripts/checkpatch.pl --strict --ignore FILE_PATH_CHANGES --git bf8dffb05bda..727cd6d8857b
# or use the check-patch-compliance.sh script:
bash kernel-checkers/check-patch-compliance.sh --kernel-src . \
  --base bf8dffb05bda45b8490ab4ee04998dfa9b8be292 \
  --head 727cd6d8857b57a05ab37154963154091c31761b

❌ tag-check

Root cause: Target branch qcom-6.18.y requires all commits to have a subject prefix tag.

Failure details:

The PR targets qcom-6.18.y, which is not qcom-next or qcom-next-staging. All commits merged into branches other than those two must start with one of the required prefix tags:

  • FROMLIST: — Posted to mailing list (lore.kernel.org)
  • FROMGIT: — Taken from a maintainer git tree
  • UPSTREAM: — Merged into Linus's mainline tree
  • BACKPORT: — Upstream patch backported with modifications
  • QCLINUX: — Vendor-only change with no upstream equivalent
  • PENDING: — Work-in-progress, not yet posted upstream
  • WORKAROUND: — Temporary fix not suitable for upstream

Commit 727cd6d8857b ("usb: dwc3: qcom: Keep USB bandwidth votes in nominal range") is missing this prefix.

Fix:

Same as the check-patch-compliance fix above — add FROMLIST: prefix to the subject line since the patch has been posted to lore.kernel.org.


Verdict

1 blocker must be fixed before merge:

The commit subject must be amended to include the FROMLIST: prefix. This is required by both check-patch-compliance and the tag-check policy for the qcom-6.18.y branch.

All other checkers passed cleanly:

  • ✅ No checkpatch style issues
  • ✅ No sparse warnings introduced
  • ✅ No UAPI changes
  • ⏭️ No DT binding or DTS changes (skipped)

Once the subject prefix is added, the PR will be ready to merge.

@akakum-qualcomm
akakum-qualcomm force-pushed the akash/dwc3-qcom-nominal-votes branch from 727cd6d to be6aabf Compare October 29, 2026 15:41
@qswat-orbit-external

Copy link
Copy Markdown

Merge Check Failed: CR Not Eligible for Merge

CR 4700998 is not eligible for merge.

The parent software image for kernel.qli.2.0 is not development complete.

Entity: kernel.qli.2.0
CR: 4700998
Reason: CR_CANNOT_MERGE

Please ensure the CR passes both CCT (ComponentChangeTasks) and ICT (Integration Change Tasks) validations.

@qlijarvis

Copy link
Copy Markdown

PR #1365 — validate-patch

PR: #1365

Verdict Issues Detailed Report
❌ 0 Full report

Final Summary

  1. Lore link present: Yes — https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/
  2. Lore link matches PR commits: ❌ No — PR implements the inverse of the lore patch (decreases bandwidth where lore increases it, and vice versa)
  3. Upstream patch status: ⏳ Decision Pending — under review with critical bug and unanswered maintainer questions; not accepted
  4. PR present in qcom-next/topics: Partial - 1/1 commit(s) only have partial integration evidence
Verdict: ❌ — click to expand

🔍 Patch Validation Report

PR: #1365 - FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Upstream commit: https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/
Verdict: ❌ FAIL

Commit Message

Check Status Note
Subject matches upstream ❌ Lore: "usb: dwc3: Update nominal max votes for qcom usb" vs PR: "FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range" — completely different subjects
Body preserves rationale ❌ PR commit message describes reducing bandwidth votes to avoid turbo range; lore describes increasing votes to improve performance — opposite intent
Fixes tag present/correct ✅ No Fixes tag in either (not a bug fix)
Authorship preserved ⚠️ Lore author: akash.kumar@oss.qualcomm.com, PR author: akakum@qti.qualcomm.com — different email domains but same person; lore author's Signed-off-by is present
Backport note (if applicable) N/A FROMLIST prefix used (not a backport)

Diff Comparison

File Status Notes
drivers/usb/dwc3/dwc3-qcom.c:47 ❌ OPPOSITE CHANGE: Lore increases USB_MEMORY_PEAK_SS_BW from 2500→5500; PR decreases it from 5500→4500
drivers/usb/dwc3/dwc3-qcom.c:48 ❌ OPPOSITE CHANGE: Lore sets APPS_USB_AVG_BW from 0→40; PR sets it from 40→0
drivers/usb/dwc3/dwc3-qcom.c:49 ❌ OPPOSITE CHANGE: Lore increases APPS_USB_PEAK_BW from 40→80; PR decreases it from 80→40

Upstream Patch Status

Community verdict: ⏳ Decision Pending — under review with critical issues raised

Evidence from lore thread:

  • Sashiko AI bot (2026-06-11): Flagged APPS_USB_AVG_BW 40 missing MBps_to_icc() macro — would pass 40 kBps instead of 40 MBps
  • Dmitry Baryshkov (maintainer, 2026-06-12): Requested MBps_to_icc(40) fix and asked for justification of APPS_USB_*_BW changes
  • No acceptance signals (no "applied", "queued", "Reviewed-by", or "Acked-by" from maintainers)
  • No rejection signals (no "NAK" or explicit rejection)

The lore patch has a critical bug (missing macro) and unanswered maintainer questions. It is not ready for merge.

Integration Presence (from integration_presence_report.md)

Commit qcom-next topics Final
1/1 partial missing PARTIAL

Status: Only partial evidence found in qcom-next; full change not verified in either qcom-next or topic branches.

Issues Found

Critical Issues:

  1. The PR patch applies the INVERSE of the lore patch

    • Lore increases bandwidth votes (2500→5500, 0→40, 40→80)
    • PR decreases bandwidth votes (5500→4500, 40→0, 80→40)
    • The commit messages describe opposite goals:
      • Lore: "Increase nominal max votes to improve DDR performance"
      • PR: "Reduce the USB memory peak vote... to keep bandwidth within nominal range"
  2. Subject line mismatch

    • Lore: "usb: dwc3: Update nominal max votes for qcom usb"
    • PR: "FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range"
    • These describe fundamentally different changes
  3. The lore patch has a critical bug

    • APPS_USB_AVG_BW 40 is missing MBps_to_icc() macro
    • Would vote for 40 kBps instead of 40 MBps (1000× too small)
    • Flagged by both Sashiko AI and maintainer Dmitry Baryshkov
    • The PR patch does NOT have this bug (it correctly uses 0, not 40)
  4. The lore patch is under review with unanswered questions

    • Maintainer asked for math explanation and justification
    • No follow-up revision posted
    • Not accepted or merged anywhere
  5. Wrong lore link for this PR

    • The PR appears to implement a different change (bandwidth reduction)
    • The linked lore patch implements bandwidth increase
    • These are conflicting approaches to the same problem

Verdict

❌ FAIL — Do not merge

This PR claims to be FROMLIST: from the linked lore patch, but implements the exact opposite change. The lore patch increases bandwidth votes to improve performance; this PR decreases them to stay within nominal range. These are mutually exclusive solutions.

Root cause: Either:

  1. The wrong lore link was provided (there may be a different lore patch that matches this PR's intent), OR
  2. The PR author changed their approach after posting to lore but reused the old lore link

Required actions:

  1. If this PR's approach (reducing bandwidth) is correct:
    • Remove the lore link (this is not a FROMLIST patch)
    • Change prefix to QCLINUX: or PENDING:
    • Provide independent justification for the bandwidth reduction
  2. If the lore approach (increasing bandwidth) is correct:
    • Revert this PR's diff to match the lore patch
    • Fix the APPS_USB_AVG_BW bug (add MBps_to_icc(40))
    • Wait for maintainer approval on lore before merging

Final Summary

  1. Lore link present: Yes — https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/
  2. Lore link matches PR commits: ❌ No — PR implements the inverse of the lore patch (decreases bandwidth where lore increases it, and vice versa)
  3. Upstream patch status: ⏳ Decision Pending — under review with critical bug and unanswered maintainer questions; not accepted
  4. PR present in qcom-next/topics: Partial — only partial evidence in qcom-next; full change not verified

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: c7896917e523946138ce654d4b6c1e33d5a8f4e5
topics remote: topics -> https://github.com/qualcomm-linux/kernel-topics
topics fetch: fetched

Commit Subject qcom-next topics Final
1/1 [PATCH] FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in 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

Copy link
Copy Markdown

PR #1365 — checker-log-analyzer

PR: #1365
Checker run: https://github.com/qualcomm-linux/kernel-config/actions/runs/40904248780

Checker Result Summary
Checker Result Summary
checkpatch ✅ No style issues
dt-binding-check ⏭️ Skipped (no DT binding changes)
dtb-check ⏭️ Skipped (no DTS changes)
sparse-check ✅ No static analysis warnings
check-uapi-headers ✅ No UAPI changes
check-patch-compliance ❌ Content mismatch + author mismatch
tag-check ✅ Valid FROMLIST: prefix present

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #1365 - FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Source: https://github.com/qualcomm-linux/kernel-config/actions/runs/40904248780

Checker Result Summary
checkpatch ✅ No style issues
dt-binding-check ⏭️ Skipped (no DT binding changes)
dtb-check ⏭️ Skipped (no DTS changes)
sparse-check ✅ No static analysis warnings
check-uapi-headers ✅ No UAPI changes
check-patch-compliance ❌ Content mismatch + author mismatch
tag-check ✅ Valid FROMLIST: prefix present

❌ check-patch-compliance

Root cause: Two distinct failures on commit be6aabf — the patch content differs from the upstream lore link, and the commit author email doesn't match the upstream author.

Failure details:

Checking commit: FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range
Change is different from the one mentioned in Link
Author mismatch:
  Original author: Akash Kumar <akash.kumar@oss.qualcomm.com>
  Commit author : Akash Kumar <akakum@qti.qualcomm.com>

Analysis:

Issue 1: Content mismatch
The patch in the PR differs from the upstream patch at:
Link: https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/

This could be due to:

  • Context line differences (line numbers shifted)
  • Additional changes not present upstream
  • Missing hunks from the upstream patch
  • Legitimate adaptation for the vendor tree

Issue 2: Author mismatch

  • Upstream author: Akash Kumar <akash.kumar@oss.qualcomm.com>
  • Commit author: Akash Kumar <akakum@qti.qualcomm.com>

The email domain changed from @oss.qualcomm.com (upstream) to @qti.qualcomm.com (internal).

Fix:

For content mismatch:

  1. Fetch the upstream patch and compare:

    b4 am --single-message -C -l -3 https://lore.kernel.org/all/20260611081831.2743464-1-akash.kumar@oss.qualcomm.com/ -o /tmp/upstream
    git format-patch -1 be6aabf904d8 --stdout > /tmp/pr-patch
  2. Compare the actual code changes (ignore context):

    diff <(awk '/^diff/,/^--$/' /tmp/pr-patch | grep -E '^[+-][^+-]') \
         <(awk '/^diff/,/^--$/' /tmp/upstream/*.mbx | grep -E '^[+-][^+-]')
  3. If identical: The mismatch is context-only (false positive) — document this in the PR description.

  4. If different: Determine if the changes are:

    • Legitimate adaptation → Document the delta in the commit message
    • Missing hunks → Add them
    • Extra changes → Remove or attribute separately

For author mismatch:

git rebase -i bf8dffb05bda   # mark commit as 'edit'
git commit --amend --author="Akash Kumar <akash.kumar@oss.qualcomm.com>"
git rebase --continue

Reproduce locally:

cd /path/to/kernel
bash ../kernel-checkers/check-patch-compliance.sh \
  --kernel-src . \
  --base bf8dffb05bda45b8490ab4ee04998dfa9b8be292 \
  --head be6aabf904d8d5daa1c991d10ab2f59c592c661f

Verdict

2 blockers must be fixed before merge:

  1. Author email mismatch — Change commit author to match upstream: akash.kumar@oss.qualcomm.com
  2. Content mismatch — Verify the patch content matches the upstream lore link. If there are legitimate adaptations, document them in the commit message. If the mismatch is context-only, note this in the PR description.

All other checkers passed cleanly. Once these two issues are resolved, the PR will be ready to merge.

@qcomlnxci

Copy link
Copy Markdown

Test Matrix

Test Case kbdev-hamoa-iot-evk-multimedia kbdev-lemans-evk-multimedia kbdev-monaco-evk-multimedia kbdev-purwa-iot-evk-multimedia kbdev-qcs615-ride-multimedia kbdev-qcs6490-rb3gen2-multimedia kbdev-qcs8300-ride-multimedia kbdev-qcs9100-ride-r3-multimedia kbdev-shikra-iqs-evk-multimedia
BT_FW_KMD_Service ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
BT_ON_OFF ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
BT_SCAN ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ❌ Fail
CPUFreq_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
CPU_affinity ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
DSP_AudioPD ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip
Ethernet_Basic_Validation ⚠️ skip ✅ Pass ✅ Pass ⚠️ skip ⚠️ skip ⚠️ skip ❌ Fail ❌ Fail ⚠️ skip
Freq_Scaling ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
GIC ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ❌ Fail
IPA ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
Interrupts ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
KVM_Driver ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ❌ Fail
KVM_EL2_DTB ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ❌ Fail
KVM_Infra ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ❌ Fail
OpenCV ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
PCIe ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ❌ Fail
Probe_Failure_Check ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail
RMNET ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
UFS_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip
USBHost ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip ❌ Fail ✅ Pass ❌ Fail
WiFi_Firmware_Driver ✅ Pass ✅ Pass ❌ Fail ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
WiFi_OnOff ✅ Pass ✅ Pass ❌ Fail ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip
adsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip
cdsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
gpdsp_remoteproc ⚠️ skip ✅ Pass ✅ Pass ⚠️ skip ⚠️ skip ⚠️ skip ✅ Pass ✅ Pass ⚠️ skip
hotplug ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
irq ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
kaslr ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
pinctrl ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
qcom_hwrng ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
rngtest ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
shmbridge ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
smmu ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ✅ Pass ✅ Pass ❌ Fail ✅ Pass
watchdog ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass
wpss_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass

@qlijarvis

Copy link
Copy Markdown

LAVA Failed Case Triage Summary

PR: #1365

Job 265128 | SoC qcs8300-ride

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265128

Failed test cases in LAVA job 265128 (SoC: qcs8300-ride).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: Three pre-existing platform-specific probe failures unrelated to the PR: (1) xhci-hcd timeout (-ETIMEDOUT) indicating USB controller hardware initialization failure on qcs8300-ride, (2) Aquantia AQR115C PHY firmware-name DT property read failure (-EINVAL) in ethernet PHY driver, and (3) cfg80211 regulatory.db firmware file missing from rootfs (-ENOENT). None of these failures are introduced by the USB dwc3-qcom bandwidth vote changes in PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365.
  3. Possible fix: Suppress this test case failure as a known platform issue for qcs8300-ride. The three probe failures are board/configuration-specific and do not indicate a kernel regression: (1) xhci-hcd -110 is a known hardware initialization timeout on this platform variant, (2) Aquantia PHY firmware-name property is optional and the PHY can operate without it, (3) regulatory.db is a userspace configuration file that should be added to the rootfs if wireless regulatory enforcement is required. If these failures must be resolved: for (1) investigate USB controller power/clock sequencing in qcs8300 DT; for (2) add firmware-name property to the PHY node if custom firmware is needed; for (3) install wireless-regdb package in the rootfs.
  4. Detail analysis attachment: failed_case_job265128_1_detailed.md
  Case 2: ** USBHost — xHCI Host Controller Probe Failure
  1. Failed case: ** USBHost — xHCI Host Controller Probe Failure
  2. Root cause: ** The xHCI host controller (xhci-hcd.0.auto) probe failed with -ETIMEDOUT (error -110) during hardware setup at 25 seconds into boot on qcs8300-ride. The controller failed to respond to initialization commands within the expected timeout window, preventing USB root hub enumeration. This is a driver probe hard-fail caused by the underlying DWC3 Qcom wrapper not establishing proper interconnect bandwidth or power state before xHCI attempts hardware handshake.
  3. Possible fix: Revert or adjust the interconnect bandwidth vote changes in PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (drivers/usb/dwc3/dwc3-qcom.c). The PR reduced USB_MEMORY_PEAK_SS_BW from 5500 MB/s to 4500 MB/s and cleared APPS_USB_AVG_BW to 0. On qcs8300, this may be starving the USB controller during probe, causing the xHCI setup timeout. Restore the original bandwidth votes or tune them specifically for qcs8300 to ensure the controller receives sufficient interconnect bandwidth during initialization.
  4. Detail analysis attachment: failed_case_job265128_2_detailed.md
  Case 3: Ethernet_Basic_Validation
  1. Failed case: Ethernet_Basic_Validation
  2. Root cause: Aquantia AQR115C PHY driver probe failed with error -22 (EINVAL) because the device tree node for the PHY is missing the required firmware-name property, preventing firmware load and PHY initialization. When the qcom-ethqos Ethernet driver attempts to bring up eth0, it cannot attach to the PHY because the PHY probe failed earlier.
  3. Possible fix: Add the firmware-name property to the Aquantia AQR115C PHY device tree node in the qcs8300-ride DTS file. The property should specify the path to the Aquantia PHY firmware file (typically firmware-name = "Rhe-05.06-Candidate9-AQR_Mediatek_23B_P5_ID45824_LCLVER1.cld"; or similar, depending on the specific firmware version available for this PHY). This is a board-level DT fix, not a PR-introduced regression.
  4. Detail analysis attachment: failed_case_job265128_3_detailed.md
  Case 4: KVM_Driver — /dev/kvm not available (platform limitation, not a kernel regression)
  1. Failed case: KVM_Driver — /dev/kvm not available (platform limitation, not a kernel regression)
  2. Root cause: The qcs8300-ride target is running as a guest VM under Gunyah hypervisor (EFI selection='el2gh', arm-pv paravirtualization active, gunyah-md-region reserved). KVM requires EL2 (hypervisor) privilege level to create /dev/kvm, but when running as a guest VM, the kernel executes at EL1 and EL2 is owned by the host hypervisor. Nested virtualization (KVM-inside-VM) is not supported in this Gunyah guest configuration, despite CONFIG_KVM=y being set. The kernel correctly detects it cannot initialize KVM and does not create /dev/kvm.
  3. Possible fix: This is expected behavior for a guest VM configuration, not a bug. To enable KVM testing: (1) flash the kernel to bare-metal mode (non-virtualized boot path without Gunyah hypervisor), or (2) if nested virtualization is required, enable it in the Gunyah hypervisor configuration and ensure the guest kernel is booted with nested virt support, or (3) skip KVM tests when the LAVA job targets a virtualized guest environment by adding a pre-flight check for /sys/hypervisor or paravirt indicators.
  4. Detail analysis attachment: failed_case_job265128_4_detailed.md
  Case 5: KVM_EL2_DTB
  1. Failed case: KVM_EL2_DTB
  2. Root cause: Could not be determined confidently from available logs.
  3. Possible fix: This is a platform/test configuration mismatch, not a PR-introduced regression (the PR modifies USB bandwidth settings only). The KVM tests should be skipped on QCS8300 Ride when running under Gunyah hypervisor, or the platform firmware/boot configuration should be changed to boot in KVM mode instead of Gunyah mode if KVM testing is required. Add a test prerequisite check to skip KVM tests when running under non-KVM hypervisors (detect via /sys/hypervisor/type or hypervisor boot messages).
  4. Detail analysis attachment: failed_case_job265128_5_detailed.md
  Case 6: KVM_Infra — KVM device node unavailable (pre-existing platform limitation)
  1. Failed case: KVM_Infra — KVM device node unavailable (pre-existing platform limitation)
  2. Root cause: /dev/kvm device node is not present because the KVM driver did not initialize during boot on the qcs8300-ride platform. CONFIG_KVM is enabled in the kernel config, but no KVM initialization messages appear in the boot log, indicating the KVM driver did not probe. This is a pre-existing platform configuration issue unrelated to the PR's USB bandwidth changes.
  3. Possible fix: This is not a PR-introduced regression. The qcs8300-ride (Monaco) platform does not have KVM support enabled in its device tree or the hardware does not support virtualization at EL2. To enable KVM: (1) verify the SoC supports ARMv8 virtualization extensions, (2) ensure the bootloader boots the kernel at EL2 (not EL1), (3) add the required KVM device tree nodes if missing, (4) verify CONFIG_ARM64_VHE or CONFIG_KVM_ARM_HOST are enabled if needed. If KVM is not expected on this platform, mark these tests as "not applicable" for qcs8300-ride in the CI test matrix.
  4. Detail analysis attachment: failed_case_job265128_6_detailed.md
  Case 7: 0_qcom-next-ci-premerge-tests
  1. Failed case: 0_qcom-next-ci-premerge-tests
  2. Root cause: LAVA infrastructure state tracking error — test runner completed successfully (received <LAVA_TEST_RUNNER EXIT> signal and "lava_test_shell seems to have completed" confirmation) but LAVA dispatcher incorrectly marked the test run as "unfinished" and failed it.
  3. Possible fix: Re-trigger the LAVA job — this is a transient LAVA dispatcher state tracking issue, not a kernel regression; if the issue persists across multiple retries, escalate to LAVA infrastructure team to investigate dispatcher state machine logic for test completion detection.
  4. Detail analysis attachment: failed_case_job265128_7_detailed.md
Job 265129 | SoC purwa-evk

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265129

Failed test cases in LAVA job 265129 (SoC: purwa-evk).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: Four pre-existing probe failures detected on purwa-evk: qcom_qseecom_uefisecapp (-EBUSY), two qcom-pcie controllers (-ENODATA indicating missing DT properties), and regulatory.db firmware (-ENOENT). These are platform configuration issues unrelated to the PR's USB bandwidth vote changes in dwc3-qcom.c.
  3. Possible fix: These probe failures are pre-existing board/DT issues, not regressions introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365. The Probe_Failure_Check test should either: (1) exclude known platform-specific probe failures for purwa-evk from the failure criteria, or (2) fix the underlying issues: add missing PCIe DT properties (clocks/regulators/resets causing -ENODATA), resolve qseecom TZ app conflict (-EBUSY), and optionally add regulatory.db to the rootfs. The PR itself is not the cause and should not be blocked by this test failure.
  4. Detail analysis attachment: failed_case_job265129_1_detailed.md
  Case 2: smmu
  1. Failed case: smmu
  2. Root cause: Video codec device aa00000.video-codec is not attached to any IOMMU group on purwa-evk platform; the device tree for this SoC does not include the required iommus property for the video codec node, causing the SMMU test to fail its critical master protection check.
  3. Possible fix: Add the missing iommus property to the video codec device tree node for purwa-evk (likely in arch/arm64/boot/dts/qcom/purwa.dtsi or board-specific DTS); this is a pre-existing platform issue unrelated to the USB bandwidth changes in PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365.
  4. Detail analysis attachment: failed_case_job265129_2_detailed.md
  Case 3: KVM_Driver — /dev/kvm device node not available
  1. Failed case: KVM_Driver — /dev/kvm device node not available
  2. Root cause: KVM driver initialization failed with "HYP mode not available" because the purwa-evk platform does not support ARM Virtualization Extensions (EL2/HYP mode). The hardware/firmware does not provide the hypervisor execution level required for KVM operation. This is a pre-existing platform limitation, not a regression introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (which only modifies USB bandwidth parameters).
  3. Possible fix: Mark KVM_Driver, KVM_EL2_DTB, and KVM_Infra tests as expected failures (XFAIL) or skip them entirely in the CI test suite for purwa-evk and other platforms that lack EL2/HYP support. Add a platform capability check to the test runner to automatically skip virtualization tests on platforms without hypervisor support.
  4. Detail analysis attachment: failed_case_job265129_3_detailed.md
  Case 4: KVM Driver Initialization Failure — HYP mode not available
  1. Failed case: KVM Driver Initialization Failure — HYP mode not available
  2. Root cause: The purwa-evk platform does not provide EL2 (hypervisor exception level) access to the Linux kernel; KVM initialization fails at probe with "HYP mode not available" (line 3814), preventing /dev/kvm device node creation and causing all KVM-dependent tests (KVM_Driver, KVM_EL2_DTB, KVM_Infra) to fail.
  3. Possible fix: This is a platform/firmware limitation, not a kernel regression. The purwa-evk board firmware boots the kernel at EL1 (non-secure supervisor mode) without EL2 access. To enable KVM: (1) verify the board supports virtualization extensions in hardware, (2) update the bootloader/firmware to boot Linux at EL2 or with VHE (Virtualization Host Extensions) enabled, or (3) exclude KVM tests from the purwa-evk CI test suite if the platform is not intended to support virtualization.
  4. Detail analysis attachment: failed_case_job265129_4_detailed.md
  Case 5: KVM Infrastructure Failure — KVM/Gunyah Hypervisor Incompatibility
  1. Failed case: KVM Infrastructure Failure — KVM/Gunyah Hypervisor Incompatibility
  2. Root cause: KVM driver initialization failed because Gunyah hypervisor is running on Purwa EVK and owns ARM EL2 (HYP mode); KVM requires exclusive EL2 access and cannot coexist with Gunyah, resulting in "HYP mode not available" error and no /dev/kvm device creation.
  3. Possible fix: Disable KVM tests for Purwa EVK platform in the LAVA job definition, as this platform runs Gunyah hypervisor which is architecturally incompatible with KVM. Add platform-specific test filtering: skip all KVM tests when target is purwa-evk or when Gunyah hypervisor is detected in boot logs.
  4. Detail analysis attachment: failed_case_job265129_5_detailed.md
  Case 6: KVM_Infra (test infrastructure limitation — platform does not support KVM)
  1. Failed case: KVM_Infra (test infrastructure limitation — platform does not support KVM)
  2. Root cause: The Purwa EVK platform does not have EL2 (Hypervisor mode) available, as indicated by the kernel message kvm [1]: HYP mode not available at boot time. CONFIG_KVM is enabled in the kernel config, but the hardware does not support virtualization, so /dev/kvm is never created. This is a platform hardware limitation, not a kernel bug or PR-introduced regression.
  3. Possible fix: Mark KVM tests as "skip" or "not applicable" for Purwa EVK platform in the LAVA test definition, or add a platform capability check before running KVM tests. The test should detect the absence of EL2 support and skip gracefully rather than reporting a failure. This is not a kernel issue that requires a code fix.
  4. Detail analysis attachment: failed_case_job265129_6_detailed.md
Job 265130 | SoC qcs615-ride

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265130

Failed test cases in LAVA job 265130 (SoC: qcs615-ride).

  Case 1: ** Probe_Failure_Check — cfg80211 regulatory.db firmware load failure (benign)
  1. Failed case: ** Probe_Failure_Check — cfg80211 regulatory.db firmware load failure (benign)
  2. Root cause: ** The cfg80211 wireless subsystem attempted to load the optional regulatory.db firmware file at boot (timestamp 8.055062s) and failed with error -2 (ENOENT — file not found). This file is not present in the rootfs /lib/firmware/ directory. WiFi functionality is unaffected because cfg80211 falls back to compiled-in regulatory rules.
  3. Possible fix: This is a false positive test failure. No fix required. The Probe_Failure_Check test should be updated to exclude regulatory.db from the failure pattern, as it is an optional file and WiFi functional tests (WiFi_Firmware_Driver, WiFi_OnOff) both passed. Alternatively, add regulatory.db to the rootfs if strict regulatory compliance is required for the target region.
  4. Detail analysis attachment: failed_case_job265130_1_detailed.md
  Case 2: smmu (test expectation issue — not a kernel failure)
  1. Failed case: smmu (test expectation issue — not a kernel failure)
  2. Root cause: The smmu test expects child devices aa00000.video-codec:video-decoder and aa00000.video-codec:video-encoder to have independent IOMMU group attachments, but the Linux kernel IOMMU subsystem only attaches the parent platform device (aa00000.video-codec) to IOMMU group 7. The child devices are virtual V4L2 video nodes created by the video codec driver and do not require separate IOMMU attachments. This is expected kernel behavior on qcs615-ride, not a regression.
  3. Possible fix: Update the smmu test logic to skip IOMMU group attachment checks for virtual child devices (video-decoder, video-encoder) that are created by V4L2 drivers. Only verify IOMMU attachment for the parent platform device (aa00000.video-codec), which is correctly attached to group 7.
  4. Detail analysis attachment: failed_case_job265130_2_detailed.md
  Case 3: KVM_Driver
  1. Failed case: KVM_Driver
  2. Root cause: KVM driver cannot initialize because the kernel is running at EL1 (kernel exception level) instead of EL2 (hypervisor exception level). The boot log shows CPU: All CPU(s) started at EL1 and KVM reports kvm [1]: HYP mode not available. On ARM64, KVM requires the kernel to run at EL2 to provide virtualization support. This is a platform/firmware boot configuration issue on qcs615-ride, not a kernel regression.
  3. Possible fix: Configure the bootloader/firmware to boot the kernel at EL2 instead of EL1. This requires updating the boot chain (ABL/UEFI) to enter the kernel at EL2. Alternatively, if EL2 boot is not supported on this platform, mark KVM tests as expected-to-fail for qcs615-ride in the LAVA test suite configuration.
  4. Detail analysis attachment: failed_case_job265130_3_detailed.md
  Case 4: KVM Driver Initialization Failure — HYP mode unavailable
  1. Failed case: KVM Driver Initialization Failure — HYP mode unavailable
  2. Root cause: The qcs615-ride board boots with the Gunyah hypervisor running at EL2, which prevents the Linux KVM driver from initializing. KVM requires exclusive control of EL2 (HYP mode) to provide virtualization, but Gunyah already owns this privilege level. The kernel message kvm [1]: HYP mode not available at boot time [3.376685] confirms KVM detected this conflict and aborted initialization, resulting in no /dev/kvm device node creation.
  3. Possible fix: This is a platform configuration issue, not a kernel regression. To enable KVM on qcs615-ride, the board must boot without the Gunyah hypervisor. Options: (1) Reconfigure the boot chain to load the kernel directly at EL2 without Gunyah, or (2) Mark KVM tests as "not applicable" for qcs615-ride in the LAVA test suite when Gunyah is present. The PR under test (USB bandwidth changes) is unrelated and does not cause this failure.
  4. Detail analysis attachment: failed_case_job265130_4_detailed.md
  Case 5: KVM_Infra
  1. Failed case: KVM_Infra
  2. Root cause: KVM initialization failed because HYP mode (EL2) is not available on the QCS615 platform, preventing /dev/kvm device node creation despite CONFIG_KVM being enabled in the kernel configuration.
  3. Possible fix: This is a platform hardware/firmware limitation, not a kernel regression. The QCS615 SoC does not support EL2 hypervisor mode required for KVM. Exclude KVM tests from the QCS615 test suite, or mark them as expected-to-skip for this platform.
  4. Detail analysis attachment: failed_case_job265130_5_detailed.md
  Case 6: KVM_Infra — /dev/kvm device node not available (pre-existing platform limitation)
  1. Failed case: KVM_Infra — /dev/kvm device node not available (pre-existing platform limitation)
  2. Root cause: KVM initialization failed during early boot with "kvm [1]: HYP mode not available" because the QCS615 platform does not support EL2 hypervisor mode, preventing /dev/kvm device node creation. CONFIG_KVM is enabled but the hardware lacks the required virtualization extensions.
  3. Possible fix: This is a pre-existing platform limitation unrelated to the PR (which only modifies USB bandwidth parameters). The KVM tests should be skipped on QCS615 targets that lack EL2 support, or the test suite should be updated to mark this as an expected skip rather than a failure when "HYP mode not available" is detected in dmesg.
  4. Detail analysis attachment: failed_case_job265130_6_detailed.md
Job 265131 | SoC shikra-iqs-evk

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265131

Failed test cases in LAVA job 265131 (SoC: shikra-iqs-evk).

  Case 1: GIC (Test Script Bug — CPU Count Mismatch)
  1. Failed case: GIC (Test Script Bug — CPU Count Mismatch)
  2. Root cause: The GIC test script assumes an 8-CPU system and attempts to parse timer interrupt counts for CPUs 4-7, but the shikra-iqs-evk platform has only 4 CPUs (0-3). The script's line 75 bash integer comparison fails when it encounters non-numeric fields ("GICv3", "Level", "arch_timer") from /proc/interrupts for non-existent CPUs, causing false test failures for CPUs 4-7 while CPUs 0-3 correctly pass.
  3. Possible fix: Update the GIC test script to dynamically detect the number of online CPUs from /sys/devices/system/cpu/online (or /proc/cpuinfo) and only validate timer interrupts for CPUs that actually exist on the target platform, rather than hardcoding an assumption of 8 CPUs.
  4. Detail analysis attachment: failed_case_job265131_1_detailed.md
  Case 2: ** Probe_Failure_Check — Pre-existing Platform Firmware and Hardware Issues
  1. Failed case: ** Probe_Failure_Check — Pre-existing Platform Firmware and Hardware Issues
  2. Root cause: ** The test detected 7 pre-existing probe failures on shikra-iqs-evk: (1) missing firmware files (ucsi_ccg ccg_secondary.cyacd2, regulatory.db) not present in rootfs, (2) TPM SPI hardware timeout (board-specific), and (3) audio codec deferred probe due to missing clock dependencies. None are introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365, which only modifies USB bandwidth votes in dwc3-qcom.c and does not affect probe-time driver initialization.
  3. Possible fix: Mark this test case as PASS (known pre-existing issues) for PR validation purposes. The probe failures are platform/infrastructure issues requiring separate fixes: (1) add missing firmware files to rootfs, (2) investigate TPM hardware connectivity on shikra-iqs-evk, (3) fix audio codec DT clock dependencies. These issues do not block PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 merge as they are unrelated to the USB bandwidth optimization changes.
  4. Detail analysis attachment: failed_case_job265131_2_detailed.md
  Case 3: ** PCIe
  1. Failed case: ** PCIe
  2. Root cause: ** Test infrastructure issue — QPS615 PCIe switch firmware file TC956X_Firmware_PCIeBridge.bin is missing from the rootfs /lib/firmware directory, and the corresponding kernel driver module tc956x_pcie_eth is not present for the running kernel. This is a board-specific test dependency failure, not a kernel functionality regression. PCIe hardware enumeration succeeded (7 devices detected, all links operational, no AER errors), confirming the PCIe controller and kernel PCIe subsystem are working correctly.
  3. Possible fix: Add the QPS615 firmware file TC956X_Firmware_PCIeBridge.bin to the Yocto rootfs recipe under /lib/firmware/, and ensure the tc956x_pcie_eth kernel module is built and included in the image for shikra-iqs-evk. This is a build/image configuration fix, not a kernel code fix. The PR changes (USB bandwidth tuning in dwc3-qcom.c) are unrelated to this failure and should not be blocked by it.
  4. Detail analysis attachment: failed_case_job265131_3_detailed.md
  Case 4: ** USBHost
  1. Failed case: ** USBHost
  2. Root cause: ** Test infrastructure configuration mismatch — shikra-iqs-evk USB ports are configured in gadget/peripheral mode (device tree or Type-C role detection), but the USBHost test expects host mode with root hubs and xHCI controller. The kernel USB subsystem is healthy with no errors; this is not a kernel regression.
  3. Possible fix: Configure the shikra-iqs-evk device tree to set dr_mode = "host" for the USB controller nodes (4c00000.usb and/or 4e00000.usb), or ensure Type-C role switching is configured to default to host mode for this test. Alternatively, skip the USBHost test for boards where USB is intentionally configured in gadget-only mode.
  4. Detail analysis attachment: failed_case_job265131_4_detailed.md
  Case 5: ** BT_SCAN (Test Environment Issue — No External Bluetooth Devices)
  1. Failed case: ** BT_SCAN (Test Environment Issue — No External Bluetooth Devices)
  2. Root cause: ** Bluetooth hardware and driver are fully functional (BT_ON_OFF passed, hci0 adapter operational, power on/off successful), but no external Bluetooth devices are present in the LAVA lab environment for the scan test to discover. The test executed 3 scan attempts with multiple fallback strategies (expect PTY, interactive bluetoothctl) and correctly reported "Discovery started" and "Discovering: yes", but the devices command returned no results because no scannable Bluetooth devices exist in RF range of the shikra-iqs-evk board.
  3. Possible fix: This is not a kernel bug. The PR (USB bandwidth vote changes in dwc3-qcom.c) does not touch Bluetooth code. Re-run the test in a LAVA environment with at least one active Bluetooth device (phone, beacon, or dedicated test device) within RF range of the board. If the LAVA lab does not support Bluetooth scan testing, mark BT_SCAN as SKIP or remove it from the mandatory test suite for this board/lab configuration.
  4. Detail analysis attachment: failed_case_job265131_5_detailed.md
  Case 6: ** KVM_Driver (Test Infrastructure Issue — Not a Kernel Bug)
  1. Failed case: ** KVM_Driver (Test Infrastructure Issue — Not a Kernel Bug)
  2. Root cause: ** The shikra-iqs-evk platform boots with Gunyah hypervisor firmware active at EL2 (confirmed by boot log: "Hypervisor cold boot, version: gunyah-mobile-98db777b6"). When a hypervisor owns EL2, the Linux KVM subsystem cannot initialize because it requires exclusive EL2 access. The kernel correctly reports kvm [1]: HYP mode not available at boot. The test expects /dev/kvm to exist but this is architecturally impossible on this platform configuration.
  3. Possible fix: Update the LAVA test definition to skip KVM_Driver, KVM_EL2_DTB, and KVM_Infra tests on platforms that boot with Gunyah or other Type-1 hypervisors. Add a pre-test check: if dmesg | grep -q "HYP mode not available" or if hypervisor boot banner is present, mark these tests as SKIP rather than FAIL.
  4. Detail analysis attachment: failed_case_job265131_6_detailed.md
  Case 7: ** KVM_EL2_DTB
  1. Failed case: ** KVM_EL2_DTB
  2. Root cause: ** Platform booted into EL1 (kernel mode) instead of EL2 (hypervisor mode). KVM on ARM64 requires the CPU to boot into EL2 to enable virtualization extensions. The bootloader/firmware on the Shikra IQS EVK board is not configured to boot into EL2, causing the kernel to report "HYP mode not available" at boot time (dmesg timestamp 3.176s).
  3. Possible fix: Configure the bootloader (ABL/UEFI) on the Shikra IQS EVK board to boot the kernel at EL2 instead of EL1. This typically requires setting a bootloader flag or updating the boot chain configuration to enable hypervisor mode. Verify the fix by checking for the absence of "HYP mode not available" in dmesg and confirming /dev/kvm device node creation after reboot.
  4. Detail analysis attachment: failed_case_job265131_7_detailed.md
  Case 8: KVM_Infra — KVM not available (platform limitation)
  1. Failed case: KVM_Infra — KVM not available (platform limitation)
  2. Root cause: The Shikra IQS EVK platform boots with Gunyah hypervisor (gunyah-mobile-98db777b6) running at EL2, which prevents Linux KVM from initializing. Kernel logs show kvm [1]: HYP mode not available at boot because EL2 is already occupied by the Gunyah hypervisor. CONFIG_KVM is enabled in the kernel config, but /dev/kvm cannot be created because the KVM subsystem cannot claim EL2.
  3. Possible fix: This is not a kernel regression introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (USB bandwidth tuning patch). The failure is a pre-existing platform configuration issue. The Shikra IQS EVK firmware includes Gunyah hypervisor by default, making nested virtualization (KVM under Gunyah) unavailable. To enable KVM testing on this platform, either: (1) flash firmware without Gunyah hypervisor (if available), or (2) exclude KVM tests from the Shikra IQS EVK test suite, or (3) use a different platform that boots Linux directly at EL2 for KVM validation.
  4. Detail analysis attachment: failed_case_job265131_8_detailed.md
  Case 9: KVM_Infra
  1. Failed case: KVM_Infra
  2. Root cause: KVM driver initialization failed with "HYP mode not available" because the shikra-iqs-evk platform is running with Gunyah hypervisor (gunyah-mobile-98db777b6) in a configuration that does not expose EL2 virtualization extensions to Linux. CONFIG_KVM is enabled in the kernel but the /dev/kvm device node is never created because kvm_arch_init() fails when it detects EL2 is not accessible from the kernel (likely running at EL1 under the hypervisor).
  3. Possible fix: This is a pre-existing platform/firmware configuration issue, not a regression introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (which only modifies USB bandwidth votes in dwc3-qcom.c). To enable KVM on this platform, the Gunyah hypervisor configuration must be changed to either: (1) boot Linux at EL2 instead of EL1, or (2) enable nested virtualization support in Gunyah to expose virtual EL2 to the guest kernel. The test should be marked as expected-fail or skipped on shikra-iqs-evk until the platform firmware is updated to support KVM.
  4. Detail analysis attachment: failed_case_job265131_9_detailed.md
Job 265132 | SoC qcs9100-ride

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265132

Failed test cases in LAVA job 265132 (SoC: qcs9100-ride).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: Two pre-existing platform issues unrelated to PR changes: (1) regulatory.db firmware file missing from rootfs (benign cfg80211 warning), and (2) Aquantia AQR115C PHY probe failure due to missing firmware-name DT property (error -22/EINVAL), preventing end0 Ethernet interface from attaching to PHY.
  3. Possible fix: For regulatory.db: benign warning, no action required (cfg80211 falls back to built-in regulatory data). For Aquantia PHY: add firmware-name property to the PHY DT node at stmmac-0:08 (23040000.ethernet MDIO bus) in qcs9100-ride device tree, or verify PHY firmware file is present in rootfs if firmware-name is already specified.
  4. Detail analysis attachment: failed_case_job265132_1_detailed.md
  Case 2: ** smmu (IOMMU group attachment validation failure)
  1. Failed case: ** smmu (IOMMU group attachment validation failure)
  2. Root cause: ** The video codec device aa00000.video-codec on qcs9100-ride failed to attach to an IOMMU group, indicating the device either failed to probe, is missing from the device tree, or the IOMMU binding is not configured correctly for this platform.
  3. Possible fix: Verify the qcs9100 device tree includes the video codec node at aa00000.video-codec with correct iommus property; if the device is intentionally disabled or not supported on qcs9100-ride, update the smmu test's critical device list to exclude it for this platform.
  4. Detail analysis attachment: failed_case_job265132_2_detailed.md
  Case 3: Ethernet_Basic_Validation
  1. Failed case: Ethernet_Basic_Validation
  2. Root cause: Aquantia AQR115C PHY driver probe failed with error -22 (EINVAL) during boot due to missing firmware-name property in device tree, preventing end0 interface from attaching to PHY when brought up administratively.
  3. Possible fix: Add the missing firmware-name property to the Aquantia AQR115C PHY node in the qcs9100-ride device tree, or configure the PHY driver to use a default firmware path if the property is optional for this hardware revision.
  4. Detail analysis attachment: failed_case_job265132_3_detailed.md
  Case 4: KVM_Driver
  1. Failed case: KVM_Driver
  2. Root cause: KVM driver initialization failed because the platform does not support EL2 (Hypervisor mode) - kernel message "kvm [1]: HYP mode not available" indicates the CPU is not running with virtualization extensions enabled or accessible, preventing /dev/kvm device node creation.
  3. Possible fix: This is a platform hardware/firmware limitation, not a kernel regression. The QCS9100 (Lemans) platform either does not have ARM virtualization extensions (EL2) enabled in the bootloader/firmware, or the SoC variant does not support KVM. Verify bootloader configuration enables EL2, or exclude KVM tests from the CI test suite for this platform if virtualization is not a supported feature.
  4. Detail analysis attachment: failed_case_job265132_4_detailed.md
  Case 5: KVM Driver Initialization Failure — HYP mode not available
  1. Failed case: KVM Driver Initialization Failure — HYP mode not available
  2. Root cause: The qcs9100-ride platform is configured to boot under Gunyah hypervisor (confirmed by boot log: "Hypervisor cold boot, version: gunyah-3fe287d48" and "Gunyah based bootup"), which runs the kernel as a guest VM without direct EL2 (Hypervisor mode) access. KVM requires native EL2 access to create /dev/kvm and enable virtualization, but cannot initialize when the kernel itself is running as a guest under another hypervisor.
  3. Possible fix: This is not a kernel bug or PR-introduced regression — it is expected behavior for this platform configuration. To enable KVM testing on qcs9100-ride: (1) reconfigure the platform to boot in native EL2 mode without Gunyah hypervisor, or (2) use nested virtualization if supported by Gunyah (requires hypervisor configuration changes), or (3) exclude KVM tests from the CI test suite for Gunyah-based boot configurations, as KVM and Gunyah are mutually exclusive virtualization solutions.
  4. Detail analysis attachment: failed_case_job265132_5_detailed.md
  Case 6: KVM_Infra
  1. Failed case: KVM_Infra
  2. Root cause: KVM infrastructure test failed because the qcs9100-ride (LeMans) platform does not support ARM Virtualization Extensions (EL2/HYP mode) — kernel message "kvm [1]: HYP mode not available" indicates hardware/firmware does not provide EL2, preventing /dev/kvm device creation despite CONFIG_KVM being enabled.
  3. Possible fix: This is a platform hardware/firmware limitation, not a kernel regression introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (USB bandwidth tuning). The qcs9100-ride board does not support virtualization. To resolve: (1) skip KVM tests on qcs9100-ride in CI job definitions, or (2) run KVM tests only on platforms with EL2 support (e.g., SA8775P, SM8550), or (3) if EL2 is expected on this platform, verify bootloader/firmware configuration enables virtualization extensions and boots kernel at EL2.
  4. Detail analysis attachment: failed_case_job265132_6_detailed.md
  Case 7: Driver Initialization Failure — KVM driver probe
  1. Failed case: Driver Initialization Failure — KVM driver probe
  2. Root cause: KVM driver fails to initialize because ARM Hypervisor mode (EL2) is not available on the qcs9100-ride platform. The kernel message kvm [1]: HYP mode not available indicates the firmware/bootloader has not enabled EL2 or secure firmware is restricting access. This is a pre-existing platform configuration issue, not related to the PR changes (USB bandwidth tuning).
  3. Possible fix: This is not a PR-introduced regression. To enable KVM on qcs9100-ride: (1) verify the board firmware/bootloader supports and enables EL2 hypervisor mode, (2) check if secure firmware (TrustZone) configuration allows EL2 access, (3) if the platform does not support virtualization extensions, mark the KVM tests as expected-to-fail for this board in the CI configuration.
  4. Detail analysis attachment: failed_case_job265132_7_detailed.md
Job 265133 | SoC lemans-evk

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265133

Failed test cases in LAVA job 265133 (SoC: lemans-evk).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: The test flagged three firmware load errors (regulatory.db, wcnhpbtfw21.tlv, hpbtfw21.tlv) as probe failures, but these are benign fallback attempts during driver initialization; Bluetooth firmware successfully loaded via fallback mechanism (BT_FW_KMD_Service and BT_ON_OFF tests both passed), and WiFi regulatory database failure is a known non-critical fallback (WiFi_Firmware_Driver test passed).
  3. Possible fix: Suppress these specific firmware fallback errors in the Probe_Failure_Check test logic, or exclude firmware load errors with error -2 (ENOENT) when the corresponding functional tests (BT_ON_OFF, WiFi_Firmware_Driver) pass; alternatively, update the test to only flag firmware errors that result in non-functional hardware.
  4. Detail analysis attachment: failed_case_job265133_1_detailed.md
  Case 2: smmu
  1. Failed case: smmu
  2. Root cause: Could not be determined confidently from available logs.
  3. Possible fix: This is a pre-existing platform issue, not a PR regression. To resolve: (1) verify the video codec device tree node at aa00000.video-codec includes a valid iommus property pointing to the correct SMMU instance; (2) confirm the video codec driver (likely qcom-venus or qcom-iris) is enabled in the kernel config and successfully probes the device; (3) if the video codec hardware is not present or functional on lemans-evk, update the test expectations to exclude this device from the critical master list for this platform.
  4. Detail analysis attachment: failed_case_job265133_2_detailed.md
  Case 3: 0_qcom-next-ci-premerge-tests
  1. Failed case: 0_qcom-next-ci-premerge-tests
  2. Root cause: The test suite reported two individual test failures: (1) Probe_Failure_Check failed due to benign firmware load errors (regulatory.db and Bluetooth firmware files not found - these are known false positives as the devices function correctly), and (2) smmu test failed because the Video codec device (aa00000.video-codec) is missing IOMMU group attachment. The video codec IOMMU attachment failure is unrelated to the PR changes (USB bandwidth tuning in dwc3-qcom.c) and represents a pre-existing platform/DT configuration issue on lemans-evk.
  3. Possible fix: The Probe_Failure_Check failure is a false positive (firmware files load via fallback paths; devices work). The smmu failure requires adding the video codec device to an IOMMU group in the lemans-evk device tree - verify the video-codec@aa00000 node has a valid iommus property pointing to the appropriate SMMU instance. This is a platform DT issue, not a regression introduced by PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365.
  4. Detail analysis attachment: failed_case_job265133_3_detailed.md
Job 265134 | SoC monaco-evk

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265134

Failed test cases in LAVA job 265134 (SoC: monaco-evk).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: Pre-existing WiFi driver probe failure — ath11k_pci PCIe WiFi driver (WCN6855 hw2.1) probe timed out with error -110 (ETIMEDOUT) during MHI firmware load, unrelated to PR changes in USB bandwidth configuration.
  3. Possible fix: This is a known pre-existing WiFi firmware/probe issue on iq-8275-evk boards with WCN6855 PCIe WiFi. The PR modifies only USB dwc3-qcom bandwidth votes and does not touch WiFi, PCIe, MHI, or interconnect paths used by ath11k. Approve the PR; track the WiFi probe timeout separately as a board/firmware/interconnect issue requiring investigation of MHI firmware load path, PCIe link training, and interconnect bandwidth for the WiFi subsystem.
  4. Detail analysis attachment: failed_case_job265134_1_detailed.md
  Case 2: WiFi_Firmware_Driver
  1. Failed case: WiFi_Firmware_Driver
  2. Root cause: ath11k_pci driver probe failed with -ETIMEDOUT (-110) because the required firmware file ath11k/WCN6855/hw2.1/nfa765/amss.bin is missing from the filesystem, causing MHI power-up to timeout during PCIe WiFi device initialization on monaco-evk.
  3. Possible fix: Add the missing firmware file ath11k/WCN6855/hw2.1/nfa765/amss.bin to the rootfs firmware directory (/lib/firmware/ath11k/WCN6855/hw2.1/nfa765/) in the build image; this is a board-specific firmware packaging issue unrelated to the PR's USB bandwidth changes.
  4. Detail analysis attachment: failed_case_job265134_2_detailed.md
  Case 3: WiFi_OnOff — WiFi Driver Probe Failure
  1. Failed case: WiFi_OnOff — WiFi Driver Probe Failure
  2. Root cause: ath11k_pci driver probe failed with error -110 (ETIMEDOUT) on monaco-evk because MHI firmware load failed (ath11k/WCN6855/hw2.1/nfa765/amss.bin returned -ENOENT); the WCN6855 WiFi chip could not be initialized without its firmware, causing the probe timeout during MHI power-up sequence.
  3. Possible fix: Add the missing WCN6855 firmware file (ath11k/WCN6855/hw2.1/nfa765/amss.bin) to the rootfs firmware directory (/lib/firmware/); this is a board/image configuration issue unrelated to PR FROMLIST: usb: dwc3: qcom: Keep USB bandwidth votes in nominal range #1365 (which only modifies USB bandwidth settings in dwc3-qcom.c).
  4. Detail analysis attachment: failed_case_job265134_3_detailed.md
  Case 4: 0_qcom-next-ci-premerge-tests
  1. Failed case: 0_qcom-next-ci-premerge-tests
  2. Root cause: WiFi driver probe failure — ath11k_pci probe failed with error -110 (timeout) while loading firmware ath11k/WCN6855/hw2.1/nfa765/amss.bin, causing WiFi_OnOff and Probe_Failure_Check tests to fail. The PR changes USB interconnect bandwidth votes which may have indirectly affected PCIe/MHI timing for the WiFi module on monaco-evk.
  3. Possible fix: Investigate whether the reduced USB bandwidth votes (USB_MEMORY_PEAK_SS_BW from 5500→4500 MB/s, APPS_USB_AVG_BW from 40→0, APPS_USB_PEAK_BW from 80→40 MB/s) are causing interconnect contention or timing changes that affect PCIe device enumeration or MHI firmware loading on monaco-evk. Verify WiFi probe succeeds on other platforms. If monaco-evk-specific, check if the board's PCIe/WiFi power sequencing or interconnect topology requires higher bandwidth votes during boot.
  4. Detail analysis attachment: failed_case_job265134_4_detailed.md
Job 265135 | SoC hamoa-evk

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265135

Failed test cases in LAVA job 265135 (SoC: hamoa-evk).

  Case 1: Probe_Failure_Check
  1. Failed case: Probe_Failure_Check
  2. Root cause: Test detected two pre-existing platform-specific probe failures unrelated to the PR: (1) qcom_qseecom_uefisecapp probe failed with -EBUSY (secure world service unavailable on hamoa-evk), and (2) regulatory.db firmware load failed with -ENOENT (file not present in rootfs). Both failures are benign: QSEE UEFI secure app is optional for this platform configuration, and WiFi functional tests (WiFi_Firmware_Driver, WiFi_OnOff) passed, confirming wireless regulatory functionality works via alternative mechanism.
  3. Possible fix: Suppress these two known benign probe failures in the Probe_Failure_Check test for hamoa-evk platform. The PR (USB bandwidth tuning in dwc3-qcom.c) does not touch QSEE or wireless regulatory code and cannot have introduced these failures. No kernel code fix required; update CI test filter to exclude these platform-specific benign probe failures.
  4. Detail analysis attachment: failed_case_job265135_1_detailed.md
  Case 2: smmu
  1. Failed case: smmu
  2. Root cause: The video codec device aa00000.video-codec is not attached to any IOMMU group on hamoa-evk, indicating missing or incorrect device tree IOMMU configuration for this device; the PR changes USB bandwidth settings and does not touch video codec or IOMMU configuration, making this a pre-existing platform issue.
  3. Possible fix: Add the missing iommus property to the video codec device tree node (aa00000.video-codec) in the hamoa-evk device tree to attach it to an appropriate IOMMU group; verify the device tree includes the correct IOMMU phandle and stream ID for the video codec hardware.
  4. Detail analysis attachment: failed_case_job265135_2_detailed.md
  Case 3: KVM_Driver
  1. Failed case: KVM_Driver
  2. Root cause: Could not be determined confidently from available logs.
  3. Possible fix: This is not a bug to fix in the kernel or PR. To enable KVM on hamoa-evk: (1) Update the board's bootloader/firmware to boot the kernel at EL2, or (2) Mark KVM tests as "skip" for hamoa-evk in the LAVA test suite configuration, as this platform does not support virtualization in its current firmware configuration.
  4. Detail analysis attachment: failed_case_job265135_3_detailed.md
  Case 4: KVM_EL2_DTB — KVM Driver Initialization Failure
  1. Failed case: KVM_EL2_DTB — KVM Driver Initialization Failure
  2. Root cause: KVM driver failed to initialize because hypervisor (EL2/HYP) mode is not available on the hamoa-evk platform. The kernel log shows kvm [1]: HYP mode not available at boot time, preventing /dev/kvm device node creation. This is a platform hardware/firmware limitation, not a kernel regression.
  3. Possible fix: This is not a PR-introduced regression (the PR only modifies USB bandwidth settings). The hamoa-evk platform does not support KVM/virtualization in hardware or firmware. Either: (1) exclude KVM tests from the hamoa-evk test suite, or (2) enable hypervisor support in the platform firmware/bootloader if the hardware supports it, or (3) mark this test as expected-fail/skip for hamoa-evk in the CI configuration.
  4. Detail analysis attachment: failed_case_job265135_4_detailed.md
  Case 5: KVM_Infra
  1. Failed case: KVM_Infra
  2. Root cause: Could not be determined confidently from available logs.
  3. Possible fix: Exclude KVM tests from the Hamoa EVK test suite, as this platform does not support virtualization. Add platform capability detection to the CI pipeline to skip virtualization tests on platforms without HYP mode support. Alternatively, run KVM tests only on platforms with confirmed virtualization support (e.g., platforms with Gunyah hypervisor or native KVM support).
  4. Detail analysis attachment: failed_case_job265135_5_detailed.md
  Case 6: KVM_Infra (Pre-existing Platform Limitation — Not PR-introduced)
  1. Failed case: KVM_Infra (Pre-existing Platform Limitation — Not PR-introduced)
  2. Root cause: The Hamoa EVK platform boots with Gunyah hypervisor (Type-1) which takes exclusive control of EL2 (Hypervisor mode), preventing KVM (Type-2 hypervisor) from initializing. Kernel message: kvm [1]: HYP mode not available.
  3. Possible fix: This is not a bug requiring a fix. KVM and Gunyah are mutually exclusive — only one can control EL2 at a time. To enable KVM testing on Hamoa EVK, the platform must be reconfigured to boot without Gunyah hypervisor (remove Gunyah from device tree and bootloader configuration). Alternatively, exclude KVM tests from the CI test suite for Gunyah-enabled platforms.
  4. Detail analysis attachment: failed_case_job265135_6_detailed.md
Job 265136 | SoC qcs6490-rb3gen2

LAVA job: https://lava-oss.qualcomm.com/scheduler/job/265136

Failed test cases in LAVA job 265136 (SoC: qcs6490-rb3gen2).

  Case 1: Probe_Failure_Check (Test Infrastructure False Positive)
  1. Failed case: Probe_Failure_Check (Test Infrastructure False Positive)
  2. Root cause: The Probe_Failure_Check test flagged two benign firmware load failures (-ENOENT) for optional hardware: (1) regulatory.db for WiFi regulatory database (WiFi functional, WiFi_OnOff test passed), and (2) renesas_usb_fw.mem for optional Renesas PCIe USB controller (platform USB functional, USBHost skipped as expected). Both failures are pre-existing on qcs6490-rb3gen2 and unrelated to the PR's USB bandwidth vote changes in dwc3-qcom driver.
  3. Possible fix: Update the Probe_Failure_Check test to suppress known benign firmware load failures for optional hardware on qcs6490-rb3gen2: add regex exclusions for regulatory.db (when WiFi functional tests pass) and renesas_usb_fw.mem (when platform USB is functional). Alternatively, configure the test to only fail on probe errors for critical platform devices, not optional PCIe peripherals.
  4. Detail analysis attachment: failed_case_job265136_1_detailed.md
  Case 2: 0_qcom-next-ci-premerge-tests
  1. Failed case: 0_qcom-next-ci-premerge-tests
  2. Root cause: LAVA test runner completed all test cases successfully (including the last test qcom_hwrng at line 6618) and printed <LAVA_TEST_RUNNER EXIT> at line 6625, but LAVA dispatcher marked the test run as "unfinished" and failed it. This is a LAVA test harness signaling issue, not a kernel regression. The only genuine test failure was Probe_Failure_Check, which detected two pre-existing probe errors (regulatory.db firmware missing, xhci-pci-renesas probe failure) unrelated to the PR's USB bandwidth vote changes in dwc3-qcom.c.
  3. Possible fix: The test infrastructure issue requires investigation of why LAVA marked a completed test run as "unfinished" despite receiving <LAVA_TEST_RUNNER EXIT>. For the Probe_Failure_Check failure: (1) regulatory.db firmware is a known benign missing-firmware warning in many kernel configs; (2) xhci-pci-renesas probe failure is board-specific and unrelated to the PR changes. Neither issue is introduced by this PR. Re-trigger the CI job to verify if the LAVA "unfinished" marking is transient infrastructure noise.
  4. Detail analysis attachment: failed_case_job265136_2_detailed.md

@sgaud-quic

Copy link
Copy Markdown
Contributor

akakum-qualcomm add mainline PR to CR.

I see some queries on upstream change, please answer those before we can merge this.

@akakum-qualcomm

Copy link
Copy Markdown
Author

4700998

akakum-qualcomm add mainline PR to CR.

I see some queries on upstream change, please answer those before we can merge this.

i have already dropped mentioned change which reviewer had objection in commit,

+#define APPS_USB_PEAK_BW MBps_to_icc(80)

@akakum-qualcomm

Copy link
Copy Markdown
Author

mainline PR : qualcomm-linux/kernel-topics#2211

@qswat-orbit-external

Copy link
Copy Markdown

Merge Check Failed: CR Not Eligible for Merge

CR 4700998 is not eligible for merge.

The parent software image for kernel.qli.2.0 is not development complete.

Entity: kernel.qli.2.0
CR: 4700998
Reason: CR_CANNOT_MERGE

Please ensure the CR passes both CCT (ComponentChangeTasks) and ICT (Integration Change Tasks) validations.

@sgaud-quic

Copy link
Copy Markdown
Contributor

mainline PR : qualcomm-linux/kernel-topics#2211

Mainline PR is not merged.

@akakum-qualcomm

Copy link
Copy Markdown
Author

Crs-Fixed: 4700998

qualcomm-linux/kernel-topics#2227

@sgaud-quic

Copy link
Copy Markdown
Contributor

Crs-Fixed: 4700998

qualcomm-linux/kernel-topics#2227

Mainline PR not merged.

@akakum-qualcomm

Copy link
Copy Markdown
Author

Crs-Fixed: 4700998

qualcomm-linux/kernel-topics#2227

Mainline PR not merged.

qualcomm-linux/kernel-topics#2227
mainline PR is merged now.

@akakum-qualcomm
akakum-qualcomm force-pushed the akash/dwc3-qcom-nominal-votes branch from be6aabf to d9590a1 Compare November 18, 2026 09:51
@qcomlnxci
qcomlnxci requested a review from a team November 18, 2026 09:54
@akakum-qualcomm
akakum-qualcomm force-pushed the akash/dwc3-qcom-nominal-votes branch from d9590a1 to 2e4c74a Compare November 18, 2026 09:56
@qcomlnxci

Copy link
Copy Markdown

Test results

Workflow run: 44280129621 attempt 1

Known failures configured: 46

Target Pass Non-pass Total
kbdev-hamoa-iot-evk-multimedia 29 7 36
kbdev-lemans-evk-multimedia 33 3 36
kbdev-monaco-evk-multimedia 32 4 36
kbdev-purwa-iot-evk-multimedia 29 7 36
kbdev-qcs615-ride-multimedia 28 8 36
kbdev-qcs6490-rb3gen2-multimedia 32 4 36
kbdev-qcs8300-ride-multimedia 29 7 36
kbdev-qcs9100-ride-r3-multimedia 29 7 36
kbdev-qrb2210-rb1-multimedia - ⛔ Not submitted -
kbdev-shikra-iqs-evk-multimedia 21 15 36

Test Matrix

Test Case kbdev-hamoa-iot-evk-multimedia kbdev-lemans-evk-multimedia kbdev-monaco-evk-multimedia kbdev-purwa-iot-evk-multimedia kbdev-qcs615-ride-multimedia kbdev-qcs6490-rb3gen2-multimedia kbdev-qcs8300-ride-multimedia kbdev-qcs9100-ride-r3-multimedia kbdev-qrb2210-rb1-multimedia kbdev-shikra-iqs-evk-multimedia
LAVA result ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ⛔ Not submitted ✅ Results
Audio_Card_Registration ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip ✅ Pass ⚠️ skip ⚠️ skip ⛔ Not submitted ⚠️ skip
BT_FW_KMD_Service ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
BT_ON_OFF ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
BT_SCAN ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
CPUFreq_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
CPU_affinity ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
DSP_AudioPD ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
Ethernet_Basic_Validation ⚠️ skip ⚠️ skip ⚠️ skip ⚠️ skip ⚠️ skip ⚠️ skip ⚠️ skip ❌ Fail ⛔ Not submitted ❌ Fail
Freq_Scaling ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
GIC ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
IPA ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
Interrupts ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
KVM_Driver ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
KVM_EL2_DTB ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
KVM_Infra ❌ Fail ✅ Pass ✅ Pass ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
OpenCV ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
PCIe ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
Probe_Failure_Check ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
RMNET ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
UFS_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
USBHost ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⚠️ skip ❌ Fail ✅ Pass ⛔ Not submitted ❌ Fail
WiFi_Firmware_Driver ✅ Pass ✅ Pass ❌ Fail ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
WiFi_OnOff ✅ Pass ✅ Pass ❌ Fail ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
adsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
cdsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
gpdsp_remoteproc ⚠️ skip ✅ Pass ✅ Pass ⚠️ skip ⚠️ skip ⚠️ skip ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
hotplug ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
irq ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
kaslr ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
pinctrl ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
qcom_hwrng ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
rngtest ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
shmbridge ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
smmu ❌ Fail ❌ Fail ✅ Pass ❌ Fail ❌ Fail ✅ Pass ✅ Pass ❌ Fail ⛔ Not submitted ✅ Pass
watchdog ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
wpss_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
Failure summary

Known failures

  • kbdev-hamoa-iot-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-hamoa-iot-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-hamoa-iot-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-hamoa-iot-evk-multimedia: smmu => fail.
  • kbdev-hamoa-iot-evk-multimedia: KVM_Driver => fail.
  • kbdev-hamoa-iot-evk-multimedia: KVM_EL2_DTB => fail.
  • kbdev-hamoa-iot-evk-multimedia: KVM_Infra => fail.
  • kbdev-lemans-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-lemans-evk-multimedia: smmu => fail.
  • kbdev-monaco-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-monaco-evk-multimedia: WiFi_Firmware_Driver => fail.
  • kbdev-monaco-evk-multimedia: WiFi_OnOff => fail.
  • kbdev-purwa-iot-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-purwa-iot-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-purwa-iot-evk-multimedia: smmu => fail.
  • kbdev-purwa-iot-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-purwa-iot-evk-multimedia: KVM_Driver => fail.
  • kbdev-purwa-iot-evk-multimedia: KVM_EL2_DTB => fail.
  • kbdev-purwa-iot-evk-multimedia: KVM_Infra => fail.
  • kbdev-qcs615-ride-multimedia: gpdsp_remoteproc => skip.
  • kbdev-qcs615-ride-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs615-ride-multimedia: smmu => fail.
  • kbdev-qcs615-ride-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs615-ride-multimedia: KVM_Driver => fail.
  • kbdev-qcs615-ride-multimedia: KVM_EL2_DTB => fail.
  • kbdev-qcs615-ride-multimedia: KVM_Infra => fail.
  • kbdev-qcs6490-rb3gen2-multimedia: gpdsp_remoteproc => skip.
  • kbdev-qcs6490-rb3gen2-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs6490-rb3gen2-multimedia: USBHost => skip.
  • kbdev-qcs6490-rb3gen2-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs8300-ride-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs8300-ride-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs8300-ride-multimedia: KVM_Driver => fail.
  • kbdev-qcs8300-ride-multimedia: KVM_EL2_DTB => fail.
  • kbdev-qcs8300-ride-multimedia: KVM_Infra => fail.
  • kbdev-qcs9100-ride-r3-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs9100-ride-r3-multimedia: smmu => fail.
  • kbdev-qcs9100-ride-r3-multimedia: Ethernet_Basic_Validation => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_Driver => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_EL2_DTB => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_Infra => fail.
  • kbdev-shikra-iqs-evk-multimedia: adsp_remoteproc => skip.
  • kbdev-shikra-iqs-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-shikra-iqs-evk-multimedia: GIC => fail.
  • kbdev-shikra-iqs-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-shikra-iqs-evk-multimedia: UFS_Validation => skip.
  • kbdev-shikra-iqs-evk-multimedia: PCIe => fail.
  • kbdev-shikra-iqs-evk-multimedia: USBHost => fail.
  • kbdev-shikra-iqs-evk-multimedia: WiFi_OnOff => skip.
  • kbdev-shikra-iqs-evk-multimedia: DSP_AudioPD => skip.
  • kbdev-shikra-iqs-evk-multimedia: KVM_Driver => fail.
  • kbdev-shikra-iqs-evk-multimedia: KVM_EL2_DTB => fail.
  • kbdev-shikra-iqs-evk-multimedia: KVM_Infra => fail.

Unexpected failures

  • kbdev-lemans-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-monaco-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs615-ride-multimedia: Audio_Card_Registration => skip.
  • kbdev-qcs8300-ride-multimedia: USBHost => fail.
  • kbdev-qcs8300-ride-multimedia: Audio_Card_Registration => skip.
  • kbdev-qcs9100-ride-r3-multimedia: Audio_Card_Registration => skip.
  • kbdev-qrb2210-rb1-multimedia did not produce a LAVA result artifact.
  • kbdev-shikra-iqs-evk-multimedia: Ethernet_Basic_Validation => fail.
  • kbdev-shikra-iqs-evk-multimedia: BT_SCAN => fail.
  • kbdev-shikra-iqs-evk-multimedia: Audio_Card_Registration => skip.

Result: blocked because unexpected failures were found.

@qcomlnxci

Copy link
Copy Markdown

Test results

Workflow run: 44280726758 attempt 1

Known failures configured: 46

Target Pass Non-pass Total
kbdev-hamoa-iot-evk-multimedia 29 7 36
kbdev-lemans-evk-multimedia 33 3 36
kbdev-monaco-evk-multimedia 33 3 36
kbdev-purwa-iot-evk-multimedia 27 7 34
kbdev-qcs615-ride-multimedia
kbdev-qcs6490-rb3gen2-multimedia 32 4 36
kbdev-qcs8300-ride-multimedia 28 8 36
kbdev-qcs9100-ride-r3-multimedia 29 7 36
kbdev-qrb2210-rb1-multimedia - ⛔ Not submitted -
kbdev-shikra-iqs-evk-multimedia 21 15 36

Test Matrix

Test Case kbdev-hamoa-iot-evk-multimedia kbdev-lemans-evk-multimedia kbdev-monaco-evk-multimedia kbdev-purwa-iot-evk-multimedia kbdev-qcs615-ride-multimedia kbdev-qcs6490-rb3gen2-multimedia kbdev-qcs8300-ride-multimedia kbdev-qcs9100-ride-r3-multimedia kbdev-qrb2210-rb1-multimedia kbdev-shikra-iqs-evk-multimedia
LAVA result ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ✅ Results ⛔ Not submitted ✅ Results
Audio_Card_Registration ✅ Pass ✅ Pass ✅ Pass ◻️ ◻️ ✅ Pass ⚠️ skip ⚠️ skip ⛔ Not submitted ⚠️ skip
BT_FW_KMD_Service ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
BT_ON_OFF ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
BT_SCAN ✅ Pass ✅ Pass ✅ Pass ❌ Fail ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
CPUFreq_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
CPU_affinity ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
DSP_AudioPD ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
Ethernet_Basic_Validation ⚠️ skip ⚠️ skip ✅ Pass ⚠️ skip ◻️ ⚠️ skip ⚠️ skip ❌ Fail ⛔ Not submitted ❌ Fail
Freq_Scaling ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
GIC ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
IPA ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
Interrupts ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
KVM_Driver ❌ Fail ✅ Pass ✅ Pass ❌ Fail ◻️ ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
KVM_EL2_DTB ❌ Fail ✅ Pass ✅ Pass ◻️ ◻️ ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
KVM_Infra ❌ Fail ✅ Pass ✅ Pass ❌ Fail ◻️ ✅ Pass ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
OpenCV ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
PCIe ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ❌ Fail
Probe_Failure_Check ❌ Fail ❌ Fail ❌ Fail ❌ Fail ◻️ ❌ Fail ❌ Fail ❌ Fail ⛔ Not submitted ❌ Fail
RMNET ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
UFS_Validation ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
USBHost ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ⚠️ skip ❌ Fail ✅ Pass ⛔ Not submitted ❌ Fail
WiFi_Firmware_Driver ✅ Pass ✅ Pass ❌ Fail ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
WiFi_OnOff ✅ Pass ✅ Pass ❌ Fail ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
adsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
cdsp_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
gpdsp_remoteproc ⚠️ skip ✅ Pass ✅ Pass ⚠️ skip ◻️ ⚠️ skip ✅ Pass ✅ Pass ⛔ Not submitted ⚠️ skip
hotplug ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
irq ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
kaslr ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
pinctrl ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
qcom_hwrng ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
rngtest ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ❌ Fail ✅ Pass ⛔ Not submitted ✅ Pass
shmbridge ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
smmu ❌ Fail ❌ Fail ✅ Pass ❌ Fail ◻️ ✅ Pass ✅ Pass ❌ Fail ⛔ Not submitted ✅ Pass
watchdog ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
wpss_remoteproc ✅ Pass ✅ Pass ✅ Pass ✅ Pass ◻️ ✅ Pass ✅ Pass ✅ Pass ⛔ Not submitted ✅ Pass
Failure summary

Known failures

  • kbdev-hamoa-iot-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-hamoa-iot-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-hamoa-iot-evk-multimedia: smmu => fail.
  • kbdev-hamoa-iot-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-hamoa-iot-evk-multimedia: KVM_Driver => fail.
  • kbdev-hamoa-iot-evk-multimedia: KVM_EL2_DTB => fail.
  • kbdev-hamoa-iot-evk-multimedia: KVM_Infra => fail.
  • kbdev-lemans-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-lemans-evk-multimedia: smmu => fail.
  • kbdev-monaco-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-monaco-evk-multimedia: WiFi_Firmware_Driver => fail.
  • kbdev-monaco-evk-multimedia: WiFi_OnOff => fail.
  • kbdev-purwa-iot-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-purwa-iot-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-purwa-iot-evk-multimedia: smmu => fail.
  • kbdev-purwa-iot-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-purwa-iot-evk-multimedia: KVM_Driver => fail.
  • kbdev-purwa-iot-evk-multimedia: KVM_Infra => fail.
  • kbdev-qcs6490-rb3gen2-multimedia: gpdsp_remoteproc => skip.
  • kbdev-qcs6490-rb3gen2-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs6490-rb3gen2-multimedia: USBHost => skip.
  • kbdev-qcs6490-rb3gen2-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs8300-ride-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs8300-ride-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-qcs8300-ride-multimedia: KVM_Infra => fail.
  • kbdev-qcs8300-ride-multimedia: KVM_Driver => fail.
  • kbdev-qcs8300-ride-multimedia: KVM_EL2_DTB => fail.
  • kbdev-qcs9100-ride-r3-multimedia: Probe_Failure_Check => fail.
  • kbdev-qcs9100-ride-r3-multimedia: smmu => fail.
  • kbdev-qcs9100-ride-r3-multimedia: Ethernet_Basic_Validation => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_Driver => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_EL2_DTB => fail.
  • kbdev-qcs9100-ride-r3-multimedia: KVM_Infra => fail.
  • kbdev-shikra-iqs-evk-multimedia: adsp_remoteproc => skip.
  • kbdev-shikra-iqs-evk-multimedia: gpdsp_remoteproc => skip.
  • kbdev-shikra-iqs-evk-multimedia: GIC => fail.
  • kbdev-shikra-iqs-evk-multimedia: Probe_Failure_Check => fail.
  • kbdev-shikra-iqs-evk-multimedia: UFS_Validation => skip.
  • kbdev-shikra-iqs-evk-multimedia: PCIe => fail.
  • kbdev-shikra-iqs-evk-multimedia: USBHost => fail.
  • kbdev-shikra-iqs-evk-multimedia: WiFi_OnOff => skip.
  • kbdev-shikra-iqs-evk-multimedia: DSP_AudioPD => skip.
  • kbdev-shikra-iqs-evk-multimedia: KVM_Driver => fail.
  • kbdev-shikra-iqs-evk-multimedia: KVM_EL2_DTB => fail.
  • kbdev-shikra-iqs-evk-multimedia: KVM_Infra => fail.

Unexpected failures

  • kbdev-lemans-evk-multimedia: Ethernet_Basic_Validation => skip.
  • kbdev-purwa-iot-evk-multimedia: BT_SCAN => fail.
  • kbdev-qcs8300-ride-multimedia: rngtest => fail.
  • kbdev-qcs8300-ride-multimedia: USBHost => fail.
  • kbdev-qcs8300-ride-multimedia: Audio_Card_Registration => skip.
  • kbdev-qcs9100-ride-r3-multimedia: Audio_Card_Registration => skip.
  • kbdev-qrb2210-rb1-multimedia did not produce a LAVA result artifact.
  • kbdev-shikra-iqs-evk-multimedia: Ethernet_Basic_Validation => fail.
  • kbdev-shikra-iqs-evk-multimedia: BT_SCAN => fail.
  • kbdev-shikra-iqs-evk-multimedia: Audio_Card_Registration => skip.

Result: blocked because unexpected failures were found.

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

usb: dwc3: Update nominal max votes for qcom usb

Missing FROMLIST in subject.

Increase the nominal max USB memory bandwidth vote to let DDR scale
higher within its nominal corner, fixing glitches seen during
simultaneous USB audio use cases.

USB_MEMORY_PEAK_SS_BW is raised from 2500 MB/s to 4500 MB/s (+2000
MB/s, ~1.8x). This ICC vote does not by itself determine DDR
frequency; the vote-to-frequency mapping is platform-specific and
comes from each target's interconnect/OPP table. On the tested
targets, the old vote kept DDR pinned at 1 GHz within the nominal
corner, even though DDR can run at 1.5 GHz (+50%) within the same
corner. At 1 GHz, glitches were observed during multiple concurrent
USB audio use cases.

Tested internally on RB3, RB4, RB8, Shikra, and other KLM targets.
With the updated vote, DDR is allowed to run at the higher nominal
frequency, and the following were measured (not merely projected):

- USB memory throughput: 125 MB/s -> 145 MB/s (+16%)
- RTD-USBC audio latency: 17.9 ms -> 16.6 ms (-7.3%)

RTD-USBC is the round-trip audio latency measured over a USB-C audio
path; lower is better. Note the +50% DDR frequency headroom does not
map linearly to the +16% throughput gain, since throughput is also
bounded by other factors along the USB/DDR path; the frequency
increase is necessary but not sufficient to explain the full gain on
its own.

Link: https://lore.kernel.org/all/20261105100626.2003771-1-akash.kumar@oss.qualcomm.com/
Signed-off-by: Akash Kumar <akash.kumar@oss.qualcomm.com>
@akakum-qualcomm

Copy link
Copy Markdown
Author

usb: dwc3: Update nominal max votes for qcom usb

Missing FROMLIST in subject.

Added.

@qcomlnxci
qcomlnxci requested a review from a team November 19, 2026 10:31
@qcomlnxci

Copy link
Copy Markdown

Test results

Workflow run: 44520007286 attempt 1

Known failures configured: 54

Target Pass Non-pass Total
kbdev-hamoa-iot-evk-multimedia - ⛔ Not submitted -
kbdev-lemans-evk-multimedia - ⛔ Not submitted -
kbdev-monaco-evk-multimedia - ⛔ Not submitted -
kbdev-purwa-iot-evk-multimedia - ⛔ Not submitted -
kbdev-qcs615-ride-multimedia - ⛔ Not submitted -
kbdev-qcs6490-rb3gen2-multimedia - ⛔ Not submitted -
kbdev-qcs8300-ride-multimedia - ⛔ Not submitted -
kbdev-qcs9100-ride-r3-multimedia - ⛔ Not submitted -
kbdev-qrb2210-rb1-multimedia - ⛔ Not submitted -
kbdev-shikra-iqs-evk-multimedia - ⛔ Not submitted -

Test Matrix

Test Case kbdev-hamoa-iot-evk-multimedia kbdev-lemans-evk-multimedia kbdev-monaco-evk-multimedia kbdev-purwa-iot-evk-multimedia kbdev-qcs615-ride-multimedia kbdev-qcs6490-rb3gen2-multimedia kbdev-qcs8300-ride-multimedia kbdev-qcs9100-ride-r3-multimedia kbdev-qrb2210-rb1-multimedia kbdev-shikra-iqs-evk-multimedia
LAVA result ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted ⛔ Not submitted
Failure summary

Unexpected failures

  • kbdev-hamoa-iot-evk-multimedia did not produce a LAVA result artifact.
  • kbdev-lemans-evk-multimedia did not produce a LAVA result artifact.
  • kbdev-monaco-evk-multimedia did not produce a LAVA result artifact.
  • kbdev-purwa-iot-evk-multimedia did not produce a LAVA result artifact.
  • kbdev-qcs615-ride-multimedia did not produce a LAVA result artifact.
  • kbdev-qcs6490-rb3gen2-multimedia did not produce a LAVA result artifact.
  • kbdev-qcs8300-ride-multimedia did not produce a LAVA result artifact.
  • kbdev-qcs9100-ride-r3-multimedia did not produce a LAVA result artifact.
  • kbdev-qrb2210-rb1-multimedia did not produce a LAVA result artifact.
  • kbdev-shikra-iqs-evk-multimedia did not produce a LAVA result artifact.

Result: blocked because unexpected failures were found.

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.

6 participants