Update scsi_bsg_mpi3mr.h

[ayukum-hyd]

No description provided.

[rahujosh]

PR #48 — validate-patch

PR: #48

Verdict	Issues	Detailed Report
❌	5	Full report

Verdict: ❌ — click to expand

🔍 Patch Validation

PR: Update scsi_bsg_mpi3mr.h — PR #48 (qualcomm-linux-stg/kernel)
Upstream commit: N/A — no upstream reference provided; change appears to be a local/vendor-only deletion with no corresponding mainline commit cited
Verdict: ❌ FAIL

Commit Message

Check	Status	Note
Subject matches upstream	❌	Subject is "Update scsi_bsg_mpi3mr.h" — a generic, non-descriptive title with no upstream equivalent found; does not follow kernel convention (subsystem: short imperative description)
Body preserves rationale	❌	Commit body is completely absent — no explanation of why struct mpi3_driver_info_layout and its use in mpi3mr_bsg_in_adpinfo are being removed
Fixes tag present/correct	❌	No Fixes: tag; if this is a bug fix or ABI correction, the tag is mandatory
Authorship preserved	⚠️	Author is ayukum-hyd <ayukum@qti.qualcomm.com> — a Qualcomm-internal alias/handle; no From: or Co-developed-by: preserving an upstream author, and the email domain is @qti.qualcomm.com (internal), not a public upstream identity
Backport note (if applicable)	❌	No [ upstream commit <sha> ] note; if this is a backport or vendor-local change derived from upstream, the lineage must be stated

Diff

File	Status	Notes
include/uapi/scsi/scsi_bsg_mpi3mr.h	⚠️	Deletes struct mpi3_driver_info_layout (9 lines) and its embedded use in struct mpi3mr_bsg_in_adpinfo — structurally coherent, but no upstream commit or rationale is cited to justify removing a UAPI-exported struct; this is a UAPI ABI-breaking change with zero justification in the commit message

Issues

• Missing commit body: The commit has no description at all. Removing a uapi struct is a significant, potentially ABI-breaking change — the rationale (e.g., "struct was never used by userspace", "replaced by X", "fixes ABI mismatch") must be documented.
• Non-descriptive subject line: "Update scsi_bsg_mpi3mr.h" does not follow Linux kernel commit message conventions. It should be something like scsi: mpi3mr: remove unused mpi3_driver_info_layout from UAPI header.
• No upstream traceability: There is no Link:, Fixes:, or [ upstream commit ] tag. It is impossible to verify whether this deletion is aligned with mainline, is a vendor-local decision, or is a regression.
• Author identity: The author ayukum-hyd is an internal Qualcomm handle. Upstream submissions require a real name and a publicly reachable email address per Documentation/process/submitting-patches.rst.
• UAPI impact undocumented: include/uapi/ headers are part of the stable kernel–userspace ABI. Removing struct mpi3_driver_info_layout without a Fixes: tag or ABI-break justification is a red flag for any stable or vendor tree.

Verdict

Reject: the commit message is entirely missing a body and rationale, the subject is non-descriptive, there is no upstream traceability for a UAPI-breaking struct removal, and the author identity does not meet upstream kernel standards — all must be corrected before this can be merged.

[rahujosh]

PR #48 — checker-log-analyzer

PR: #48
Checker run: https://github.com/qualcomm-linux-stg/kernel-config-test/actions/runs/23713793043

Checker	Result	Summary
Checker	Result	Summary
checkpatch	⏭️ N/A	No log present; not triggered for this run
dt-binding-check	⏭️ N/A	No log present; not triggered for this run
dtb-check	✅ Pass	"No changes in Devicetree" — skipped cleanly
sparse-check	⏭️ N/A	No log present; not triggered for this run
check-uapi-headers	❌ Fail	ABI-breaking deletion in include/uapi/scsi/scsi_bsg_mpi3mr.h
check-patch-compliance	⏭️ N/A	No log present; not triggered for this run
tag-check	⏭️ N/A	No log present; not triggered for this run
qcom-next-check	⏭️ N/A	No log present; not triggered for this run

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #48 — "Update scsi_bsg_mpi3mr.h" (ayukum-hyd)
Source: https://github.com/qualcomm-linux-stg/kernel-config-test/actions/runs/23713793043
Overall CI status: ❌ failure (reported back to commit 1d53e785568005804ccff3500117c9758bae78d9)

Checker	Result	Summary
checkpatch	⏭️ N/A	No log present; not triggered for this run
dt-binding-check	⏭️ N/A	No log present; not triggered for this run
dtb-check	✅ Pass	"No changes in Devicetree" — skipped cleanly
sparse-check	⏭️ N/A	No log present; not triggered for this run
check-uapi-headers	❌ Fail	ABI-breaking deletion in include/uapi/scsi/scsi_bsg_mpi3mr.h
check-patch-compliance	⏭️ N/A	No log present; not triggered for this run
tag-check	⏭️ N/A	No log present; not triggered for this run
qcom-next-check	⏭️ N/A	No log present; not triggered for this run

---

❌ check-uapi-headers

Root cause: The PR deletes struct mpi3_driver_info_layout and its embedded member driver_info from include/uapi/scsi/scsi_bsg_mpi3mr.h, shrinking struct mpi3mr_bsg_in_adpinfo from 1344 bits (168 bytes) to 384 bits (48 bytes) — a backwards-incompatible ABI change in a user-facing UAPI header.

Failure details:

==== ABI differences detected in include/scsi/scsi_bsg_mpi3mr.h
     from 9448598b22c5 -> 1d53e7855680 ====
    [C] 'struct mpi3mr_bsg_in_adpinfo' changed:
      type size changed from 1344 to 384 (in bits)
      1 data member deletion:
        'mpi3_driver_info_layout driver_info', at offset 384 (in bits)
        at scsi_bsg_mpi3mr.h:146:1

error - 1/941 UAPI headers compatible with arm64 appear _not_ to be
        backwards compatible

What the patch does:

• Removes the entire struct mpi3_driver_info_layout definition (11 lines).
• Removes the struct mpi3_driver_info_layout driver_info member from struct mpi3mr_bsg_in_adpinfo.
• This is a UAPI ABI break: any userspace binary that was compiled against the old header and reads/writes driver_info will silently access wrong memory or crash.

Fix options:

1. Do not delete — if the struct is still used by any in-tree or out-of-tree userspace (e.g. mpi3mrctl), keep it and mark it deprecated with a comment instead.
2. Deprecate, don't remove — add __attribute__((deprecated)) or a /* deprecated */ comment and keep the struct for at least one kernel release cycle.
3. If removal is intentional — it must go through the proper UAPI deprecation process: announce on the mailing list, give userspace time to adapt, and coordinate with the scsi maintainer. The commit message must explain why the ABI break is justified.

Reproduce locally:

./scripts/check-uapi.sh -b 1d53e785568005804ccff3500117c9758bae78d9 \
                        -p 9448598b22c50c8a5bb77a9103e2d49f134c9578 \
                        -l uapi_errors.log

---

Verdict

1 blocker to fix: check-uapi-headers fails because the PR makes a backwards-incompatible deletion from a UAPI header (struct mpi3_driver_info_layout + its embedded use in struct mpi3mr_bsg_in_adpinfo). UAPI structs must not be removed without a proper deprecation cycle. The PR is not ready to merge until this ABI break is resolved.

[rahujosh]

PR #48 — validate-patch

PR: #48

Verdict	Issues	Detailed Report
❌	0	Full report

Verdict: ❌ — click to expand

🔍 Patch Validation

PR: #48 — "Update scsi_bsg_mpi3mr.h"
Upstream commit: N/A — no upstream reference provided; network unavailable to verify independently
Verdict: ❌ FAIL

---

Commit Message

Check	Status	Note
Subject matches upstream	❌	"Update scsi_bsg_mpi3mr.h" is a non-descriptive filename-only subject; kernel convention requires <subsystem>: <imperative verb phrase>, e.g. scsi: mpi3mr: remove mpi3_driver_info_layout from UAPI header
Body preserves rationale	❌	Commit body is completely absent — no explanation of why the struct is removed, no reference to a firmware/spec change, no bug being fixed
Fixes tag present/correct	❌	No Fixes: tag; if this is a bug fix (e.g. ABI correction or dead-code removal) the tag is required; if it is a deliberate feature removal, the rationale must be in the body
Authorship preserved	⚠️	From: ayukum-hyd <ayukum@qti.qualcomm.com> — the From name is a machine/username alias (ayukum-hyd), not a real name as required by the DCO and kernel Signed-off-by policy (Documentation/process/submitting-patches.rst §14)
Backport note (if applicable)	❌	No Link: tag pointing to an upstream commit; no cherry-picked from note; impossible to verify this is a faithful backport without it

---

Diff

File	Status	Notes
include/uapi/scsi/scsi_bsg_mpi3mr.h	❌	Removes struct mpi3_driver_info_layout from a UAPI header — this is a userspace ABI-breaking change; once exported, UAPI structs cannot be deleted without a documented upstream justification and ack from the UAPI maintainer
include/uapi/scsi/scsi_bsg_mpi3mr.h	❌	Removes driver_info field from struct mpi3mr_bsg_in_adpinfo — changes the binary layout of a UAPI struct, breaking any userspace binary that uses BSG_MPI3MR_OPCODE_ADPINFO
include/uapi/scsi/scsi_bsg_mpi3mr.h	⚠️	Orphaned kdoc comment: the @driver_release_date / @driver_capabilities kernel-doc lines (context lines 16–18 in the patch) remain in the file after the struct they document is deleted, leaving dangling documentation

---

Issues

1. UAPI ABI breakage (critical): include/uapi/scsi/scsi_bsg_mpi3mr.h is a userspace-visible header. Removing struct mpi3_driver_info_layout and the driver_info field from mpi3mr_bsg_in_adpinfo silently changes the size and layout of the BSG adapter-info response structure. Any compiled userspace tool (e.g. mpi3mrctl) that was built against the old header will silently misparse the response. This requires an explicit upstream justification, a check-uapi-headers pass, and maintainer sign-off.

2. Missing commit rationale: There is zero explanation of why these symbols are removed. Is this a dead-code cleanup? A firmware ABI realignment? A security fix? Without a body, reviewers and bisectors cannot assess correctness or regression risk.

3. Non-real author name: The From: / Signed-off-by: name ayukum-hyd is a username, not a legal name. The DCO (Developer Certificate of Origin) and kernel policy require a real name (first + last). This must be corrected before the patch can be accepted upstream.

4. No upstream traceability: There is no Link: tag, no cherry picked from commit <sha>, and no reference to a mailing-list thread. It is impossible to determine whether this change has upstream acceptance or is a local-only modification.

5. Orphaned kdoc: After the deletion, the kernel-doc comment block that documents @driver_release_date and @driver_capabilities is left dangling in the file with no struct to attach to. This will produce kernel-doc warnings and confuse documentation generators.

---

Verdict

Reject. The patch makes an undocumented, ABI-breaking deletion from a UAPI header with no commit rationale, no upstream traceability, and a non-conforming author name — all three must be corrected before this can be considered for merge.

[rahujosh]

PR #48 — checker-log-analyzer

PR: #48
Checker run: https://github.com/qualcomm-linux-stg/kernel-config-test/actions/runs/23713793043

Checker	Result	Summary
Checker	Result	Summary
checkpatch	⏭️ N/A	No log present; not triggered for this run
dt-binding-check	⏭️ N/A	No log present; not triggered for this run
dtb-check	✅	No DT changes detected; skipped cleanly
sparse-check	⏭️ N/A	No log present; not triggered for this run
check-uapi-headers	❌	ABI-breaking deletion in struct mpi3mr_bsg_in_adpinfo
check-patch-compliance	⏭️ N/A	No log present; not triggered for this run
tag-check	⏭️ N/A	No log present; not triggered for this run
qcom-next-check	⏭️ N/A	No log present; not triggered for this run

Detailed report: Full report

Checker analysis — click to expand

🤖 CI Checker Analysis (checker-log-analyzer)

PR: #48 — "Update scsi_bsg_mpi3mr.h" (ayukum-hyd)
Source: https://github.com/qualcomm-linux-stg/kernel-config-test/actions/runs/23713793043

Checker	Result	Summary
checkpatch	⏭️ N/A	No log present; not triggered for this run
dt-binding-check	⏭️ N/A	No log present; not triggered for this run
dtb-check	✅	No DT changes detected; skipped cleanly
sparse-check	⏭️ N/A	No log present; not triggered for this run
check-uapi-headers	❌	ABI-breaking deletion in struct mpi3mr_bsg_in_adpinfo
check-patch-compliance	⏭️ N/A	No log present; not triggered for this run
tag-check	⏭️ N/A	No log present; not triggered for this run
qcom-next-check	⏭️ N/A	No log present; not triggered for this run

---

❌ check-uapi-headers

Root cause: The patch deletes struct mpi3_driver_info_layout and removes the driver_info member from struct mpi3mr_bsg_in_adpinfo in the UAPI header include/uapi/scsi/scsi_bsg_mpi3mr.h, shrinking the struct from 1344 bits (168 bytes) to 384 bits (48 bytes) — a backwards-incompatible ABI change.

Failure details:

==== ABI differences detected in include/scsi/scsi_bsg_mpi3mr.h
     from 9448598b22c5 -> 1d53e7855680 ====
    [C] 'struct mpi3mr_bsg_in_adpinfo' changed:
      type size changed from 1344 to 384 (in bits)
      1 data member deletion:
        'mpi3_driver_info_layout driver_info', at offset 384 (in bits)
        at scsi_bsg_mpi3mr.h:146:1

error - 1/941 UAPI headers compatible with arm64 appear _not_ to be
        backwards compatible

What the patch does:

• Deletes the entire struct mpi3_driver_info_layout definition (11 lines removed from include/uapi/scsi/scsi_bsg_mpi3mr.h)
• Removes the struct mpi3_driver_info_layout driver_info member from struct mpi3mr_bsg_in_adpinfo

Why this is a blocker: UAPI headers are a stable kernel-to-userspace ABI contract. Removing a struct member shrinks the struct size and breaks any existing userspace application (e.g., mpi3mrctl) that was compiled against the old header and passes mpi3mr_bsg_in_adpinfo via ioctl. This is a hard ABI regression.

Fix options:

1. Preferred — deprecate, don't delete: Mark mpi3_driver_info_layout and the driver_info member as deprecated using a comment or __attribute__((deprecated)), but keep them in the header. Userspace must not be broken.
2. If the struct was never shipped in a released kernel: Confirm the field was never part of a stable release (check git log --follow on the header). If it was only ever in a staging/vendor tree, document that justification clearly in the commit message and request a checker override.
3. If removal is intentional and justified: The commit message must explain why the ABI break is acceptable (e.g., the struct was never exposed to real userspace), and the PR author should request an explicit override: kernel-checker from an authorized reviewer (shashim-quic or sgaud-quic).

Reproduce locally:

./scripts/check-uapi.sh -b 1d53e785568005804ccff3500117c9758bae78d9 \
                        -p 9448598b22c50c8a5bb77a9103e2d49f134c9578 \
                        -l uapi_errors.log

---

Verdict

1 blocker: check-uapi-headers fails due to an ABI-breaking deletion of struct mpi3_driver_info_layout and its embedded member from the UAPI header — either restore the struct/member, or obtain an explicit checker override with justification before merging.