fix(reflinktree): retry transient FICLONE EINVAL and add diagnostics

[s0up4200]

Summary

• retry Linux reflink clone attempts on transient EINVAL in addition to EAGAIN
• include filesystem diagnostics (srcDev, dstDev, srcFsType, dstFsType) in reflink failure messages to speed up mount mismatch debugging
• keep clone-range fallback behavior unchanged

Validation

• docker run --rm -v "$PWD":/src -w /src golang:1.25 go test ./pkg/reflinktree -v
• make lint
• make test
• make build

Summary by CodeRabbit

• Bug Fixes

  • Improved Linux file-clone robustness with automatic retries for transient clone errors and clearer fallback handling.
  • Error messages now include filesystem diagnostics (device IDs and filesystem types) to aid troubleshooting.

• Tests

  • Added tests covering retry behavior, retry exhaustion, backoff timing, and verification that diagnostic info appears in error messages.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

---

Walkthrough

Adds retry logic and filesystem-device diagnostics to Linux reflink clone operations; updates error paths to include diagnostics and introduces tests verifying EINVAL retry behavior and diagnostic inclusion on EXDEV.

Changes

Cohort / File(s)	Summary
Reflink implementation pkg/reflinktree/reflink_linux.go	Added retry logic using shouldRetryCloneError() (treating EAGAIN and EINVAL as retryable), integrated diagnostic collection (cloneDiagnostics(), deviceID(), filesystemType()), augmented error messages with diagnostics, and adjusted control flow around FICLONERANGE fallback.
Tests pkg/reflinktree/reflink_linux_test.go	Added tests for EINVAL retry behavior (TestCloneFile_RetriesEINVAL, TestCloneFile_EINVALRetriesExhausted) with sleep tracking, and a test asserting filesystem diagnostics appear in errors on EXDEV; preserved existing FICLONERANGE fallback test.

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~22 minutes

Possibly related PRs

• fix(reflinktree): fall back to FICLONERANGE when FICLONE unsupported #1221: Related changes to FICLONERANGE fallback and clone control flow in reflink_linux.go.
• fix(reflink): retry EAGAIN clones #1360: Related modifications adding clone retry behavior and corresponding tests in the same area.

Suggested labels

codex

Poem

🐰 I hopped through syscalls, sniffed each mount and dev,
Retried on EINVAL, then EAGAIN said "revive!"
I told you where things live—src, dst, fs type too,
Diagnostics in pawprints, so clones know what to do.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately captures the main changes: adding retry logic for transient EINVAL errors and including diagnostic context in error messages.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

• 📝 Generate docstrings (stacked PR)
• 📝 Generate docstrings (commit on current branch)

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch codex/reflink-einval-retry

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In `@pkg/reflinktree/reflink_linux.go`:
- Around line 151-161: The linter flags an unnecessary conversion in deviceID:
syscall.Stat_t.Dev is already uint64, so remove the redundant uint64(...) cast
and call strconv.FormatUint(sys.Dev, 10) directly; update the return in deviceID
to use sys.Dev without the conversion and ensure strconv is imported/used.