perf(quickemu): improve disk I/O defaults (cache,aio)

[flexiondotorg]

• add cache=writeback and aio=threads to DRIVE_OPTIMISATIONS
• retain existing optimisations: discard=unmap,detect-zeroes=unmap
• improve write performance (host writeback caching) and async I/O parallelism via thread-based AIO
• follows Proxmox/libvirt QEMU disk best practices; researched for macOS but applies to all guests

Type of change

• New feature (non-breaking change which adds functionality)

Checklist:

• I have performed a self-review of my code

[cubic-dev-ai]

1 issue found across 1 file

Confidence score: 3/5

• Defaulting to cache=writeback in quickemu can increase data-loss risk on host failure because writes may be acknowledged before reaching storage
• Score reflects a concrete user-impact risk (data integrity) with moderate severity, even though scope is limited to disk cache defaults
• Pay close attention to quickemu - default disk cache behavior could lead to data loss on crashes

Prompt for AI agents (all issues)

Check if these issues are valid — if so, understand the root cause of each and fix them.


<file name="quickemu">

<violation number="1" location="quickemu:1892">
P2: Setting `cache=writeback` as a default for all VM disks increases the risk of data loss on host failures (docs warn that writeback reports writes complete before they hit storage). Consider leaving cache unspecified or using a safer mode by default and making writeback opt‑in.</violation>
</file>

_{Reply with feedback, questions, or to request a fix. Tag @cubic-dev-ai to re-run a review.}

[cubic-dev-ai]

P2: Setting cache=writeback as a default for all VM disks increases the risk of data loss on host failures (docs warn that writeback reports writes complete before they hit storage). Consider leaving cache unspecified or using a safer mode by default and making writeback opt‑in.

Prompt for AI agents

Check if this issue is valid — if so, understand the root cause and fix it. At quickemu, line 1892:

<comment>Setting `cache=writeback` as a default for all VM disks increases the risk of data loss on host failures (docs warn that writeback reports writes complete before they hit storage). Consider leaving cache unspecified or using a safer mode by default and making writeback opt‑in.</comment>

<file context>
@@ -1887,8 +1887,9 @@ function vm_boot() {
-    DRIVE_OPTIMISATIONS="discard=unmap,detect-zeroes=unmap"
+    # Optimise disk I/O: enable TRIM/discard, zero detection for thin provisioning,
+    # writeback caching and threaded async I/O
+    DRIVE_OPTIMISATIONS="discard=unmap,detect-zeroes=unmap,cache=writeback,aio=threads"
 
     if [[ "${boot}" == *"efi"* ]]; then
</file context>

Suggested change

	DRIVE_OPTIMISATIONS="discard=unmap,detect-zeroes=unmap,cache=writeback,aio=threads"
	DRIVE_OPTIMISATIONS="discard=unmap,detect-zeroes=unmap,aio=threads"