Update volume APIs to match docker arguments by kvega005 · Pull Request #40181 · microsoft/WSL

kvega005 · 2026-04-14T22:25:00Z

Summary of the Pull Request

PR Checklist

Closes: Link to issue #xxx
Communication: I've discussed this with core contributors already. If work hasn't been agreed, this work might be rejected
Tests: Added/updated if needed and all pass
Localization: All end user facing strings can be localized
Dev docs: Added/updated if needed
Documentation updated: If checked, please file a pull request on our docs repo and link it here: #xxx

Detailed Description of the Pull Request / Additional comments

Validation Steps Performed

…eManagement

Copilot

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.

Copilot

Pull request overview

Copilot reviewed 17 out of 17 changed files in this pull request and generated 8 comments.

Copilot · 2026-04-15T01:18:55Z

src/windows/common/wslutil.cpp

+std::map<std::string, std::string> wsl::windows::common::wslutil::ParseKeyValuePairs(const KeyValuePair* pairs, ULONG count, LPCSTR reservedKey)
+{
+    std::map<std::string, std::string> result;
+
+    for (ULONG i = 0; i < count; i++)
+    {
+        THROW_HR_IF_NULL_MSG(E_INVALIDARG, pairs[i].Key, "Key at index %lu is null", i);
+        THROW_HR_IF_NULL_MSG(E_INVALIDARG, pairs[i].Value, "Value at index %lu is null", i);


ParseKeyValuePairs dereferences pairs[i] without validating pairs itself. If a caller passes count > 0 with a null pointer (e.g., via COM misuse or a future call site bug), this will AV. Add an explicit guard at the start like “if (count > 0) THROW_HR_IF_NULL(E_INVALIDARG, pairs);” (or return empty only when count == 0). This also protects WSLCSession::CreateVolume which immediately calls this helper on Options->DriverOpts/Labels.

Copilot · 2026-04-15T01:18:55Z

src/windows/wslcsession/WSLCSession.cpp

+    auto lock = m_lock.lock_shared();
+    std::lock_guard volumesLock(m_volumesLock);
+
+    THROW_HR_IF(HRESULT_FROM_WIN32(ERROR_INVALID_STATE), !m_dockerClient);
+    docker_schema::PruneVolumeResult pruneResult;
+    try
+    {
+        pruneResult = m_dockerClient->PruneVolumes(filters);
+    }


PruneVolumes holds m_volumesLock while making the Docker HTTP call, which can block all other volume operations for the duration of the network request. Consider calling m_dockerClient->PruneVolumes(filters) before acquiring m_volumesLock, then take m_volumesLock only to reconcile m_volumes with VolumesDeleted. This reduces contention and lowers the chance of lock-induced latency spikes.

Copilot · 2026-04-15T01:18:55Z

src/windows/wslcsession/WSLCSession.cpp

+    WSL_LOG("VolumesPruned", TraceLoggingValue(Results->VolumesCount, "PrunedCount"));
+
+    auto volumeCount = pruneResult.VolumesDeleted.has_value() ? pruneResult.VolumesDeleted->size() : 0;
+    auto output = wil::make_unique_cotaskmem<WSLCVolumeName[]>(volumeCount);
+    for (size_t i = 0; i < volumeCount; ++i)
+    {
+        strcpy_s(output[i], pruneResult.VolumesDeleted.value()[i].c_str());
+    }
+
+    Results->Volumes = output.release();
+    Results->VolumesCount = static_cast<ULONG>(volumeCount);
+    Results->SpaceReclaimed = pruneResult.SpaceReclaimed;
+


Two concrete issues here: (1) the log uses Results->VolumesCount before it’s populated (it will always log 0 due to the earlier ZeroMemory). Log volumeCount (or log after setting Results->VolumesCount). (2) strcpy_s’s return value is ignored; if a Docker-reported volume name exceeds WSLC_MAX_VOLUME_NAME_LENGTH, this will fail silently and you’ll return an empty string element. Check the return and fail the call (or skip/clip with an explicit policy) if copying doesn’t succeed.

Suggested change

WSL_LOG("VolumesPruned", TraceLoggingValue(Results->VolumesCount, "PrunedCount"));

auto volumeCount = pruneResult.VolumesDeleted.has_value() ? pruneResult.VolumesDeleted->size() : 0;

auto output = wil::make_unique_cotaskmem<WSLCVolumeName[]>(volumeCount);

for (size_t i = 0; i < volumeCount; ++i)

{

strcpy_s(output[i], pruneResult.VolumesDeleted.value()[i].c_str());

}

Results->Volumes = output.release();

Results->VolumesCount = static_cast<ULONG>(volumeCount);

Results->SpaceReclaimed = pruneResult.SpaceReclaimed;

auto volumeCount = pruneResult.VolumesDeleted.has_value() ? pruneResult.VolumesDeleted->size() : 0;

auto output = wil::make_unique_cotaskmem<WSLCVolumeName[]>(volumeCount);

for (size_t i = 0; i < volumeCount; ++i)

{

const errno_t copyResult = strcpy_s(output[i], pruneResult.VolumesDeleted.value()[i].c_str());

RETURN_HR_IF(HRESULT_FROM_WIN32(ERROR_INSUFFICIENT_BUFFER), copyResult != 0);

}

Results->Volumes = output.release();

Results->VolumesCount = static_cast<ULONG>(volumeCount);

Results->SpaceReclaimed = pruneResult.SpaceReclaimed;

WSL_LOG("VolumesPruned", TraceLoggingValue(Results->VolumesCount, "PrunedCount"));

src/windows/wslcsession/WSLCVhdVolume.cpp

Copilot · 2026-04-15T01:18:56Z

src/windows/wslcsession/WSLCVhdVolume.h

+enum class DetachVolumeFlags
+{
+    None = 0,
+    DeleteVhd = 1,
+};


DetachVolumeFlags is modeled like a flags enum, but it’s currently used as a scalar (and there are no flag operators). Either (a) make it a simple enum (non-flags) and name it accordingly, or (b) define flag operators (e.g., DEFINE_ENUM_FLAG_OPERATORS) and treat it as a bitmask consistently (see also the equality check in Detach()).

Copilot · 2026-04-15T01:18:56Z

src/windows/wslcsession/WSLCVhdVolume.cpp

    m_virtualMachine.DetachDisk(m_lun);
    m_attached = false;
+
+    if (flags == DetachVolumeFlags::DeleteVhd)


If DetachVolumeFlags is intended to be a bitmask, checking with == will behave incorrectly once additional flags are added (e.g., DeleteVhd | Force). Use a bit-test instead (after adding flag operators), or keep the type as a non-flags enum if only one mode is intended.

Suggested change

if (flags == DetachVolumeFlags::DeleteVhd)

if ((static_cast<unsigned int>(flags) & static_cast<unsigned int>(DetachVolumeFlags::DeleteVhd)) != 0)

Copilot · 2026-04-15T01:18:57Z

localization/strings/en-US/Resources.resw

  </data>
  <data name = "MessageWslcInvalidVolumeOptions" xml:space = "preserve" >
-    <value>Invalid volume options: '{}'</value>
+    <value>Missing required option: {}</value>


The resource key name MessageWslcInvalidVolumeOptions no longer matches the updated text (“Missing required option…”). This makes the API harder to understand/maintain and increases the chance of incorrect reuse elsewhere. Consider introducing a new localized resource with a matching name (and updating call sites), or revert this one to a generic “invalid options” message and add a dedicated “missing option” message.

Suggested change

<value>Missing required option: {}</value>

<value>Invalid volume options: {}</value>

Copilot · 2026-04-15T01:18:57Z

src/windows/wslcsession/WSLCSession.cpp

 }
 CATCH_RETURN();

+HRESULT WSLCSession::PruneVolumes(const WSLCPruneVolumesOptions* Options, WSLCPruneVolumesResults* Results)


PruneVolumes introduces new behavior (filter translation, calling Docker prune, and reconciling m_volumes + deleting VHDs) but there are no corresponding updates in test/windows/WSLCTests.cpp in this PR. Add at least one test covering: (1) pruning removes WSLC-managed volumes from ListVolumes, and (2) Detach(DeleteVhd) path deletes the underlying VHD file when Docker reports the volume as deleted.

OneBlue

Change looks great! Couple minor comments

OneBlue · 2026-04-15T01:57:07Z

src/windows/wslcsession/WSLCVhdVolume.cpp

+        THROW_HR_WITH_USER_ERROR_IF(E_INVALIDARG, Localization::MessageWslcInvalidVolumeOptions("SizeBytes"), it == DriverOpts.end());
+
+        auto& value = it->second;
+        THROW_HR_WITH_USER_ERROR_IF(E_INVALIDARG, Localization::MessageInvalidSize(value), value.empty() || value[0] == '-');


Is this check required ? I'd expect ToUInt64() to fail if value was empty or negative

OneBlue · 2026-04-15T02:05:54Z

test/windows/WSLCTests.cpp

+        WSLCDriverOption driverOpts[] = {{"SizeBytes", "1073741824"}};
+        volumeOptions.DriverOpts = driverOpts;
+        volumeOptions.DriverOptsCount = ARRAYSIZE(driverOpts);
+        WSLCVolumeInformation volInfo{};


nit: Here we're leaking the pointers returned by CreateVolume().

We might want to create a helper similar to [this][https://github.com/microsoft/WSL/blob/f60ec519d5add21a450d9f9d14a0239f7a99de5d/src/windows/common/wslutil.h#L130) to make it easier for callers to properly deallocate the volume info

OneBlue · 2026-04-15T02:19:10Z

src/windows/service/inc/wslc.idl

-    [unique] LPCSTR Options;
+    [unique] LPCSTR Name;
+    [unique] LPCSTR Driver;
+    [unique, size_is(DriverOptsCount)] const WSLCDriverOption* DriverOpts;


Given that DriverOpts and Labels require memory allocation & deallocation, I would recommend removing them from this structure to make things easier for the caller (especially since they're not needed for the simple wslc volume ls case).

If the caller needs to see the driver options or labels, they can call Inspect()

OneBlue · 2026-04-15T02:22:46Z

src/windows/wslcsession/WSLCVhdVolume.cpp

-        WSLCVolumeMetadata metadata{};
-        metadata.Type = WSLCVhdVolumeType;
+        BYTE randomBytes[32];
+        THROW_IF_NTSTATUS_FAILED(BCryptGenRandom(nullptr, randomBytes, sizeof(randomBytes), BCRYPT_USE_SYSTEM_PREFERRED_RNG));


nit: Given that we don't need cryptographically secure random bits, I'd recommend using std::generate here (example:

WSL/src/windows/common/WslCoreConfig.cpp

Lines 166 to 171 in acd3cd4

const std::independent_bits_engine<std::default_random_engine, 16, unsigned short> random;

// independent_bits_engine doesn't support unsigned char on MSVC

auto* macAddressShort = reinterpret_cast<unsigned short*>(MacAddress.data());

std::generate(macAddressShort, macAddressShort + 3, random);

)

Update volume APIs to match docker arguments

9c1d922

Copilot AI review requested due to automatic review settings April 14, 2026 22:25

kvega005 added 2 commits April 14, 2026 15:28

Merge remote-tracking branch 'origin/feature/wsl-for-apps' into volum…

fd79dbb

…eManagement

Fix formatting

b84f6b0

Copilot AI reviewed Apr 14, 2026

View reviewed changes

fix formatting

6dc1385

kvega005 requested a review from Copilot April 15, 2026 01:11

Copilot AI reviewed Apr 15, 2026

View reviewed changes

Copilot started reviewing on behalf of kvega005 April 15, 2026 01:22 View session

kvega005 marked this pull request as ready for review April 15, 2026 01:46

kvega005 requested a review from a team as a code owner April 15, 2026 01:46

OneBlue reviewed Apr 15, 2026

View reviewed changes

Copilot started reviewing on behalf of kvega005 April 15, 2026 03:05 View session

	if (flags == DetachVolumeFlags::DeleteVhd)
	if ((static_cast<unsigned int>(flags) & static_cast<unsigned int>(DetachVolumeFlags::DeleteVhd)) != 0)

	<value>Missing required option: {}</value>
	<value>Invalid volume options: {}</value>

	const std::independent_bits_engine<std::default_random_engine, 16, unsigned short> random;

	// independent_bits_engine doesn't support unsigned char on MSVC
	auto* macAddressShort = reinterpret_cast<unsigned short*>(MacAddress.data());
	std::generate(macAddressShort, macAddressShort + 3, random);

Conversation

kvega005 commented Apr 14, 2026

Summary of the Pull Request

PR Checklist

Detailed Description of the Pull Request / Additional comments

Validation Steps Performed

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

OneBlue left a comment

Choose a reason for hiding this comment

Uh oh!

OneBlue Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

OneBlue Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

OneBlue Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

OneBlue Apr 15, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants