Search criteria
Related vulnerabilities
GHSA-FW38-PC54-JVX9
Vulnerability from github – Published: 2026-06-05 16:40 – Updated: 2026-06-05 16:40Summary
The account-data trie syncers leak bounded throttler slots on error paths in syncDataTrie(). Each failed trie sync permanently consumes one slot from
the NumGoRoutinesThrottler, and the slot is never returned unless the sync succeeds or the root hash was already present.
I confirmed this on the current default branch develop at commit 9640d63 (observed on May 20, 2026). I also confirmed the bug with a runtime PoC
using the real timeout path in trieSyncer.StartSyncing(): two timed-out sync attempts are enough to exhaust a throttler with capacity 2.
This affects the epoch bootstrap path because syncUserAccountsState() and syncKappAccountsState() create bounded throttlers and abort bootstrap
immediately if the syncer returns an error. Once enough trie-root sync attempts fail, the syncer cannot make forward progress and bootstrap fails.
## Affected Components
data/syncer/userAccountsSyncer.godata/syncer/kappAccountsSyncer.godata/trie/sync.gocore/throttler/numGoRoutinesThrottler.gocore/bootstrap/process.go
## Affected Version
Verified on:
- develop HEAD 9640d63
Please check whether the same code is present in supported 1.7.x releases.
## Suggested Severity
High
## Vulnerability Details
### Root Cause
Both account-data syncers call StartProcessing() before creating / starting the trie syncer, but they only call EndProcessing() on the success path
and on the duplicate-root early return.
userAccountsSyncer.syncDataTrie():
```go func (u *userAccountsSyncer) syncDataTrie(rootHash []byte, ssh data.SyncStatisticsHandler, ctx context.Context) error { u.throttler.StartProcessing()
u.syncerMutex.Lock()
if _, ok := u.dataTries[string(rootHash)]; ok {
u.syncerMutex.Unlock()
u.throttler.EndProcessing()
return nil
}
dataTrie, err := trie.NewTrie(...)
if err != nil {
u.syncerMutex.Unlock()
return err
}
trieSyncer, err := trie.NewTrieSyncer(arg)
if err != nil {
u.syncerMutex.Unlock()
return err
}
u.syncerMutex.Unlock()
err = trieSyncer.StartSyncing(rootHash, ctx)
if err != nil {
return err
}
u.throttler.EndProcessing()
return nil
}
The same bug exists in kappAccountsSyncer.syncDataTrie().
### Missing slot release paths
After StartProcessing(), the following error paths return without EndProcessing():
1. trie.NewTrie(...) returns an error
2. trie.NewTrieSyncer(...) returns an error
3. trieSyncer.StartSyncing(...) returns an error
### Why this matters
NumGoRoutinesThrottler is a strict bounded counter:
func (ngrt *NumGoRoutinesThrottler) CanProcess() bool { valCounter := atomic.LoadInt32(&ngrt.counter) return valCounter < ngrt.max }
func (ngrt *NumGoRoutinesThrottler) StartProcessing() { atomic.AddInt32(&ngrt.counter, 1) }
func (ngrt *NumGoRoutinesThrottler) EndProcessing() { atomic.AddInt32(&ngrt.counter, -1) }
Once leaked, a slot remains consumed for the lifetime of that throttler instance.
The parent loops in both syncers wait for capacity before starting the next account-data trie sync:
for !u.throttler.CanProcess() { select { case <-time.After(timeBetweenRetries): continue case <-ctx.Done(): return common.ErrTimeIsOut } }
So after enough failures, further roots stop progressing and the sync operation eventually returns time is out.
### Bootstrap impact
Epoch bootstrap uses these syncers directly and aborts on any error:
err = e.syncUserAccountsState(e.epochStartMeta.Header.TrieRoot) if err != nil { return nil, nil, err }
err = e.syncKappAccountsState(e.epochStartMeta.Header.KAppsTrieRoot) if err != nil { return nil, nil, err }
The throttlers for these paths are real bounded throttlers created from numConcurrentTrieSyncers.
## Proof of Concept
I verified the bug with the real timeout path, not only with a canceled context.
The PoC below uses:
- a real NumGoRoutinesThrottler with capacity 2
- a real trieSyncer.StartSyncing()
- an empty trie-node cache and a request handler that never supplies nodes
- a short sync timeout (1s) so StartSyncing() returns trie.ErrTimeIsOut
After the first failed sync, one slot remains leaked.
After the second failed sync, the throttler is exhausted.
### PoC test
package syncer
import ( "context" "testing" "time"
commonmock "github.com/klever-io/klever-go/common/mock"
corethrottler "github.com/klever-io/klever-go/core/throttler"
"github.com/klever-io/klever-go/data"
"github.com/klever-io/klever-go/data/trie"
triestats "github.com/klever-io/klever-go/data/trie/statistics"
"github.com/stretchr/testify/require"
)
func newBaseSyncerForTimeoutPOC(t testing.T) baseAccountsSyncer { t.Helper()
storageManager, err := trie.NewTrieStorageManagerWithoutPruning(commonmock.NewMemDbMock())
require.NoError(t, err)
return &baseAccountsSyncer{
hasher: commonmock.HasherMock{},
marshalizer: &commonmock.MarshalizerMock{},
trieSyncers: make(map[string]data.TrieSyncer),
dataTries: make(map[string]data.Trie),
trieStorageManager: storageManager,
requestHandler: &commonmock.RequestHandlerStub{},
timeout: time.Second,
cacher: commonmock.NewCacherStub(),
maxTrieLevelInMemory: 5,
name: "timeout-poc",
maxHardCapForMissingNodes: 1,
}
}
func TestPOC_UserAccountsSyncer_LeaksThrottlerSlotOnTrieTimeout(t *testing.T) { thr, err := corethrottler.NewNumGoRoutinesThrottler(2) require.NoError(t, err)
s := &userAccountsSyncer{
baseAccountsSyncer: newBaseSyncerForTimeoutPOC(t),
throttler: thr,
}
err = s.syncDataTrie([]byte("missing-root-1"), triestats.NewTrieSyncStatistics(), context.Background())
require.ErrorIs(t, err, trie.ErrTimeIsOut)
require.True(t, thr.CanProcess())
err = s.syncDataTrie([]byte("missing-root-2"), triestats.NewTrieSyncStatistics(), context.Background())
require.ErrorIs(t, err, trie.ErrTimeIsOut)
require.False(t, thr.CanProcess())
}
func TestPOC_KappAccountsSyncer_LeaksThrottlerSlotOnTrieTimeout(t *testing.T) { thr, err := corethrottler.NewNumGoRoutinesThrottler(2) require.NoError(t, err)
s := &kappAccountsSyncer{
baseAccountsSyncer: newBaseSyncerForTimeoutPOC(t),
throttler: thr,
}
err = s.syncDataTrie([]byte("missing-root-1"), triestats.NewTrieSyncStatistics(), context.Background())
require.ErrorIs(t, err, trie.ErrTimeIsOut)
require.True(t, thr.CanProcess())
err = s.syncDataTrie([]byte("missing-root-2"), triestats.NewTrieSyncStatistics(), context.Background())
require.ErrorIs(t, err, trie.ErrTimeIsOut)
require.False(t, thr.CanProcess())
}
### Command used
go test ./data/syncer -run 'TestPOC_(User|Kapp)AccountsSyncer_LeaksThrottlerSlotOnTrieTimeout' -count=1
### Result
ok github.com/klever-io/klever-go/data/syncer 4.005s
This confirms the leak with the real timeout path from trieSyncer.StartSyncing().
## Impact
An attacker who can repeatedly cause trie-node sync failures or timeouts during bootstrap can consume the bounded sync throttler until no capacity
remains.
Once enough slots are leaked:
- additional account-data trie sync attempts stop making progress
- the parent loop waits until context timeout
- SyncAccounts() fails
- epoch bootstrap fails
This is a core node availability issue. It affects fresh/restarting nodes and validators that need to bootstrap or resync state.
This is not a theoretical issue:
- StartSyncing() performs network-dependent trie-node retrieval
- it already has explicit timeout / failure paths
- the leaked throttler slots are confirmed by runtime PoC
## Recommended Fix
Release the slot with defer immediately after StartProcessing() and cancel the defer only if ownership is intentionally transferred, which is not the
case here.
Example fix pattern:
func (u *userAccountsSyncer) syncDataTrie(rootHash []byte, ssh data.SyncStatisticsHandler, ctx context.Context) error { u.throttler.StartProcessing() defer u.throttler.EndProcessing()
u.syncerMutex.Lock()
defer u.syncerMutex.Unlock()
if _, ok := u.dataTries[string(rootHash)]; ok {
return nil
}
dataTrie, err := trie.NewTrie(...)
if err != nil {
return err
}
trieSyncer, err := trie.NewTrieSyncer(arg)
if err != nil {
return err
}
u.trieSyncers[string(rootHash)] = trieSyncer
return trieSyncer.StartSyncing(rootHash, ctx)
} ``` The same pattern should be applied to:
- data/syncer/userAccountsSyncer.go
- data/syncer/kappAccountsSyncer.go
## References
- data/syncer/userAccountsSyncer.go
- data/syncer/kappAccountsSyncer.go
- data/trie/sync.go
- core/throttler/numGoRoutinesThrottler.go
- core/bootstrap/process.go
- SECURITY.md
{
"affected": [
{
"package": {
"ecosystem": "Go",
"name": "github.com/klever-io/klever-go"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "1.7.18"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-49343"
],
"database_specific": {
"cwe_ids": [
"CWE-400",
"CWE-772"
],
"github_reviewed": true,
"github_reviewed_at": "2026-06-05T16:40:40Z",
"nvd_published_at": null,
"severity": "MODERATE"
},
"details": "## Summary\n\n The account-data trie syncers leak bounded throttler slots on error paths in `syncDataTrie()`. Each failed trie sync permanently consumes one slot from\n the `NumGoRoutinesThrottler`, and the slot is never returned unless the sync succeeds or the root hash was already present.\n\n I confirmed this on the current default branch `develop` at commit `9640d63` (observed on May 20, 2026). I also confirmed the bug with a runtime PoC\n using the real timeout path in `trieSyncer.StartSyncing()`: two timed-out sync attempts are enough to exhaust a throttler with capacity `2`.\n\n This affects the epoch bootstrap path because `syncUserAccountsState()` and `syncKappAccountsState()` create bounded throttlers and abort bootstrap\n immediately if the syncer returns an error. Once enough trie-root sync attempts fail, the syncer cannot make forward progress and bootstrap fails.\n\n ## Affected Components\n\n - `data/syncer/userAccountsSyncer.go`\n - `data/syncer/kappAccountsSyncer.go`\n - `data/trie/sync.go`\n - `core/throttler/numGoRoutinesThrottler.go`\n - `core/bootstrap/process.go`\n\n ## Affected Version\n\n Verified on:\n - `develop` HEAD `9640d63`\n\n Please check whether the same code is present in supported `1.7.x` releases.\n\n ## Suggested Severity\n\n High\n\n ## Vulnerability Details\n\n ### Root Cause\n\n Both account-data syncers call `StartProcessing()` before creating / starting the trie syncer, but they only call `EndProcessing()` on the success path\n and on the duplicate-root early return.\n\n `userAccountsSyncer.syncDataTrie()`:\n\n ```go\n func (u *userAccountsSyncer) syncDataTrie(rootHash []byte, ssh data.SyncStatisticsHandler, ctx context.Context) error {\n u.throttler.StartProcessing()\n\n u.syncerMutex.Lock()\n if _, ok := u.dataTries[string(rootHash)]; ok {\n u.syncerMutex.Unlock()\n u.throttler.EndProcessing()\n return nil\n }\n\n dataTrie, err := trie.NewTrie(...)\n if err != nil {\n u.syncerMutex.Unlock()\n return err\n }\n\n trieSyncer, err := trie.NewTrieSyncer(arg)\n if err != nil {\n u.syncerMutex.Unlock()\n return err\n }\n\n u.syncerMutex.Unlock()\n\n err = trieSyncer.StartSyncing(rootHash, ctx)\n if err != nil {\n return err\n }\n\n u.throttler.EndProcessing()\n return nil\n }\n\n The same bug exists in kappAccountsSyncer.syncDataTrie().\n```\n ### Missing slot release paths\n\n After StartProcessing(), the following error paths return without EndProcessing():\n\n 1. trie.NewTrie(...) returns an error\n 2. trie.NewTrieSyncer(...) returns an error\n 3. trieSyncer.StartSyncing(...) returns an error\n\n ### Why this matters\n\n NumGoRoutinesThrottler is a strict bounded counter:\n```\n func (ngrt *NumGoRoutinesThrottler) CanProcess() bool {\n valCounter := atomic.LoadInt32(\u0026ngrt.counter)\n return valCounter \u003c ngrt.max\n }\n\n func (ngrt *NumGoRoutinesThrottler) StartProcessing() {\n atomic.AddInt32(\u0026ngrt.counter, 1)\n }\n\n func (ngrt *NumGoRoutinesThrottler) EndProcessing() {\n atomic.AddInt32(\u0026ngrt.counter, -1)\n }\n\n Once leaked, a slot remains consumed for the lifetime of that throttler instance.\n\n The parent loops in both syncers wait for capacity before starting the next account-data trie sync:\n\n for !u.throttler.CanProcess() {\n select {\n case \u003c-time.After(timeBetweenRetries):\n continue\n case \u003c-ctx.Done():\n return common.ErrTimeIsOut\n }\n }\n```\n So after enough failures, further roots stop progressing and the sync operation eventually returns time is out.\n\n ### Bootstrap impact\n\n Epoch bootstrap uses these syncers directly and aborts on any error:\n```\n err = e.syncUserAccountsState(e.epochStartMeta.Header.TrieRoot)\n if err != nil {\n return nil, nil, err\n }\n\n err = e.syncKappAccountsState(e.epochStartMeta.Header.KAppsTrieRoot)\n if err != nil {\n return nil, nil, err\n }\n```\n The throttlers for these paths are real bounded throttlers created from numConcurrentTrieSyncers.\n\n ## Proof of Concept\n\n I verified the bug with the real timeout path, not only with a canceled context.\n\n The PoC below uses:\n\n - a real NumGoRoutinesThrottler with capacity 2\n - a real trieSyncer.StartSyncing()\n - an empty trie-node cache and a request handler that never supplies nodes\n - a short sync timeout (1s) so StartSyncing() returns trie.ErrTimeIsOut\n\n After the first failed sync, one slot remains leaked.\n After the second failed sync, the throttler is exhausted.\n\n ### PoC test\n```\n package syncer\n\n import (\n \"context\"\n \"testing\"\n \"time\"\n\n commonmock \"github.com/klever-io/klever-go/common/mock\"\n corethrottler \"github.com/klever-io/klever-go/core/throttler\"\n \"github.com/klever-io/klever-go/data\"\n \"github.com/klever-io/klever-go/data/trie\"\n triestats \"github.com/klever-io/klever-go/data/trie/statistics\"\n \"github.com/stretchr/testify/require\"\n )\n\n func newBaseSyncerForTimeoutPOC(t *testing.T) *baseAccountsSyncer {\n t.Helper()\n\n storageManager, err := trie.NewTrieStorageManagerWithoutPruning(commonmock.NewMemDbMock())\n require.NoError(t, err)\n\n return \u0026baseAccountsSyncer{\n hasher: commonmock.HasherMock{},\n marshalizer: \u0026commonmock.MarshalizerMock{},\n trieSyncers: make(map[string]data.TrieSyncer),\n dataTries: make(map[string]data.Trie),\n trieStorageManager: storageManager,\n requestHandler: \u0026commonmock.RequestHandlerStub{},\n timeout: time.Second,\n cacher: commonmock.NewCacherStub(),\n maxTrieLevelInMemory: 5,\n name: \"timeout-poc\",\n maxHardCapForMissingNodes: 1,\n }\n }\n\n func TestPOC_UserAccountsSyncer_LeaksThrottlerSlotOnTrieTimeout(t *testing.T) {\n thr, err := corethrottler.NewNumGoRoutinesThrottler(2)\n require.NoError(t, err)\n\n s := \u0026userAccountsSyncer{\n baseAccountsSyncer: newBaseSyncerForTimeoutPOC(t),\n throttler: thr,\n }\n\n err = s.syncDataTrie([]byte(\"missing-root-1\"), triestats.NewTrieSyncStatistics(), context.Background())\n require.ErrorIs(t, err, trie.ErrTimeIsOut)\n require.True(t, thr.CanProcess())\n\n err = s.syncDataTrie([]byte(\"missing-root-2\"), triestats.NewTrieSyncStatistics(), context.Background())\n require.ErrorIs(t, err, trie.ErrTimeIsOut)\n require.False(t, thr.CanProcess())\n }\n\n func TestPOC_KappAccountsSyncer_LeaksThrottlerSlotOnTrieTimeout(t *testing.T) {\n thr, err := corethrottler.NewNumGoRoutinesThrottler(2)\n require.NoError(t, err)\n\n s := \u0026kappAccountsSyncer{\n baseAccountsSyncer: newBaseSyncerForTimeoutPOC(t),\n throttler: thr,\n }\n\n err = s.syncDataTrie([]byte(\"missing-root-1\"), triestats.NewTrieSyncStatistics(), context.Background())\n require.ErrorIs(t, err, trie.ErrTimeIsOut)\n require.True(t, thr.CanProcess())\n\n err = s.syncDataTrie([]byte(\"missing-root-2\"), triestats.NewTrieSyncStatistics(), context.Background())\n require.ErrorIs(t, err, trie.ErrTimeIsOut)\n require.False(t, thr.CanProcess())\n }\n```\n ### Command used\n```\n go test ./data/syncer -run \u0027TestPOC_(User|Kapp)AccountsSyncer_LeaksThrottlerSlotOnTrieTimeout\u0027 -count=1\n```\n ### Result\n```\n ok github.com/klever-io/klever-go/data/syncer 4.005s\n```\n This confirms the leak with the real timeout path from trieSyncer.StartSyncing().\n\n ## Impact\n\n An attacker who can repeatedly cause trie-node sync failures or timeouts during bootstrap can consume the bounded sync throttler until no capacity\n remains.\n\n Once enough slots are leaked:\n\n - additional account-data trie sync attempts stop making progress\n - the parent loop waits until context timeout\n - SyncAccounts() fails\n - epoch bootstrap fails\n\n This is a core node availability issue. It affects fresh/restarting nodes and validators that need to bootstrap or resync state.\n\n This is not a theoretical issue:\n\n - StartSyncing() performs network-dependent trie-node retrieval\n - it already has explicit timeout / failure paths\n - the leaked throttler slots are confirmed by runtime PoC\n\n ## Recommended Fix\n\n Release the slot with defer immediately after StartProcessing() and cancel the defer only if ownership is intentionally transferred, which is not the\n case here.\n\n Example fix pattern:\n```\n func (u *userAccountsSyncer) syncDataTrie(rootHash []byte, ssh data.SyncStatisticsHandler, ctx context.Context) error {\n u.throttler.StartProcessing()\n defer u.throttler.EndProcessing()\n\n u.syncerMutex.Lock()\n defer u.syncerMutex.Unlock()\n\n if _, ok := u.dataTries[string(rootHash)]; ok {\n return nil\n }\n\n dataTrie, err := trie.NewTrie(...)\n if err != nil {\n return err\n }\n\n trieSyncer, err := trie.NewTrieSyncer(arg)\n if err != nil {\n return err\n }\n\n u.trieSyncers[string(rootHash)] = trieSyncer\n return trieSyncer.StartSyncing(rootHash, ctx)\n }\n```\n The same pattern should be applied to:\n\n - data/syncer/userAccountsSyncer.go\n - data/syncer/kappAccountsSyncer.go\n\n ## References\n\n - data/syncer/userAccountsSyncer.go\n - data/syncer/kappAccountsSyncer.go\n - data/trie/sync.go\n - core/throttler/numGoRoutinesThrottler.go\n - core/bootstrap/process.go\n - SECURITY.md",
"id": "GHSA-fw38-pc54-jvx9",
"modified": "2026-06-05T16:40:40Z",
"published": "2026-06-05T16:40:40Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/klever-io/klever-go/security/advisories/GHSA-fw38-pc54-jvx9"
},
{
"type": "PACKAGE",
"url": "https://github.com/klever-io/klever-go"
},
{
"type": "WEB",
"url": "https://github.com/klever-io/klever-go/releases/tag/v1.7.18"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
],
"summary": "Klever-Go KVM: Throttler slot leak in trie account-data sync causes epoch bootstrap / state sync DoS"
}