Files
seaweedFS/weed/iam/policy/policy_engine.go
Chris Lu ee3813787e feat(s3api): Implement S3 Policy Variables (#8039)
* feat: Add AWS IAM Policy Variables support to S3 API

Implements policy variables for dynamic access control in bucket policies.

Supported variables:
- aws:username - Extracted from principal ARN
- aws:userid - User identifier (same as username in SeaweedFS)
- aws:principaltype - IAMUser, IAMRole, or AssumedRole
- jwt:* - Any JWT claim (e.g., jwt:preferred_username, jwt:sub)

Key changes:
- Added PolicyVariableRegex to detect ${...} patterns
- Extended CompiledStatement with DynamicResourcePatterns, DynamicPrincipalPatterns, DynamicActionPatterns
- Added Claims field to PolicyEvaluationArgs for JWT claim access
- Implemented SubstituteVariables() for variable replacement from context and JWT claims
- Implemented extractPrincipalVariables() for ARN parsing
- Updated EvaluateConditions() to support variable substitution
- Comprehensive unit and integration tests

Resolves #8037

* feat: Add LDAP and PrincipalAccount variable support

Completes future enhancements for policy variables:

- Added ldap:* variable support for LDAP claims
  - ldap:username - LDAP username from claims
  - ldap:dn - LDAP distinguished name from claims
  - ldap:* - Any LDAP claim

- Added aws:PrincipalAccount extraction from ARN
  - Extracts account ID from principal ARN
  - Available as ${aws:PrincipalAccount} in policies

Updated SubstituteVariables() to check LDAP claims
Updated extractPrincipalVariables() to extract account ID
Added comprehensive tests for new variables

* feat(s3api): implement IAM policy variables core logic and optimization

* feat(s3api): integrate policy variables with S3 authentication and handlers

* test(s3api): add integration tests for policy variables

* cleanup: remove unused policy conversion files

* Add S3 policy variables integration tests and path support

- Add comprehensive integration tests for policy variables
- Test username isolation, JWT claims, LDAP claims
- Add support for IAM paths in principal ARN parsing
- Add tests for principals with paths

* Fix IAM Role principal variable extraction

IAM Roles should not have aws:userid or aws:PrincipalAccount
according to AWS behavior. Only IAM Users and Assumed Roles
should have these variables.

Fixes TestExtractPrincipalVariables test failures.

* Security fixes and bug fixes for S3 policy variables

SECURITY FIXES:
- Prevent X-SeaweedFS-Principal header spoofing by clearing internal
  headers at start of authentication (auth_credentials.go)
- Restrict policy variable substitution to safe allowlist to prevent
  client header injection (iam/policy/policy_engine.go)
- Add core policy validation before storing bucket policies

BUG FIXES:
- Remove unused sid variable in evaluateStatement
- Fix LDAP claim lookup to check both prefixed and unprefixed keys
- Add ValidatePolicy call in PutBucketPolicyHandler

These fixes prevent privilege escalation via header injection and
ensure only validated identity claims are used in policy evaluation.

* Additional security fixes and code cleanup

SECURITY FIXES:
- Fixed X-Forwarded-For spoofing by only trusting proxy headers from
  private/localhost IPs (s3_iam_middleware.go)
- Changed context key from "sourceIP" to "aws:SourceIp" for proper
  policy variable substitution

CODE IMPROVEMENTS:
- Kept aws:PrincipalAccount for IAM Roles to support condition evaluations
- Removed redundant STS principaltype override
- Removed unused service variable
- Cleaned up commented-out debug logging statements
- Updated tests to reflect new IAM Role behavior

These changes prevent IP spoofing attacks and ensure policy variables
work correctly with the safe allowlist.

* Add security documentation for ParseJWTToken

Added comprehensive security comments explaining that ParseJWTToken
is safe despite parsing without verification because:
- It's only used for routing to the correct verification method
- All code paths perform cryptographic verification before trusting claims
- OIDC tokens: validated via validateExternalOIDCToken
- STS tokens: validated via ValidateSessionToken

Enhanced function documentation with clear security warnings about
proper usage to prevent future misuse.

* Fix IP condition evaluation to use aws:SourceIp key

Fixed evaluateIPCondition in IAM policy engine to use "aws:SourceIp"
instead of "sourceIP" to match the updated extractRequestContext.

This fixes the failing IP-restricted role test where IP-based policy
conditions were not being evaluated correctly.

Updated all test cases to use the correct "aws:SourceIp" key.

* Address code review feedback: optimize and clarify

PERFORMANCE IMPROVEMENT:
- Optimized expandPolicyVariables to use regexp.ReplaceAllStringFunc
  for single-pass variable substitution instead of iterating through
  all safe variables. This improves performance from O(n*m) to O(m)
  where n is the number of safe variables and m is the pattern length.

CODE CLARITY:
- Added detailed comment explaining LDAP claim fallback mechanism
  (checks both prefixed and unprefixed keys for compatibility)
- Enhanced TODO comment for trusted proxy configuration with rationale
  and recommendations for supporting cloud load balancers, CDNs, and
  complex network topologies

All tests passing.

* Address Copilot code review feedback

BUG FIXES:
- Fixed type switch for int/int32/int64 - separated into individual cases
  since interface type switches only match the first type in multi-type cases
- Fixed grammatically incorrect error message in types.go

CODE QUALITY:
- Removed duplicate Resource/NotResource validation (already in ValidateStatement)
- Added comprehensive comment explaining isEnabled() logic and security implications
- Improved trusted proxy NOTE comment to be more concise while noting limitations

All tests passing.

* Fix test failures after extractSourceIP security changes

Updated tests to work with the security fix that only trusts
X-Forwarded-For/X-Real-IP headers from private IP addresses:

- Set RemoteAddr to 127.0.0.1 in tests to simulate trusted proxy
- Changed context key from "sourceIP" to "aws:SourceIp"
- Added test case for untrusted proxy (public RemoteAddr)
- Removed invalid ValidateStatement call (validation happens in ValidatePolicy)

All tests now passing.

* Address remaining Gemini code review feedback

CODE SAFETY:
- Deep clone Action field in CompileStatement to prevent potential data races
  if the original policy document is modified after compilation

TEST CLEANUP:
- Remove debug logging (fmt.Fprintf) from engine_notresource_test.go
- Remove unused imports in engine_notresource_test.go

All tests passing.

* Fix insecure JWT parsing in IAM auth flow

SECURITY FIX:
- Renamed ParseJWTToken to ParseUnverifiedJWTToken with explicit security warnings.
- Refactored AuthenticateJWT to use the trusted SessionInfo returned by ValidateSessionToken
  instead of relying on unverified claims from the initial parse.
- Refactored ValidatePresignedURLWithIAM to reuse the robust AuthenticateJWT logic, removing
  duplicated and insecure manual token parsing.

This ensures all identity information (Role, Principal, Subject) used for authorization
decisions is derived solely from cryptographically verified tokens.

* Security: Fix insecure JWT claim extraction in policy engine

- Refactored EvaluatePolicy to accept trusted claims from verified Identity instead of parsing unverified tokens
- Updated AuthenticateJWT to populate Claims in IAMIdentity from verified sources (SessionInfo/ExternalIdentity)
- Updated s3api_server and handlers to pass claims correctly
- Improved isPrivateIP to support IPv6 loopback, link-local, and ULA
- Fixed flaky distributed_session_consistency test with retry logic

* fix(iam): populate Subject in STSSessionInfo to ensure correct identity propagation

This fixes the TestS3IAMAuthentication/valid_jwt_token_authentication failure by ensuring the session subject (sub) is correctly mapped to the internal SessionInfo struct, allowing bucket ownership validation to succeed.

* Optimized isPrivateIP

* Create s3-policy-tests.yml

* fix tests

* fix tests

* tests(s3/iam): simplify policy to resource-based \ (step 1)

* tests(s3/iam): add explicit Deny NotResource for isolation (step 2)

* fixes

* policy: skip resource matching for STS trust policies to allow AssumeRole evaluation

* refactor: remove debug logging and hoist policy variables for performance

* test: fix TestS3IAMBucketPolicyIntegration cleanup to handle per-subtest object lifecycle

* test: fix bucket name generation to comply with S3 63-char limit

* test: skip TestS3IAMPolicyEnforcement until role setup is implemented

* test: use weed mini for simpler test server deployment

Replace 'weed server' with 'weed mini' for IAM tests to avoid port binding issues
and simplify the all-in-one server deployment. This improves test reliability
and execution time.

* security: prevent allocation overflow in policy evaluation

Add maxPoliciesForEvaluation constant to cap the number of policies evaluated
in a single request. This prevents potential integer overflow when allocating
slices for policy lists that may be influenced by untrusted input.

Changes:
- Add const maxPoliciesForEvaluation = 1024 to set an upper bound
- Validate len(policies) < maxPoliciesForEvaluation before appending bucket policy
- Use append() instead of make([]string, len+1) to avoid arithmetic overflow
- Apply fix to both IsActionAllowed policy evaluation paths
2026-01-16 11:12:28 -08:00

1419 lines
40 KiB
Go

package policy
import (
"context"
"encoding/json"
"fmt"
"net"
"path/filepath"
"regexp"
"strconv"
"strings"
"sync"
"time"
)
// Effect represents the policy evaluation result
type Effect string
const (
EffectAllow Effect = "Allow"
EffectDeny Effect = "Deny"
)
// Package-level regex cache for performance optimization
var (
regexCache = make(map[string]*regexp.Regexp)
regexCacheMu sync.RWMutex
policyVariablePattern = regexp.MustCompile(`\$\{([^}]+)\}`)
safePolicyVariables = map[string]bool{
// AWS standard identity variables
"aws:username": true,
"aws:userid": true,
"aws:PrincipalArn": true,
"aws:PrincipalAccount": true,
"aws:principaltype": true,
"aws:FederatedProvider": true,
"aws:PrincipalServiceName": true,
// SAML identity variables
"saml:username": true,
"saml:sub": true,
"saml:aud": true,
"saml:iss": true,
// OIDC/JWT identity variables
"oidc:sub": true,
"oidc:aud": true,
"oidc:iss": true,
// JWT identity variables
"jwt:preferred_username": true,
"jwt:sub": true,
"jwt:iss": true,
"jwt:aud": true,
// AWS request context (not from headers)
"aws:SourceIp": true,
"aws:SecureTransport": true,
"aws:CurrentTime": true,
"s3:prefix": true,
"s3:delimiter": true,
"s3:max-keys": true,
}
)
// PolicyEngine evaluates policies against requests
type PolicyEngine struct {
config *PolicyEngineConfig
initialized bool
store PolicyStore
}
// PolicyEngineConfig holds policy engine configuration
type PolicyEngineConfig struct {
// DefaultEffect when no policies match (Allow or Deny)
DefaultEffect string `json:"defaultEffect"`
// StoreType specifies the policy store backend (memory, filer, etc.)
StoreType string `json:"storeType"`
// StoreConfig contains store-specific configuration
StoreConfig map[string]interface{} `json:"storeConfig,omitempty"`
}
// PolicyDocument represents an IAM policy document
type PolicyDocument struct {
// Version of the policy language (e.g., "2012-10-17")
Version string `json:"Version"`
// Id is an optional policy identifier
Id string `json:"Id,omitempty"`
// Statement contains the policy statements
Statement []Statement `json:"Statement"`
}
// Statement represents a single policy statement
type Statement struct {
// Sid is an optional statement identifier
Sid string `json:"Sid,omitempty"`
// Effect specifies whether to Allow or Deny
Effect string `json:"Effect"`
// Principal specifies who the statement applies to (optional in role policies)
Principal interface{} `json:"Principal,omitempty"`
// NotPrincipal specifies who the statement does NOT apply to
NotPrincipal interface{} `json:"NotPrincipal,omitempty"`
// Action specifies the actions this statement applies to
Action StringList `json:"Action"`
// NotAction specifies actions this statement does NOT apply to
NotAction StringList `json:"NotAction,omitempty"`
// Resource specifies the resources this statement applies to
Resource StringList `json:"Resource"`
// NotResource specifies resources this statement does NOT apply to
NotResource StringList `json:"NotResource,omitempty"`
// Condition specifies conditions for when this statement applies
Condition map[string]map[string]interface{} `json:"Condition,omitempty"`
}
// StringList handles fields that can be a string or a list of strings
type StringList []string
// UnmarshalJSON implements custom unmarshalling for StringList
func (sl *StringList) UnmarshalJSON(data []byte) error {
var s string
if err := json.Unmarshal(data, &s); err == nil {
*sl = []string{s}
return nil
}
var sa []string
if err := json.Unmarshal(data, &sa); err == nil {
*sl = sa
return nil
}
return fmt.Errorf("invalid string list")
}
// EvaluationContext provides context for policy evaluation
type EvaluationContext struct {
// Principal making the request (e.g., "user:alice", "role:admin")
Principal string `json:"principal"`
// Action being requested (e.g., "s3:GetObject")
Action string `json:"action"`
// Resource being accessed (e.g., "arn:aws:s3:::bucket/key")
Resource string `json:"resource"`
// RequestContext contains additional request information
RequestContext map[string]interface{} `json:"requestContext,omitempty"`
}
// EvaluationResult contains the result of policy evaluation
type EvaluationResult struct {
// Effect is the final decision (Allow or Deny)
Effect Effect `json:"effect"`
// MatchingStatements contains statements that matched the request
MatchingStatements []StatementMatch `json:"matchingStatements,omitempty"`
// EvaluationDetails provides detailed evaluation information
EvaluationDetails *EvaluationDetails `json:"evaluationDetails,omitempty"`
}
// StatementMatch represents a statement that matched during evaluation
type StatementMatch struct {
// PolicyName is the name of the policy containing this statement
PolicyName string `json:"policyName"`
// StatementSid is the statement identifier
StatementSid string `json:"statementSid,omitempty"`
// Effect is the effect of this statement
Effect Effect `json:"effect"`
// Reason explains why this statement matched
Reason string `json:"reason,omitempty"`
}
// EvaluationDetails provides detailed information about policy evaluation
type EvaluationDetails struct {
// Principal that was evaluated
Principal string `json:"principal"`
// Action that was evaluated
Action string `json:"action"`
// Resource that was evaluated
Resource string `json:"resource"`
// PoliciesEvaluated lists all policies that were evaluated
PoliciesEvaluated []string `json:"policiesEvaluated"`
// ConditionsEvaluated lists all conditions that were evaluated
ConditionsEvaluated []string `json:"conditionsEvaluated,omitempty"`
}
// PolicyStore defines the interface for storing and retrieving policies
type PolicyStore interface {
// StorePolicy stores a policy document (filerAddress ignored for memory stores)
StorePolicy(ctx context.Context, filerAddress string, name string, policy *PolicyDocument) error
// GetPolicy retrieves a policy document (filerAddress ignored for memory stores)
GetPolicy(ctx context.Context, filerAddress string, name string) (*PolicyDocument, error)
// DeletePolicy deletes a policy document (filerAddress ignored for memory stores)
DeletePolicy(ctx context.Context, filerAddress string, name string) error
// ListPolicies lists all policy names (filerAddress ignored for memory stores)
ListPolicies(ctx context.Context, filerAddress string) ([]string, error)
}
// NewPolicyEngine creates a new policy engine
func NewPolicyEngine() *PolicyEngine {
return &PolicyEngine{}
}
// Initialize initializes the policy engine with configuration
func (e *PolicyEngine) Initialize(config *PolicyEngineConfig) error {
if config == nil {
return fmt.Errorf("config cannot be nil")
}
if err := e.validateConfig(config); err != nil {
return fmt.Errorf("invalid configuration: %w", err)
}
e.config = config
// Initialize policy store
store, err := e.createPolicyStore(config)
if err != nil {
return fmt.Errorf("failed to create policy store: %w", err)
}
e.store = store
e.initialized = true
return nil
}
// InitializeWithProvider initializes the policy engine with configuration and a filer address provider
func (e *PolicyEngine) InitializeWithProvider(config *PolicyEngineConfig, filerAddressProvider func() string) error {
if config == nil {
return fmt.Errorf("config cannot be nil")
}
if err := e.validateConfig(config); err != nil {
return fmt.Errorf("invalid configuration: %w", err)
}
e.config = config
// Initialize policy store with provider
store, err := e.createPolicyStoreWithProvider(config, filerAddressProvider)
if err != nil {
return fmt.Errorf("failed to create policy store: %w", err)
}
e.store = store
e.initialized = true
return nil
}
// validateConfig validates the policy engine configuration
func (e *PolicyEngine) validateConfig(config *PolicyEngineConfig) error {
if config.DefaultEffect != "Allow" && config.DefaultEffect != "Deny" {
return fmt.Errorf("invalid default effect: %s", config.DefaultEffect)
}
if config.StoreType == "" {
config.StoreType = "filer" // Default to filer store for persistence
}
return nil
}
// createPolicyStore creates a policy store based on configuration
func (e *PolicyEngine) createPolicyStore(config *PolicyEngineConfig) (PolicyStore, error) {
switch config.StoreType {
case "memory":
return NewMemoryPolicyStore(), nil
case "", "filer":
// Check if caching is explicitly disabled
if config.StoreConfig != nil {
if noCache, ok := config.StoreConfig["noCache"].(bool); ok && noCache {
return NewFilerPolicyStore(config.StoreConfig, nil)
}
}
// Default to generic cached filer store for better performance
return NewGenericCachedPolicyStore(config.StoreConfig, nil)
case "cached-filer", "generic-cached":
return NewGenericCachedPolicyStore(config.StoreConfig, nil)
default:
return nil, fmt.Errorf("unsupported store type: %s", config.StoreType)
}
}
// createPolicyStoreWithProvider creates a policy store with a filer address provider function
func (e *PolicyEngine) createPolicyStoreWithProvider(config *PolicyEngineConfig, filerAddressProvider func() string) (PolicyStore, error) {
switch config.StoreType {
case "memory":
return NewMemoryPolicyStore(), nil
case "", "filer":
// Check if caching is explicitly disabled
if config.StoreConfig != nil {
if noCache, ok := config.StoreConfig["noCache"].(bool); ok && noCache {
return NewFilerPolicyStore(config.StoreConfig, filerAddressProvider)
}
}
// Default to generic cached filer store for better performance
return NewGenericCachedPolicyStore(config.StoreConfig, filerAddressProvider)
case "cached-filer", "generic-cached":
return NewGenericCachedPolicyStore(config.StoreConfig, filerAddressProvider)
default:
return nil, fmt.Errorf("unsupported store type: %s", config.StoreType)
}
}
// IsInitialized returns whether the engine is initialized
func (e *PolicyEngine) IsInitialized() bool {
return e.initialized
}
// AddPolicy adds a policy to the engine (filerAddress ignored for memory stores)
func (e *PolicyEngine) AddPolicy(filerAddress string, name string, policy *PolicyDocument) error {
if !e.initialized {
return fmt.Errorf("policy engine not initialized")
}
if name == "" {
return fmt.Errorf("policy name cannot be empty")
}
if policy == nil {
return fmt.Errorf("policy cannot be nil")
}
if err := ValidatePolicyDocument(policy); err != nil {
return fmt.Errorf("invalid policy document: %w", err)
}
return e.store.StorePolicy(context.Background(), filerAddress, name, policy)
}
// Evaluate evaluates policies against a request context (filerAddress ignored for memory stores)
func (e *PolicyEngine) Evaluate(ctx context.Context, filerAddress string, evalCtx *EvaluationContext, policyNames []string) (*EvaluationResult, error) {
if !e.initialized {
return nil, fmt.Errorf("policy engine not initialized")
}
if evalCtx == nil {
return nil, fmt.Errorf("evaluation context cannot be nil")
}
result := &EvaluationResult{
Effect: Effect(e.config.DefaultEffect),
EvaluationDetails: &EvaluationDetails{
Principal: evalCtx.Principal,
Action: evalCtx.Action,
Resource: evalCtx.Resource,
PoliciesEvaluated: policyNames,
},
}
var matchingStatements []StatementMatch
explicitDeny := false
hasAllow := false
// Evaluate each policy
for _, policyName := range policyNames {
policy, err := e.store.GetPolicy(ctx, filerAddress, policyName)
if err != nil {
continue // Skip policies that can't be loaded
}
// Evaluate each statement in the policy
for _, statement := range policy.Statement {
if e.statementMatches(&statement, evalCtx) {
match := StatementMatch{
PolicyName: policyName,
StatementSid: statement.Sid,
Effect: Effect(statement.Effect),
Reason: "Action, Resource, and Condition matched",
}
matchingStatements = append(matchingStatements, match)
if statement.Effect == "Deny" {
explicitDeny = true
} else if statement.Effect == "Allow" {
hasAllow = true
}
}
}
}
result.MatchingStatements = matchingStatements
// AWS IAM evaluation logic:
// 1. If there's an explicit Deny, the result is Deny
// 2. If there's an Allow and no Deny, the result is Allow
// 3. Otherwise, use the default effect
if explicitDeny {
result.Effect = EffectDeny
} else if hasAllow {
result.Effect = EffectAllow
}
return result, nil
}
// EvaluateTrustPolicy evaluates a trust policy document directly (without storing it)
// This is used for AssumeRole/AssumeRoleWithWebIdentity trust policy validation
func (e *PolicyEngine) EvaluateTrustPolicy(ctx context.Context, trustPolicy *PolicyDocument, evalCtx *EvaluationContext) (*EvaluationResult, error) {
if !e.initialized {
return nil, fmt.Errorf("policy engine not initialized")
}
if evalCtx == nil {
return nil, fmt.Errorf("evaluation context cannot be nil")
}
if trustPolicy == nil {
return nil, fmt.Errorf("trust policy cannot be nil")
}
result := &EvaluationResult{
Effect: Effect(e.config.DefaultEffect),
EvaluationDetails: &EvaluationDetails{
Principal: evalCtx.Principal,
Action: evalCtx.Action,
Resource: evalCtx.Resource,
PoliciesEvaluated: []string{"trust-policy"},
},
}
var matchingStatements []StatementMatch
explicitDeny := false
hasAllow := false
// Evaluate each statement in the trust policy
for _, statement := range trustPolicy.Statement {
if e.statementMatches(&statement, evalCtx) {
match := StatementMatch{
PolicyName: "trust-policy",
StatementSid: statement.Sid,
Effect: Effect(statement.Effect),
Reason: "Principal, Action, and Condition matched",
}
matchingStatements = append(matchingStatements, match)
if statement.Effect == "Deny" {
explicitDeny = true
} else if statement.Effect == "Allow" {
hasAllow = true
}
}
}
result.MatchingStatements = matchingStatements
// AWS IAM evaluation logic:
// 1. If there's an explicit Deny, the result is Deny
// 2. If there's an Allow and no Deny, the result is Allow
// 3. Otherwise, use the default effect
if explicitDeny {
result.Effect = EffectDeny
} else if hasAllow {
result.Effect = EffectAllow
}
return result, nil
}
// statementMatches checks if a statement matches the evaluation context
func (e *PolicyEngine) statementMatches(statement *Statement, evalCtx *EvaluationContext) bool {
// Check principal match (for trust policies)
// If Principal field is present, it must match
if statement.Principal != nil {
if !e.matchesPrincipal(statement.Principal, evalCtx) {
return false
}
}
// Check action match
if !e.matchesActions(statement.Action, evalCtx.Action, evalCtx) {
return false
}
// Check resource match (optional for trust policies)
// For STS trust policy evaluations (AssumeRole*), resource matching should be skipped
// Trust policies typically don't include Resource, and enforcing resource matching
// here may cause valid trust statements to be rejected.
if strings.HasPrefix(evalCtx.Action, "sts:") {
// Skip resource checks for trust policy evaluation
} else if len(statement.Resource) > 0 {
if !e.matchesResources(statement.Resource, evalCtx.Resource, evalCtx) {
return false
}
}
// Check conditions
if !e.matchesConditions(statement.Condition, evalCtx) {
return false
}
return true
}
// matchesActions checks if any action in the list matches the requested action
func (e *PolicyEngine) matchesActions(actions []string, requestedAction string, evalCtx *EvaluationContext) bool {
for _, action := range actions {
if awsIAMMatch(action, requestedAction, evalCtx) {
return true
}
}
return false
}
// matchesResources checks if any resource in the list matches the requested resource
func (e *PolicyEngine) matchesResources(resources []string, requestedResource string, evalCtx *EvaluationContext) bool {
for _, resource := range resources {
if awsIAMMatch(resource, requestedResource, evalCtx) {
return true
}
}
return false
}
// matchesPrincipal checks if the principal in the statement matches the evaluation context
// This is used for trust policy evaluation (e.g., AssumeRole, AssumeRoleWithWebIdentity)
func (e *PolicyEngine) matchesPrincipal(principal interface{}, evalCtx *EvaluationContext) bool {
// Handle plain string principal (e.g., "*" or "arn:aws:iam::...")
if principalStr, ok := principal.(string); ok {
// Check wildcard FIRST before context validation
// This allows "*" to work without requiring context
if principalStr == "*" {
return true
}
// For non-wildcard string principals, we'd need specific matching logic
// For now, treat as a match if it equals the principal in context
if contextPrincipal, exists := evalCtx.RequestContext["principal"]; exists {
if contextPrincipalStr, ok := contextPrincipal.(string); ok {
return principalStr == contextPrincipalStr
}
}
return false
}
// Handle structured principal (e.g., {"Federated": "*"} or {"AWS": "arn:..."})
if principalMap, ok := principal.(map[string]interface{}); ok {
// For each principal type (Federated, AWS, Service, etc.)
for principalType, principalValue := range principalMap {
// Get the context key for this principal type
contextKey := getPrincipalContextKey(principalType)
if !e.evaluatePrincipalValue(principalValue, evalCtx, contextKey) {
return false
}
}
return true
}
// Unknown principal format
return false
}
// evaluatePrincipalValue evaluates a principal value against the evaluation context
// This handles wildcards, arrays, and context matching
func (e *PolicyEngine) evaluatePrincipalValue(principalValue interface{}, evalCtx *EvaluationContext, contextKey string) bool {
// Handle single string value
if principalStr, ok := principalValue.(string); ok {
// Check wildcard FIRST before context validation
// This allows {"Federated": "*"} to work without requiring context
if principalStr == "*" {
return true
}
// Then check against context
contextValue, exists := evalCtx.RequestContext[contextKey]
if !exists {
return false
}
contextStr, ok := contextValue.(string)
if !ok {
return false
}
return principalStr == contextStr
}
// Handle array of strings - convert to []interface{} for unified handling
var principalArray []interface{}
switch arr := principalValue.(type) {
case []interface{}:
principalArray = arr
case []string:
principalArray = make([]interface{}, len(arr))
for i, v := range arr {
principalArray[i] = v
}
default:
return false
}
if len(principalArray) > 0 {
for _, item := range principalArray {
if itemStr, ok := item.(string); ok {
// Wildcard in array allows any value
if itemStr == "*" {
return true
}
}
}
// If no wildcard found, check against context
contextValue, exists := evalCtx.RequestContext[contextKey]
if !exists {
return false
}
contextStr, ok := contextValue.(string)
if !ok {
return false
}
// Check if any array item matches the context
for _, item := range principalArray {
if itemStr, ok := item.(string); ok {
if itemStr == contextStr {
return true
}
}
}
}
return false
}
// getPrincipalContextKey returns the context key for a given principal type
// Uses AWS-compatible context keys for maximum compatibility
func getPrincipalContextKey(principalType string) string {
switch principalType {
case "Federated":
// For federated identity (OIDC/SAML), use the standard AWS context key
// This is typically populated with the identity provider ARN or URL
return "aws:FederatedProvider"
case "AWS":
// For AWS principals (IAM users/roles), use the principal ARN
return "aws:PrincipalArn"
case "Service":
// For AWS service principals
return "aws:PrincipalServiceName"
default:
// For any other principal type, use aws: prefix for compatibility
return "aws:Principal" + principalType
}
}
// matchesConditions checks if all conditions are satisfied
func (e *PolicyEngine) matchesConditions(conditions map[string]map[string]interface{}, evalCtx *EvaluationContext) bool {
if len(conditions) == 0 {
return true // No conditions means always match
}
for conditionType, conditionBlock := range conditions {
if !e.evaluateConditionBlock(conditionType, conditionBlock, evalCtx) {
return false
}
}
return true
}
// evaluateConditionBlock evaluates a single condition block
func (e *PolicyEngine) evaluateConditionBlock(conditionType string, block map[string]interface{}, evalCtx *EvaluationContext) bool {
switch conditionType {
// IP Address conditions
case "IpAddress":
return e.evaluateIPCondition(block, evalCtx, true)
case "NotIpAddress":
return e.evaluateIPCondition(block, evalCtx, false)
// String conditions
case "StringEquals":
return e.EvaluateStringCondition(block, evalCtx, true, false)
case "StringNotEquals":
return e.EvaluateStringCondition(block, evalCtx, false, false)
case "StringLike":
return e.EvaluateStringCondition(block, evalCtx, true, true)
case "StringNotLike":
return e.EvaluateStringCondition(block, evalCtx, false, true)
case "StringEqualsIgnoreCase":
return e.evaluateStringConditionIgnoreCase(block, evalCtx, true, false)
case "StringNotEqualsIgnoreCase":
return e.evaluateStringConditionIgnoreCase(block, evalCtx, false, false)
case "StringNotLikeIgnoreCase":
return e.evaluateStringConditionIgnoreCase(block, evalCtx, false, true)
// Numeric conditions
case "NumericEquals":
return e.evaluateNumericCondition(block, evalCtx, "==")
case "NumericNotEquals":
return e.evaluateNumericCondition(block, evalCtx, "!=")
case "NumericLessThan":
return e.evaluateNumericCondition(block, evalCtx, "<")
case "NumericLessThanEquals":
return e.evaluateNumericCondition(block, evalCtx, "<=")
case "NumericGreaterThan":
return e.evaluateNumericCondition(block, evalCtx, ">")
case "NumericGreaterThanEquals":
return e.evaluateNumericCondition(block, evalCtx, ">=")
// Date conditions
case "DateEquals":
return e.evaluateDateCondition(block, evalCtx, "==")
case "DateNotEquals":
return e.evaluateDateCondition(block, evalCtx, "!=")
case "DateLessThan":
return e.evaluateDateCondition(block, evalCtx, "<")
case "DateLessThanEquals":
return e.evaluateDateCondition(block, evalCtx, "<=")
case "DateGreaterThan":
return e.evaluateDateCondition(block, evalCtx, ">")
case "DateGreaterThanEquals":
return e.evaluateDateCondition(block, evalCtx, ">=")
// Boolean conditions
case "Bool":
return e.evaluateBoolCondition(block, evalCtx)
// Null conditions
case "Null":
return e.evaluateNullCondition(block, evalCtx)
default:
// Unknown condition types default to false (more secure)
return false
}
}
// evaluateIPCondition evaluates IP address conditions
func (e *PolicyEngine) evaluateIPCondition(block map[string]interface{}, evalCtx *EvaluationContext, shouldMatch bool) bool {
sourceIP, exists := evalCtx.RequestContext["aws:SourceIp"]
if !exists {
return !shouldMatch // If no IP in context, condition fails for positive match
}
sourceIPStr, ok := sourceIP.(string)
if !ok {
return !shouldMatch
}
sourceIPAddr := net.ParseIP(sourceIPStr)
if sourceIPAddr == nil {
return !shouldMatch
}
for key, value := range block {
if key == "aws:SourceIp" {
ranges, ok := value.([]string)
if !ok {
continue
}
for _, ipRange := range ranges {
if strings.Contains(ipRange, "/") {
// CIDR range
_, cidr, err := net.ParseCIDR(ipRange)
if err != nil {
continue
}
if cidr.Contains(sourceIPAddr) {
return shouldMatch
}
} else {
// Single IP
if sourceIPStr == ipRange {
return shouldMatch
}
}
}
}
}
return !shouldMatch
}
// EvaluateStringCondition evaluates string-based conditions
func (e *PolicyEngine) EvaluateStringCondition(block map[string]interface{}, evalCtx *EvaluationContext, shouldMatch bool, useWildcard bool) bool {
// Iterate through all condition keys in the block
for conditionKey, conditionValue := range block {
// Get the context values for this condition key
contextValues, exists := evalCtx.RequestContext[conditionKey]
if !exists {
// If the context key doesn't exist, condition fails for positive match
if shouldMatch {
return false
}
continue
}
// Convert context value to string slice
var contextStrings []string
switch v := contextValues.(type) {
case string:
contextStrings = []string{v}
case []string:
contextStrings = v
case []interface{}:
for _, item := range v {
if str, ok := item.(string); ok {
contextStrings = append(contextStrings, str)
}
}
default:
// Convert to string as fallback
contextStrings = []string{fmt.Sprintf("%v", v)}
}
// Convert condition value to string slice
var expectedStrings []string
switch v := conditionValue.(type) {
case string:
expectedStrings = []string{v}
case []string:
expectedStrings = v
case []interface{}:
for _, item := range v {
if str, ok := item.(string); ok {
expectedStrings = append(expectedStrings, str)
} else {
expectedStrings = append(expectedStrings, fmt.Sprintf("%v", item))
}
}
default:
expectedStrings = []string{fmt.Sprintf("%v", v)}
}
// Evaluate the condition using AWS IAM-compliant matching
conditionMet := false
for _, expected := range expectedStrings {
for _, contextValue := range contextStrings {
if useWildcard {
// Use AWS IAM-compliant wildcard matching for StringLike conditions
// This handles case-insensitivity and policy variables
if awsIAMMatch(expected, contextValue, evalCtx) {
conditionMet = true
break
}
} else {
// For StringEquals/StringNotEquals, also support policy variables but be case-sensitive
expandedExpected := expandPolicyVariables(expected, evalCtx)
if expandedExpected == contextValue {
conditionMet = true
break
}
}
}
if conditionMet {
break
}
}
// For shouldMatch=true (StringEquals, StringLike): condition must be met
// For shouldMatch=false (StringNotEquals): condition must NOT be met
if shouldMatch && !conditionMet {
return false
}
if !shouldMatch && conditionMet {
return false
}
}
return true
}
// ValidatePolicyDocument validates a policy document structure
func ValidatePolicyDocument(policy *PolicyDocument) error {
return ValidatePolicyDocumentWithType(policy, "resource")
}
// ValidateTrustPolicyDocument validates a trust policy document structure
func ValidateTrustPolicyDocument(policy *PolicyDocument) error {
return ValidatePolicyDocumentWithType(policy, "trust")
}
// ValidatePolicyDocumentWithType validates a policy document for specific type
func ValidatePolicyDocumentWithType(policy *PolicyDocument, policyType string) error {
if policy == nil {
return fmt.Errorf("policy document cannot be nil")
}
if policy.Version == "" {
return fmt.Errorf("version is required")
}
if len(policy.Statement) == 0 {
return fmt.Errorf("at least one statement is required")
}
for i, statement := range policy.Statement {
if err := validateStatementWithType(&statement, policyType); err != nil {
return fmt.Errorf("statement %d is invalid: %w", i, err)
}
}
return nil
}
// validateStatement validates a single statement (for backward compatibility)
func validateStatement(statement *Statement) error {
return validateStatementWithType(statement, "resource")
}
// validateStatementWithType validates a single statement based on policy type
func validateStatementWithType(statement *Statement, policyType string) error {
if statement.Effect != "Allow" && statement.Effect != "Deny" {
return fmt.Errorf("invalid effect: %s (must be Allow or Deny)", statement.Effect)
}
if len(statement.Action) == 0 {
return fmt.Errorf("at least one action is required")
}
// Trust policies don't require Resource field, but resource policies do
if policyType == "resource" {
if len(statement.Resource) == 0 {
return fmt.Errorf("at least one resource is required")
}
} else if policyType == "trust" {
// Trust policies should have Principal field
if statement.Principal == nil {
return fmt.Errorf("trust policy statement must have Principal field")
}
// Trust policies typically have specific actions
validTrustActions := map[string]bool{
"sts:AssumeRole": true,
"sts:AssumeRoleWithWebIdentity": true,
"sts:AssumeRoleWithCredentials": true,
}
for _, action := range statement.Action {
if !validTrustActions[action] {
return fmt.Errorf("invalid action for trust policy: %s", action)
}
}
}
return nil
}
// matchResource checks if a resource pattern matches a requested resource
// Uses hybrid approach: simple suffix wildcards for compatibility, filepath.Match for complex patterns
func matchResource(pattern, resource string) bool {
if pattern == resource {
return true
}
// Handle simple suffix wildcard (backward compatibility)
if strings.HasSuffix(pattern, "*") {
prefix := pattern[:len(pattern)-1]
return strings.HasPrefix(resource, prefix)
}
// For complex patterns, use filepath.Match for advanced wildcard support (*, ?, [])
matched, err := filepath.Match(pattern, resource)
if err != nil {
// Fallback to exact match if pattern is malformed
return pattern == resource
}
return matched
}
// awsIAMMatch performs AWS IAM-compliant pattern matching with case-insensitivity and policy variable support
func awsIAMMatch(pattern, value string, evalCtx *EvaluationContext) bool {
// Step 1: Substitute policy variables (e.g., ${aws:username}, ${saml:username})
expandedPattern := expandPolicyVariables(pattern, evalCtx)
// Step 2: Handle special patterns
if expandedPattern == "*" {
return true // Universal wildcard
}
// Step 3: Case-insensitive exact match
if strings.EqualFold(expandedPattern, value) {
return true
}
// Step 4: Handle AWS-style wildcards (case-insensitive)
if strings.Contains(expandedPattern, "*") || strings.Contains(expandedPattern, "?") {
return AwsWildcardMatch(expandedPattern, value)
}
return false
}
// expandPolicyVariables substitutes AWS policy variables in the pattern
func expandPolicyVariables(pattern string, evalCtx *EvaluationContext) string {
if evalCtx == nil || evalCtx.RequestContext == nil {
return pattern
}
// Use pre-compiled regexp for efficient single-pass substitution
result := policyVariablePattern.ReplaceAllStringFunc(pattern, func(match string) string {
// Extract variable name from ${variable}
variable := match[2 : len(match)-1]
// Only substitute if variable is in the safe allowlist
if !safePolicyVariables[variable] {
return match // Leave unsafe variables as-is
}
// Get value from request context
if value, exists := evalCtx.RequestContext[variable]; exists {
if str, ok := value.(string); ok {
return str
}
}
// Variable not found or not a string, leave as-is
return match
})
return result
}
// getContextValue safely gets a value from the evaluation context
func getContextValue(evalCtx *EvaluationContext, key, defaultValue string) string {
if value, exists := evalCtx.RequestContext[key]; exists {
if str, ok := value.(string); ok {
return str
}
}
return defaultValue
}
// AwsWildcardMatch performs case-insensitive wildcard matching like AWS IAM
func AwsWildcardMatch(pattern, value string) bool {
// Create regex pattern key for caching
// First escape all regex metacharacters, then replace wildcards
regexPattern := regexp.QuoteMeta(pattern)
regexPattern = strings.ReplaceAll(regexPattern, "\\*", ".*")
regexPattern = strings.ReplaceAll(regexPattern, "\\?", ".")
regexPattern = "^" + regexPattern + "$"
regexKey := "(?i)" + regexPattern
// Try to get compiled regex from cache
regexCacheMu.RLock()
regex, found := regexCache[regexKey]
regexCacheMu.RUnlock()
if !found {
// Compile and cache the regex
compiledRegex, err := regexp.Compile(regexKey)
if err != nil {
// Fallback to simple case-insensitive comparison if regex fails
return strings.EqualFold(pattern, value)
}
// Store in cache with write lock
regexCacheMu.Lock()
// Double-check in case another goroutine added it
if existingRegex, exists := regexCache[regexKey]; exists {
regex = existingRegex
} else {
regexCache[regexKey] = compiledRegex
regex = compiledRegex
}
regexCacheMu.Unlock()
}
return regex.MatchString(value)
}
// matchAction checks if an action pattern matches a requested action
// Uses hybrid approach: simple suffix wildcards for compatibility, filepath.Match for complex patterns
func matchAction(pattern, action string) bool {
if pattern == action {
return true
}
// Handle simple suffix wildcard (backward compatibility)
if strings.HasSuffix(pattern, "*") {
prefix := pattern[:len(pattern)-1]
return strings.HasPrefix(action, prefix)
}
// For complex patterns, use filepath.Match for advanced wildcard support (*, ?, [])
matched, err := filepath.Match(pattern, action)
if err != nil {
// Fallback to exact match if pattern is malformed
return pattern == action
}
return matched
}
// evaluateStringConditionIgnoreCase evaluates string conditions with case insensitivity
func (e *PolicyEngine) evaluateStringConditionIgnoreCase(block map[string]interface{}, evalCtx *EvaluationContext, shouldMatch bool, useWildcard bool) bool {
for key, expectedValues := range block {
contextValue, exists := evalCtx.RequestContext[key]
if !exists {
if !shouldMatch {
continue // For NotEquals, missing key is OK
}
return false
}
contextStr, ok := contextValue.(string)
if !ok {
return false
}
contextStr = strings.ToLower(contextStr)
matched := false
// Handle different value types
switch v := expectedValues.(type) {
case string:
expectedStr := strings.ToLower(v)
if useWildcard {
matched, _ = filepath.Match(expectedStr, contextStr)
} else {
matched = expectedStr == contextStr
}
case []interface{}:
for _, val := range v {
if valStr, ok := val.(string); ok {
expectedStr := strings.ToLower(valStr)
if useWildcard {
if m, _ := filepath.Match(expectedStr, contextStr); m {
matched = true
break
}
} else {
if expectedStr == contextStr {
matched = true
break
}
}
}
}
}
if shouldMatch && !matched {
return false
}
if !shouldMatch && matched {
return false
}
}
return true
}
// evaluateNumericCondition evaluates numeric conditions
func (e *PolicyEngine) evaluateNumericCondition(block map[string]interface{}, evalCtx *EvaluationContext, operator string) bool {
for key, expectedValues := range block {
contextValue, exists := evalCtx.RequestContext[key]
if !exists {
return false
}
contextNum, err := parseNumeric(contextValue)
if err != nil {
return false
}
matched := false
// Handle different value types
switch v := expectedValues.(type) {
case string:
expectedNum, err := parseNumeric(v)
if err != nil {
return false
}
matched = compareNumbers(contextNum, expectedNum, operator)
case float64:
matched = compareNumbers(contextNum, v, operator)
case int:
matched = compareNumbers(contextNum, float64(v), operator)
case int64:
matched = compareNumbers(contextNum, float64(v), operator)
case []interface{}:
for _, val := range v {
expectedNum, err := parseNumeric(val)
if err != nil {
continue
}
if compareNumbers(contextNum, expectedNum, operator) {
matched = true
break
}
}
default:
}
if !matched {
return false
}
}
return true
}
// evaluateDateCondition evaluates date conditions
func (e *PolicyEngine) evaluateDateCondition(block map[string]interface{}, evalCtx *EvaluationContext, operator string) bool {
for key, expectedValues := range block {
contextValue, exists := evalCtx.RequestContext[key]
if !exists {
return false
}
contextTime, err := parseDateTime(contextValue)
if err != nil {
return false
}
matched := false
// Handle different value types
switch v := expectedValues.(type) {
case string:
expectedTime, err := parseDateTime(v)
if err != nil {
return false
}
matched = compareDates(contextTime, expectedTime, operator)
case []interface{}:
for _, val := range v {
expectedTime, err := parseDateTime(val)
if err != nil {
continue
}
if compareDates(contextTime, expectedTime, operator) {
matched = true
break
}
}
}
if !matched {
return false
}
}
return true
}
// evaluateBoolCondition evaluates boolean conditions
func (e *PolicyEngine) evaluateBoolCondition(block map[string]interface{}, evalCtx *EvaluationContext) bool {
for key, expectedValues := range block {
contextValue, exists := evalCtx.RequestContext[key]
if !exists {
return false
}
contextBool, err := parseBool(contextValue)
if err != nil {
return false
}
matched := false
// Handle different value types
switch v := expectedValues.(type) {
case string:
expectedBool, err := parseBool(v)
if err != nil {
return false
}
matched = contextBool == expectedBool
case bool:
matched = contextBool == v
case []interface{}:
for _, val := range v {
expectedBool, err := parseBool(val)
if err != nil {
continue
}
if contextBool == expectedBool {
matched = true
break
}
}
}
if !matched {
return false
}
}
return true
}
// evaluateNullCondition evaluates null conditions
func (e *PolicyEngine) evaluateNullCondition(block map[string]interface{}, evalCtx *EvaluationContext) bool {
for key, expectedValues := range block {
_, exists := evalCtx.RequestContext[key]
expectedNull := false
switch v := expectedValues.(type) {
case string:
expectedNull = v == "true"
case bool:
expectedNull = v
}
// If we expect null (true) and key exists, or expect non-null (false) and key doesn't exist
if expectedNull == exists {
return false
}
}
return true
}
// Helper functions for parsing and comparing values
// parseNumeric parses a value as a float64
func parseNumeric(value interface{}) (float64, error) {
switch v := value.(type) {
case float64:
return v, nil
case float32:
return float64(v), nil
case int:
return float64(v), nil
case int64:
return float64(v), nil
case string:
return strconv.ParseFloat(v, 64)
default:
return 0, fmt.Errorf("cannot parse %T as numeric", value)
}
}
// compareNumbers compares two numbers using the given operator
func compareNumbers(a, b float64, operator string) bool {
switch operator {
case "==":
return a == b
case "!=":
return a != b
case "<":
return a < b
case "<=":
return a <= b
case ">":
return a > b
case ">=":
return a >= b
default:
return false
}
}
// parseDateTime parses a value as a time.Time
func parseDateTime(value interface{}) (time.Time, error) {
switch v := value.(type) {
case string:
// Try common date formats
formats := []string{
time.RFC3339,
"2006-01-02T15:04:05Z",
"2006-01-02T15:04:05",
"2006-01-02 15:04:05",
"2006-01-02",
}
for _, format := range formats {
if t, err := time.Parse(format, v); err == nil {
return t, nil
}
}
return time.Time{}, fmt.Errorf("cannot parse date: %s", v)
case time.Time:
return v, nil
default:
return time.Time{}, fmt.Errorf("cannot parse %T as date", value)
}
}
// compareDates compares two dates using the given operator
func compareDates(a, b time.Time, operator string) bool {
switch operator {
case "==":
return a.Equal(b)
case "!=":
return !a.Equal(b)
case "<":
return a.Before(b)
case "<=":
return a.Before(b) || a.Equal(b)
case ">":
return a.After(b)
case ">=":
return a.After(b) || a.Equal(b)
default:
return false
}
}
// parseBool parses a value as a boolean
func parseBool(value interface{}) (bool, error) {
switch v := value.(type) {
case bool:
return v, nil
case string:
return strconv.ParseBool(v)
default:
return false, fmt.Errorf("cannot parse %T as boolean", value)
}
}