Files
seaweedfs/weed/iam/sts/session_claims.go
T
ad89f003a5 s3: end a catalog-vended session with the access key it was vended for (#11945)
* s3: end a catalog-vended session with the access key it was vended for

A table credential the Iceberg catalog vends is a self-contained STS session:
it stays valid until it expires (an hour by default) even after the access key
that loaded the table is deleted or deactivated, or its identity is removed
or disabled. An operator revoking a leaked key cannot cut off the table
credentials that key already obtained.

Sessions are stateless JWTs, so there is no revocation list to consult. This
change follows the same pattern as role sessions, which are already refused
once their role's ID no longer matches (checkSessionRoleBinding): a vended
session carries the access key it was vended for (src_ak, recorded by the
previous change), and validateSTSSessionToken checks it against the identity
table held in memory. A session of provider iceberg-catalog that carries a
source key is refused (InvalidAccessKeyId, as for an unknown key) once its
identity is gone or disabled, or the key is no longer one of that identity's
active, unexpired credentials.

Scope:
- Only catalog-vended sessions are checked; web-identity, AssumeRole and
  every other STS session are validated as before.
- The vendor binds a session to the caller's key only when the caller is a
  stored identity holding that key. A caller that signed with another STS
  session's key, an OIDC user, or a request that recorded no key gets an
  unbound session, as today; binding it would refuse every request made with
  it. A session minted before this change carries no key and is unbound.
- The check is one map lookup under the read lock and a scan of one
  identity's keys (about 11 ns, 0 allocations, with 1,000 identities).

Tests (s3api_iceberg_vended_session_test.go) mint real sessions through the
catalog vendor and validate them the way a data-path request is validated:
deactivating, expiring or deleting key A ends A's session while the session
vended through key B on the same identity keeps working; disabling or
deleting the identity ends it; callers that are not a stored identity holding
the key get unbound sessions that keep working; other STS sessions are not
checked. The first two tests and the third fail without this change.

* s3: keep the vended-session key binding when the key dies before vend

vendedSessionKeyLive gated the binding, so a key revoked between the
request's authentication and the vend skipped it and minted an unbound
session that stayed usable until it expired. Every catalog auth path
pairs the identity name in context with the key that authenticated, so
a stored-identity principal always means the request signed with that
identity's own key. Bind whenever the principal is a stored identity: a
session bound to a now-dead key is refused on its first request, which
is what the binding is for.

* s3: gate the vended-session key binding on the authenticated caller type

lookupByIdentityName(principal) named only that some identity exists, not
that the request signed with a stored key: an STS session or OIDC principal
that collides with an identity name got a session bound to its temporary
key, which vendedSessionKeyLive then refused on every request. Verifiers
now record the verified key as a stored credential, and the vendor binds
only when that marker is present — keeping the binding even when the key
was revoked between authentication and the vend.

* s3: enforce the vended key binding under a dedicated claim

Sessions minted before the binding existed carry the caller's key in the
audit-only SourceAccessKeyId — including temporary STS keys that can never
resolve in the identity table. Enforcing on that field would end live table
credentials at upgrade; the binding now travels under BoundAccessKeyId,
leaving legacy sessions unbound as before.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* s3: bind vended sessions to a credential fingerprint

The key id alone cannot identify a credential: an operator can delete a
key and re-create it under the same id with a different secret, reviving
sessions minted for the original. Verifiers now record a fingerprint of
the secret at authentication; the session dies unless it still matches.

Generated with [Devin](https://devin.ai)

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>

* s3: keep recording the caller's key on every catalog session

SourceAccessKeyId was unconditional audit metadata before the binding
existed; gating it on the stored-credential marker dropped the record
for callers that authenticated another way. With the binding under its
own claim there is no ambiguity left, so the field records the caller's
key again regardless of how it authenticated.

* s3: key the vended-session credential fingerprint with the signing key

The bound-credential fingerprint travels inside the vended session
token. As a plain SHA-256 of the secret it gave any token holder an
offline guessing oracle for weak secrets, and CodeQL flags it twice.

Store the verified secret in the request marker instead and HMAC it
with the STS signing key at vend and at validation. The claim can no
longer be mined, and the per-request auth path stops hashing secrets.

---------

Co-authored-by: Javier de la Torre <jatorre@lakehousebox.com>
Co-authored-by: Chris Lu <chrislusf@users.noreply.github.com>
Co-authored-by: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
2026-10-23 03:58:06 -07:00

290 lines
10 KiB
Go

package sts
import (
"crypto/sha256"
"encoding/base64"
"strings"
"time"
"github.com/golang-jwt/jwt/v5"
"github.com/seaweedfs/seaweedfs/weed/glog"
)
// ComputeParentUser returns a stable per-identity hash derived from the OIDC
// (sub, iss) tuple. Only the (sub, iss) pair is guaranteed stable across token
// refreshes per OpenID Connect Core 1.0 §5.7, so any per-user state (audit
// logs, quotas) must key off this value rather than the access-key or session
// id. The hash is base64-rawurl-encoded SHA-256 over "openid:<sub>:<iss>" so
// it stays filesystem-safe and bounded in length for storage in audit paths.
func ComputeParentUser(sub, iss string) string {
if sub == "" || iss == "" {
return ""
}
h := sha256.Sum256([]byte("openid:" + sub + ":" + iss))
return base64.RawURLEncoding.EncodeToString(h[:])
}
// IsParentUserID reports whether s has the shape ComputeParentUser produces —
// a base64url-encoded SHA-256 — i.e. an already issuer-scoped id rather than a
// bare subject string.
func IsParentUserID(s string) bool {
b, err := base64.RawURLEncoding.DecodeString(s)
return err == nil && len(b) == sha256.Size
}
// identityClaimPriority is the order in which ResolveIdentityClaim looks for a
// human-readable, authoritative identity attribute in an STS request context.
// The context is populated at federation time from the validated OIDC token
// (see AssumeRoleWithWebIdentity), so every entry here is server-asserted and
// not client-supplied. preferred_username/email/name are conventional OIDC
// user claims; sub is the always-present stable subject identifier and the
// final fallback so a federated session never audits as fully anonymous.
var identityClaimPriority = []string{"preferred_username", "email", "name", "sub"}
// ResolveIdentityClaim returns the most human-readable authoritative identity
// claim available in ctx, or "" when none is present. ctx is the STS request
// context (sessionInfo.RequestContext) populated from the validated OIDC token
// at federation time. Non-string values are skipped so a structured claim
// never leaks into an audit-facing field. Whitespace-only values are treated
// as absent so a blank preferred claim does not mask a usable email or sub.
func ResolveIdentityClaim(ctx map[string]interface{}) string {
if len(ctx) == 0 {
return ""
}
for _, key := range identityClaimPriority {
if v, ok := ctx[key].(string); ok {
if trimmed := strings.TrimSpace(v); trimmed != "" {
return trimmed
}
}
}
return ""
}
// STSSessionClaims represents comprehensive session information embedded in JWT tokens
// This eliminates the need for separate session storage by embedding all session
// metadata directly in the token itself - enabling true stateless operation
type STSSessionClaims struct {
jwt.RegisteredClaims
// Session identification
SessionId string `json:"sid"` // session_id (abbreviated for smaller tokens)
SessionName string `json:"snam"` // session_name (abbreviated for smaller tokens)
TokenType string `json:"typ"` // token_type
// Role information
RoleArn string `json:"role"` // role_arn
RoleId string `json:"rid,omitempty"` // unique ID of the assumed role
AssumedRole string `json:"assumed"` // assumed_role_user
Principal string `json:"principal"` // principal_arn
// Authorization data
Policies []string `json:"pol,omitempty"` // policies (abbreviated)
// SessionPolicy contains inline session policy JSON (optional)
SessionPolicy string `json:"spol,omitempty"`
// Identity provider information
IdentityProvider string `json:"idp"` // identity_provider
ExternalUserId string `json:"ext_uid"` // external_user_id
ProviderIssuer string `json:"prov_iss"` // provider_issuer
// Request context (optional, for policy evaluation)
RequestContext map[string]interface{} `json:"req_ctx,omitempty"`
// Session metadata
AssumedAt time.Time `json:"assumed_at"` // when role was assumed
MaxDuration int64 `json:"max_dur,omitempty"` // maximum session duration in seconds
// ParentUser is a stable hash of (sub, iss) for tokens minted from an OIDC
// identity. It survives token rotation since only the (sub, iss) tuple is
// guaranteed stable per OpenID Connect Core 1.0. Empty for non-federated
// session types.
ParentUser string `json:"puid,omitempty"`
// SourceAccessKeyId is the access key the caller authenticated with when
// a SeaweedFS service minted the session for it (AssumeRoleForPrincipal);
// empty for every other session type.
SourceAccessKeyId string `json:"src_ak,omitempty"`
// BoundAccessKeyId is the stored credential the session is bound to: the
// session ends when that key is gone. It is a claim of its own because
// tokens minted before the binding existed carry SourceAccessKeyId as an
// audit field, never a binding — enforcing those would break them.
BoundAccessKeyId string `json:"bind_ak,omitempty"`
BoundAccessKeyFingerprint string `json:"bind_fp,omitempty"`
}
// NewSTSSessionClaims creates new STS session claims with all required information
func NewSTSSessionClaims(sessionId, issuer string, expiresAt time.Time) *STSSessionClaims {
now := time.Now()
return &STSSessionClaims{
RegisteredClaims: jwt.RegisteredClaims{
Issuer: issuer,
Subject: sessionId,
IssuedAt: jwt.NewNumericDate(now),
ExpiresAt: jwt.NewNumericDate(expiresAt),
NotBefore: jwt.NewNumericDate(now),
// jti = sessionId. The session id is already a unique random
// identifier, so reusing it as the JWT id avoids a second secret
// while still giving the revocation layer a stable lookup key.
ID: sessionId,
},
SessionId: sessionId,
TokenType: TokenTypeSession,
AssumedAt: now,
}
}
// ToSessionInfo converts JWT claims back to SessionInfo structure
// This enables seamless integration with existing code expecting SessionInfo
func (c *STSSessionClaims) ToSessionInfo(credGen *CredentialGenerator) *SessionInfo {
var expiresAt time.Time
if c.ExpiresAt != nil {
expiresAt = c.ExpiresAt.Time
}
var credentials *Credentials
if credGen != nil {
creds, err := credGen.GenerateTemporaryCredentials(c.SessionId, expiresAt)
if err != nil {
glog.Warningf("Failed to generate credentials for STS session %s: %v", c.SessionId, err)
} else {
credentials = creds
}
}
return &SessionInfo{
SessionId: c.SessionId,
SessionName: c.SessionName,
RoleArn: c.RoleArn,
RoleId: c.RoleId,
AssumedRoleUser: c.AssumedRole,
Principal: c.Principal,
Policies: c.Policies,
SessionPolicy: c.SessionPolicy,
ExpiresAt: expiresAt,
IdentityProvider: c.IdentityProvider,
ExternalUserId: c.ExternalUserId,
ProviderIssuer: c.ProviderIssuer,
RequestContext: c.RequestContext,
ParentUser: c.ParentUser,
// The access key the session was minted for, if recorded
SourceAccessKeyId: c.SourceAccessKeyId,
// The stored credential the session is bound to, if any
BoundAccessKeyId: c.BoundAccessKeyId,
BoundAccessKeyFingerprint: c.BoundAccessKeyFingerprint,
// Provide the Subject (sub) from registered claims
Subject: c.Subject,
Credentials: credentials,
}
}
// IsValid checks if the session claims are valid (not expired, etc.)
func (c *STSSessionClaims) IsValid() bool {
now := time.Now()
// Check expiration
if c.ExpiresAt != nil && c.ExpiresAt.Before(now) {
return false
}
// Check not-before
if c.NotBefore != nil && c.NotBefore.After(now) {
return false
}
// Ensure required fields are present
if c.SessionId == "" || c.RoleArn == "" || c.Principal == "" {
return false
}
return true
}
// GetSessionId returns the session identifier
func (c *STSSessionClaims) GetSessionId() string {
return c.SessionId
}
// GetExpiresAt returns the expiration time
func (c *STSSessionClaims) GetExpiresAt() time.Time {
if c.ExpiresAt != nil {
return c.ExpiresAt.Time
}
return time.Time{}
}
// WithRoleInfo sets role-related information in the claims
func (c *STSSessionClaims) WithRoleInfo(roleArn, assumedRole, principal string) *STSSessionClaims {
c.RoleArn = roleArn
c.AssumedRole = assumedRole
c.Principal = principal
return c
}
// WithPolicies sets the policies associated with this session
// WithRoleId binds the session to the assumed role's unique ID. An empty ID
// leaves the session unbound (roles that predate role IDs).
func (c *STSSessionClaims) WithRoleId(roleId string) *STSSessionClaims {
c.RoleId = roleId
return c
}
func (c *STSSessionClaims) WithPolicies(policies []string) *STSSessionClaims {
c.Policies = policies
return c
}
// WithSessionPolicy sets the inline session policy JSON for this session
func (c *STSSessionClaims) WithSessionPolicy(policy string) *STSSessionClaims {
c.SessionPolicy = policy
return c
}
// WithIdentityProvider sets identity provider information
func (c *STSSessionClaims) WithIdentityProvider(providerName, externalUserId, providerIssuer string) *STSSessionClaims {
c.IdentityProvider = providerName
c.ExternalUserId = externalUserId
c.ProviderIssuer = providerIssuer
return c
}
// WithRequestContext sets request context for policy evaluation
func (c *STSSessionClaims) WithRequestContext(ctx map[string]interface{}) *STSSessionClaims {
c.RequestContext = ctx
return c
}
// WithMaxDuration sets the maximum session duration
func (c *STSSessionClaims) WithMaxDuration(duration time.Duration) *STSSessionClaims {
c.MaxDuration = int64(duration.Seconds())
return c
}
// WithSessionName sets the session name
func (c *STSSessionClaims) WithSessionName(sessionName string) *STSSessionClaims {
c.SessionName = sessionName
return c
}
// WithSourceAccessKeyId records the access key the session was minted for.
func (c *STSSessionClaims) WithSourceAccessKeyId(accessKeyId string) *STSSessionClaims {
c.SourceAccessKeyId = accessKeyId
return c
}
// WithBoundAccessKey binds the session to a stored credential: the key id it
// was vended for and a fingerprint of its secret at mint time.
func (c *STSSessionClaims) WithBoundAccessKey(accessKeyId, fingerprint string) *STSSessionClaims {
c.BoundAccessKeyId = accessKeyId
c.BoundAccessKeyFingerprint = fingerprint
return c
}
// WithParentUser sets the stable per-identity hash for the session. See
// ComputeParentUser for the derivation rule.
func (c *STSSessionClaims) WithParentUser(parentUser string) *STSSessionClaims {
c.ParentUser = parentUser
return c
}