mirror of
https://github.com/coder/coder.git
synced 2026-10-31 11:02:50 +08:00
## Summary `httpapi.Read` decoded request bodies with no size limit, so a single request could allocate memory without bound. This adds a 4 MiB default ceiling, leaves the endpoints that legitimately need more explicitly exempted, and counts the rejections so a limit set too tight is visible. This is the first of three PRs split out of #28048, covering the endpoints that answer in `codersdk.Response` shape. The OAuth2 decode paths (RFC 6749, RFC 7591) and the SCIM ones (RFC 7644) answer in their own error shapes and follow in separate PRs, along with the lint rule that pins the invariant. Closes PLAT-463. Remediates SEC-416 (CWE-770, CVSS 7.5) and SEC-392. ## Problem `httpapi.Read` calls `json.NewDecoder(r.Body).Decode(value)` with no ceiling, and no middleware in the chain bounds body size. The exposure is pre-authentication: login, OTP, and first-user creation all read a body before any authorization decision is reached. The existing rate limiter bounds request *rate*, which is orthogonal to the memory a single admitted request may consume. ## Fix `Read` is split into `Read` and `ReadLimit`. `ReadLimit` wraps `r.Body` in an `http.MaxBytesReader` and keeps the existing decode and validate logic; `Read` delegates to it with a new `DefaultMaxRequestBodyBytes` of 4 MiB, which covers the 124 remaining non-test callers at a single site. `http.MaxBytesReader` composes as tightest-wins, so the handlers that pre-wrapped their own bodies pass their limit to `ReadLimit` rather than wrapping, and each keeps its previous ceiling byte for byte. That matters most for the bulk secrets import at `8 * MaxSecretsFileBytes`: an unconditional wrap inside `Read` would have silently halved it to the default. `TestImportUserSecretsBodyLargerThanDefaultLimit` is the regression guard for that specific failure, and `TestMaxBytesReaderNesting` pins the composition behavior the whole requirement rests on. Every rejection site calls `httpapi.RecordRequestBodyLimit`, which names the limit that tripped on the request's existing log line and marks the request so `coderd_api_requests_too_large_total{reason="request_body"}` counts body rejections apart from the 413s coderd answers for other causes, such as agent log storage overflow. A limit set too tight for a legitimate payload therefore surfaces without waiting for a user report. The limit is a constant rather than a deployment option: an operator raising it to unblock something would reopen the vulnerability as configuration, where a security scan will not find it. A legitimate 413 is answered with a targeted `ReadLimit` on that endpoint. ## Behavior change `POST /api/v2/files` now answers 413 rather than 400 when a request body exceeds `HTTPFileMaxBytes`. It installed that bound already but reported the rejection as a read failure, which leaked the stdlib `http: request body too large` string through `Detail` and kept the largest limit in the tree off the metric. The separate 413 for an oversized expanded archive is unchanged. The task log snapshot endpoint now answers 413 rather than 400 when its 64 KiB cap is exceeded. Routing it through `ReadLimit` also changes its decode-failure message from "Failed to decode request payload." to "Request body must be valid JSON.", which is what every other endpoint answers. Its tests are updated to match both. `coderd_api_requests_too_large_total` is new, so there is no existing query to migrate. It counts the 413s coderd answers, labeled `method`, `path`, and `reason`. `reason="request_body"` is a rejection by one of the limits above; `reason="other"` is a 413 that has nothing to do with body size, such as agent log storage overflow. ## Reading this The commits are ordered to be read in sequence. Commits 1 and 2 are the security fix; commits 3 to 5 are the observability consequences, and commit 3 is the one that touches dashboards. Commit 7 documents the limit on the REST API reference index. Commits 6 and 8 add and revert an exhaustive `@Failure 413` annotation pass, which buried the fix under its regenerated swagger, and cancel out.
394 lines
13 KiB
Go
394 lines
13 KiB
Go
package agentfiles
|
|
|
|
import (
|
|
"archive/tar"
|
|
"context"
|
|
"encoding/json"
|
|
"errors"
|
|
"io"
|
|
"io/fs"
|
|
"net/http"
|
|
"os"
|
|
"path/filepath"
|
|
"slices"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/bmatcuk/doublestar/v4"
|
|
"golang.org/x/xerrors"
|
|
|
|
"cdr.dev/slog/v3"
|
|
"github.com/coder/coder/v2/agent/agentcontext"
|
|
"github.com/coder/coder/v2/coderd/httpapi"
|
|
"github.com/coder/coder/v2/codersdk/workspacesdk"
|
|
)
|
|
|
|
const (
|
|
bundleFilesRequestMaxBytes = 64 * 1024
|
|
// bundleFilesWriteTimeout gives slow links well over the server's 20s
|
|
// WriteTimeout to stream the archive.
|
|
bundleFilesWriteTimeout = 5 * time.Minute
|
|
|
|
tarBlockSize = 512
|
|
)
|
|
|
|
// defaultBundleFilesLimits caps a single collection. Tar headers, block
|
|
// padding, and manifest file entries are charged against MaxTotalBytes,
|
|
// so it approximately bounds the response size.
|
|
var defaultBundleFilesLimits = workspacesdk.BundleFilesLimits{
|
|
MaxFiles: 10000,
|
|
MaxBytesPerFile: 10 * 1024 * 1024,
|
|
MaxTotalBytes: 100 * 1024 * 1024,
|
|
}
|
|
|
|
var errBundleFilesFileLimit = xerrors.New("bundle files file count limit reached")
|
|
|
|
// HandleBundleFiles streams a tar archive of the requested workspace
|
|
// files. Environment variables in paths are expanded in the agent's
|
|
// environment; paths must then be absolute or start with ~/, which
|
|
// resolves against the agent user's home directory.
|
|
func (api *API) HandleBundleFiles(w http.ResponseWriter, r *http.Request) {
|
|
var req workspacesdk.BundleFilesRequest
|
|
if !httpapi.ReadLimit(r.Context(), w, r, bundleFilesRequestMaxBytes, &req) {
|
|
return
|
|
}
|
|
|
|
home, err := api.envInfo.HomeDir()
|
|
if err != nil {
|
|
api.logger.Error(r.Context(), "get user home dir", slog.Error(err))
|
|
httpapi.InternalServerError(w, xerrors.Errorf("get user home dir: %w", err))
|
|
return
|
|
}
|
|
|
|
if err := http.NewResponseController(w).SetWriteDeadline(time.Now().Add(bundleFilesWriteTimeout)); err != nil {
|
|
api.logger.Warn(r.Context(), "extend bundle files write deadline", slog.Error(err))
|
|
}
|
|
|
|
clientCtx := r.Context()
|
|
ctx, cancel := context.WithTimeout(clientCtx, bundleFilesWriteTimeout)
|
|
defer cancel()
|
|
|
|
w.Header().Set("Content-Type", "application/x-tar")
|
|
w.WriteHeader(http.StatusOK)
|
|
if err := collectBundleFiles(ctx, clientCtx, home, req, w, api.bundleFilesLimits); err != nil {
|
|
api.logger.Error(clientCtx, "collect bundle files", slog.Error(err))
|
|
}
|
|
}
|
|
|
|
// collectBundleFiles streams a tar with the requested files under files/
|
|
// and a manifest.json describing the collection. Per-path problems are
|
|
// recorded in the manifest, not fatal. ctx bounds the collection;
|
|
// clientCtx is the request context.
|
|
func collectBundleFiles(ctx, clientCtx context.Context, home string, req workspacesdk.BundleFilesRequest, w io.Writer, limits workspacesdk.BundleFilesLimits) error {
|
|
manifest := workspacesdk.BundleFilesManifest{Requested: req.Paths, Limits: limits}
|
|
paths := req.Paths
|
|
|
|
home, err := filepath.Abs(home)
|
|
if err != nil {
|
|
// Collect nothing; the archive still carries the manifest.
|
|
appendManifestError(&manifest, "", "", "resolve home directory: "+err.Error())
|
|
paths = nil
|
|
}
|
|
|
|
tw := tar.NewWriter(w)
|
|
c := &bundleFilesCollector{
|
|
tw: tw,
|
|
clientCtx: clientCtx,
|
|
home: home,
|
|
limits: limits,
|
|
manifest: &manifest,
|
|
seenPaths: map[string]struct{}{},
|
|
remainingBytes: limits.MaxTotalBytes,
|
|
}
|
|
for _, requested := range paths {
|
|
if !c.collectPattern(ctx, requested) {
|
|
break
|
|
}
|
|
}
|
|
if clientCtx.Err() != nil {
|
|
// The client is gone; there is nobody to receive the manifest.
|
|
return xerrors.Errorf("client disconnected: %w", clientCtx.Err())
|
|
}
|
|
|
|
manifestJSON, err := json.MarshalIndent(manifest, "", " ")
|
|
if err != nil {
|
|
return xerrors.Errorf("marshal manifest: %w", err)
|
|
}
|
|
err = tw.WriteHeader(&tar.Header{
|
|
Name: "manifest.json",
|
|
Mode: 0o644,
|
|
Size: int64(len(manifestJSON)),
|
|
ModTime: time.Now(),
|
|
})
|
|
if err != nil {
|
|
return xerrors.Errorf("create manifest in archive: %w", err)
|
|
}
|
|
if _, err := tw.Write(manifestJSON); err != nil {
|
|
return xerrors.Errorf("write manifest: %w", err)
|
|
}
|
|
if err := tw.Close(); err != nil {
|
|
return xerrors.Errorf("close archive: %w", err)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// bundleFilesCollector streams matched files into an archive while
|
|
// enforcing limits and recording per-path problems in the manifest.
|
|
// Collect methods return false once a global limit ends collection.
|
|
type bundleFilesCollector struct {
|
|
tw *tar.Writer
|
|
clientCtx context.Context
|
|
home string
|
|
limits workspacesdk.BundleFilesLimits
|
|
manifest *workspacesdk.BundleFilesManifest
|
|
seenPaths map[string]struct{}
|
|
remainingBytes int64
|
|
filesWritten int
|
|
}
|
|
|
|
func (c *bundleFilesCollector) collectPattern(ctx context.Context, requested string) bool {
|
|
if ctx.Err() != nil {
|
|
return c.stopCanceled(requested, "")
|
|
}
|
|
if c.filesWritten >= c.limits.MaxFiles {
|
|
return c.stop(requested, "", "file count limit reached")
|
|
}
|
|
|
|
matches, matchesTruncated, err := bundleFileMatches(ctx, c.home, requested, c.limits.MaxFiles-c.filesWritten)
|
|
if err != nil {
|
|
if ctx.Err() != nil {
|
|
return c.stopCanceled(requested, "")
|
|
}
|
|
appendManifestError(c.manifest, requested, "", err.Error())
|
|
return true
|
|
}
|
|
if len(matches) == 0 {
|
|
appendManifestError(c.manifest, requested, "", "no matches")
|
|
return true
|
|
}
|
|
if matchesTruncated {
|
|
c.manifest.Truncated = true
|
|
appendManifestError(c.manifest, requested, "", "file count limit reached")
|
|
}
|
|
|
|
for _, abs := range matches {
|
|
if !c.collectFile(ctx, requested, abs) {
|
|
return false
|
|
}
|
|
}
|
|
return true
|
|
}
|
|
|
|
func (c *bundleFilesCollector) collectFile(ctx context.Context, requested string, abs string) bool {
|
|
if ctx.Err() != nil {
|
|
return c.stopCanceled(requested, abs)
|
|
}
|
|
if c.filesWritten >= c.limits.MaxFiles {
|
|
return c.stop(requested, abs, "file count limit reached")
|
|
}
|
|
// Each entry costs a tar header block before any data fits.
|
|
if c.remainingBytes <= tarBlockSize {
|
|
return c.stop(requested, abs, "total byte limit reached")
|
|
}
|
|
if _, ok := c.seenPaths[abs]; ok {
|
|
return true
|
|
}
|
|
c.seenPaths[abs] = struct{}{}
|
|
|
|
// Stat before open: opening a FIFO would block. Stat follows symlinks,
|
|
// so a directly requested symlink collects its target.
|
|
info, err := os.Stat(abs)
|
|
if err != nil {
|
|
reason := "stat path: " + err.Error()
|
|
if errors.Is(err, fs.ErrNotExist) {
|
|
reason = "does not exist"
|
|
}
|
|
appendManifestError(c.manifest, requested, abs, reason)
|
|
return true
|
|
}
|
|
if !info.Mode().IsRegular() {
|
|
appendManifestError(c.manifest, requested, abs, "not a regular file: "+fileModeTypeName(info.Mode()))
|
|
return true
|
|
}
|
|
|
|
bytesToWrite := min(info.Size(), c.limits.MaxBytesPerFile, c.remainingBytes-tarBlockSize)
|
|
entry := workspacesdk.BundleFilesManifestEntry{
|
|
Requested: requested,
|
|
Path: abs,
|
|
ArchivePath: BundleFilesArchivePath(abs),
|
|
Size: info.Size(),
|
|
ModTime: info.ModTime(),
|
|
BytesWritten: bytesToWrite,
|
|
Truncated: bytesToWrite < info.Size(),
|
|
}
|
|
c.manifest.Truncated = c.manifest.Truncated || entry.Truncated
|
|
if err := writeBundleFileEntry(c.tw, abs, entry); err != nil {
|
|
appendManifestError(c.manifest, requested, abs, err.Error())
|
|
return true
|
|
}
|
|
c.manifest.Files = append(c.manifest.Files, entry)
|
|
// The last file may overshoot the budget by under a block; the bound
|
|
// is approximate, not exact.
|
|
entryJSON, _ := json.Marshal(entry)
|
|
c.remainingBytes -= tarEntrySize(bytesToWrite) + int64(len(entryJSON))
|
|
c.filesWritten++
|
|
return true
|
|
}
|
|
|
|
// stop marks the manifest truncated, records the reason, and halts
|
|
// collection.
|
|
func (c *bundleFilesCollector) stop(requested string, filePath string, reason string) bool {
|
|
c.manifest.Truncated = true
|
|
appendManifestError(c.manifest, requested, filePath, reason)
|
|
return false
|
|
}
|
|
|
|
// stopCanceled halts collection after the collection context ended: a
|
|
// timeout is recorded in the manifest and the archive is finished, while a
|
|
// client disconnect makes the caller abort without a manifest.
|
|
func (c *bundleFilesCollector) stopCanceled(requested string, filePath string) bool {
|
|
if c.clientCtx.Err() != nil {
|
|
return false
|
|
}
|
|
return c.stop(requested, filePath, "exceeded maximum collection time")
|
|
}
|
|
|
|
// bundleFileMatches expands requested against home and returns matching
|
|
// cleaned absolute paths. Non-glob paths return a single candidate without
|
|
// checking existence; the caller reports missing files on stat.
|
|
func bundleFileMatches(ctx context.Context, home string, requested string, maxMatches int) ([]string, bool, error) {
|
|
// Env vars expand from the agent environment and ~ resolves against
|
|
// the agent home, matching the agent's expandPathToAbs. Glob patterns
|
|
// never exist on disk, so canonicalization keeps them lexical.
|
|
abs, err := agentcontext.CanonicalizePathIn(home, os.ExpandEnv(requested))
|
|
if err != nil {
|
|
return nil, false, err
|
|
}
|
|
if !strings.ContainsAny(abs, "*?{[") {
|
|
return []string{abs}, false, nil
|
|
}
|
|
|
|
base, pattern := doublestar.SplitPattern(filepath.ToSlash(abs))
|
|
matches := make([]string, 0, min(maxMatches, 64))
|
|
// WithNoFollow avoids symlink cycles. Checking the limit before the
|
|
// append keeps matches from growing past maxMatches.
|
|
err = doublestar.GlobWalk(bundleFilesFS{ctx: ctx, fsys: os.DirFS(base)}, pattern, func(match string, _ fs.DirEntry) error {
|
|
if len(matches) >= maxMatches {
|
|
return errBundleFilesFileLimit
|
|
}
|
|
matches = append(matches, filepath.Join(base, filepath.FromSlash(match)))
|
|
return nil
|
|
}, doublestar.WithFilesOnly(), doublestar.WithNoFollow())
|
|
matchesTruncated := errors.Is(err, errBundleFilesFileLimit)
|
|
if err != nil && !matchesTruncated {
|
|
return nil, false, xerrors.Errorf("glob pattern: %w", err)
|
|
}
|
|
// doublestar does not guarantee ordering, so sort for a deterministic
|
|
// archive.
|
|
slices.Sort(matches)
|
|
return matches, matchesTruncated, nil
|
|
}
|
|
|
|
// bundleFilesFS cancels a glob walk once the request context ends. Only
|
|
// Open is implemented; the fs.ReadDir and fs.Stat helpers fall back to it,
|
|
// so every filesystem operation of the walk passes the context check.
|
|
type bundleFilesFS struct {
|
|
ctx context.Context
|
|
fsys fs.FS
|
|
}
|
|
|
|
func (f bundleFilesFS) Open(name string) (fs.File, error) {
|
|
if err := f.ctx.Err(); err != nil {
|
|
return nil, err
|
|
}
|
|
return f.fsys.Open(name)
|
|
}
|
|
|
|
// BundleFilesArchivePath maps a cleaned absolute path to its archive entry
|
|
// name: files/ plus the path with the leading separator trimmed and any
|
|
// Windows drive colon dropped, keeping the name fs.ValidPath-safe.
|
|
func BundleFilesArchivePath(abs string) string {
|
|
p := strings.TrimPrefix(filepath.ToSlash(abs), "/")
|
|
if len(p) >= 2 && p[1] == ':' {
|
|
p = p[:1] + p[2:]
|
|
}
|
|
return "files/" + p
|
|
}
|
|
|
|
// fileModeTypeName names the type of a non-regular file.
|
|
func fileModeTypeName(mode fs.FileMode) string {
|
|
switch {
|
|
case mode.IsDir():
|
|
return "directory"
|
|
case mode&fs.ModeSymlink != 0:
|
|
return "symlink"
|
|
case mode&fs.ModeNamedPipe != 0:
|
|
return "named pipe"
|
|
case mode&fs.ModeSocket != 0:
|
|
return "socket"
|
|
case mode&fs.ModeDevice != 0, mode&fs.ModeCharDevice != 0:
|
|
return "device"
|
|
default:
|
|
return "irregular file"
|
|
}
|
|
}
|
|
|
|
// writeBundleFileEntry writes the last entry.BytesWritten bytes of the
|
|
// file at abs to the archive at entry.ArchivePath. A file that shrinks
|
|
// after stat is zero-padded to the declared size, since a short entry
|
|
// would corrupt every entry after it; the short read is still an error.
|
|
func writeBundleFileEntry(tw *tar.Writer, abs string, entry workspacesdk.BundleFilesManifestEntry) error {
|
|
f, err := os.Open(abs)
|
|
if err != nil {
|
|
return xerrors.Errorf("open file: %w", err)
|
|
}
|
|
defer f.Close()
|
|
|
|
if entry.BytesWritten < entry.Size {
|
|
if _, err := f.Seek(entry.Size-entry.BytesWritten, io.SeekStart); err != nil {
|
|
return xerrors.Errorf("seek tail: %w", err)
|
|
}
|
|
}
|
|
err = tw.WriteHeader(&tar.Header{
|
|
Name: entry.ArchivePath,
|
|
Mode: 0o644,
|
|
Size: entry.BytesWritten,
|
|
ModTime: entry.ModTime,
|
|
})
|
|
if err != nil {
|
|
return xerrors.Errorf("create archive entry: %w", err)
|
|
}
|
|
n, err := io.Copy(tw, io.LimitReader(f, entry.BytesWritten))
|
|
if err == nil && n < entry.BytesWritten {
|
|
err = io.ErrUnexpectedEOF
|
|
}
|
|
if err != nil {
|
|
if _, padErr := io.CopyN(tw, zeroReader{}, entry.BytesWritten-n); padErr != nil {
|
|
return xerrors.Errorf("pad short entry: %w", padErr)
|
|
}
|
|
return xerrors.Errorf("copy file: %w", err)
|
|
}
|
|
return nil
|
|
}
|
|
|
|
// tarEntrySize returns the archive bytes a file entry consumes: one
|
|
// header block plus the data rounded up to whole blocks.
|
|
func tarEntrySize(dataBytes int64) int64 {
|
|
return tarBlockSize + (dataBytes+tarBlockSize-1)/tarBlockSize*tarBlockSize
|
|
}
|
|
|
|
type zeroReader struct{}
|
|
|
|
func (zeroReader) Read(p []byte) (int, error) {
|
|
clear(p)
|
|
return len(p), nil
|
|
}
|
|
|
|
func appendManifestError(m *workspacesdk.BundleFilesManifest, requested string, filePath string, reason string) {
|
|
m.Errors = append(m.Errors, workspacesdk.BundleFilesManifestError{
|
|
Requested: requested,
|
|
Path: filePath,
|
|
Reason: reason,
|
|
})
|
|
}
|