1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
|
This patch comes from upstream. It corresponds to a patch applied to
the generated C source code for llhttp included in Node.js 14.16.0
(see commit 641f786bb1a1f6eb1ff8750782ed939780f2b31a). That commit
fixes CVE-2020-8287. With this patch, the output of our
llhttp-bootstrap package matches the files included in Node.js 14.16.0
exactly.
commit e9b36ea64709c35ca66094d5cf3787f444029601
Author: Fedor Indutny <fedor@indutny.com>
Date: Sat Oct 10 19:56:01 2020 -0700
http: unset `F_CHUNKED` on new `Transfer-Encoding`
Duplicate `Transfer-Encoding` header should be a treated as a single,
but with original header values concatenated with a comma separator. In
the light of this, even if the past `Transfer-Encoding` ended with
`chunked`, we should be not let the `F_CHUNKED` to leak into the next
header, because mere presence of another header indicates that `chunked`
is not the last transfer-encoding token.
diff --git a/src/llhttp/http.ts b/src/llhttp/http.ts
index f4f1a6e..0a0c365 100644
--- a/src/llhttp/http.ts
+++ b/src/llhttp/http.ts
@@ -460,11 +460,19 @@ export class HTTP {
.match([ ' ', '\t' ], n('header_value_discard_ws'))
.otherwise(checkContentLengthEmptiness);
+ // Multiple `Transfer-Encoding` headers should be treated as one, but with
+ // values separate by a comma.
+ //
+ // See: https://tools.ietf.org/html/rfc7230#section-3.2.2
+ const toTransferEncoding = this.unsetFlag(
+ FLAGS.CHUNKED,
+ 'header_value_te_chunked');
+
n('header_value_start')
.otherwise(this.load('header_state', {
[HEADER_STATE.UPGRADE]: this.setFlag(FLAGS.UPGRADE, fallback),
[HEADER_STATE.TRANSFER_ENCODING]: this.setFlag(
- FLAGS.TRANSFER_ENCODING, 'header_value_te_chunked'),
+ FLAGS.TRANSFER_ENCODING, toTransferEncoding),
[HEADER_STATE.CONTENT_LENGTH]: n('header_value_content_length_once'),
[HEADER_STATE.CONNECTION]: n('header_value_connection'),
}, 'header_value'));
@@ -847,6 +855,11 @@ export class HTTP {
return span.start(span.end(this.node(next)));
}
+ private unsetFlag(flag: FLAGS, next: string | Node): Node {
+ const p = this.llparse;
+ return p.invoke(p.code.and('flags', ~flag), this.node(next));
+ }
+
private setFlag(flag: FLAGS, next: string | Node): Node {
const p = this.llparse;
return p.invoke(p.code.or('flags', flag), this.node(next));
diff --git a/test/request/transfer-encoding.md b/test/request/transfer-encoding.md
index a7d1681..b0891d6 100644
--- a/test/request/transfer-encoding.md
+++ b/test/request/transfer-encoding.md
@@ -353,6 +353,38 @@ off=106 headers complete method=3 v=1/1 flags=200 content_length=0
off=106 error code=15 reason="Request has invalid `Transfer-Encoding`"
```
+## POST with `chunked` and duplicate transfer-encoding
+
+<!-- meta={"type": "request", "noScan": true} -->
+```http
+POST /post_identity_body_world?q=search#hey HTTP/1.1
+Accept: */*
+Transfer-Encoding: chunked
+Transfer-Encoding: deflate
+
+World
+```
+
+```log
+off=0 message begin
+off=5 len=38 span[url]="/post_identity_body_world?q=search#hey"
+off=44 url complete
+off=54 len=6 span[header_field]="Accept"
+off=61 header_field complete
+off=62 len=3 span[header_value]="*/*"
+off=67 header_value complete
+off=67 len=17 span[header_field]="Transfer-Encoding"
+off=85 header_field complete
+off=86 len=7 span[header_value]="chunked"
+off=95 header_value complete
+off=95 len=17 span[header_field]="Transfer-Encoding"
+off=113 header_field complete
+off=114 len=7 span[header_value]="deflate"
+off=123 header_value complete
+off=125 headers complete method=3 v=1/1 flags=200 content_length=0
+off=125 error code=15 reason="Request has invalid `Transfer-Encoding`"
+```
+
## POST with `chunked` before other transfer-coding (lenient)
TODO(indutny): should we allow it even in lenient mode? (Consider disabling
|