<div dir="auto">Thanks reviewing!</div><div dir="auto"><br></div><div dir="auto">I agree with your early return strategy and I would reconsider that condition below.</div><div dir="auto"><br></div><div dir="auto"><div dir="auto"># HG changeset patch</div><div dir="auto"># User Yugo Horie <<a href="mailto:u5.horie@gmail.com">u5.horie@gmail.com</a>></div><div dir="auto"># Date 1677107390 -32400</div><div dir="auto"># Thu Feb 23 08:09:50 2023 +0900</div><div dir="auto"># Node ID a3ca45d39fcfd32ca92a6bd25ec18b6359b90f1a</div><div dir="auto"># Parent f4653576ffcd286bed7229e18ee30ec3c713b4de</div><div dir="auto">Core: restrict the rule of utf-8 decode.</div><div dir="auto"><br></div><div dir="auto">The first byte being above 0xf8 which is referred to 5byte</div><div dir="auto">over length older utf-8 becomes invalid.</div><div dir="auto">Even the range of the first byte from 0xf5 to</div><div dir="auto">0xf7 is valid in the term of the codepoint decoding.</div><div dir="auto">See <a href="https://datatracker.ietf.org/doc/html/rfc3629#section-4">https://datatracker.ietf.org/doc/html/rfc3629#section-4</a>.</div><div dir="auto"><br></div><div dir="auto">diff -r f4653576ffcd -r a3ca45d39fcf src/core/ngx_string.c</div><div dir="auto">--- a/src/core/ngx_string.c Thu Feb 23 07:56:44 2023 +0900</div><div dir="auto">+++ b/src/core/ngx_string.c Thu Feb 23 08:09:50 2023 +0900</div><div dir="auto">@@ -1363,8 +1363,12 @@</div><div dir="auto"> uint32_t u, i, valid;</div><div dir="auto"> </div><div dir="auto"> u = **p;</div><div dir="auto">-</div><div dir="auto">- if (u >= 0xf0) {</div><div dir="auto">+ if (u >= 0xf8) {</div><div dir="auto">+</div><div dir="auto">+ (*p)++;</div><div dir="auto">+ return 0xffffffff;</div><div dir="auto">+</div><div dir="auto">+ } else if (u >= 0xf0) {</div><div dir="auto"> </div><div dir="auto"> u &= 0x07;</div><div dir="auto"> valid = 0xffff;</div></div><div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Thu, Feb 23, 2023 at 6:41 Maxim Dounin <<a href="mailto:mdounin@mdounin.ru">mdounin@mdounin.ru</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-style:solid;padding-left:1ex;border-left-color:rgb(204,204,204)">Hello!<br>
<br>
On Thu, Feb 23, 2023 at 12:10:27AM +0900, u5h wrote:<br>
<br>
> # HG changeset patch<br>
> # User Yugo Horie <<a href="mailto:u5.horie@gmail.com" target="_blank">u5.horie@gmail.com</a>><br>
> # Date 1677077775 -32400<br>
> # Wed Feb 22 23:56:15 2023 +0900<br>
> # Node ID 1a9487706c6af90baf2ed770db29f689c3850721<br>
> # Parent cffaf3f2eec8fd33605c2a37814f5ffc30371989<br>
> core: return error when the first byte is above 0xf5 in utf-8<br>
> <br>
> * see <a href="https://datatracker.ietf.org/doc/html/rfc3629#section-4" rel="noreferrer" target="_blank">https://datatracker.ietf.org/doc/html/rfc3629#section-4</a><br>
> <br>
<br>
Thanks for catching, this needs to be addressed.<br>
<br>
The existing code accepts invalid UTF-8 sequences with the first <br>
byte being 11111xxx (>= 0xf8), and interprets them as if the first <br>
byte was 11110xxx (as a 4-byte sequence). This is going to <br>
misinterpret future UTF-8 if it will be ever expanded to 5-byte <br>
sequences (unlikely though), and might cause other problems <br>
(including security ones, as the RFC rightfully warns, though also <br>
unlikely in this particular case).<br>
<br>
A more verbose commit log might be beneficial here, as well as <br>
the one matching nginx style for commit logs. See <br>
<a href="http://nginx.org/en/docs/contributing_changes.html" rel="noreferrer" target="_blank">http://nginx.org/en/docs/contributing_changes.html</a> for some basic <br>
tips.<br>
<br>
> diff -r cffaf3f2eec8 -r 1a9487706c6a src/core/ngx_string.c<br>
> --- a/src/core/ngx_string.c Thu Feb 02 23:38:48 2023 +0300<br>
> +++ b/src/core/ngx_string.c Wed Feb 22 23:56:15 2023 +0900<br>
> @@ -1364,7 +1364,7 @@<br>
> <br>
> u = **p;<br>
> <br>
> - if (u >= 0xf0) {<br>
> + if (u < 0xf5 && u >= 0xf0) {<br>
> <br>
> u &= 0x07;<br>
> valid = 0xffff;<br>
<br>
The suggested change doesn't look correct to me.<br>
<br>
In particular, the "u == 0xf5" (as well as 0xf6 and 0xf7) case is <br>
valid in terms of the codepoint decoding, though resulting value <br>
will be in the invalid range as per the comment above the <br>
function. So it does not really need any special handling, except <br>
may be to simplify things. The interesting part here is "u >= 0xf8" <br>
(which is complimentary to "u &= 0x07" being used in "if").<br>
<br>
Further, as the condition is written, "u >= 0xf5" is now excluded <br>
from the first "if", but will be matched by the "if (u >= 0xe0)" <br>
below, and misinterpreted as a 3-byte sequence. This looks worse <br>
than the existing behaviour.<br>
<br>
I would rather consider something like:<br>
<br>
diff --git a/src/core/ngx_string.c b/src/core/ngx_string.c<br>
--- a/src/core/ngx_string.c<br>
+++ b/src/core/ngx_string.c<br>
@@ -1364,7 +1364,12 @@ ngx_utf8_decode(u_char **p, size_t n)<br>
<br>
u = **p;<br>
<br>
- if (u >= 0xf0) {<br>
+ if (u >= 0xf8) {<br>
+<br>
+ (*p)++;<br>
+ return 0xffffffff;<br>
+<br>
+ } else if (u >= 0xf0) {<br>
<br>
u &= 0x07;<br>
valid = 0xffff;<br>
<br>
<br>
-- <br>
Maxim Dounin<br>
<a href="http://mdounin.ru/" rel="noreferrer" target="_blank">http://mdounin.ru/</a><br>
_______________________________________________<br>
nginx-devel mailing list<br>
<a href="mailto:nginx-devel@nginx.org" target="_blank">nginx-devel@nginx.org</a><br>
<a href="https://mailman.nginx.org/mailman/listinfo/nginx-devel" rel="noreferrer" target="_blank">https://mailman.nginx.org/mailman/listinfo/nginx-devel</a><br>
</blockquote></div></div>