From: Rafael Sadowski Subject: Re: smtpd: validate the UTF-8-iness of AUTH PLAIN To: Omar Polo Cc: tech@openbsd.org Date: Thu, 12 Nov 2026 07:31:48 +0100 On Mon Nov 09, 2026 at 03:54:08PM +0100, Omar Polo wrote: > rfc4616 (section 2 in particular) [0] describes the AUTH PLAIN method, and > requires username and password to be UTF-8 encoded strings of unicode > characters. I can't see many advantages in doing that for passwords (apart from the fact that non-compliant RFC4616 clients might encounter issues), but I'd do it anyway! > > [0]: https://www.rfc-editor.org/info/rfc4616/#section-2 > > At the moment we're not really caring, but some of this data can be > passed not only to auth_userokay(), but to external programs (via > tables) as well. So I think the extra validation make sense. > > I didn't find something upon a quick grep in src/ to validate utf8 so I > *didn't* wrote my own, but rather re-used another one that I've > previously used across several projects. it's small, compact, and built > on a state machine so it's nice =) > > we could also make use of it in other parts of smtpd in the future is my > hope. relayd/httpd could need something like that too. > > thoughs? > > > diff /home/op/w/smtpd > path + /home/op/w/smtpd > commit - cf11025edfb764b317335d997f83f8ef4bee792d > blob - a4923ed2706b368d48fec3e7e77fcf8e3f50d824 > file + smtp_session.c > --- smtp_session.c > +++ smtp_session.c > @@ -32,6 +32,7 @@ > #include "smtpd.h" > #include "log.h" > #include "rfc5322.h" > +#include "utf8d.h" > > #define SMTP_LINE_MAX 65535 > #define DATA_HIWAT 65535 > @@ -1912,6 +1913,18 @@ smtp_proceed_rollback(struct smtp_session *s, const ch > smtp_enter_state(s, STATE_HELO); > } > > +static int > +is_valid_utf8(const char *s) > +{ > + uint32_t cp, state = UTF8_ACCEPT; > + > + for (; *s; ++s) { > + if (utf8_decode(&state, &cp, (unsigned char)*s) == UTF8_REJECT) > + return 0; > + } > + return (state == UTF8_ACCEPT); > +} > + > static void > smtp_rfc4954_auth_plain(struct smtp_session *s, char *arg) > { > @@ -1956,6 +1969,9 @@ smtp_rfc4954_auth_plain(struct smtp_session *s, char * > if (pass[strcspn(pass, "\r\n")] != '\0') > goto abort; > > + if (!is_valid_utf8(user) || !is_valid_utf8(pass)) > + goto abort; > + > m_create(p_lka, IMSG_SMTP_AUTHENTICATE, 0, 0, -1); > m_add_id(p_lka, s->id); > m_add_string(p_lka, s->listener->authtable); > commit - cf11025edfb764b317335d997f83f8ef4bee792d > blob - /dev/null > file + utf8d.h (mode 644) > --- /dev/null > +++ utf8d.h > @@ -0,0 +1,55 @@ > +/* > + * Copyright (c) 2008-2009 Bjoern Hoehrmann > + * > + * Permission is hereby granted, free of charge, to any person obtaining a > + * copy of this software and associated documentation files (the "Software"), > + * to deal in the Software without restriction, including without limitation > + * the rights to use, copy, modify, merge, publish, distribute, sublicense, > + * and/or sell copies of the Software, and to permit persons to whom the > + * Software is furnished to do so, subject to the following conditions: > + * > + * The above copyright notice and this permission notice shall be included in > + * all copies or substantial portions of the Software. > + * > + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR > + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, > + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL > + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER > + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING > + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER > + * DEALINGS IN THE SOFTWARE. > + */ > + > +// See http://bjoern.hoehrmann.de/utf-8/decoder/dfa/ for details. > + > +#define UTF8_ACCEPT 0 > +#define UTF8_REJECT 1 > + > +static const uint8_t utf8d[] = { > + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 00..1f > + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 20..3f > + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 40..5f > + 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, // 60..7f > + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9,9, // 80..9f > + 7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7, // a0..bf > + 8,8,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2, // c0..df > + 0xa,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x3,0x4,0x3,0x3, // e0..ef > + 0xb,0x6,0x6,0x6,0x5,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8,0x8, // f0..ff > + 0x0,0x1,0x2,0x3,0x5,0x8,0x7,0x1,0x1,0x1,0x4,0x6,0x1,0x1,0x1,0x1, // s0..s0 > + 1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0,1,1,1,1,1,0,1,0,1,1,1,1,1,1, // s1..s2 > + 1,2,1,1,1,1,1,2,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,2,1,1,1,1,1,1,1,1, // s3..s4 > + 1,2,1,1,1,1,1,1,1,2,1,1,1,1,1,1,1,1,1,1,1,1,1,3,1,3,1,1,1,1,1,1, // s5..s6 > + 1,3,1,1,1,1,1,3,1,3,1,1,1,1,1,1,1,3,1,1,1,1,1,1,1,1,1,1,1,1,1,1, // s7..s8 > +}; > + > +static uint32_t inline > +utf8_decode(uint32_t* state, uint32_t* codep, uint32_t byte) { > + uint32_t type = utf8d[byte]; byte is used as an index in utf8d, so any value > 255 reads out of bounds? I think byte should be uint8_t and you can save the cast in the caller. > + > + *codep = (*state != UTF8_ACCEPT) ? > + (byte & 0x3fu) | (*codep << 6) : > + (0xff >> type) & (byte); > + > + *state = utf8d[256 + *state*16 + type]; > + return *state; > +} > > >