On Tue, Nov 17, 2020 at 03:40:39PM -0300, K R wrote:
> >Synopsis: sftp URI destination can't handle IPv6 addresses
> >Category: system amd64
> >Environment:
> System : OpenBSD 6.8
> Details : OpenBSD 6.8-current (GENERIC) #173: Mon Nov 16
> 22:16:04 MST 2020
> [email protected]:
> /usr/src/sys/arch/amd64/compile/GENERIC
>
> Architecture: OpenBSD.amd64
> Machine : amd64
> >Description:
> sftp(1) won't work if destination is written in URI format:
>
> sftp://[user@]host[:port][/path]
>
> and host is an IPv6 address, enclosed in square brackets.
>
> >How-To-Repeat:
>
> # sftp sftp://user@[::1]:2222/
> usage: sftp [-46AaCfNpqrv] [-B buffer_size] [-b batchfile] [-c cipher]
> [-D sftp_server_path] [-F ssh_config] [-i identity_file]
> [-J destination] [-l limit] [-o ssh_option] [-P port]
> [-R num_requests] [-S program] [-s subsystem | sftp_server]
> destination
>
> # sftp -P 2222 user@[::1]/
>
> works as expected.
i have just looked into this issue and confirmed,
ssh:// scp:// URI formats are affected since those
are using same code path, misc.c:859:parse_uri
i changed the code just for debugging and the errstr
gives the clue:
misc.c:
909 if (!valid_domain(host, 0, &errstr)) {
910 err("%s", errstr); // XXX: debug, fill and give me
errstr please
911 goto out;
912 }
snap01$ ./sftp sftp://[1001:19f0:9002:1b3a:5400:2ff:feb3:7ed3]:22
sftp: domain name "1001:19f0:9002:1b3a:5400:2ff:feb3:7ed3" contains
invalid characters: Undefined error: 0
snap01$ ./sftp sftp://[2001:19f0:9002:1b3a:5400:2ff:feb3:7ed3]:22
sftp: domain name "2001:19f0:9002:1b3a:5400:2ff:feb3:7ed3" contains
invalid characters: Undefined error: 0
looking at valid_domain() now with a clue:
2116 if (c != '.' && c != '-' && !isalnum(c) &&
2117 c != '_') /* technically invalid, but common */ {
2118 snprintf(errbuf, sizeof(errbuf), "domain name "
2119 "\"%.100s\" contains invalid characters",
name);
2120 goto bad;
2121 }
see above, IPv6 char ':' fails,
> >Fix:
Accordingly to
https://tools.ietf.org/html/draft-ietf-secsh-scp-sftp-ssh-uri-04
"host" part of the URI referenced at
https://tools.ietf.org/html/rfc3986#section-3.2.2
it can be an IPv6/IPv4 address or a domain name itself, so the most
obvious solution should be to remove the valid_domain() check from
parse_uri(), but RFC also says that registered names and hexadecimal
addresses should be in lowercase for the sake of uniformity, while
only using uppercase letters for percent-encodings.
Looking for thoughts,
