- dovecot - dovecot.org

[Dovecot] 0.99.10.7-rc1
by Timo Sirainen 13 Jul '04

13 Jul '04

http://dovecot.org/rc/ Yet another 0.99 release. If it works, tomorrow comes the real one: + Added outlook-pop3-no-nuls workaround to fix Outlook hang in mails with NULs. + Config file lines can now contain quoted strings ("value ") - If client didn't finish downloading a single mail in 30 seconds, Dovecot closed the connection. This was supposed to work so that if client hasn't read data at all in 30 seconds, it's disconnected. - Maildir: LIST now doesn't skip symlinks Strange that I heard about the 30sec bug yesterday for the first time. It's been there forever. Guess people have fast internet connections nowadays :)

2 4

[Dovecot] postfix patch does not compile on freebsd 4
by Kevin 13 Jul '04

13 Jul '04

i'm trying to use timo's postfix-2.1.2-dovecot-auth.patch on freebsd4 with postfix-2.1.4 the problem is that smtpd_sasl_glue.c tries to include stdint.h which isn't present with fbsd4. is there another lib that'll provide the same functionality? i've tried using inttypes.h but i don't know of any other similar ones stdint is in fbsd5 apparently but i'm still scared of it have any ideas?

2 1

Re: [Dovecot] ldap auth problems
by Karl Latiss 13 Jul '04

13 Jul '04

That didn't seem to help either. In the end I re-compiled the Fedora Core 2 release from SRC RPM and that seems to be working. I'm not sure what the difference is. On Wed, 2004-07-07 at 18:42, Timo Sirainen wrote: > On Mon, 2004-07-05 at 12:17, Karl Latiss wrote: > > [pid 24674] close(8 <unfinished ...> > > [pid 24674] <... close resumed> ) = 0 > > [pid 24674] --- SIGSEGV (Segmentation fault) @ 0 (0) --- > > Process 24674 detached > .. > > Hmm. Doesn't really help. Getting gdb backtrace would show where it > crashes, but I'm not sure if there's any easy way to do it. > > One would be to put sleep(10); to beginning of src/auth/main.c, then > start dovecot and "gdb /usr/local/libexec/dovecot/dovecot-auth pid" > where pid being the running dovecot-auth's process id. then "c" to > continue, wait for crash and "bt" to get backtrace.

1 0

[Dovecot] mailbox problem
by Jean-Francois Theroux 12 Jul '04

12 Jul '04

Hi guys, I use a virtual mailbox setup with dovecot/postfix. Which reside under /virtual. I get an error message which doesn't return anything interesting in Google. I get this: imap(user(a)domain.com): File isn't in mbox format: /virtual/domain.com/user/imapd/Trash I have no idea on how to fix this. I seem to be able to delete messages though. But I'd like to repair this, since I like to have clean logs ;) Thanks for your time, -- Jean-Francois Theroux Systems administrator 514.726.3732 PrivalODC http://www.privalodc.comp

2 1

[Dovecot] 1.0-test28
by Timo Sirainen 12 Jul '04

12 Jul '04

http://dovecot.org/test/ - Fixed some crashes with cache file. THREAD and SORT especially crashed it. - Fixed the SSL infinite looping in login processes - Recent flags are working again with maildir. Almost with mbox too. - Maildir code moves mail into cur/ directory directly if the mail contains flags, so we don't violate maildir spec anymore by placing them into new/ with the flags. Although.. now that I think about it, that breaks recent flags, as only mails in new/ are treated as being recent.. Well, I guess next version implements my another idea for it: Store flags only into index and keep the file in new/, flags will be added when the file is moved to cur/. Non-Dovecot MUAs don't see the flags in the mean time though. - SIGHUPing dovecot master process finally reloads configuration completely and without crashing - Added --exec-mail parameter to dovecot. It can be used to launch imap/pop3 processes with configuration from dovecot.conf - Changed client_workarounds to imap_client_workarounds and pop3_client_workarounds. Added the first pop3 workaround: outlook-no-nuls

1 0

[Dovecot] Debian woody backport of 0.99.10.6 now available on backports.org
by Jaldhar H. Vyas 10 Jul '04

10 Jul '04

Well, the backports.org people finally installed my backport of the latest stable version of dovecot. Please remove the Braincells apt source form /etc/apt/sources.list and add these lines instead. deb http://www.backports.org/debian stable dovecot openldap2 postgresql openssl deb-src http://www.backports.org/debian stable dovecot (If you are using other backports you can merge dovecot etc. into an existing line.) I have also updated the wiki. If there are any other places which mention the old location, please change them. -- Jaldhar H. Vyas <jaldhar(a)debian.org> La Salle Debain - http://www.braincells.com/debian/

1 1

[Dovecot] NUL characters in POP3 conversation cause Outlook to hang
by Matthew Reimer 10 Jul '04

10 Jul '04

We're being bitten an Outlook bug [1,2] that is triggered by a NUL character in a message being sent through POP3. I see that dovecot's imap implementation converts 0x00 to 0x80, but this conversion is not done for POP3. Would it be possible to add an option to do this mapping for POP3 as well? Matt [1] http://xforce.iss.net/xforce/xfdb/15859 [2] http://www.derkeiler.com/Mailing-Lists/NT-Bugtraq/2004-04/0014.html

2 1

[Dovecot] pop3 usernames
by Vlad Mazek 10 Jul '04

10 Jul '04

Hello, I would like to allow users to login as user(a)domain.tld instead of providing their login name; I'm coming from the sendmail/qpopper land and am trying to migrate to postfix/maildir/dovecot one step at a time and I've been able to find a way to do most of what I want except for the above requirement. Do I have to use a SQL server or is there any way to easilly map an email address to a mailbox or even a maildir? Thank you. -Vlad

2 1

[Dovecot] permission problem ??
by Joachim Dostal 09 Jul '04

09 Jul '04

version dovecot-1.0-test27: Jul 9 21:49:07 server dovecot: IMAP(testtest): mprotect() failed with index file /home/testtest/mail/.imap/INBOX/dovecot.index: Permission denied with version 0.99.10.6 i have no such troubles ... ? tx4hlp, joachim

2 3

[Dovecot] Re : Permission denied
by Frédéric Sapin 09 Jul '04

09 Jul '04

Hello Timo, I'm using linux Fedora Core 2. The permissions are : drwxrwxrwt 2 root mail 32 2004-06-16 04:45 /var/spool/mail/ These permissions are made buy the command chmod a+rwxt /var/spool/mail. So if someone wants to erase the /var/spool/mail directory, it's possible unfortunately. If there are theses permissions the user can receive his mails but it's dangerous I think. And if I modify the dovecot.conf file : mail_extra_groups = mail and if the permissions are the initial permission, and if I enter chmod +t /var/spool/mail the user can't receive his mail. Thanks Frederic

4 6

[Dovecot] bug in openssl proxy code (1.0-test24) ?
by Andrey Panin 09 Jul '04

09 Jul '04

Hi all, If client connects to pop3 server, issues STLS command and then drops connection dovecot starts flooding syslog with this message repeated endlessly: Jul 9 14:31:40 smtp dovecot: pop3-login: SSL_accept() syscall failed: EOF [80.254.111.9] Best regards. -- Andrey Panin | Linux and UNIX system administrator pazke(a)donpac.ru | PGP key: wwwkeys.pgp.net

2 3

[Dovecot] passing remote ip to pam
by dean gaudet 09 Jul '04

09 Jul '04

to improve forensic log info i want to set the PAM_RHOST value to the remote ip (which pam logs as rhost=foo in failure messages). i didn't look to see if anything has been done in this way on CVS because i'm still on 0.99.10.6. below is a bit of a hack. in some sense the remote_ip might make more sense in the AUTH_LOGIN_REQUEST_NEW packet rather than the continue packet... but that looked like i'd have to change more code :) btw -- is there anything which stops mech_plain_auth_continue from doing a somewhat unbounded pstrdup if you send "a\0b" for the auth string? -dean diff -rpu dovecot-0.99.10.6.deborig/debian/changelog dovecot-0.99.10.6/debian/changelog --- dovecot-0.99.10.6.deborig/debian/changelog 2004-07-08 18:11:55.000000000 -0700 +++ dovecot-0.99.10.6/debian/changelog 2004-07-08 18:34:40.000000000 -0700 @@ -1,3 +1,9 @@ +dovecot (0.99.10.6-3.dg1) unstable; urgency=low + + * hack a solution to pass PAM_RHOST the remote ip + + -- dean gaudet <dean(a)arctic.org> Thu, 8 Jul 2004 18:34:26 -0700 + dovecot (0.99.10.6-3) unstable; urgency=low * Patched so dovecot follows symlinks to directories again. diff -rpu dovecot-0.99.10.6.deborig/src/auth/auth-login-interface.h dovecot-0.99.10.6/src/auth/auth-login-interface.h --- dovecot-0.99.10.6.deborig/src/auth/auth-login-interface.h 2003-05-18 05:26:28.000000000 -0700 +++ dovecot-0.99.10.6/src/auth/auth-login-interface.h 2004-07-08 18:32:22.000000000 -0700 @@ -1,6 +1,8 @@ #ifndef __AUTH_LOGIN_INTERFACE_H #define __AUTH_LOGIN_INTERFACE_H +#include "network.h" + /* max. size for auth_login_request_continue.data[] */ #define AUTH_LOGIN_MAX_REQUEST_DATA_SIZE 4096 @@ -57,6 +59,8 @@ struct auth_login_request_continue { enum auth_login_request_type type; /* AUTH_LOGIN_REQUEST_CONTINUE */ unsigned int id; + struct ip_addr remote_ip; + size_t data_size; /* unsigned char data[]; */ }; diff -rpu dovecot-0.99.10.6.deborig/src/auth/mech-plain.c dovecot-0.99.10.6/src/auth/mech-plain.c --- dovecot-0.99.10.6.deborig/src/auth/mech-plain.c 2003-05-18 05:26:28.000000000 -0700 +++ dovecot-0.99.10.6/src/auth/mech-plain.c 2004-07-08 18:25:14.000000000 -0700 @@ -21,6 +21,7 @@ mech_plain_auth_continue(struct auth_req size_t i, count, len; auth_request->callback = callback; + auth_request->remote_ip = request->remote_ip; /* authorization ID \0 authentication ID \0 pass. we'll ignore authorization ID for now. */ diff -rpu dovecot-0.99.10.6.deborig/src/auth/mech.h dovecot-0.99.10.6/src/auth/mech.h --- dovecot-0.99.10.6.deborig/src/auth/mech.h 2003-05-18 05:26:28.000000000 -0700 +++ dovecot-0.99.10.6/src/auth/mech.h 2004-07-08 18:31:42.000000000 -0700 @@ -1,6 +1,7 @@ #ifndef __MECH_H #define __MECH_H +#include "network.h" #include "auth-login-interface.h" struct login_connection; @@ -11,6 +12,7 @@ typedef void mech_callback_t(struct auth struct auth_request { pool_t pool; char *user; + struct ip_addr remote_ip; struct login_connection *conn; unsigned int id; diff -rpu dovecot-0.99.10.6.deborig/src/auth/passdb-pam.c dovecot-0.99.10.6/src/auth/passdb-pam.c --- dovecot-0.99.10.6.deborig/src/auth/passdb-pam.c 2003-11-08 06:17:51.000000000 -0800 +++ dovecot-0.99.10.6/src/auth/passdb-pam.c 2004-07-08 18:32:52.000000000 -0700 @@ -166,13 +166,23 @@ static int pam_userpass_conv(int num_msg return PAM_SUCCESS; } -static int pam_auth(pam_handle_t *pamh, const char *user, const char **error) +static int pam_auth(pam_handle_t *pamh, const char *user, + const struct ip_addr *remote_ip, const char **error) { void *item; int status; + const char *addr; *error = NULL; + if ((addr = net_ip2addr(remote_ip)) + && (status = pam_set_item(pamh, PAM_RHOST, addr)) + != PAM_SUCCESS) { + *error = t_strdup_printf("pam_set_item(PAM_RHOST, %s) failed: %s", + addr, pam_strerror(pamh, status)); + return status; + } + if ((status = pam_authenticate(pamh, 0)) != PAM_SUCCESS) { *error = t_strdup_printf("pam_authenticate(%s) failed: %s", user, pam_strerror(pamh, status)); @@ -205,7 +215,8 @@ static int pam_auth(pam_handle_t *pamh, static void pam_verify_plain_child(const char *service, const char *user, - const char *password, int fd) + const char *password, const struct ip_addr *remote_ip, + int fd) { pam_handle_t *pamh; struct pam_userpass userpass; @@ -228,7 +239,7 @@ pam_verify_plain_child(const char *servi str = t_strdup_printf("pam_start(%s) failed: %s", user, pam_strerror(pamh, status)); } else { - status = pam_auth(pamh, user, &str); + status = pam_auth(pamh, user, remote_ip, &str); if ((status2 = pam_end(pamh, status)) == PAM_SUCCESS) { /* FIXME: check for PASSDB_RESULT_UNKNOWN_USER somehow? */ @@ -353,7 +364,7 @@ pam_verify_plain(struct auth_request *re if (pid == 0) { (void)close(fd[0]); - pam_verify_plain_child(service, request->user, password, fd[1]); + pam_verify_plain_child(service, request->user, password, &request->remote_ip, fd[1]); _exit(0); } diff -rpu dovecot-0.99.10.6.deborig/src/imap-login/client-authenticate.c dovecot-0.99.10.6/src/imap-login/client-authenticate.c --- dovecot-0.99.10.6.deborig/src/imap-login/client-authenticate.c 2003-11-24 11:50:55.000000000 -0800 +++ dovecot-0.99.10.6/src/imap-login/client-authenticate.c 2004-07-08 18:21:19.000000000 -0700 @@ -140,7 +140,7 @@ static void login_callback(struct auth_r case 0: /* continue */ ptr = buffer_get_data(client->plain_login, &size); - auth_continue_request(request, ptr, size); + auth_continue_request(request, ptr, size, &client->common.ip); buffer_set_used_size(client->plain_login, 0); break; @@ -270,7 +270,8 @@ static void client_auth_input(void *cont } else { auth_continue_request(client->common.auth_request, buffer_get_data(buf, NULL), - buffer_get_used_size(buf)); + buffer_get_used_size(buf), + &client->common.ip); } /* clear sensitive data */ diff -rpu dovecot-0.99.10.6.deborig/src/login-common/auth-connection.c dovecot-0.99.10.6/src/login-common/auth-connection.c --- dovecot-0.99.10.6.deborig/src/login-common/auth-connection.c 2003-05-26 08:27:13.000000000 -0700 +++ dovecot-0.99.10.6/src/login-common/auth-connection.c 2004-07-08 18:33:38.000000000 -0700 @@ -335,13 +335,21 @@ int auth_init_request(enum auth_mech mec } void auth_continue_request(struct auth_request *request, - const unsigned char *data, size_t data_size) + const unsigned char *data, size_t data_size, + const struct ip_addr *remote_ip) { struct auth_login_request_continue auth_request; /* send continued request to auth */ auth_request.type = AUTH_LOGIN_REQUEST_CONTINUE; auth_request.id = request->id; + if (remote_ip) { + auth_request.remote_ip = *remote_ip; + } + else { + memset(&auth_request.remote_ip, + 0, sizeof(auth_request.remote_ip)); + } auth_request.data_size = data_size; if (o_stream_send(request->conn->output, &auth_request, diff -rpu dovecot-0.99.10.6.deborig/src/login-common/auth-connection.h dovecot-0.99.10.6/src/login-common/auth-connection.h --- dovecot-0.99.10.6.deborig/src/login-common/auth-connection.h 2003-02-09 23:56:23.000000000 -0800 +++ dovecot-0.99.10.6/src/login-common/auth-connection.h 2004-07-08 18:20:58.000000000 -0700 @@ -48,7 +48,8 @@ int auth_init_request(enum auth_mech mec const char **error); void auth_continue_request(struct auth_request *request, - const unsigned char *data, size_t data_size); + const unsigned char *data, size_t data_size, + const struct ip_addr *remote_ip); void auth_abort_request(struct auth_request *request); diff -rpu dovecot-0.99.10.6.deborig/src/pop3-login/client-authenticate.c dovecot-0.99.10.6/src/pop3-login/client-authenticate.c --- dovecot-0.99.10.6.deborig/src/pop3-login/client-authenticate.c 2003-11-24 11:52:09.000000000 -0800 +++ dovecot-0.99.10.6/src/pop3-login/client-authenticate.c 2004-07-08 18:21:32.000000000 -0700 @@ -141,7 +141,7 @@ static void login_callback(struct auth_r case 0: ptr = buffer_get_data(client->plain_login, &size); - auth_continue_request(request, ptr, size); + auth_continue_request(request, ptr, size, &client->common.ip); buffer_set_used_size(client->plain_login, 0); break; @@ -259,7 +259,8 @@ static void client_auth_input(void *cont } else { auth_continue_request(client->common.auth_request, buffer_get_data(buf, NULL), - buffer_get_used_size(buf)); + buffer_get_used_size(buf), + &client->common.ip); } /* clear sensitive data */

2 1

[Dovecot] Permission denied
by Frédéric Sapin 09 Jul '04

09 Jul '04

Hello all, I'm a new user of dovecot and I've got a permission denied problem. I've read the VarMailDotLock wiki paper but there is still a security problem. If I enter : chmod a+rwxt /var/spool/mail it works but I think it's very dangerous to have thaht rights. So I've tried to edit the dovecot.conf file and to change the mail_extra_groups = mail and to change the rights : chmod +t /var/spool/mail but it doesn't work ! The mails are in /var/spool/mail/unix_user_name Could you please help me to configure the /var/spool/mail to have a good security and to allow people to receive their mails ? Thanks Frederic

2 1

[Dovecot] a yum install with mysql support??
by alex＠alution.com 09 Jul '04

09 Jul '04

Hello, Well im guessing i have already answered this question myself. Im using FC1 and installed dovecot the easy way (yum install dovecot) It has stated that it installed version 0.99.10.5 and from what I have looked at in the mailing list archives, this version SHOULD have mysql authentication support. However what I don´t know is if by installing the yum package it will allow this support. Im guessing that I will have to install it the real way, as if I take a look in the logs, it states... ============== dovecot-auth: "Jul 09 00:37:38 "Fatal: Unknown userdb type 'mysql' dovecot: "Jul 09 00:37:38 "Error: Auth process died too early - shutting down dovecot: "Jul 09 00:37:38 "Error: child 31532 (auth) returned error 89 ============== Am i correct in thinking this? Or is there something else a miss?? Thanks Alex

1 0

Re: [Dovecot] 1.0-test26
by Matthew Reimer 09 Jul '04

09 Jul '04

Timo Sirainen wrote: > On 9.7.2004, at 00:57, Mark E. Mallett wrote: > >> On Fri, Jul 09, 2004 at 12:34:04AM +0300, Timo Sirainen wrote: >> >>> http://dovecot.org/test/ >>> >>> - Fixed APPEND hanging if the whole message was sent in one TCP packet >>> (ie. fixes the "Sent mail" hangs) >> >> >> Curious-- how does the application-level code care about TCP packets? > > > It reads all of it into buffer, then later tries to read more even > though everything was already in that buffer so it gets stuck at > waiting. Dovecot's istream-limit wrapper stream specifically was a > problem. It called parent stream's read() function without checking > first if there was already data in it's buffer. > > These wrapper streams have been quite difficult in general. mbox support > is implemented using one of those too and it took quite a while to get > it working correctly. I guess I should write some comments there how > exactly they should be implemented. Is this a problem in 0.99 as well? Matt

2 1

[Dovecot] 1.0-test26
by Timo Sirainen 09 Jul '04

09 Jul '04

http://dovecot.org/test/ - Fixed APPEND hanging if the whole message was sent in one TCP packet (ie. fixes the "Sent mail" hangs) - Some changes in MySQL defaults - Working APOP auth - Fixes/improvements for cache file code. Still isn't as smart as it could, but it seems to work again. The code is nearing a state where plugins could add their own data into cache file.

2 3

Re: [Dovecot] Logging "issue" with PAM authentication
by Steve Johnson 07 Jul '04

07 Jul '04

Great, thanks a lot, solved the problem. No more annoying logs every 2-5 seconds :) Steve Johnson On Wed, 2004-07-07 at 14:39, Ray Miller wrote: > On Wed, Jul 07, 2004 at 02:21:04PM -0400, Steve Johnson wrote: > > > > I've tried with just the following: > > > > auth sufficient /lib/security/$ISA/pam_ldap.so use_first_pass > > password sufficient /lib/security/$ISA/pam_ldap.so use_authtok > > session optional /lib/security/$ISA/pam_ldap.so > > [...] > > In both of these situations, users can't authenticate. Is there > > something I'm doing wrong? I'm not that familiar with pam, so it's a > > very plausible cause! :) > > I'm fairly sure you don't need the password line, but that's not > likely to be causing your problem - more likely it's the use_first_pass > option. Try: > > auth required /lib/security/$ISA/pam_ldap.so > account required /lib/security/$ISA/pam_ldap.so > password required /lib/security/$ISA/pam_deny.so > session required /lib/security/$ISA/pam_ldap.so

1 0

[Dovecot] Peer Config Review
by Eric Kahklen 07 Jul '04

07 Jul '04

I am trying to setup a mail server using Postfix and Squirrelmail. The mail server will ONLY use webmail for access. I will not be using any email clients other than Squirrelmail. I've got the server working in a testing environment. Could people who are dovecot savy look over the config I have set? I've only included the directives that are enabled with no comments. I want a simple but secure backup mail server for about 75 users. Thank you, Eric base_dir = /var/run/dovecot/ protocols = imap login_dir = /var/run/dovecot/login login = imap login_executable = /usr/libexec/dovecot/imap-login mbox_locks = fcntl imap_executable = /usr/libexec/dovecot/imap imap_modules = /usr/lib/dovecot/imap auth = default auth_mechanisms = plain auth_userdb = passwd auth_passdb = pam auth_executable = /usr/libexec/dovecot/dovecot-auth auth_user = root --

1 0

[Dovecot] Logging "issue" with PAM authentication
by Steve Johnson 07 Jul '04

07 Jul '04

Hi, We currently have a postfix/dovecot setup using nss_ldap with PAM for authentication. Everything is working fine, but there are excessive "error" messages in /var/log/messages that I'd like to prevent from happening. Since auth_userdb defaults to passwd and that our users are not in the passwd file, we get the following message everytime a user logs in: Jul 7 13:34:37 hostname dovecot(pam_unix)[32206]: check pass; user unknown Jul 7 13:34:37 hostname dovecot(pam_unix)[32206]: authentication failure; logname= uid=0 euid=0 tty= ruser= rhost= We of course also get the input that a user has logged in, in /var/log/maillog: imap-login: Jul 07 13:34:37 Info: Login: username [xxx.xxx.xxx.xxx] I've tried changing auth_verbose to no, and I still get those messages. Is there any way to prevent this from happening? Thanks a lot, Steve Johnson

2 2

[Dovecot] mbox:~/%u-mail/
by justin 07 Jul '04

07 Jul '04

Is it possible to use something like: default_mail_env = mbox:~/%u-mail/:INBOX=/var/mail/%u such that for a user: user1 with a home directory: /home/share the imap files would then be stored in: /home/share/user1-mail/ ???

2 1

[Dovecot] UID Error on Test 24 and 25
by Doug Eubanks 07 Jul '04

07 Jul '04

I started testing Dovecot 1.0 Test 22! I loved the performance improvements, althought dovecot has been very fast for us from the beginning. However, now I am testing test 24 and 25, and I am seeing these errors in the dovecot log and squirrelmail is giving me a ERROR : Connection dropped by imap-server. Query: STATUS "Mailing Lists.SpamAssassin" (MESSAGES UNSEEN RECENT) This is the dovecot error from the log file dovecot: Jul 07 08:52:33 Info: imap-login: Login: admin [208.*.*.*] dovecot: Jul 07 08:52:33 Error: IMAP(admin): file mail-index-view.c: line 286 (_view_lookup_uid_range): assertion failed: (first_uid > 0) dovecot: Jul 07 08:52:33 Error: child 12469 (imap) killed with signal 6 The user I am loggin in with has a uid and gid of 501, not 0! I am using passwd and shadow auth with Maildirs. If anyone has any pointers, they would greatly be appreciated. Sincerely, Doug Eubanks doug(a)simflex.com Post Hoc, Ergo Propter Hoc *** DISCLAIMER *** This e-mail and any attachments thereto may contain information, which is confidential and/or protected by intellectual property rights and are intended for the sole use of the recipient(s) named above. Any use of the information contained herein (including, but not limited to, total or partial reproduction, communication or distribution in any form) by persons other than the designated recipient(s) is prohibited. If you have received this e-mail in error, please notify the sender either by telephone or by e-mail and delete the material from any computer. Thank you for your cooperation.

2 1

[Dovecot] 1.0-test25 with new cache file code
by Timo Sirainen 07 Jul '04

07 Jul '04

http://dovecot.org/test/ Pretty much the only change since test24 is a redesign of how cache file writing works. Reading works as before - without locks. Before (in 1.0-tests, 0.99 was worse) we used to just try to lock it. If it failed, we just didn't update cache. If it didn't fail, the lock could have been kept a long time. For example if a user was downloading a 10MB mail, the lock could have been kept the whole time and no-one else could have modified the cache. The old behaviour wasn't actually too bad, as normally it didn't matter if something couldn't be written to cache. It would have been the next time, all it did was cost some I/O. The new behaviour is more reliable as the locks are kept only a very short time. It works by first writing the updates only into internal 32kB buffer. Once the buffer gets full, or the transaction completes, it's written to disk. When buffer gets full, space is reserved from the cache file for the current transaction. The amount of space is more than is needed for writing the internal buffer, so we don't have to lock the cache file every time the buffer gets full (and so we don't need a huge buffer). Once transaction completes, we lock the cache again and try to free the extra reserved space we have left. If multiple processes weren't updating the cache file at the same time, the freeing should be as simple as decreasing "used_file_size" field in header. Alternatively we'll update a linked list of "holes" in the file. When reserving space, we first try to use those holes instead of appending to end of file. Pointers to new cache records are also written to transaction log at the end of cache transaction. They'll get written to index at next sync after which they become visible to other processes. Each cache record contains a pointer to previous record, so there can be multiple cache records for a single message. The bad thing with the new logic is that nothing prevents two processes from writing the same cache records and so using twice as much space as necessary. But I don't think it's much of a problem. They'll go away at next cache compression. These changes also make it quite simple to make cache file work with shared filesystems. I'd just need to implement some kind of internal caching buffer which remembers what parts of file are read into memory. These changes also make it possible to easily and reliably store X-UIDL headers into cache file, where the POP3 UIDL command could quickly get them. Not in this release yet, though..

3 2

[Dovecot] appenduid?
by Johannes Berg 06 Jul '04

06 Jul '04

Hi, Does dovecot support APPENDUID? If not, would it be possible to add? johannes

2 1

Re: [Dovecot] [PATCH, RFC] add APOP authentication mechanism
by Timo Sirainen 06 Jul '04

06 Jul '04

On 1.7.2004, at 16:26, Andrey Panin wrote: >> + if (strcmp(auth_request->protocol, "POP3")) { >> + if (verbose) >> + i_info("apop(%s): wrong protocol %s", >> get_log_prefix(auth_request), >> + auth_request->protocol); >> + mech_auth_finish(auth_request, NULL, 0, FALSE); >> + return TRUE; >> + } >> >> I don't think there's really any point in checking this? Maybe someone >> really wants to use APOP with another protocol :) > > This will be RFC violation isn't it ? (I'll send a public reply, the last question could need input from others) Well, it's not compatible with any RFC of course, but I don't think it would violate anything? I'd just rather like to avoid unnecessary code bloat. Well, except now that I thought about the logic instead of just the code, there are problems. APOP prevents replay attacks by sending different timestamp in greeting, but you're sending the greeting to dovecot-auth as user input to APOP authentication mechanism. So, a user could simply say "AUTH APOP" and send wanted timestamp as input. One way to fix this would be to disallow APOP mechanism to be used with AUTH command in pop3-login and require that the request was done with POP3 protocol. But I think that's pretty ugly kludge. It also requires that processes connecting to dovecot-auth are trusted, which is something I've tried to avoid as much as possible. So dovecot-auth doesn't really know if it's pop3-login which connects to it.. Better fix would be to get part of the timestamp from dovecot-auth itself. It could send a growing ID number as handshake which would be used as part of the timestamp, and APOP mechanism processing would check that timestamp really contains the ID. So timestamp would be <authid.unixtime@hostname>. That also makes the PID go away, which was another thing I would have suggested. POP3 RFC also says: It is conjectured that use of the APOP command provides origin identification and replay protection for a POP3 session. Accordingly, a POP3 server which implements both the PASS and APOP commands should not allow both methods of access for a given user; that is, for a given mailbox name, either the USER/PASS command sequence or the APOP command is allowed, but not both. But I don't know if there's any nice way to do it, except perhaps by defining APOP password scheme which would actually be plaintext but wouldn't work with anything but APOP authentication. Do any other servers prevent it either?

2 5

[Dovecot] Debian dovecot news
by Jaldhar H. Vyas 04 Jul '04

04 Jul '04

I uploaded 0.99.10.6 to sid. Owing to the importance of the dot lock fix, I've tried to have this pushed to sarge quickly. Let's see how that goes. I've prepared a backport of 0.99.10.6 to woody which should be showing up on www.backports.org soon hopefully. In the meantime, if you desperately need it and don't mind working out dependencies yourself, you can find it at http://www.braincells.com/backports/ I uploaded 1.0test18 to http://src.braincells.com/dovecot-test/ (I'll do test20 tomorrow) I told Timo I'd port dovecot's gnutls support from gnutls7 to gnutls10 but I have absolutely no time for it so if someone would volunteer, contact me, I'd really appreciate it. -- Jaldhar H. Vyas <jaldhar(a)debian.org> La Salle Debain - http://www.braincells.com/debian/

4 7