xref: /freebsd/contrib/sendmail/KNOWNBUGS (revision 3cbb4cc200f8a0ad7ed08233425ea54524a21f1c)
1
2
3	     K N O W N   B U G S   I N   S E N D M A I L
4
5
6The following are bugs or deficiencies in sendmail that we are aware of
7but which have not been fixed in the current release.  You probably
8want to get the most up to date version of this from ftp.sendmail.org
9in /pub/sendmail/KNOWNBUGS.  For descriptions of bugs that have been
10fixed, see the file RELEASE_NOTES (in the root directory of the sendmail
11distribution).
12
13This list is not guaranteed to be complete.
14
15* Header values which are too long may be truncated.
16
17  If a value of a structured header is longer than 256 (MAXNAME)
18  characters then it may be truncated during output. For example,
19  if a single address in the To: header is longer than 256 characters
20  then it will be truncated which may result in a syntactically
21  invalid address.
22
23* Delivery to programs that generate too much output may cause problems
24
25  If e-mail is delivered to a program which generates too much
26  output, then sendmail may issue an error:
27
28  timeout waiting for input from local during Draining Input
29
30  Make sure that the program does not generate output beyond a
31  status message (corresponding to the exit status).  This may
32  require a wrapper around the actual program to redirect output
33  to /dev/null.
34
35  Such a problem has been reported for bulk_mailer.
36
37* Null bytes are not handled properly in headers.
38
39  Sendmail should handle full binary data.  As it stands, it handles
40  all values in the body, but not 0x00 in the header.  Changing
41  this would require a major restructuring of the code -- for
42  example, almost no C library support could be used to handle
43  strings.
44
45* Header checks are not called if header value is too long or empty.
46
47  If the value of a header is longer than 1250 (MAXNAME + MAXATOM - 6)
48  characters or it contains a single word longer than 256 (MAXNAME)
49  characters then no header check is done even if one is configured for
50  the header.
51
52* Header lines which are too long will be split incorrectly.
53
54  Header lines which are longer than 2045 characters will be split
55  but some characters might be lost.  Fix: obey RFC (2)822 and do not
56  send lines that are longer than 1000 characters.
57
58* milter communication fails if a single header is larger than 64K.
59
60  If a single header is larger than 64KB (which is not possible in the
61  default configuration) then it cannot be transferred in one block to
62  libmilter and hence the communication fails.  This can be avoided by
63  increasing the constant MILTER_CHUNK_SIZE in
64  include/libmilter/mfdef.h and recompiling sendmail, libmilter, and
65  all (statically linked) milters (or by using undocumented compile
66  time options: _FFR_MAXDATASIZE/_FFR_MDS_NEGOTIATE; you have to
67  read the source code in order to use these properly).
68
69* Sender addresses whose domain part cause a temporary A record lookup
70  failure but have a valid MX record will be temporarily rejected in
71  the default configuration.  Solution: fix the DNS at the sender side.
72  If that's not easy to achieve, possible workarounds are:
73  - add an entry to the access map:
74	dom.ain	OK
75  - (only for advanced users) replace
76
77# Resolve map (to check if a host exists in check_mail)
78Kresolve host -a<OKR> -T<TEMP>
79
80   with
81
82# Resolve map (to check if a host exists in check_mail)
83Kcanon host -a<OKR> -T<TEMP>
84Kdnsmx dns -R MX -a<OKR> -T<TEMP>
85Kresolve sequence dnsmx canon
86
87
88* Duplicate error messages.
89
90  Sometimes identical, duplicate error messages can be generated.  As
91  near as I can tell, this is rare and relatively innocuous.
92
93* Misleading error messages.
94
95  If an illegal address is specified on the command line together
96  with at least one valid address and PostmasterCopy is set, the
97  DSN does not contain the illegal address, but only the valid
98  address(es).
99
100* \231 considered harmful.
101
102  Header addresses that have the \231 character (and possibly others
103  in the range \201 - \237) behave in odd and usually unexpected ways.
104
105* AuthRealm for Cyrus SASL may not work as expected. The man page
106  and the actual usage for sasl_server_new() seem to differ.
107  Feedback for the "correct" usage is welcome, a patch to match
108  the description of the man page is in contrib/AuthRealm.p0.
109
110* accept() problem on SVR4.
111
112  Apparently, the sendmail daemon loop (doing accept()s on the network)
113  can get into a weird state on SVR4; it starts logging ``SYSERR:
114  getrequests: accept: Protocol Error''.  The workaround is to kill
115  and restart the sendmail daemon.  We don't have an SVR4 system at
116  Berkeley that carries more than token mail load, so I can't validate
117  this.  It is likely to be a glitch in the sockets emulation, since
118  "Protocol Error" is not possible error code with Berkeley TCP/IP.
119
120  I've also had someone report the message ``sendmail: accept:
121  SIOCGPGRP failed errno 22'' on an SVR4 system.  This message is
122  not in the sendmail source code, so I assume it is also a bug
123  in the sockets emulation.  (Errno 22 is EINVAL "Invalid Argument"
124  on all the systems I have available, including Solaris 2.x.)
125  Apparently, this problem is due to linking -lc before -lsocket;
126  if you are having this problem, check your Makefile.
127
128* accept() problem on Linux.
129
130  The accept() in sendmail daemon loop can return ETIMEDOUT.  An
131  error is reported to syslog:
132
133  Jun  9 17:14:12 hostname sendmail[207]: NOQUEUE: SYSERR(root):
134			getrequests: accept: Connection timed out
135
136  "Connection timed out" is not documented as a valid return from
137  accept(2) and this was believed to be a bug in the Linux kernel.
138  Later information from the Linux kernel group states that Linux
139  2.0 kernels follow RFC1122 while sendmail follows the original BSD
140  (now POSIX 1003.1g draft) specification.  The 2.1.X and later kernels
141  will follow the POSIX draft.
142
143* Excessive mailing list nesting can run out of file descriptors.
144
145  If you have a mailing list that includes lots of other mailing
146  lists, each of which has a separate owner, you can run out of
147  file descriptors.  Each mailing list with a separate owner uses
148  one open file descriptor (prior to 8.6.6 it was three open
149  file descriptors per list).  This is particularly egregious if
150  you have your connection cache set to be large.
151
152* Connection caching breaks if you pass the port number as an argument.
153
154  If you have a definition such as:
155
156	  Mport,          P=[IPC], F=kmDFMuX, S=11/31, R=21,
157			  M=2100000, T=DNS/RFC822/SMTP,
158			  A=IPC [127.0.0.1] $h
159
160  (i.e., where $h is the port number instead of the host name) the
161  connection caching code will break because it won't notice that
162  two messages addressed to different ports should use different
163  connections.
164
165* ESMTP SIZE underestimates the size of a message
166
167  Sendmail makes no allowance for headers that it adds, nor does it
168  account for the SMTP on-the-wire \r\n expansion.  It probably doesn't
169  allow for 8->7 bit MIME conversions either.
170
171* Client ignores SIZE parameter.
172
173  When sendmail acts as client and the server specifies a limit
174  for the mail size, sendmail will ignore this and try to send the
175  mail anyway.  The server will usually reject the MAIL command
176  which specifies the size of the message and hence this problem
177  is not significant.
178
179* Paths to programs being executed and the mode of program files are
180  not checked.  Essentially, the RunProgramInUnsafeDirPath and
181  RunWritableProgram bits in the DontBlameSendmail option are always
182  set.  This is not a problem if your system is well managed (that is,
183  if binaries and system directories are mode 755 instead of something
184  foolish like 777).
185
186* 8-bit data in GECOS field
187
188  If the GECOS (personal name) information in the passwd file contains
189  8-bit characters, those characters can be included in the message
190  header, which can cause problems when sending SMTP to hosts that
191  only accept 7-bit characters.
192
193* 8->7 bit MIME conversion
194
195  When sendmail is doing 8->7 bit MIME conversions, and the message
196  contains certain MIME body types that cannot be converted to 7-bit,
197  sendmail will pass the message as 8-bit.
198
199* 7->8 bit MIME conversion
200
201  If a message that is encoded as 7-bit MIME is converted to 8-bit and
202  that message when decoded is illegal (e.g., because of long lines or
203  illegal characters), sendmail can produce an illegal message.
204
205* MIME encoded full name phrases in the From: header
206
207  If a full name phrase includes characters from MustQuoteChars, sendmail
208  will quote the entire full name phrase.  If MustQuoteChars includes
209  characters which are not special characters according to STD 11 (RFC
210  822), this quotation can interfere with MIME encoded full name phrases.
211  By default, sendmail includes the single quote character (') in
212  MustQuoteChars even though it is not listed as a special character in
213  STD 11.
214
215* bestmx map with -z flag truncates the list of MX hosts
216
217  A bestmx map configured with the -z flag will truncate the list
218  of MX hosts.  This prevents creation of strings which are too
219  long for ruleset parsing.  This can have an adverse effect on the
220  relay_based_on_MX feature.
221
222* Saving to ~sender/dead.letter fails if su'ed to root
223
224  If ErrorMode is set to print and an error in sending mail occurs,
225  the normal action is to print a message to the screen and append
226  the message to a dead.letter file in the sender's home directory.
227  In the case where the sender is using su to act as root, the file
228  safety checks prevent sendmail from saving the dead.letter file
229  because the sender's uid and the current real uid do not match.
230
231* Berkeley DB 2.X race condition with fcntl() locking
232
233  There is a race condition for Berkeley DB 2.X databases on
234  operating systems which use fcntl() style locking, such as
235  Solaris.  Sendmail locks the map before calling db_open() to
236  prevent others from modifying the map while it is being opened.
237  Unfortunately, Berkeley DB opens the map, closes it, and then
238  reopens it.  fcntl() locking drops the lock when any file
239  descriptor pointing to the file is closed, even if it is a
240  different file descriptor than the one used to initially lock
241  the file.  As a result there is a possibility that entries in a
242  map might not be found during a map rebuild.  As a workaround,
243  you can use makemap to build a map with a new name and then
244  "mv" the new db file to replace the old one.
245
246  Sleepycat Software has added code to avoid this race condition to
247  Berkeley DB versions after 2.7.5.
248
249* File open timeouts not available on hard mounted NFS file systems
250
251  Since SIGALRM does not interrupt an RPC call for hard mounted
252  NFS file systems, it is impossible to implement a timeout on a file
253  open operation.  Therefore, while the NFS server is not responding,
254  attempts to open a file on that server will hang.  Systems with
255  local mail delivery and NFS hard mounted home directories should be
256  avoided, as attempts to open the forward files could hang.
257
258* Race condition for delivery to set-user-ID files
259
260  Sendmail will deliver to a file if the file is owned by the DefaultUser
261  or has the set-user-ID bit set.  Unfortunately, some systems clear that bit
262  when a file is modified.  Sendmail compensates by resetting the file mode
263  back to it's original settings.  Unfortunately, there's still a
264  permission failure race as sendmail checks the permissions before locking
265  the file.  This is unavoidable as sendmail must verify the file is safe
266  to open before opening it.  A file can not be locked until it is open.
267
268* MAIL_HUB always takes precedence over LOCAL_RELAY
269
270  Despite the information in the documentation, MAIL_HUB ($H) will always
271  be used if set instead of LOCAL_RELAY ($R).  This will be fixed in a
272  future version.
273
274