KNOWNBUGS revision 90792
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* Delivery to programs that generate too much output may cause problems 16 17 If e-mail is delivered to a program which generates too much 18 output, then sendmail may issue an error: 19 20 timeout waiting for input from local during Draining Input 21 22 Make sure that the program does not generate output beyond a 23 status message (corresponding to the exit status). This may 24 require a wrapper around the actual program to redirect output 25 to /dev/null. 26 27 Such a problem has been reported for bulk_mailer. 28 29* Null bytes are not handled properly in headers. 30 31 Sendmail should handle full binary data. As it stands, it handles 32 all values in the body, but only 0x01-0x80 and 0xA0-0xFF in 33 the header. Notably missing is 0x00, which would require a major 34 restructuring of the code -- for example, almost no C library support 35 could be used to handle strings. 36 37* Header checks are not called if header value is too long or empty. 38 39 If the value of a header is longer than 1250 (MAXNAME + MAXATOM - 6) 40 characters or it contains a single word longer than 256 (MAXNAME) 41 characters then no header check is done even if one is configured for 42 the header. 43 44* Duplicate error messages. 45 46 Sometimes identical, duplicate error messages can be generated. As 47 near as I can tell, this is rare and relatively innocuous. 48 49* Misleading error messages. 50 51 If an illegal address is specified on the command line together 52 with at least one valid address and PostmasterCopy is set, the 53 DSN does not contain the illegal address, but only the valid 54 address(es). 55 56* \231 considered harmful. 57 58 Header addresses that have the \231 character (and possibly others 59 in the range \201 - \237) behave in odd and usually unexpected ways. 60 61* accept() problem on SVR4. 62 63 Apparently, the sendmail daemon loop (doing accept()s on the network) 64 can get into a weird state on SVR4; it starts logging ``SYSERR: 65 getrequests: accept: Protocol Error''. The workaround is to kill 66 and restart the sendmail daemon. We don't have an SVR4 system at 67 Berkeley that carries more than token mail load, so I can't validate 68 this. It is likely to be a glitch in the sockets emulation, since 69 "Protocol Error" is not possible error code with Berkeley TCP/IP. 70 71 I've also had someone report the message ``sendmail: accept: 72 SIOCGPGRP failed errno 22'' on an SVR4 system. This message is 73 not in the sendmail source code, so I assume it is also a bug 74 in the sockets emulation. (Errno 22 is EINVAL "Invalid Argument" 75 on all the systems I have available, including Solaris 2.x.) 76 Apparently, this problem is due to linking -lc before -lsocket; 77 if you are having this problem, check your Makefile. 78 79* accept() problem on Linux. 80 81 The accept() in sendmail daemon loop can return ETIMEDOUT. An 82 error is reported to syslog: 83 84 Jun 9 17:14:12 hostname sendmail[207]: NOQUEUE: SYSERR(root): 85 getrequests: accept: Connection timed out 86 87 "Connection timed out" is not documented as a valid return from 88 accept(2) and this was believed to be a bug in the Linux kernel. 89 Later information from the Linux kernel group states that Linux 90 2.0 kernels follow RFC1122 while sendmail follows the original BSD 91 (now POSIX 1003.1g draft) specification. The 2.1.X and later kernels 92 will follow the POSIX draft. 93 94* Excessive mailing list nesting can run out of file descriptors. 95 96 If you have a mailing list that includes lots of other mailing 97 lists, each of which has a separate owner, you can run out of 98 file descriptors. Each mailing list with a separate owner uses 99 one open file descriptor (prior to 8.6.6 it was three open 100 file descriptors per list). This is particularly egregious if 101 you have your connection cache set to be large. 102 103* Connection caching breaks if you pass the port number as an argument. 104 105 If you have a definition such as: 106 107 Mport, P=[IPC], F=kmDFMuX, S=11/31, R=21, 108 M=2100000, T=DNS/RFC822/SMTP, 109 A=IPC [127.0.0.1] $h 110 111 (i.e., where $h is the port number instead of the host name) the 112 connection caching code will break because it won't notice that 113 two messages addressed to different ports should use different 114 connections. 115 116* ESMTP SIZE underestimates the size of a message 117 118 Sendmail makes no allowance for headers that it adds, nor does it 119 account for the SMTP on-the-wire \r\n expansion. It probably doesn't 120 allow for 8->7 bit MIME conversions either. 121 122* Client ignores SIZE parameter. 123 124 When sendmail acts as client and the server specifies a limit 125 for the mail size, sendmail will ignore this and try to send the 126 mail anyway. The server will usually reject the MAIL command 127 which specifies the size of the message and hence this problem 128 is not significant. 129 130* Paths to programs being executed and the mode of program files are 131 not checked. Essentially, the RunProgramInUnsafeDirPath and 132 RunWritableProgram bits in the DontBlameSendmail option are always 133 set. This is not a problem if your system is well managed (that is, 134 if binaries and system directories are mode 755 instead of something 135 foolish like 777). 136 137* 8-bit data in GECOS field 138 139 If the GECOS (personal name) information in the passwd file contains 140 8-bit characters, those characters can be included in the message 141 header, which can cause problems when sending SMTP to hosts that 142 only accept 7-bit characters. 143 144* 8->7 bit MIME conversion 145 146 When sendmail is doing 8->7 bit MIME conversions, and the message 147 contains certain MIME body types that cannot be converted to 7-bit, 148 sendmail will strip the message to 7-bit. 149 150* 7->8 bit MIME conversion 151 152 If a message that is encoded as 7-bit MIME is converted to 8-bit and 153 that message when decoded is illegal (e.g., because of long lines or 154 illegal characters), sendmail can produce an illegal message. 155 156* MIME encoded full name phrases in the From: header 157 158 If a full name phrase includes characters from MustQuoteChars, sendmail 159 will quote the entire full name phrase. If MustQuoteChars includes 160 characters which are not special characters according to STD 11 (RFC 161 822), this quotation can interfere with MIME encoded full name phrases. 162 By default, sendmail includes the single quote character (') in 163 MustQuoteChars even though it is not listed as a special character in 164 STD 11. 165 166* bestmx map with -z flag truncates the list of MX hosts 167 168 A bestmx map configured with the -z flag will truncate the list 169 of MX hosts. This prevents creation of strings which are too 170 long for ruleset parsing. This can have an adverse effect on the 171 relay_based_on_MX feature. 172 173* Saving to ~sender/dead.letter fails if su'ed to root 174 175 If ErrorMode is set to print and an error in sending mail occurs, 176 the normal action is to print a message to the screen and append 177 the message to a dead.letter file in the sender's home directory. 178 In the case where the sender is using su to act as root, the file 179 safety checks prevent sendmail from saving the dead.letter file 180 because the sender's uid and the current real uid do not match. 181 182* Berkeley DB 2.X race condition with fcntl() locking 183 184 There is a race condition for Berkeley DB 2.X databases on 185 operating systems which use fcntl() style locking, such as 186 Solaris. Sendmail locks the map before calling db_open() to 187 prevent others from modifying the map while it is being opened. 188 Unfortunately, Berkeley DB opens the map, closes it, and then 189 reopens it. fcntl() locking drops the lock when any file 190 descriptor pointing to the file is closed, even if it is a 191 different file descriptor than the one used to initially lock 192 the file. As a result there is a possibility that entries in a 193 map might not be found during a map rebuild. As a workaround, 194 you can use makemap to build a map with a new name and then 195 "mv" the new db file to replace the old one. 196 197 Sleepycat Software has added code to avoid this race condition to 198 Berkeley DB versions after 2.7.5. 199 200* File open timeouts not available on hard mounted NFS file systems 201 202 Since SIGALRM does not interrupt an RPC call for hard mounted 203 NFS file systems, it is impossible to implement a timeout on a file 204 open operation. Therefore, while the NFS server is not responding, 205 attempts to open a file on that server will hang. Systems with 206 local mail delivery and NFS hard mounted home directories should be 207 avoided, as attempts to open the forward files could hang. 208 209* Race condition for delivery to set-user-ID files 210 211 Sendmail will deliver to a fail if the file is owned by the DefaultUser 212 or has the set-user-ID bit set. Unfortunately, some systems clear that bit 213 when a file is modified. Sendmail compensates by resetting the file mode 214 back to it's original settings. Unfortunately, there's still a 215 permission failure race as sendmail checks the permissions before locking 216 the file. This is unavoidable as sendmail must verify the file is safe 217 to open before opening it. A file can not be locked until it is open. 218 219$Revision: 8.54 $, Last updated $Date: 2001/12/17 16:07:51 $ 220