This patch, apparently, came from the freebsd syslogd.
The patch was subsequently lost although the old Fedora RPM can still
be found.
I took that patch, the current freebsd sources and a dash of salt to
produce a new patch which I have attached.
The patch does two things: it will arrange for syslogd to listen on an
IPv6 socket in addition to the IPv4 socket and it uses getaddrinfo()
and getnameinfo() rather than gethostby[name|addr] and getservbyname.
It would be possible to use a single socket that accepts both IPv6 and
IPv4 datagrams, but this would mean that numeric IPv4 addresses would
be prefixed with ::ffff and this could cause some confusion -- and I'm
also not sure that everything that sysklogd runs on has a dual stack
like that.
I've also introduced the -4, -6 and -A with the same meanings that the
freebsd and rsyslog variants do.
This will affect local messages sent from a different timezone. Also
removed code due to the changed semantics. This was inspired by
Anders Henke from Schlund + Partner AG.
From: Solar Designer <solar@openwall.com>
1. Ensures that "len" is not placed in a register and as such can't be
clobbered by longjmp(). With the particular code, it does not really
matter whether it is clobbered or not, but this avoids the gcc warning.
2. Makes endtty() the signal handler only after the variable that
function uses is initialized. In the original code, the signal
handler was setup too early and if there would be SIGALRM before
control reaches setjmp(), syslogd would segfault (if not worse).
Basically, this is a minor correctness patch.
guaranteed to be available when the child is forked, hence, fixing a
race condition. This used to create problems with UML and fast
machines. Thanks to Jon Burgess <Jon_Burgess@eur.3com.com>
. Changed SOCK_STREAM to SOCK_DGRAM in syslog.c
. klogd will only change the console log level if `-c' is supplied
. syslogd.c by Bill Nottingham <notting@redhat.com>
Um, if the directory is invalid, the bind() call in
create_unix_socket fails. Without the return -1, we return the
invalid fd that we just closed. When syslogd then starts
listening, select goes into a hard loop getting EBADF, IIRC.
. klogd.c by Troels Walsted Hansen <troels@thule.no>
I found a bug in the sysklogd package version 1.4. When it
encounters a zero byte in the kernel logging output, the text
parser enters a busy loop. I came upon it when the 3c59x driver
from kernel 2.4.0 started outputting two zero bytes for the product
code of my laptop's 3Com card. It could be argued that the kernel
should never output zero bytes in the logging info, but obviously
that will happen from time to time.
I fear this bug might be considered a security issue as well, if
the kernel can be coerced to output a zero byte somehow, all kernel
logging will stop.
Wolfgang Oertl <Wolfgang.Oertl@uibk.ac.at> had a similar bugfix
idea
. klogd.c by Thomas Roessler <roessler@does-not-exist.org>
Additionally, the patch prevents LogLine from being invoked with a
negative counter as an argument.
Removed unixm/unix domain sockets and switch to Datagram Unix
Sockets. This should remove one possibility to play DoS with
syslogd. Thanks to Olaf Kirch <okir@caldera.de> for the patch.
Fixed bug in printchopped() that caused syslogd to emit
kern.emerg messages when splitting long lines. Thanks to
Daniel Jacobowitz <dan@debian.org> for the fix.
Fixed some bugs in printline() code that did not escape
control characters '\177' through '\237' and contained a
single-byte buffer overflow. Thanks to Solar Designer
<solar@false.com>.
* Removed superflous call to utmpname() (closes: Bug#34733)
* Added --ignore-size and limit for rotating log files at 2MB to
syslogd-listfiles (closes: Bug#36198)
* Changed Greg's e-mail address to represent his current one.
* Shortened line length for kernel logging slightly.
* Corrected return value of AddModule (closes: Bug#30093)
* Finally fixed an error with `-a' processing, thanks to Topi Miettinen
<tom@medialab.sonera.net> (closes: Bug#30462)
