had been granted access to the portmapper via hosts.{allow,deny} could use
PMAPPROC_CALLIT to call PMAPPROC_{SET,UNSET} to (un)register services as if
they were running on the local host.
The new code disallows all indirect calls to the portmapper except for
PMAPPROC_NULL unless the -i (insecure) flag has been specified.
While there, add a new flag, -p (paranoid) which also disallows indirect calls
to a small number of other services, including key parts of NFS and NIS. This
code hardcodes the services to be disallowed, and is thus somewhat of a hack,
but will serve for the time being (until portmap is replaced by rpcbind as part
of fvdl's current rpc work, due to happen before 1.5).
Problem pointed out by Frank van der Linden <fvdl@netbsd.org>, solution determined
in discussion with Frank van der Linden and with Bill Sommerfeld <sommerfeld@netbsd.org>.
Some inspiration drawn from the (less general) handling of this problem in Wietse
Venema's libwrap'ed portmap.
use of non-exported function __ivaliduser{,_sa}().
we cannot make __ivaliduser{,_sa}() static yet, since doing that would choke
compiled lpd binaries. we should do it on next libc major version bump.
added a memo on lib/libc/shlib_version.
while here, do some whitespace/const cleanup, convert to use addentry(),
g/c section[] (now uses buf[] directly) - 10 character limit for section
name is gone
- decrease warning level on missing rtadvd.conf (actually, the file
can be omitted)
- strict prototype
- gather stats better, emit stats on SIGUSR1 to /var/run
+ Use _PATH_GROUP and _PATH_MASTERPASSWD (from OpenBSD)
+ Use -G group1,group2,group3 for multiple groups in useradd and usermod
(pointed out by Matt Green, and also changed in OpenBSD, but done more
efficiently here)
+ is_number should not be inside #ifdef EXTENSIONS (from OpenBSD)
+ clear up yet another usage message (for user(8) and group(8)) - noticed
in passing, unknown if fixed anywhere else
support the address family (like including "tcp6" in inetd.conf, on
non-IPv6 kernel).
was:
inetd[185]: ftp/tcp6: *: hostname nor servname provided, or not known
now:
inetd[315]: ftp/tcp6: *: the address family is not supported by the kernel
1. if there is a colon present, use that as a separator for user:group
2. if there is no colon, attempt to convert the arg into a username,
searching backwards in the string for a '.' for us.er.group
3. if the arg doesn't match a username and has a '.' in it, split it
up and try user.group
package matching a certain pattern. Examples:
yui# cd /usr/pkgsrc/packages/i386ELF/All/
yui# ls unzip*
unzip-5.40.tgz unzip-5.41.tgz
yui# pkg_admin lsall 'unzip*'
unzip-5.40.tgz
unzip-5.41.tgz
yui# pkg_admin lsall 'unzip>=5.40'
unzip-5.40.tgz
unzip-5.41.tgz
yui# pkg_admin lsall 'unzip>=5.41'
unzip-5.41.tgz
yui# pkg_admin lsbest 'unzip>=5.40'
unzip-5.41.tgz
yui# pkg_admin lsall /usr/pkgsrc/packages/i386ELF/All/'{mit,unproven}-pthread*'
/usr/pkgsrc/packages/i386ELF/All/mit-pthreads-1.60b6.tgz
This adds a shell/user-interface to pkg-patterns, which are a superset
of sh/csh patterns and can't be expanded by any shell.
a static once-generated version instead. We know we have IPv6
headers available here.
The probing was problematical for several reasons:
o it probed the host headers, not the headers in the build or DESTDIR
tree (could be fixed in another way)
o the probe_ipv6 script mucks with PATH, which would be problematical
for cross compilation.
contents of that header (the only file that includes it compiles to the
same object code on multiple architectures with or without including
<ieeefp.h>), so remove all references to it.
Fix sent to NTP maintainers - they will probably implement this change
after the immenient 4.1.0 release, but don't want to change it so close
to the release date.
- isakmp: print CERT and SIG payload. fix IPsec-AH algorithm type.
- rt6: avoid duplicated IPv6 src/dst.
sync with tcpdump.org.
XXX we need to think about future synchronization with tcpdump.org...
default nis-domain "";
would cause a NULL pointer deref while writing out the lease into
the persistent database if the server didn't include an nis-domain
option in the reply.
From: hiro@takechi.org
XXX checkremote() should be improved. gethostname -> getaddrinfo is
not the right thing to do, we cannot assume DNS FQDNs is configured
as hostname. if the goal here is to check if it is really remote or not,
getifaddrs() is the way to go.
struct dirent *, rather than non-const. this makes scandir(3) the
same as the scandir implementations in libiberty and glibc, and the
select function has no need to modify the dirent.
Handling all kinds of wildcards properly would be hard, though it should
handle some cases better now. Esp. ones with '-' in the pkg(!) name, and/or
ones with dewer relational versions. I.e. the teTeX-share>1.97 case should
work now.
after a "-" with the "more liberal pattern", but also anything after
any pkg wildcard.
That way if someone has teTeX-share-1.0 installed and pkg_adds
teTeX-1.0.7 (which wants teTeX-share>1.0.2), it will be caught.
XXX This "quick depends pre-check" would be nice to be done in
bsd.pkg.mk as well
dependencies if they _can_ be installed. I.e. if a package wants
version X installed, but version Y is already installed, pkg_adding
that required pkg will blow up later (probably some pkg_add recursions
down, given what we keep in the depends list). Now, it stopps right
away:
noon# pkg_add /usr/pkgsrc/packages/i386ELF/All/xdaemon-1.0.tgz
pkg_add: pkg `xteddy-1.*' required, but `xteddy-2.0.1' found installed.
Please resolve this conflict!
The idea of this is from Thomas Klausner, further inputs from Alistair
Crooks.
* allow pkg names without versions given to "pkg_admin check"
* Use sizeof() instead of hardcoding the buffers' size in some places
dig/host/whatever assumes that it is using BIND8 code. mixing BIND4 in
libc with BIND8 code will result in very strange behavior, or program panics.
it is not necessary for dig/host/whatever to obey /etc/nsswitch.conf, actually
dig(1) is explicit about it.
now dist/bind is almost clean BIND822p5, with the following exception:
- /etc/irs.conf will never be visited when running BIND8 toolchain,
to make it less complex. the search order for BIND8 toolchain is
defined in dist/bind/lib/irs/gen.c:default_map_rules().
and usr.sbin/bind compiles them in BSD make framework, with no tricks at all.
- Move SEE ALSO section before the HISTORY section as described in
mdoc(7) with:
+ removing traing dot.
+ spliting files into FILES section.
- Refer pathnames using .Pa (rather .Ar or .Xr).
this changes function prototype for __ivalid*.
This commit breaks binary compatibility for __ivalid*. I believe this can be
forgiven due to the following:
- this is not really exported function. no function prototype is in headers.
function name starts with underbars. No third-party applications are
expected to use it.
- the function was introduced very recently, when rcmd.c was made IPv6-ready.
- the only customer in NetBSD tree is lpd.
(lpd.c) sync with rcmd.c change.
on each of them, and don't bother listening on them. this allows messages
to be forwarded even with -s enabled.
XXX: not 100% sure if this is susceptable to a DoS, e.g someone filling
up kernel buffers with a backlog of packets not being read, but shutdown(2)
should prevent that even if it possibly doesn't at this time.
* minor KNF
dev/vndvar.h or dev/ccdvar.h so that struct pool is always available
regardless of multiple inclusion. Actually, ccdconfig.c compiles without
this change but ...
per mount point, the way ``try to export and unexport an unnecessary
mount point later'' may leave unwanted information in the kernel, and
doesn't work if a mount point is default exported or publically exported.
To fix the race condtion, we need another way.
version 0 RMT protocol, the query of status makes no sense (send back, in
binary, the mtget structure), but systems sometimes use it anyway. And when
they use it, they get upset if they get back something that is a different
size from what they expect. The original mtget structure in BSD 4.3 is
24 bytes in size, and that seems to be what they can cope with.
The correct fix would be to go to Version 1 RMT, but that's a horse of
another color...
