o avoid dying in daemon mode for some uncommon, but recoverable, errors
o close leaking file descriptors for CGI and daemon mode
o handle poll errors properly
o don't try to handle more than one request per process yet
this covers PR#38489 and PR#40079, as well some some issues reported
privately.
o make bozohttpd internally more modular, preparing the way
to handle more than one request per process
o fix http-auth, set $REMOTE_USER not $REMOTEUSER. also fix
cgi-bin with cvsweb, from Holger Weiss <holger@CIS.FU-Berlin.DE>
o fix an uninitialised variable use in daemon mode
o fix ssl mode with newer OpenSSL
o mmap large files in manageable sizes so we can serve any size file
o refactor url processing to handle query strings correctly for CGI
from Sergey Katsev at Coyote Point
o add If-Modified-Since support, from Joerg Sonnenberger
<joerg@netbsd.org>
o many more manual fixes, from NetBSD
both the "data" and "prefix" elements are identical. For example, the
two lines
amiga/boot (8) - system bootstrapping procedures
amd64/boot (8) - system bootstrapping procedures
both appear with data = "boot (8) - system bootstrapping procedures" but
with different values for prefix, and we do not want to reject one of
them as a duplicate.
1) Fix an issue where because of the reordering of transform_request
and process_cgi, the cgi-script name was being cut off
by one character (transform_request for some reason
cuts off the leading '/' for the file name as part of its
processing). As an 'easy' fix, simply re-add the '/' to the
front of the filename, which means that we don't have to
mess with the logic that sometimes adds +1 and sometimes doesn't.
2) Work around ridiculous bug in PHP reported by lukem in 2004,
but stubbornly never fixed by the PHP maintainers:
Change the SCRIPT_NAME and SCRIPT_FILENAME variables
to contain the file name only, not the query if one exists.
Having the query in SCRIPT_FILENAME causes php-cgi to not
work, as per the bug information here:
http://bugs.php.net/bug.php?id=28227
3) Fix a memory leak because URL wasn't being freed.
CGI handling, including bin/40355 . There are two main changes here:
1) call process_cgi() after transform_request(), not before. Now it is
possible to have a default cgi handler catch a request for a path that
was produced by transformation, e.g. by index generation -- so now the
index can be "generated" by a CGI if that is what the user desires.
2) More clearly distinguish "file" from "query" portions of the request
URL, so we do not feed ?-suffixed "arguments" to plain files, fail to
match filename extensions due to ?-suffixes, etc.
After this change, there are only two cases which use the "query"
portion of the request (the portion after the ?):
a) A redirect issued by HTTPD will redirect to the new file, but
with the same query string.
b) process_cgi() will, of course continue to use the query string.
- look in src/compat/lib/libc for libc_pic.
- define RTLD_ARCH_SUBDIR to "i386" or "sparc" for amd64 and sparc64
builds, respectively
if RTLD_ARCH_SUBDIR is defined, add this path before
RTLD_DEFAULT_LIBRARY_PATH in the default search path.
Install so first file installed matches this correct manual page
name.
(Both manual pages are still installed. And only ntalkd binary is
installed. So no change.)
checks whether user as connecting from host would be granted
access by ftpusers(5).
Support IPv6 in the host directive of ftpusers(5).
(May resolve PR 26555)
Both features from Rudolf Cejka <cejkar@fit.vutbr.cz>
(FreeBSD's tnftpd port maintainer).
This prevents CSRF-like attacks, when a web browser is used to access
an ftp server.
Reported by Maksymilian Arciemowicz <cxib@securityreason.com>.
Fix mostly derived from OpenBSD, written by Moritz Jodeit <moritz@OpenBSD.org>
act like the other versions.
In _rtld_bind, if the result is 0, call _rtld_die.
Initialize _rtld_sym_zero.st_value to -_rtld_objself.maprelocbase. Now when
the symbol is resolved, st_value + maprelocbase will equal 0 and the above
check in _rtld_bind will fire and a call to NULL will be avoided.
(the utmpentry.c code), specifically with respect to who owns them and
when to free them. Now they're owned by utmpentry.c, only. Abolish the
freeutentries() function, which was the wrong abstraction; add instead
endutentries(), which flushes out the internally managed memory.
Update callers as necessary. Some (e.g. talkd) had been leaking memory;
others (e.g. syslogd) had been accidentally freeing and reloading utmp
more often than necessary. There are a couple untidy bits in users and
rwhod that someone should look after sometime, maybe.
Fixes PR bin/35131, which was about talkd's memory leak.
run through copy-on-write. Call fscow_run() with valid data where possible.
The LP_UFSCOW hack is no longer needed to protect ffs_copyonwrite() against
endless recursion.
- Add a flag B_MODIFY to bread(), breada() and breadn(). If set the caller
intends to modify the buffer returned.
- Always run copy-on-write on buffers returned from ffs_balloc().
- Add new function ffs_getblk() that gets a buffer, assigns a new blkno,
may clear the buffer and runs copy-on-write. Process possible errors
from getblk() or fscow_run(). Part of PR kern/38664.
Welcome to 4.99.63
Reviewed by: YAMAMOTO Takashi <yamt@netbsd.org>
EX_TEMPFAIL (cannot lookup user through NSS, may be because of an
unreachable NIS or LDAP server), or EX_OSERR (other file or memory
related problems).
Sendmail and Postfix should honour EX_TEMPFAIL and EX_OSERR by
queueing the message.
This avoids problems with transcient NSS errors, where the MTA may
successfully lookup a user at a time while mail.local can fail a
few secoonds later.
Since it doesn't have case, treat it like a digit. While here, avoid
setting the associated flag variable by incrementing it.
Fixes PR bin/33521; patch based on one from Dieter Roelants.
performed by the pam_conv (PAM conversation) callback, which then
getline()s the PASS reply internally. This involves calling
auth_pam() from user() and caching the result to use later in pass().
This allows the PAM modules to present a different password prompt
dialog if necesary. For example:
Name (localhost:lukem):
331 User lukem accepted, provide password [ otp-md4 89 xxxx12345 ].
versus
Name (localhost:lukem): root
331 User root accepted, provide password.
This is independent of (and effectively exclusive to) USE_SKEY support.
Previously ftpd with USE_SKEY=yes would provide the skey prompt
if the user had an skey configured, even if /etc/pam.d/ftpd didn't
have pam_skey in use.
I.e., ftpd shouldn't need special support for custom password prompts
(such as skey) if PAM is in use.
o fix some cgi header processing, from <thelsdj@gmail.com>
o add simple Range: header processing, from <bad@bsd.de>
o man page fixes, from NetBSD
o clean up various parts, from NetBSD
o prefix some function names with "bozo"
o align directory indexing <hr> markers
o clean up some code GCC4 grumbled about
for PT_PHDR headers when the program has been loader to a vaddr other than
then one specified (a randomized one), and modifying the relocation base
address appropriately (idea from elad)
- Apply patch from J.T. Conklin to execute .init/.fini functions in order.
- Support DF_1_INITFIRST and mark libc with DF_1_INITFIRST. Shared libs
should be recording a dependency on libc, but it's too late to do that.
Ok christos@.
This adds reachover Makefiles to build and install the atf tools.
Some are public, thus installed in /usr/bin, and others are internal
and therefore installed in /usr/libexec.
- remove unused write_str var
+ move header parsing to new function separation of parsing vs processing
+ alter some variable names to avoid confusion between header value and
header name (caused breakage with previous rev)
(i.e. free(): warning: modified (chunk-) pointer))
+ don't leak memory for ptr actually returned from strdup()
+ don't strdup() if we don't have to
+ don't break without free() if we did strdup()
+ as well as freeing h_value also free the header.
avoid wasting OS flag bits. In the future we'll probably use fileassoc to
achieve this (once there is a way to make fileassoc persistent) or in the
shorter term libelf, so that we can add and remove the note on demand instead
of burning bits on each binary. Of course since this is a tool, this means
that we'll need to think about how to handle libelf...
FORTIFY_SOURCE feature of libssp, thus checking the size of arguments to
various string and memory copy and set functions (as well as a few system
calls and other miscellany) where known at function entry. RedHat has
evidently built all "core system packages" with this option for some time.
This option should be used at the top of Makefiles (or Makefile.inc where
this is used for subdirectories) but after any setting of LIB.
This is only useful for userland code, and cannot be used in libc or in
any code which includes the libc internals, because it overrides certain
libc functions with macros. Some effort has been made to make USE_FORT=yes
work correctly for a full-system build by having the bsd.sys.mk logic
disable the feature where it should not be used (libc, libssp iteself,
the kernel) but no attempt has been made to build the entire system with
USE_FORT and doing so will doubtless expose numerous bugs and misfeatures.
Adjust the system build so that all programs and libraries that are setuid,
directly handle network data (including serial comm data), perform
authentication, or appear likely to have (or have a history of having)
data-driven bugs (e.g. file(1)) are built with USE_FORT=yes by default,
with the exception of libc, which cannot use USE_FORT and thus uses
only USE_SSP by default. Tested on i386 with no ill results; USE_FORT=no
per-directory or in a system build will disable if desired.
Some statistics:
base + COMBRELOC
$ time (for i in `seq 100`;do noatun --help>/dev/null;done)
(; for i in `seq 100`; do; noatun --help > /dev/null; done; ) 148.64s
user 4.82s system 99% cpu 2:33.93 total
base + DF caching:
$ time (for i in `seq 100`;do noatun --help>/dev/null;done)
(; for i in `seq 100`; do; noatun --help > /dev/null; done; ) 151.15s
user 5.53s system 99% cpu 2:37.23 total
base:
$ time (for i in `seq 100`;do noatun --help>/dev/null;done)
(; for i in `seq 100`; do; noatun --help > /dev/null; done; ) 492.36s user
5.34s system 99% cpu 8:19.17 total
bin/ls sources to libutil:
o Bump libutil minor version number
o Fix uses to include <util.h> to pick up the function definitions
o Fix most uses of flags_to_string() to release the now-malloc()ed result
for set{u,g}id binaries, so that in case they are playing with set{u,g}id
and exec'ing other binaries they don't get affected by the
LD_{PRELOAD,DEBUG,LIBRARY_PATH} environment setup. We leave LD_BIND_NOW alone.
There are no binaries affected in the base system.
