summaryrefslogtreecommitdiff
path: root/libexec
AgeCommit message (Collapse)Author
2008-07-28Only optimize away plt relocation when obj_base == 0. CorrectsKurt Miller
pie relocation. Assembly debugging help from kettenis@. okay kettenis@ miod@
2008-07-24release the bind lock before resetting the signal maskPhilip Guenthe
ok kurt@
2008-07-16Fix several of the 'large offset' PLT entries. With help from kurt.Dale Rahn
ok kurt@
2008-07-11shut off an annoying gcc warning.Reyk Floeter
ok beck@
2008-07-11write a syslog debug message when a new GREY or TRAPPED entry is addedReyk Floeter
to the database. ok beck@
2008-07-11greyreader() shouldn't return, so tell us if it doesPeter Hessler
ok beck@, reyk@
2008-06-30Include file order must be "monitor.h" before "extern.h" otherwise enumAnders Magnusson
auth_ret is referenced before it is declared, which is not allowed by C99. Ok krw@, millert@, gilles@
2008-06-27Remove again the old _dl_md_reloc() calls-through-plt workaround: it isMiod Vallat
no longer necessary now that libgcc is built correctly, and it caused the elfbug regress test to fail.
2008-06-26First pass at removing clauses 3 and 4 from NetBSD licenses.Ray Lai
Not sure what's more surprising: how long it took for NetBSD to catch up to the rest of the BSDs (including UCB), or the amount of code that NetBSD has claimed for itself without attributing to the actual authors. OK deraadt@
2008-06-25Make ldd grok pie binaries.Kurt Miller
2008-06-13Pass relocated Elf_Dyn * to _dl_finalize_object(). Needed for pieKurt Miller
executables. Debugged and okay drahn@
2008-06-13Completing the executable elf_object_t data exposed a bug inKurt Miller
dl_iterate_phdr(3) which was returning the load base not the object base. Problem noted in the misc/exceptions regress by David Krause. okay drahn@
2008-06-12- rename _dl_boot argument loff to dyn_loff to differentiate it fromKurt Miller
exe_loff - calculate minva, maxva, exe_loff for the executable for passing to _dl_finialize_object and completing the executable elf_object_t data. this fixes ldd output for the exe object and prepares for PIE support. - calculate the interpreter path pointer in a way that allows for the possibility that PT_INTERP comes before PT_PHDR - fix an off by one in the init of dl_data input and okay drahn@
2008-06-08When prebind_newfile() is called the input file is replaced with the newDale Rahn
file, thus leaking the file descriptor, close the input file when it done.
2008-05-26sort flags on both the synopsis and usage.Igor Sobrado
ok jmc@
2008-05-24IPv6 support for standalone mode with assistance from millert@ and deraadt@Brad Smith
Tested by brad@ and sobrado@ ok deraadt@
2008-05-22version 2 of spamd sync protocol for two reasons:Theo de Raadt
1) ip addresses were accidentally being sent in host-byte order, which caused compatibility problems (spotted by jbg) 2) the sub-headers in the frame were not natively aligned, thus timeout values were incorrectly sent by 64-bit machines ok beck
2008-05-17documentation tweaks.Igor Sobrado
ok (some time ago) jmc@
2008-05-09As was done in dhcpd, insert pad blocks after sub-messages to keep theTheo de Raadt
structures aligned to 16 byte boundaries for maximum portability (somewhat similar to CMSG's). Old spamd's and new spamd's can still communicate with this change. It would be nice if someone tested spamd on 64-bit unaligned architectures (with and without this change) ok beck
2008-05-07check if the received buffer of a sync element is even big enough toReyk Floeter
hold the header with length field. ok deraadt@ beck@
2008-05-05Constantly fill in the program header pointer and count in elf_object_tKurt Miller
for all objects which simplifies phdr usage in a few places. "go for it" drahn@
2008-04-24compare strings with reality for expected behavior; ok millert@Todd T. Fries
2008-04-13Use arc4random_buf() when requesting more than a single word of outputDamien Miller
Use arc4random_uniform() when the desired random number upper bound is not a power of two ok deraadt@ millert@
2008-04-12yes, all; girishvenkatachalamTheo de Raadt
2008-04-11Correct a bug where _dl_error was set to DL_NOT_FOUND when a shared libKurt Miller
was found but failed to load. Adjust _dl_load_shlib() to only attempt to load the first matching shared lib which allows the correct _dl_error to propagate out upon load failure. okay drahn@
2008-04-09Improve support for shared libs linked at non-zero addreses:Kurt Miller
- rename private values in struct elf_object to better describe their meaning: s/load_offs/obj_base/ "object's address '0' base" s/load_addr/load_base/ "The base address of the loadable segments" - gdb needs the obj_base value so swap positions with load_base in struct elf_object - fix a few occurrences of where load_base was used instead of obj_base. With help and okay drahn@
2008-04-08should not call setpwent(); millert agreesTheo de Raadt
2008-04-03Use the libgcc version of these functions, no point in having extra copies.Dale Rahn
2008-04-02Use the proper define for this address, not a incorrect (on 64bit) define.Dale Rahn
ok kurt@
2008-03-28typo, ok beck@Alexander von Gernler
2008-03-24msg_controllen has to be CMSG_SPACE so that the kernel can account forTheo de Raadt
each cmsg_len (ie. msg_controllen = sum of CMSG_ALIGN(cmsg_len). This works now that kernel fd passing has been fixed to accept a bit of sloppiness because of this ABI repair. lots of discussion with kettenis
2008-03-15Repair more msg_controllen dealing with structures or arrays ofTheo de Raadt
descriptors; ok hshoexer, also looked at by kettenis and henning
2008-03-15Repair the simple cases for msg_controllen where it should just beTheo de Raadt
CMSG_SIZE(sizeof(int)), not sizeof(buffer) which may be larger because of alignment; ok kettenis hshoexer
2008-03-13Correct CMSG_SPACE and CMSG_LEN usage everywhere in the tree. Due toTheo de Raadt
an extensive discussion with otto, kettenis, millert, and hshoexer
2008-02-24Use rounded down address when changing memory protections on the GOT, suchDale Rahn
that we don't accidentally touch the page after the GOT. Bug found on sparc64 affected other platforms as well. ok kettenis@
2008-02-24Use rounded down address when changing memory protections on the GOT, suchMark Kettenis
that we don't accidentally touch the page after the GOT. ok drahn@
2008-02-09Remove an old workaround that was needed in SCCS times (%M% expanding to theMarc Balmer
filename). ok mikeb, millert
2008-01-08simplify the suggested pf ruleset; ok beckJason McIntyre
2008-01-02Update commented out kerberos bits in case someone wants to add missingTodd C. Miller
kerb5 support to login_chpass.
2007-12-27Add a missing length check of received spamd sync packet headers.Reyk Floeter
From Diego Giagio (dgiagio at gmail) with input from tedu@
2007-12-14Add radius-port login.conf variable to allow people to configure aTodd C. Miller
non-standard port name or number for use when connecting to radiusd.
2007-11-29- get synopsis/usage() rightJason McIntyre
- improve options text no one objected...
2007-11-27Fix dlsym w/special handles when called from an executable that has beenKurt Miller
linked without -E. dlsym issues reported by Pedro Martelletto <pedro at ambientworks.net>. okay tedu@
2007-11-27Make _dl_md_reloc_got() report errors, and take them into account.Miod Vallat
Fixes PR #5579.
2007-11-27tweak previous;Jason McIntyre
2007-11-27add missing header for inet_ntoa.Charles Longeau
"sure" deraadt@
2007-11-26add logging option, ok deraadt@Jolan Luff
2007-11-03PR 5621 - spamd doesn't clear the last entry from an empty TRAP list,Bob Beck
noticed and patch from Piotr Sikora <piotr@sikora.nu>
2007-09-26From Tamas TEVESZ:Ray Lai
- Use strcspn to strip \r and \n. OK chl.
2007-09-26From Tamas TEVESZ:Ray Lai
- atoi -> strtonum with proper bounds check - Use vfprintf and vsyslog in the appropriate places. OK millert.