summaryrefslogtreecommitdiff
path: root/lib
AgeCommit message (Collapse)Author
2018-02-06Do not call freeaddrinfo() with a NULL parameter.Alexander Bluhm
OK jsing@
2018-02-06make sure that all error paths are correctly handled in asr_run_sync()Eric Faurot
and that the result is always properly set when the functions returns. fix issues spotted by claudio@. ok claudio@
2018-02-06recognise DLT_PPP_SERIAL.David Gwynne
shlib version doesnt need a bump before no new symbols were added or removed. part of a larger diff that was ok sthen@
2018-02-05Constants which can go into .rodataTheo de Raadt
2018-02-05Do not bother NULLing pointers in memory that is freed immediately after.Joel Sing
2018-02-05Be consistent with the goto label names used in libtls code.Joel Sing
No change to generated assembly.
2018-02-03Simple USBPcap parser for tcpdump(8). Raw dumps can be nicely analysedMartin Pieuchot
in wireshark. ok deraadt@, dlg@
2018-01-30keep in sync with ld.so malloc.cOtto Moerbeek
2018-01-30word fix; from edgar pettijohnJason McIntyre
2018-01-28- An error in the multithreaded case could print the wrong function nameOtto Moerbeek
- Start with a full page of struct region_info's - Save an mprotect in the init code: allocate 3 pages with none and make the middle page r/w instead of a r/w allocation and two calls to make the guard pages none
2018-01-28Initialize variables to avoid compiler warningsKinichiro Inoguchi
ok jsing@
2018-01-27Complete the TLS extension handling rewrite for the server-side.Joel Sing
This removes ssl_parse_clienthello_tlsext() and allows the CBS to be passed all the way through from ssl3_get_client_hello(). The renegotation check gets pulled up into ssl3_get_client_hello() which is where other such checks exist. The TLS extension parsing now also ensures that we do not get duplicates of any known extensions (the old pre-rewrite code only did this for some extensions). ok inoguchi@
2018-01-27Clarify the comment re the F5 EC curves extension bug.Joel Sing
Also reference the knowledge base article instead of a discussion thread.
2018-01-27Convert ssl3_put_cipher_by_char() to CBB.Joel Sing
While here make the CBS usage in ssl3_get_cipher_by_char() more consistent with other code. ok inoguchi@
2018-01-26- do not junk pages returned by free_bytes(), all freed chunks are alreadyOtto Moerbeek
junked - freezero(): only clear requested size
2018-01-24Make the NEON codepaths conditional on __STRICT_ALIGNMENT not beingMark Kettenis
defined as they rely on unaligned access. ok joel@
2018-01-23use END_STRONG. Yes, everything in the sh ecosystem is different..Theo de Raadt
2018-01-21Implement ffs(3) using the CLZ instructions which has been available everMark Kettenis
since ARMv5. Should be much faster but more importantly it removes the data table from .text which could introduce unwanted ROP gadgets. Based on changes in Android/Bionic by Elliott Hughes. ok patrick@
2018-01-19Use static inline such that the functions are guaranteed to be present whenMark Kettenis
compiled with clang. ok visa@
2018-01-19Partial revert of rev. 1.7, fixes build with clangJeremie Courreges-Anglas
The error was: --8<-- ffs.S:57:2: error: unsupported relocation on symbol adrne r2, .L_ffs_table ^ -->8-- Minimal workaround suggested by deraadt@, ok guenther@ kettenis@
2018-01-18Wow, it cannot even be in .data...Theo de Raadt
2018-01-18Zap the rotor, it was a wrong idea. Cluebat applied by kshe whoOtto Moerbeek
came also up with this diff. Simple, no bias and benchmarks show the extra random calls disappear in te measurement noise.
2018-01-18Sorry, the PIC macros cannot reach to .rodata....Theo de Raadt
2018-01-18Move to ffs(3) for bitmask scanning. I played with this earlier,Otto Moerbeek
but at that time ffs function calls were generated instead of the compiler inlining the code. Now that ffs is marked protected in libc this is handled better. Thanks to kshe who prompted me to look at this again.
2018-01-18Instead of trying to handle ffs() with the normal rename-mark-hidden-and-aliasPhilip Guenther
dance, mark it protected. This works better for both gcc and clang: gcc blocks overriding of internal calls, while clang permits inlining again. ok otto@
2018-01-16On armv7, remove symbols that are also present in our libc. Because of theMark Kettenis
EABI aliases these can lead to duplicate symbol errors, especially when mixing code compiled with gcc and clang. The intention is to revert this once we convert armv7 to hardfloat. ok patrick@, guenther@
2018-01-14Add two more ARM EABI aliases that I missed in the previous libc minorMark Kettenis
bump. "just go ahead -- crank" deraadt@
2018-01-13tweak previous;Jason McIntyre
2018-01-13introduce a filter called EVFILT_DEVICE that can be used to notifyRobert Nagy
listeners of device state changes. currently only supports NOTE_CHANGE that will be used by drm(4) ok kettenis@
2018-01-12Adjust references for sysctl(3) to sysctl(2)Theo de Raadt
2018-01-12sysctl(3) can now be renamed to sysctl(2)Theo de Raadt
2018-01-08Move SC_* defines into <machine/setjmp.h>, and rename them,Visa Hankala
so that <machine/signal.h> is not needed in setjmp.S. Suggested by kettenis@ long ago, OK deraadt@
2018-01-08optimization and some cleanup; mostly from kshe (except the unmap() part)Otto Moerbeek
2018-01-08TIOCUCNTL is now allowed by promise "tty".Martin Pieuchot
Reminded by jmc@
2018-01-07amd64_{get,set}_fsbase() are superfluous and unused; delete them.Philip Guenther
Delete unused/never-implemented sysarch defines and structs while here. ports check and ok naddy@ ok deraadt@ mlarkin@
2018-01-07On OpenBSD/armv7 we deliberately trap unaligned access. UnfortunatelyMark Kettenis
the assembly code in libcrypto assumes unaligned access is allowed for ARMv7. Make these paths conditional on __STRICT_ALIGNMENT not being defined and define __STRICT_ALIGNMENT in arm_arch.h for OpenBSD. ok tom@
2010-10-01import OpenSSL-1.0.0aDamien Miller
2018-01-06Remove dead gettimeofday calls and timevals.cheloha
ok jca@
2018-01-02some grammar fixes; from dholland@netbsd, -r1.68Jason McIntyre
2018-01-02Stop assuming <sys/file.h> will pull in fcntl.h when _KERNEL is defined.Philip Guenther
ok millert@ sthen@
2018-01-01Fix wrong ETHERTYPE_IPV6denis
OK jca@ kettenis@ job@
2018-01-01Only init chunk_info once, plus some moving of code to group related functions.Otto Moerbeek
2017-12-27step one in avoiding unneccesary init of chunk_info;Otto Moerbeek
some cleanup; tested by sthen@ on a ports build
2017-12-26Update to compiler-rt 5.0.1.Patrick Wildt
ok kettenis@
2017-12-26Missed in previousJeremie Courreges-Anglas
Spotted by patrick@, thanks
2017-12-26Use a version script to limit exported symbolsJeremie Courreges-Anglas
ok ratchov@ kettenis@
2017-12-26Add ARM EABI runtime aliases to the GCC runtime functions that we includeMark Kettenis
in libc. ok patrick@, jsg@, guenther@
2017-12-25memcpy from the right place. at this point, the used variable is notTed Unangst
relevant. from Mark Karpilovskij. ok millert
2017-12-24Fix one possible buffer overflow and one underflow. Also some minorTodd C. Miller
cleanups. From Jan Kokemueller. OK deraadt@
2017-12-18Add call to fuse_remove_signal_handlers(3) in fuse_teardown(3) to do thehelg
opposite of fuse_setup(3). ok mpi@