summaryrefslogtreecommitdiffstats
path: root/lib/libc/regex/regexec.c (unfollow)
Commit message (Collapse)AuthorFilesLines
2017-10-04as a result of previous change, vftype and ftype always have the samederaadt1-10/+9
value so simplify.
2017-10-046.2-current, back to workderaadt1-3/+3
2017-10-04Follow the pattern set by copy*/pcb_onfault: when xrstor faults, returnguenther4-13/+23
from the trap to a 'resume' address to effectively make xrstor_user() return an error indication, then do the FPU cleanup and trap generation from there where we can get access to the original, userspace trapframe. The original fix tried to handle the trap while on the wrong trapframe, leaking kernel addresses and possibly leading to double faults. Problem pointed out by abluhm@ ok deraadt@ mikeb@
2017-10-03Nested xrstor handled leaks a kernel address into sigval, but potentialderaadt1-1/+2
user address is unavailable so use 0 instead, since the typical cases are coredump or handing to a SIGBUS handler which cannot recover. Issue observed by bluhm ok bluhm, mlarkin, guenther said "Foo!"
2017-10-03The xrstor instruction will fault if the provided xstate data, whichguenther3-5/+26
is under userspace control via sigreturn, fails various consistency checks. Rather than trying to replicate the CPU's hardwired checks in C code, handle it like iretq: check in trap() whether a fault is from the problem instruction and handle it there. CPU behavior and the potential issue pointed out on Linux kernel-hardening ok mikeb@ deraadt@
2017-10-03Disable msi on the Yukon 88E8036 and 88E8053 as it triggers watchdog timeoutsajacoutot1-1/+9
rendering the network unusable. tested by myself committing on behalf of jsg@ ok deraadt@
2017-10-03work around config(8) modifying bss, using __attribute__ ((section(".data"))).deraadt1-2/+2
From yasuoka. I really want this fixed in a different way after release is finished.
2017-10-02Fix PermitOpen crash; spotted by benno@, ok dtucker@ deraadt@djm4-5/+21
2017-10-02Also ignore "Can't assign requested address" error when sendto(2)bluhm1-1/+2
to UDP loghost fails. Otherwise syslogd(8) would no longer send to this destination after the error occured once. tested by Rivo Nurges; OK millert@ benno@ deraadt@
2017-10-01set MAXSSIZ to the same value as on all other architectures (32MB)naddy2-4/+4
ok deraadt@
2017-10-01Align Rx buffers to prevent unaligned access in the network stack.kettenis1-2/+3
ok patrick@, deraadt@
2017-09-30openssh-7.6; ok deraadt@djm1-2/+2
2017-09-30The property that points the the attached PHY is called "phy-handle" insteadkettenis1-2/+2
of "phy". ok patrick@, deraadt@
2017-09-30Add some more boundary checks and prevent an attempt to divide by zerojsg1-1/+17
to resolve some additional crashes found by afl. ok mpi@ deraadt@
2017-09-30Add preliminary kabylake support to inteldrm(4) by backporting the relevantrobert24-156/+447
commits from linux-4.8.x. The changes are quiet minimal due to the fact that kabylake and skylake share most of the code because they are both gen9 graphics. This was tested by many and was also in snapshots for a while. ok kettenis@
2017-09-29When fault indicates PGEX_W, only tell uvm we need a writeable page..deraadt1-2/+2
don't request R+W. Issue observed by jsing in a go test of some sort. ok mlarkin
2017-09-29don't be silent on auto-allocation failure in write mode. ok millert@ krw@otto3-11/+15
2017-09-29update to unbound-1.6.6, ok floriansthen55-324/+1397
2017-09-29Rev. 1.2 enabled UTF-8 by default.schwarze1-0/+1
During subsequent Perl updates, all the documentation changes etc. got carried along, but the actual code change was deleted in Rev. 1.3 and never restored. Restore it now. Bug found by bentley@; OK afresh1@ bentley@.
2017-09-29Check that the end of sections do not exceed the filesize for bothjsg2-13/+18
symtab and sections. Corrects behaviour that led to crashes found via afl. ok mpi@
2017-09-29SPI values 0-255 are reserved, so change the example to use a valid one.phessler1-3/+3
fixes parsing of the example configuration OK benno@
2017-09-29New ddb(4) command: kill.mpi4-5/+36
Send an uncatchable SIGABRT to the process specified by the pid argument. Useful in case of CPU exhaustion to kill the DoSing process and generate a core for later inspection. ok phessler@, visa@, kettenis@, miod@
2017-09-29Do not segfault when the string table is invalid or not present.mpi1-13/+34
Based on a diff from jsg@. Found independently by jsg@ with afl and markj@FreeBSD with a port.
2017-09-29Mention that U-Boot images for Allwinner and Rockchip targets contain ajsg2-4/+6
default DTB.
2017-09-29With arm64 targets U-Boot looks for dtbs in a directory named after thejsg1-3/+5
vendor of the SoC.
2017-09-28Correct the timeout used when select-timeout iskrw1-4/+5
set to a non-zero value in dhclient.conf(5). Fixes the bsd.rd upgrade issue reported by Eivinde Eide via misc@.
2017-09-28make clang include a FILE symbol for .(s|S) filesrobert1-1/+6
This is mostly needed by syspatch at the moment to be to be able to re-link in the same order as the original libraries were linked with by relying on the readelf(1) and without this .(s|S) assembly files were not getting a file directive. A bug reports has been filed as well: https://bugs.llvm.org/show_bug.cgi?id=34019 ok deraadt@
2017-09-28The DELAY() recently added to iwn(4) to make association more reliable couldstsp1-5/+8
be triggered by incoming management frames, such as DEAUTH frames. As krw@ found out, a system receiving many such frames would become unresponsive. Ensure this extra DELAY() only happens if we decide to send an AUTH frame. ok mpi@ krw@
2017-09-28When option DDB_STRUCTINFO was removed from the kernel, too much frompatrick14-14/+168
the kernel makefiles was removed which lead to C files not being re- compiled when a dependency changed. This can lead to panics in odd places and generally a broken system. ok deraadt@
2017-09-27config -e's adds new devices to an extra locator array, which is accountedderaadt4-39/+43
using a variable which started at 0 and hence got planted in .bss. Refactor that variable to begin at the maximum and decrease, such that it lands in .data instead. ok jsg
2017-09-27Document how ioctl(2) LIOCSFD on /dev/klog registers a socket pairbluhm2-3/+21
to receive sendsyslog(2) messages. discussed with martijn@; OK jmc@ deraadt@
2017-09-27Unserstand DW_OP_bregN locations.mpi1-1/+10
Sync with readdwarf(1).
2017-09-27tweak EposeAuthinfo;jmc1-3/+5
diff from lars nooden tweaked by sthen; ok djm dtucker
2017-09-27guenther sleep-commited the version without #ifdefsderaadt2-2/+6
2017-09-27Xr; from jan staryjmc1-2/+3
2017-09-27amd64 needs FS.base values (the TCB pointer) to be validated, as noncanonicalguenther4-5/+15
addresses will cause a fault on load by the kernel. Problem observed by Maxime Villard ok kettenis@ deraadt@
2017-09-26UsePrivilegeSeparation is gone, stop trying to test it.dtucker1-18/+12
2017-09-26Prevent null pointer dereference when probing channelsmikeb1-3/+3
Account for the case when wdc is attached to the ISA bus and performs channel probing using a dummy structure that lacks the back pointer to the controller's softc. Bug reported and fix tested by Andrew Daugherity, thanks! OK phessler, jsg, krw, deraadt
2017-09-26- pfctl always prints warning when flushes rulesetsashan1-2/+2
OK mikeb@
2017-09-26bump version in advance of final releasebcook1-3/+3
2017-09-26Ignore file offset values in section headers that exceed the length ofjsg2-8/+11
the file. Avoids a crash found with afl. ok mpi@
2017-09-26Return the correct error value in generate().mpi1-3/+3
From Mark Johnston, markj@FreeBSD
2017-09-26Use quoted #includes for files kdump's mkioctls indirectly includes sojsg2-3/+3
mkioctls can find the files and not error out as drm headers are not installed to /usr/include. This results in more complete ioctl coverage in kdump (not just for drm). ok kettenis@
2017-09-26disable POOL_DEBUG for releasejsg1-2/+2
ok deraadt@
2017-09-26bump wo 2.6.2bcook1-3/+3
2017-09-25sendsyslog should take a const char * everywhere.espie5-11/+11
okay bluhm@, deraadt@
2017-09-25Calling waitpid(pid,...) where pid is either uninitialized orkrw1-5/+7
-1 because vflork() failed is bad. Initialize pid to -1 and call waitpid() only when pid != -1. Uninitialized use of pid found by clang. Suggestion of -1 from millert@. ok millert@
2017-09-25the Alea I works fine with urng(4) so update the manpage accordinglyjasper1-7/+3
2017-09-25If tls_config_parse_protocols() is called with a NULL pointer, return thejsing1-1/+4
default protocols instead of crashing - this makes the behaviour more useful and mirrors what we already do in tls_config_set_ciphers() et al.
2017-09-25Annotate some API-side memory leaks for future resolution.jsing1-1/+3
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.