summaryrefslogtreecommitdiff
path: root/sys/kern
AgeCommit message (Collapse)Author
1998-02-05rename ddb.panic_ddb, and add ddb.console. Now you can stop console ddb entryTheo de Raadt
with a sysctl. There will be architectures and drivers that lack function, and I trust the maintainers of those will forget to add the code..
1998-02-03bad types; wileyc@sekiya.twics.co.jpTheo de Raadt
1998-01-24NULL derefNiklas Hallqvist
1998-01-21support expanded panicstr; and add vsprintf backTheo de Raadt
1998-01-21Prevent a NULL-deref in the autoconf_verbose caseNiklas Hallqvist
1998-01-20typoNiklas Hallqvist
1998-01-20Provide a way to debug probes. set autoconf_verbose via DDB, orNiklas Hallqvist
option AUTOCONF_VERBOSE=1 in your kernel config.
1998-01-15Re-enabled write protection on read-only file system mounts. Thanks toConstantine Sapuntzakis
Niels for pointing this one out.
1998-01-11Fix a couple spinlock references. More code motion in vfs_subr.cConstantine Sapuntzakis
1998-01-10A couple more splbio()s in vfs_bio plus moving around a couple functions.Constantine Sapuntzakis
1998-01-10Broke up vfs_subr.c which was getting a bit huge. We now have seperate filesConstantine Sapuntzakis
for the syncer daemon as well as default VOP_*.
1998-01-09core dumps disabled for SUID programs is default again.Constantine Sapuntzakis
1998-01-09Get rid of dumping across symlinks. Races can cause this to be problematicConstantine Sapuntzakis
from a security standpoint (i.e. arbitrary files on the FS can be wiped out).
1998-01-09A better fix for the mkdir ("path/") bug. This fix strips the trailing slashesConstantine Sapuntzakis
(except in the case where the path is just /////) in the following three cases: 1) The path in mkdir 2) The destination path in rename if the source was a directory 3) The destination path in link if the source was a directory Note #3 isn't strictly necessary since most of our file systems don't support hard links of directories anyway.
1998-01-09Back out changes to lookup( ). This breaks a whole bunch of other systemConstantine Sapuntzakis
calls (like open, rename, link, symlink). Problem should probably be fixed other ways.
1998-01-09Fix for mkdir ("path/") problemConstantine Sapuntzakis
1998-01-08Soft updates bug fix: Set the flags on the buffer describing our intentionsConstantine Sapuntzakis
before we call bioops.io_start. However, don't move buffer memory to parent until bioops.io_start has had a chance to do its thing (otherwise, io_start will be very disappointed went it tries to read the buffer :) Thanks to Todd T. Fries for finding this one!
1998-01-06so_linger is in secondsTheo de Raadt
1998-01-04indentTheo de Raadt
1998-01-04Added support for enable/disable attr value. eg disable irq 10. -mojMats O Jansson
1998-01-02make {f,}chflags(.., -1) return EINVAL because -1 is VNOVAL; problem noted ↵Theo de Raadt
by provos
1998-01-02syncTheo de Raadt
1998-01-02merge recent netbsd changesTheo de Raadt
1998-01-02make mmap() return void *, add MAP_FAILEDTheo de Raadt
1997-12-30make it compile w/ GPROFMichael Shalayeff
1997-12-29sysctl ddb.panic_ddb; felix@mamba.pond.sub.orgTheo de Raadt
1997-12-28Do not lose the first char in msgbufNiklas Hallqvist
1997-12-28I should be more careful about what I check in. Undo earlier commenting-outConstantine Sapuntzakis
of security check. Restricted access to /dev/memat higher secure levels now enabled.
1997-12-26Get rid of string literals for file system names and replace withConstantine Sapuntzakis
MOUNT_* macros. Duhhh.
1997-12-16evaluate length before doing compat stuff which might overwrite theNiels Provos
length field we are going to evaluate. From FreeBSD, originally submitted by: pb@fasterix.freenix.org (Pierre Beyssac)
1997-12-12Fixed spelling in a comment.gene
1997-12-11do not clear setuid/setgid for *chmod(-1,-1); pointed out by provosTheo de Raadt
1997-12-10extra ()Theo de Raadt
1997-12-10Remove special case for O_TRUNC that is irelevant now thatTodd C. Miller
we disallow O_TRUNC with O_RDONLY.
1997-12-10try again; proactively check that O_TRUNC has FWRITETheo de Raadt
1997-12-10only permit O_TRUNC if FWRITE is set; chatted about by millert and myself ↵Theo de Raadt
for nearly an hour
1997-12-09fix O_TRUNC and O_{SH,EX}LOCK properly; problem found by provosTheo de Raadt
1997-12-08add kern.nosuidcoredump to lock things out even moreTheo de Raadt
1997-11-30fix TIOCGPGRP in sunos emulationTheo de Raadt
1997-11-24Fix non-DIAGNOSTIC (and non-COMPAT*) compilationNiklas Hallqvist
1997-11-24fix memory leak; hbriceno@lcs.mit.eduTheo de Raadt
1997-11-23indentMichael Shalayeff
1997-11-23nuke prehistoric disk statisticsMichael Shalayeff
1997-11-20Fix for open(2) when O_TRUNC and O_??LOCK are used together. ITodd C. Miller
ran the fix by Kirk, who said it was an OK solution and pointed out two problems.
1997-11-18Use 0177777, not 07777 as the mask to check for EINVAL. This allowsTodd C. Miller
one to use the st_mode field in struct stat as an argument to f?chmod().
1997-11-17kleink: When fstat(2)ing a file descriptor of a local communications domainTheo de Raadt
socket, fill the socket's creation time into the stat structure's st_[acm]time fields: POSIX requires this behavior for pipe(2). N.B.: updating the st_[am]time fields when reading/writing the pipe is neither required nor implemented, though.
1997-11-17only consider 9 bits for the umaskTheo de Raadt
1997-11-17make chmod() and fchmod() return EINVAL for bad mode bitsTheo de Raadt
1997-11-17indentTheo de Raadt
1997-11-15for shutdown(2), if "how" is not 0-2, return EINVALTheo de Raadt