summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2005-02-24fix a buffer-done race that makes mirroring work now;Michael Shalayeff
use pool instead of malloc/free and this fixes new buffer mode; implement DIOCGPDINFO for disklabel -d; tested by henning@ millert@ fries@
2005-02-24do not even ask for btlb on 8k cpus as pdc pretends there is some but then ↵Michael Shalayeff
it does not work; found w/ Moritz Kiese <mbk@mytum.de> on c180; add a debug print and fix pcxu entry
2005-02-24add wax@uturn; found on c180Michael Shalayeff
2005-02-24Add two missing breaks and suddenly using filter has no strange side effects.Claudio Jeker
Sometimes it is good to look at the rib output of a test box.
2005-02-24Close files after we are done with them; from mpech@Todd C. Miller
2005-02-24Remove the kroute nexthop specific code as there is no need for this inClaudio Jeker
ospfd. While there clean up some other no longer needed stuff. OK norby@
2005-02-24disable the SA dpd timer on sa_free(). this avoid a raceMarkus Friedl
between DPD and initial contact (double free); ok hshoexer
2005-02-24dpd_find_sa: only ready SAs, otherwise DPD will delete SAs owned byMarkus Friedl
an exchange; ok hshoexer
2005-02-24add carp(4) to SEE ALSO;Jason McIntyre
from freebsd -r1.4;
2005-02-24const correctness, string literals are const char *, no functional changeDaniel Hartmeier
from Joerg Sonnenberger joerg(at)britannica(dot)bec(dot)de
2005-02-24ok, last commit was partly wrong: an(4) uses BSS by default,Jason McIntyre
so adjust the surrounding text accordingly; verified by and ok dlg@
2005-02-24Add -N switch to select port for udpencap. Thus it's possible to run multipleHans-Joerg Hoexer
isakmpds on different ports specified with -p and -N. ok ho@
2005-02-24List some adapters known to work. ok damien@Jonathan Gray
2005-02-24Note the requirement of PCI 2.2 (3.3V signalling) on all PCI ral devicesJonathan Gray
we've come across so far. ok damien@
2005-02-24Store result of snprintf() in a signed type and check it for -1.Jonathan Gray
Do not depend on errno being set for snprintf, use errx() instead of err(). ok otto@, earlier version ok deraadt@ some time ago
2005-02-24remove docompress() completely in case of -DSMALL andMoritz Jodeit
replace it with a better error message. ok millert@ henning@
2005-02-24Add tedu's copy* test, turned into a regress test by me.Otto Moerbeek
2005-02-24fix pr4036, don't set p2->p_cpu to nullTed Unangst
ok deraadt miod markus
2005-02-24Document the 'source-track' stateful tracking option. Heavy influenceJoel Knight
from dhartmei and henning. ok dhartmei@ henning@ jmc@ jaredy@
2005-02-24Never respond to 0-length UDP packets. Reduces, but does not eliminateChad Loder
probability that isakmp service will be detected during port scans. OK hoexer@
2005-02-23use a string format argument instead of nothing in the exampleRobert Nagy
ok otto@
2005-02-23add all supported regulation domains. the forced COUNTRYCODE in ar5xxxReyk Floeter
is now disabled by default to use the value found in EEPROM. ok deraadt@
2005-02-23shrinkDavid Gwynne
ok drahn@
2005-02-23syncTheo de Raadt
2005-02-23palm zire 31; pr 4113, jamesw@bsdhosting.co.zaTheo de Raadt
2005-02-23do not permit a suspend if the lid is closed; uwe okTheo de Raadt
2005-02-23Make no distinction between standby and suspend requests for now.Uwe Stuehler
2005-02-23A healthy dose of 'die static'.Dale Rahn
2005-02-23Add missing copyright.Dale Rahn
2005-02-23fix copyright, use pxa2x0_gpio_intr_string()Dale Rahn
2005-02-23Move code from zaurus machdep to arm/xscale to live with the interruptDale Rahn
handler where it belongs code had been de-inlined to the wrong file.
2005-02-22Interrupt naming as provided by other interrupt handlers.Dale Rahn
2005-02-22more error checking on protocol calls when running in client modeJean-Francois Brousseau
2005-02-22use the file's name, and not its full path, when building the path toJean-Francois Brousseau
the RCS file in the repository spotted by pvalchev@
2005-02-22unused variableJean-Francois Brousseau
2005-02-22get rid of code that shouldn't have made it in thereJean-Francois Brousseau
2005-02-22Split the cvs daemon into two different processes: cvsd and cvsd-child.Jean-Francois Brousseau
The cvsd process only takes care of managing children and repository locks, while the child handles all requests made to access or modify the contents of the repository. This will allow us to take advantage of randomized library remapping and new propolice cookies, and minimizes the amount of code that can run with elevated privileges. Still work in progress. suggested by deraadt@
2005-02-22- Shift the brightness scale so that 0% still keeps the backlight on.Uwe Stuehler
- Separate wsconsctl display.backlight from display.brightness parameter. - Set brightness to approximately 40% on startup, since we cannot read the current brightness level yet. - Restore the current brightness on resume. ok drahn@
2005-02-22call zkbd_hinge() at attach time, to init sc_hinge; drahn okTheo de Raadt
2005-02-22- support -c and -s to retrieve the module listJean-Francois Brousseau
- silently accept most of the other options supported by checkout - do not abandon all hope if the `expand-modules' request fails
2005-02-22Be sure to attach the console keyboard, if any, before prompting for anMiod Vallat
alternate root device if boot -a.
2005-02-22Initial suspend/resume code with additional powerhooks. Enter/exitUwe Stuehler
suspend mode with power button or zzz. May not work for everyone yet. ok drahn@ and deraadt@
2005-02-22set the standard request socket path to /var/run/cvsd.sockJean-Francois Brousseau
from pvalchev@
2005-02-22fix wrong line break in comment.Moritz Jodeit
ok hshoexer@
2005-02-22build delete messages for isakmp sa correctly.Hans-Joerg Hoexer
ok and help ho@
2005-02-22add hme(4) RX TCP/UDP checksum offloadBrad Smith
Based on a diff posted to sparc@ by Pyun YongHyeon
2005-02-22typoAlexander Yurchenko
2005-02-22regenMichael Shalayeff
2005-02-22sv map fix; from Jan Johansson <janj+openbsd@wenf.org>Michael Shalayeff
2005-02-22based on hinge buttons, do screen blanking. uwe and i agree thatTheo de Raadt
later on we should look at debouncing the buttons, as the current code may lead to rapid display on/off cycles, which might be hard on the battery; uwe ok