summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2008-01-29When relayd(8) accepts control it creates `connfd' descriptor with acceptReyk Floeter
function. Close it when ctl_conn allocation fails. From Igor Zinovik (zinovik at cs.karelia dot ru)
2008-01-29More read/write result checking fixes to avoid unsigned comparisons vsKenneth R Westerback
-1. ok henning@ beck@ ray@
2008-01-29There is no need to xstrdup() optarg.Tobias Stoeckmann
2008-01-29Fixed function name in fatal calls. While at it, adjusted style.Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-29Properly free dynamically allocated memory on error path.Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-29Properly free dynamically allocated memory. Also skip needless checks.Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-29Properly check if an argument has been supplied for "Removed"Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-29Add support for VIA VT8237S SATA.Jonathan Gray
From Antonio Marques <froz@icix.org> in PR 5718.
2008-01-29regenJonathan Gray
2008-01-29Free resources now that we don't always exit after an unsuccessfulPierre-Yves Ritschard
parsing of the configuration. From Igor Zinovik <zinovik at cs.karelia.ru> ok thib@ and me.
2008-01-29Add VIA VT8237S SATA.Jonathan Gray
From Antonio Marques <froz@icix.org> in PR 5718.
2008-01-29use printf instead of echo, which is more portable;Jason McIntyre
issue noted by Matthew Szudzik ok marco
2008-01-29hosts files may contain ipv6 addresses too; update proposed byJason McIntyre
Bruce Cran in freebsd pr docs/120024; ok claudio
2008-01-29fix race in makefileMarc Espie
2008-01-29fix another make -j race, avoid rebuilding some archives twice in parallel.Marc Espie
okay millert@, went thru several full builds including >500 perl packages...
2008-01-29fix Makefile racesMarc Espie
2008-01-28Actually support checkout's -l.Tobias Stoeckmann
OK xsa@
2008-01-28Properly free "rev".Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-28Zapped some variables which made the source harder to read (and to verify).Tobias Stoeckmann
> Diff from Igor Zinovik
2008-01-28checkout's options -m and -F are mutually exclusive. Also watch out toTobias Stoeckmann
properly free arguments if options are supplied multiple times. OK ray@ > Diff from Igor Zinovik
2008-01-28spelling in comment; ok mbalmer@ jmc@Kevin Steves
2008-01-28Properly handle -R -- although it is default, it may be overwritten mit -lTobias Stoeckmann
in front of it. OK joris@, xsa@ > Diff from Igor Zinovik
2008-01-28Properly align stack such that code that uses SSE2 instructions doesn't crash.Mark Kettenis
ok espie@, beck@
2008-01-28Reconnect a client socket after three consecutive send failures.Marco Pfatschbacher
This allows recovery after an IP address change (e.g. on dialup links). Also move the update of "nextaction" timeout below the deadline check. OK henning@
2008-01-27Remove outdated sections about link0 option.Mark Kettenis
2008-01-27Add TODO's to comment.Marcus Glocker
2008-01-27Oops forgot one function when adding 16 byte scsi commands.Marco Peereboom
2008-01-27no hostap support for this device;Jason McIntyre
2008-01-27minor tweak to CAVEATS;Jason McIntyre
2008-01-26the kids want I/O;Jason McIntyre
2008-01-26- shuffle the page a little, to make it read betterJason McIntyre
- prefer I/O to IO
2008-01-26Remove duplicate protosMarco Peereboom
2008-01-26Fix some debug printsMarco Peereboom
2008-01-26regenClaudio Jeker
2008-01-26Add Radeon HD 2600 IDs. Diff by brad@ tested by myself.Claudio Jeker
2008-01-26document the supported disciplines in a little more detail;Jason McIntyre
from marco and myself
2008-01-26replace reference to deprecated @pkgcfl keyword by @conflictLandry Breuil
ok espie@
2008-01-26getuint() doesn't use its partno argument. So eliminate it. NoKenneth R Westerback
functional change. Noticed by lint.
2008-01-26Create the automatic tables at the base of the anchor stack rather thanRyan Thomas McBride
in the inline anchor. Fixes optimizer bug where automatic table creation in inline anchors fails because rules are now loaded after optimization and no transaction has been opened for the anchor. bug reported by Henrik Johansen ok henning dhartmei
2008-01-26Unify i386 and amd64 lapic code, and calibrate lapic timer with interruptsMark Kettenis
disabled (as suggested by mickey). ok krw@, marco@
2008-01-26- use command prompts for commandsJason McIntyre
- a few other minor fixes ok marco
2008-01-26Set the return value instead of returning right away so that the DMABrad Smith
memory is actually synched. ok dlg@
2008-01-26remove a bad space; from he@netbsdJason McIntyre
2008-01-26SAS1078 is a megaraid, not a fusion-mpt, so dont attach to itDavid Gwynne
2008-01-26sync dma mem back to the cpu on poll failure.David Gwynne
from mickey via pr5714
2008-01-26despatch -> dispatchDavid Gwynne
disparate spellink fix from mickey
2008-01-26newline when printing about a failure to attach.David Gwynne
2008-01-25Add support for some other CPU's. Thank you toTobias Weingartner
jahrens at centtech for the information. ok otto@, hshoexer@, beck@
2008-01-25Fix my screwup and don't return NULL from sort_paritions.Kenneth R Westerback
Problem found by ckuethe@, slightly different fix than the one proposed by otto@ applied.
2008-01-25"read(..., ..., sizeof Y) < sizeof Y" is a dangerous idiom because itKenneth R Westerback
does an unsigned comparison and read() can return -1. Use '!=' instead of '<' since read() can't return more than 'sizeof Y'. Not perfect (that would require a separate test for -1) but a very common usage. We don't actually compile this code so there is no functional change. Diff fixed & ok millert@