summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Blindly commit an entry for a 'Castlewood ORB2-E', which is a 2.2gigbillf1999-06-042-2/+12
| | | | | | removable disk. Submitted by: Troy Settle <st@i-Plus.net>
* Add the 3Dfx Voodoo Bansheemharo1999-06-041-1/+9
|
* Apply patch for PR #12008: remember to reset _yp_enabled flag inwpaul1999-06-041-0/+2
| | | | | | | endpwent() so we don't trip over a NULL db pointer later. PR: 12008 Submitted by: Valentin Netchayev <netch@lucky.net>
* Plug a mbuf leak in tcp_usr_send(). pru_send() routines are expectedpeter1999-06-042-8/+38
| | | | | | | | | | | | | | | | | | | | | | | | | to either enqueue or free their mbuf chains, but tcp_usr_send() was dropping them on the floor if the tcpcb/inpcb has been torn down in the middle of a send/write attempt. This has been responsible for a wide variety of mbuf leak patterns, ranging from slow gradual leakage to rather rapid exhaustion. This has been a problem since before 2.2 was branched and appears to have been fixed in rev 1.16 and lost in 1.23/1.28. Thanks to Jayanth Vijayaraghavan <jayanth@yahoo-inc.com> for checking (extensively) into this on a live production 2.2.x system and that it was the actual cause of the leak and looks like it fixes it. The machine in question was loosing (from memory) about 150 mbufs per hour under load and a change similar to this stopped it. (Don't blame Jayanth for this patch though) An alternative approach to this would be to recheck SS_CANTSENDMORE etc inside the splnet() right before calling pru_send() after all the potential sleeps, interrupts and delays have happened. However, this would mean exposing knowledge of the tcp stack's reset handling and removal of the pcb to the generic code. There are other things that call pru_send() directly though. Problem originally noted by: John Plevyak <jplevyak@inktomi.com>
* Remove a rather bogus search path reference..peter1999-06-033-3/+3
|
* GC unused prototypepeter1999-06-031-2/+1
|
* Quieten the bt_isa_probe() messages since they get a bit much whenpeter1999-06-033-9/+12
| | | | the isa probe has gone hunting for a card on it's own.
* remove references to isa_device, it's no longer associated with interrupts.peter1999-06-032-10/+2
|
* Check and warn about unknown devices in the config file along the samepeter1999-06-031-2/+12
| | | | | lines as the pseudo-device checking. Previously 'device xyz0' would have been silently accepted without comment.
* Remove VM86 option.kato1999-06-031-5/+4
|
* Sync with sys/i386/i386/machdep.c revision 1.339.kato1999-06-032-590/+1118
| | | | | | | | New function getmemsize_pc98 is added in this commit, since PC98 is quite different in obtaining memory size from IBM-PC. Many lines of this function is shareable with IBM-PC's getmemsize function, but sharing needs many #ifdef PC98 statements. Therefore, I gave up sharing code with IBM-PC's and just added new function.
* Fix some MP sequence number comparison bogons that are tickled bybrian1999-06-031-6/+8
| | | | | | | having different speed links in a bundle. This would manifest itself by having the link occasionally hang, but revive when a new connection is made.... Make ``show mp'' a bit prettier.
* Sync with sys/i386/conf/files.i386 and options.i386 revisions 1.246kato1999-06-034-8/+6
| | | | and 1.115, respectively.
* Mention that 8x8 font is required for 43, 50 and 60 line text modes,yokota1999-06-031-8/+57
| | | | and that font files can be specified in /etc/rc.conf[.local].
* Make it clear that PS/2 type mice will work only with "auto" andyokota1999-06-031-2/+14
| | | | | "ps/2" protocol types, and that serial mouse protocols shouldn't be specified for these mice.
* Fix PS/2 MouseMan+ protocol. We have been looking at a wrong placeyokota1999-06-033-6/+6
| | | | for the sign bit for roller movement!
* Use MACHINE_ARCH instead of MACHINE to check x86 architecture.kato1999-06-031-2/+2
| | | | Pointed out by: Akio Morita <amorita@meadow.scphys.kyoto-u.ac.jp>
* getvfsbyname() returns zero on success.ru1999-06-031-2/+2
| | | | | PR: 12000 Submitted by: Anatoly A. Orehovsky <tolik@mpeks.tomsk.su>
* Oops, quieten a compiler warning.brian1999-06-021-1/+2
|
* Mention physical and sync logging in the ``set log''brian1999-06-021-3/+3
| | | | usage message.
* Backed out the previous backout. We may as well have a version of ucmpdi2.cbde1999-06-021-2/+2
| | | | that actually compiles, and I use it for i386's with 64 bit longs.
* Fixed bitrot in synopsis (a missing include).bde1999-06-021-1/+2
|
* o Alter the mbuf type as it's processed by different layers.brian1999-06-0224-127/+241
| | | | | | | | o Show more information about missing MP fragments in ``show mp''. o Do away with mbuf_Log(). It was showing mbuf stats twice on receipt of LCP/CCP/IPCP packets.... ???!!? o Pre-allocate a bit extra when creating LQR packets to avoid having to allocate another mbuf in mbuf_Prepend().
* Back out last commit (did I get all of them?)newton1999-06-022-0/+6
| | | | Submitted by: Kazutaka YOKOTA <yokota@zodiac.mech.utsunomiya-u.ac.jp>
* Remove vinum from the build until Greg reviews phk's cdevsw changes.jb1999-06-021-2/+3
| | | | Preferred by: phk (rather than committing the patch without review).
* Spelling corrections for dummynet.ru1999-06-021-4/+5
| | | | Reviewed by: des,luigi
* Don't print backspaces in first column since they backup to the prev line.hoek1999-06-021-5/+12
|
* Introduce the ``keep-session'' option. Refer to the manbrian1999-06-026-18/+99
| | | | | | page for details. This allows MP over non-tty devices where the original ppp process must not exit (such as sshd-spawned ppp sessions).
* Unbreak memory sizing for SMP.jlemon1999-06-013-23/+33
|
* || vs && confusion in cdevsw_add().dt1999-06-011-2/+2
|
* Remove an unused variable.dt1999-06-012-4/+2
|
* cdevsw_add(&cn_devsw), not cdevsw_add(&dev).dt1999-06-011-3/+2
|
* Grok gtags too.hoek1999-06-017-42/+425
|
* Increase the length of an individual device name to LINE_LEN.brian1999-06-014-11/+15
| | | | | Adjust the base physical device name correctly after a link transfer (allowing correct multilink callbacks).
* Change the warning message issued if pci_map_port() fails. I think thiswpaul1999-06-011-5/+5
| | | | | | | | | | happens if you have a BIOS with a 'Plug & Play OS' setting and you leave it set to 'Yes.' This is wrong for FreeBSD (and LoseNT): it should be set to 'No.' Apparently it's still possible to map the iobase of the NIC and have the card work by reading the config space manually (which is what the driver does if pci_map_port() fails) but we need to warn the user to do fix their machine anyway. Anyway, warn the user to check the 'Plug & Play OS' setting in their BIOS if mapping the io space fails.
* Introduce the makebdev() function, it does the same as the makedev()phk1999-06-016-25/+33
| | | | function for now, but that will change.
* Null commit; note that there is a new memory sizing routine that usesjlemon1999-06-012-2/+2
| | | | | | | the BIOS calls to determine the memory configuration. This should fix problems with >64M for good. Reviewed by: Mike Smith
* Unifdef VM86.jlemon1999-06-0150-977/+812
| | | | Reviewed by: silence on on -current
* Make vm86 a standard componentjlemon1999-06-017-35/+12
| | | | Reviewed by: silence on on -current
* Unifdef VM86jlemon1999-06-014-25/+9
| | | | Reviewed by: silence on on -current
* Updated to support new Meteor/PPB cards (PCI-PCI Bridge).roger1999-06-011-0/+31
| | | | | | | | The driver now identifies the IBM PCI-PCI Bridge fitted to newer Matrox cards and initialises it. Sumitted by: Anton Berezin <tobez@plab.ku.dk> The Protein Laboratory, University of Copenhagen
* Shorten a detour around dev_t to get a udev_t created.phk1999-06-011-3/+2
|
* Make async/sync/physical/hdlc dumps prettier by showing printablebrian1999-06-011-12/+23
| | | | characters at the end of the line in hexdump style.
* Fixed locations of include files.kato1999-06-014-10/+10
|
* Get rid of segfaults in a set-uid program.ru1999-06-011-8/+15
| | | | | PR: 11823 Reviewed by: des
* Sync with sys/i386/isa/clock.c revision 1.136.kato1999-06-013-66/+72
|
* Damn, forgot this: time -> when.ru1999-06-011-2/+2
|
* Correct dumb mistake in previous commit.dfr1999-06-012-4/+4
|
* Sync header line with newsyslog(8).ru1999-06-011-3/+3
| | | | | PR: 11511 Reviewed by: des
* Ignore the quota.user and quota.group files rather than thebrian1999-06-011-2/+3
| | | | ``quotas'' file when cleaning /tmp.
OpenPOWER on IntegriCloud