summaryrefslogtreecommitdiffstats
path: root/sys
Commit message (Collapse)AuthorAgeFilesLines
* Change this to do what it should have done from the start.phk1995-07-121-3/+3
| | | | | Add argument for buffer for output. Fix manuals.
* Optimize the zeroing of quotes. b_to_q() was always taking about 20% ofbde1995-07-112-64/+92
| | | | | its time zeroing quotes although quotes are not very common. E.g., there are never any quotes in output queues.
* Rearrange the inner loop of scanc() to get better code on i*86'sbde1995-07-111-2/+10
| | | | | | | | | | | (on an i486, 10 cycles (+ cache misses) instead of 15). The change should be a no-op if the compiler is any good. The best possible i*86 code for the same algorithm is only 1 more cycle faster on i486's so I don't want to bother implementing an assembler version. scanc() is a bottleneck for OPOST processing. It is naturally about 4 times as slow as bcopy() on 32-bit systems.
* Fix races in scstart(). q_to_b() wasn't called at spltty(), so therebde1995-07-116-84/+42
| | | | | | | | | | | | | | were two races: - q_to_b() might unexpectedly return 0 (e.g, after a keyboard signal flushes the output queue and isn't echoed). ansi_put() interprets 0 bytes as 4GB... - more output (e.g. for echoes) might arrive afer q_to_b() returns 0. Then scstart() returns presumably and the new output might not be handled for a long time. Remove unused function scxint(). Fix prototypes (foo() isn't a prototype).
* Speed up the inner loop of ansi_put() by a few percent.bde1995-07-113-6/+30
| | | | | | | syscons' output is now only about 4-5 times slower than I want. It loses a factor of 2 for scrolling output by unnecessarily copying the screen buffer, a factor of 4/3 for dumb OPOST processing, and a factor of 3/2 for clist processing.
* Enable pcvt in LINT and don't generate a compile time error if sysconsbde1995-07-114-18/+18
| | | | | and pcvt are both configured when LINT is defined. There will be a link time error instead. This is to test building of pcvt more often.
* Release summary: (detailed descriptions in Edit History in matcd.c)jkh1995-07-113-68/+150
| | | | | | | | | | | | | | | | | | | | | | | | Adds support for non-Sound Blaster host adapters, including those distributed by Reveal, Lasermate, IBM, Media Vision, Crystal and others. The driver automatically senses the correct adapter type and you can have both in the system at the same time. (This change should eliminate a few complaints.) Corrected bit-masking problem that prevented use on SB Vibra-16 boards. Declared some internal data and functions static that should have been that way all along. Documentation changes reflect the new hardware support and change the appearance version to 2.0.5 (was 2.1). Nice and tidy. :-) Beta testers have verified functionality on SB16, Vibra-16, Media Vision and Reveal adapters. -Wall still shows no warnings. Frank Durda IV uhclem%nemesis@fw.ast.com Submitted by: Frank Durda IV <uhclem%nemesis@fw.ast.com>
* ICMP messages received from broken hosts which reply to multicast packetswollman1995-07-101-4/+8
| | | | | | | were mistakenly delivered, rather than getting thrown out, which caused substantial lossage. Submitted by: Bill Fenner <fenner@parc.xerox.com>
* tcp_input.c - keep track of how many times a route contained a cached rttwollman1995-07-104-17/+61
| | | | | | | | | or ssthresh that we were able to use tcp_var.h - declare tcpstat entries for above; declare tcp_{send,recv}space in_rmx.c - fill in the MTU and pipe sizes with the defaults TCP would have used anyway in the absence of values here
* When adding a route, set rt_ifa and rt_ifp a little earlier so thatwollman1995-07-101-4/+13
| | | | the protocol-specific add routine can examine it if desired.
* swapout_threads() -> swapout_procs().dg1995-07-103-6/+6
|
* Increased global RSS limit to total RAM.dg1995-07-101-2/+2
|
* Killed bogus sparc #ifdef.dg1995-07-101-10/+1
|
* Fixed panic that occurs on certain firewall rejected packets that wasdg1995-07-094-9/+8
| | | | | | | caused by dtom() being used on an mbuf cluster. The fix involves passing around the mbuf pointer. Submitted by: Bill Fenner
* PR #kern/572:joerg1995-07-091-6/+8
| | | | | | | | | | | >Synopsis: Booting w/scsi tape in drive causes first use to fail Booting with a tape in a SCSI tape drive will cause the first use of the tape to fail with the following message: st0: bad request, must be between 0 and 0. Submitted by: mpp@legarto.minn.net (Mike Pritchard)
* Move some struct definitions outside of struct's, so their scopes forjoerg1995-07-091-35/+40
| | | | | | C++ will match the scopes for C. Submitted by: Warner Losh
* Moved call to VOP_GETATTR() out of vnode_pager_alloc() and into the placesdg1995-07-097-54/+77
| | | | | | that call vnode_pager_alloc() so that a failure return can be dealt with. This fixes a panic seen on NFS clients when a file being opened is deleted on the server before the open completes.
* This implements enough of the BSDI extensions to the net-2 ogetkerninfo()peter1995-07-091-1/+112
| | | | | | | | | | | syscall to allow applications linked against their libc's uname() to work. Netscape 1.1N being a prime example, which prints: "uname() failed. cant tell what system we're running on". This change is a little ugly, but that's mainly because of the "interesting" semantics of the BSDI extension. Since ogetkerninfo() is only enabled by COMPAT_43, Netscape will only be affected on kernels with that option (eg: "GENERIC") Reviewed by: davidg
* Correct a typo in a comment.ats1995-07-081-2/+2
|
* Several constants have been missing in the SysV IPC header files thatjoerg1995-07-082-7/+14
| | | | | | | | are recommended by the literature. Fixes PR# misc/575: constants in sys/sem.h... Submitted by: daveho@infocom.com (Dave Hovemeyer)
* PR #kern/600: PPP does not pay attention to IPTOS_LOWDELAYjoerg1995-07-081-1/+10
| | | | | | | | | | Kernel PPP doesn't pay attention to IPTOS_LOWDELAY, but uses a table of port numbers, which isn't a generic method. The following patch fixes this (the table is still used, but in addition PPP queues the packet in fastq if IPTOS_LOWDELAY is set. Obtained from: Tatu Ylonen <ylo@cs.hut.fi> Submitted by: Heikki Suonsivu <hsu@clinet.fi>
* Improve negative usecount diagnostic a little.dg1995-07-082-8/+8
|
* Added missing splx() in DIAGNOSTIC code.dg1995-07-082-2/+4
| | | | Suggested by enami@sys.ptg.sony.co.jp.
* The generated VCALL always uses the first vp which in the case of /link/dg1995-07-074-9/+127
| | | | | | | | | might not be handled by the same FS as the directory (e.g. special device files)...so it must be special-cased. This bug is seen when doing "ln /dev/console /dev/foo" or equivilent and first appeared after I fixed the argument order of VOP_LINK. YUCK! There really needs to be a way of specifying what vp to use in the VCALL; doing this could fix the strategy and bwrite special-cases, too.
* Put back second case of CC handler in NCR script.se1995-07-071-3/+6
|
* Use a consistent blocksize for sizing bufs to avoid panicing the bio system.dfr1995-07-074-14/+46
|
* Worked around a bug with if.c setting the interface up even when we don'tdg1995-07-071-1/+14
| | | | want it to.
* Modified joerg's last change to only set the interface "up" when settingdg1995-07-061-4/+9
| | | | the address if the device is a SLIP device (i.e. "attached").
* Fixed an object allocation race condition that was causing a "objectdg1995-07-062-23/+46
| | | | | | deallocated too many times" panic when using NFS. Reviewed by: John Dyson
* Multiplex the soft tty interrupt some more to support the cy driver.bde1995-07-051-1/+5
| | | | | This should be configured better, perhaps by providing a software interrupt and mask bit to go with every hardware interrupt.
* Fix error logging:bde1995-07-053-12/+30
| | | | | - get the timeout countdown right - report everything before turning timeouts off.
* Rewrite:bde1995-07-057-3896/+6942
| | | | | | | | | | | - use pseudo-dma - provide the same features and interface as sio - support multiple boards - fix bugs. Some compile-time configuration constants are set to support higher speeds and Cyclom-16Y's at a 30% relative cost in efficiency. Cyclom-16Y support is untested.
* Killed "maxmem" declaration. We don't have that variable in FreeBSD.dg1995-07-051-2/+1
|
* Protected entire epioctl routine with splimp(). In this case, it is betterdg1995-07-052-2/+12
| | | | | | | | form to do this than it is relying on individual subroutines (the logic in epioctl is itself very minimal). Ideally, unnecessary splimp()'s should now be removed if they exist; I'll leave this for a later date (a complete code review of the driver needs to be done). Fixes a bug I noticed that would show up when ifconfig'ing the interface down.
* Add PCI Id for the 3940 controller.gibbs1995-07-041-2/+9
|
* Dan Eischen's serial eeprom code.gibbs1995-07-042-0/+227
|
* First pass cleanup of this driver. This pass does not include the sequencergibbs1995-07-044-71/+399
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | optimizations I have been working on yet, but does bring in some bug fixes and performance improvments that were easy to regression test: Setup the data fifo threshold and bus off timing correctly for 27/284x cards. Users of these adapters with fast periferals (greater than 5MB/s) will notice a big performance difference. (Sometimes as large as going from 3.7->8.3MB/s). Fix handling of the active target flags. Some of the outbs where missing the base offset in the abort code. The abort code still needs lots of work. Support 3940 controllers, but only with 16 SCBs for now. Eventually I'll add support for all 255, but I need to find a tester for the code first since we have to enable the cards external SRAM to do this. Add Dan Eischen's serial eeprom reading facilities. This allows the 2940 adapters to pull additional information left over from SCSI-Select right out out of the configuration seeprom. If the BIOS is disabled on 274x controllers, reset all target parameters to there defaults since you can't rely on what is stored in scratch ram. Report motherboard controllers as such. Stick the first SG address and count into the SCB data and count areas for all transfers in preparation of a later sequencer optimization. Keep track of which targets can are allowed to have the disconnection priveledge since this will be handled by the kernel driver in the future. If a target issues a message reject in response to a tagged message, disable tagged queuing for that target. Some seagates say they can do tagged queuing, but lie, and its a shame to have to disable tagged queuing on all devices just because you have one that can't cope.
* Add entry for i386/scsi/93cx6.c, the file that handles serial eepromgibbs1995-07-042-2/+4
| | | | | routines for the aic7xxx driver. If and when other drivers start to access similar serial eeproms, this file should probably be moved.
* Change SCB_LIST_NULL to 0xff from 0x10 to allow for 255 commands. Thisgibbs1995-07-041-7/+8
| | | | | | | | is needed for 3940 support. Have tagged commands look to see if a target is "busy" with a non tagged command before executing. This prevents overlapped tagged and non tagged commands which can happen since request sense commands are not tagged.
* Added some spaces for KNF. Moved some zero-initialized pointers into thedg1995-07-041-14/+11
| | | | kernel's .bss.
* This is the end result of about a dozen passes through this code to fixdg1995-07-041-729/+696
| | | | | | incorrect indents, a variety of poor coding practices such as comparing pointers to constants ('0'), poor code structuring, etc, etc. This brings the code up to the minimum standards for inclusion in FreeBSD.
* Define TRUE and FALSE.dg1995-07-041-0/+7
|
* 1) Removed bogus #includedg1995-07-041-46/+25
| | | | | | | 2) Rewrote "bad_packet" code to be less buggy and more readable. 3) Removed a pile of goto's; the code is now somewhat less reminiscent of a certain Italian pasta. 4) Changed all boolean returns of "0" and "1" to FALSE/TRUE.
* Slightly modify my previous change to return EINVAL instead ofjoerg1995-07-021-4/+2
| | | | | | EFAULT. Submitted by: Peter Wemm
* Revision 1.21 of if_sl.c broke the traditional behaviour thatjoerg1995-07-021-1/+4
| | | | | | assigning an address to an interface automatically marks this interface IFF_UP. The fix corrects this (and closes PR sys/577). This is consistent with the way ethernet interfaces are being handled.
* I saw a very low-key commit message on the netbsd mailing lists andjoerg1995-07-011-2/+4
| | | | | | | figured out what the problem was.. Anyway, I rate it as "highly serious". Submitted by: peter@haywire.DIALix.COM (Peter Wemm)
* The PCI config mechanism 1 test failed for the Intel Aries.se1995-06-306-12/+12
| | | | | | Make it less strict ... Submitted by: NIIMI Satoshi <sa2c@and.or.jp>
* Keep track of the number of samples through the srtt filter so that wewollman1995-06-295-21/+23
| | | | | | know better when to cache values in the route, rather than relying on a heuristic involving sequence numbers that broke when tcp_sendspace was increased to 16k.
* Removed "GATEWAY" consideration when calculating number of mbuf clusters.dg1995-06-292-10/+2
| | | | It now always uses the value that was used for the GATEWAY case.
* Killed "TIMEZONE" and "DST" options. They have been forced to 0 by configdg1995-06-293-7/+7
| | | | for more than a year now. Moved the declaration of 'tz' into kern_time.c.
OpenPOWER on IntegriCloud