Commit graph

790 commits

Author SHA1 Message Date
Jordan K. Hubbard
3578b06c6a Fix a compilation-stopping typo; whoops, David - more sleep! :-)
Submitted by:	jkh
1994-08-08 16:45:11 +00:00
Garrett Wollman
57034e7407 Run-time configuration of VFS update interval. Old UPDATE_INTERVAL
configuration option is no longer supported.
1994-08-08 15:41:08 +00:00
David Greenman
e80859e4d6 Last commit for this file should have read:
Reduced maximum transfer size by one to allow for catching a too large
condition correctly. Do single block I/O if the size is too large.

Submitted by:	Bruce Evans
1994-08-08 13:56:46 +00:00
David Greenman
7d7bb69d41 Detect the "EOF" condition. Specifically, end of partition.
Submitted by:	Bruce Evans
1994-08-08 13:53:55 +00:00
David Greenman
8867d2f19c Added support for SIOCSIFMTU. 1994-08-08 13:33:16 +00:00
David Greenman
4e68ceaba0 Process tracing code. Written by Sean Eric Fagan.
Submitted by:	Sean Eric Fagan
1994-08-08 13:00:27 +00:00
David Greenman
90fd8c3866 Added ioctl support for SIOCSIFMTU. 1994-08-08 12:09:04 +00:00
David Greenman
660255695f On second thought, better restrict the mtu to between 72-65535...strange
things happen otherwise.
1994-08-08 11:43:44 +00:00
David Greenman
75ee03cb27 Enforce the mtu to between the range 1-65535 before calling the driver
ioctl routine.
1994-08-08 10:58:30 +00:00
David Greenman
a7028af7f0 Added ioctl support for SIOCGIFMTU and SIOCSIFMTU. These set the per-
interface MTU.
1994-08-08 10:49:26 +00:00
David Greenman
f847d045bf Take out multicast support for the moment. 1994-08-08 09:14:37 +00:00
David Greenman
3f6254c656 Increased ARG_MAX even further...to 64k bytes. 1994-08-08 09:12:43 +00:00
David Greenman
866dba7305 Changed B_AGE policy to work correctly in a world with relatively large
buffer caches. The old policy generally ended up caching nothing.
1994-08-08 09:11:44 +00:00
Garrett Wollman
4849be9c13 Define a sysctl MIB variable for the YP domain name. 1994-08-08 00:30:12 +00:00
Garrett Wollman
71919dabc4 Increase ARG_MAX so that `make clean' in src/lib/libc works again.
(Adding YP pushed it over the limit.)
1994-08-07 23:00:50 +00:00
Garrett Wollman
86b9a9cc2d Use the header files that are compatible with the code just moved over
from 1.1.5.
1994-08-07 18:41:02 +00:00
David Greenman
8339815ff9 Made pmap_kenter "TLB safe". ...and then removed all the pmap_updates that
are no longer needed because of this.
1994-08-07 14:53:27 +00:00
David Greenman
a481f20029 Provide support for upcoming merged VM/buffer cache, and fixed a few bugs
that haven't appeared to manifest themselves (yet).

Submitted by:	John Dyson
1994-08-07 13:10:43 +00:00
David Greenman
c564966bf0 Don't kremove process VM pages (oops!). This was the cause of the instability
that was introduced last night.

Submitted by:	 John Dyson
1994-08-07 03:31:52 +00:00
David Greenman
cf1344f6fc Implemented "fast" mbuf macros. a small number of mbufs are cached in
a linked list for fast allocation/free. Improves TCP performance by
about 20%.

Submitted by:	John Dyson
1994-08-06 11:26:16 +00:00
David Greenman
539d9ba017 Don't strip the kernel. 1994-08-06 10:52:15 +00:00
David Greenman
c87801fe54 Fixed various prototype problems with the pmap functions and the subsequent
problems that fixing them caused.
1994-08-06 10:25:50 +00:00
David Greenman
ae92ea446f Incorporated 1.1.5 improvements to the bounce buffer code (i.e. make it
actually work), and additionally improved it's performance via new pmap
routines and "pbuf" allocation policy.

Submitted by:	John Dyson
1994-08-06 09:20:56 +00:00
David Greenman
ed3f895438 Made the tmpstk start at tmpstk. Not doing so causes problems for the
debugger.

Submitted by:	John Dyson
1994-08-06 09:17:44 +00:00
David Greenman
16f62314cd Incorporated post 1.1.5 work from John Dyson. This includes performance
improvements via the new routines pmap_qenter/pmap_qremove and pmap_kenter/
pmap_kremove. These routine allow fast mapping of pages for those
architectures that have "normal" MMUs. Also included is a fix to the
pageout daemon to properly check a queue end condition.

Submitted by:	John Dyson
1994-08-06 09:15:42 +00:00
David Greenman
93f6448c49 Implemented support for the "ps_strings" structure (grrrr...) for use in
the userland library libkvm.
1994-08-06 09:06:31 +00:00
David Greenman
f720dc2ceb Enabled page table preloading of cached objects.
Submitted by:	John Dyson
1994-08-06 09:00:50 +00:00
David Greenman
0d2afceedd Process scheduling changes - adapted from FreeBSD 1.1.5. Basically,
charge scheduling CPU of child process to the parent and have child
inherit scheduling CPU from parent on fork. Makes a **big** difference
in the feel of the system to interactive users.

Submitted by:	John Dyson
1994-08-06 07:15:04 +00:00
David Greenman
ee4834308a Added my own kernel configuration file. This can be used as a place to
start for people trying to get 2.0 running.
1994-08-06 06:36:55 +00:00
Garrett Wollman
f5c789f5a9 Delete redundant #ifdef __i386__, be consistent about idempotency
protection.

Submitted by:	Bruce Evans
1994-08-05 14:36:04 +00:00
David Greenman
9671146f8b Hack to get netstat compiling...from 1.1.5. 1994-08-05 12:33:30 +00:00
David Greenman
a03460f16e Converted 'vmunix' to 'kernel'. 1994-08-05 09:28:55 +00:00
David Greenman
1e74887068 Renamed this thing, and change 'vmunix' to 'kernel' internally. 1994-08-05 09:25:31 +00:00
Garrett Wollman
dba7a33ecc Install RPC headers from include, like they always should have been. 1994-08-04 20:39:34 +00:00
David Greenman
86cadec5dc Added assembly versions of ffs() and bcmp(). 1994-08-04 19:51:01 +00:00
David Greenman
ee06dc6013 Inlined insque and remque. 1994-08-04 19:46:57 +00:00
David Greenman
ce921c90fd Fixed bug that would cause free memory reserves to be depleted and cause a
panic in some cases.
Submitted by:	John Dyson
1994-08-04 19:43:13 +00:00
David Greenman
d6a6c0f60c Added some code that was accidently left out early in the 1.x -> 2.0 VM
system conversion.
Submitted by:	John Dyson
1994-08-04 19:40:47 +00:00
Garrett Wollman
50f44fa07d Move ieeefp.h over, and put it in the correct subdirectory this time.
Submitted by:	Andrew Moore
1994-08-04 19:16:37 +00:00
David Greenman
0f8a0292a2 Updated to include improvements from FreeBSD 1.1.5. Fixed brokeness
with multicast support and BPF.
1994-08-04 17:42:35 +00:00
David Greenman
35b0a40abb Nuke redefinition of insque and remque. 1994-08-04 06:15:10 +00:00
David Greenman
15e1ca20e4 Nuked #if 0'd _insque and _remque routines - they are now inlined in
cpufunc.h.
1994-08-04 06:10:27 +00:00
David Greenman
b1b7658158 Made NFS attribute cache timeouts kernel config file tunable via
NFS_MINATTRTIMO and NFS_MAXATTRTIMO.
1994-08-04 06:03:46 +00:00
David Greenman
bbc0ec5284 Integrated VM system improvements/fixes from FreeBSD-1.1.5. 1994-08-04 03:06:48 +00:00
David Greenman
d319b932ae Changed occurrances of "itrunc" to "ffs_truncate" to make Bruce happy. 1994-08-03 08:19:35 +00:00
David Greenman
d23d07ef95 Merged in post-1.1.5 work done by myself and John Dyson. This includes:
me:
1) TLB flush optimization that effectively eliminates half of all of the
   TLB flushes. This works by only flushing the TLB when a page is "present"
   in memory (i.e. the valid bit is set in the page table entry). See section
   5.3.5 of the Intel 386 Programmer's Reference Manual.
2) The handling of "CMAP" has been improved to catch attempts at multiple
   simultaneous use.

John:
1) Added pmap_qenter/pmap_qremove functions for fast mapping of pages into
   the kernel. This is for future optimizations and support for the upcoming
   merged VM/buffer cache.

Reviewed by:	John Dyson
1994-08-03 02:45:30 +00:00
David Greenman
5bb6db475f Added MAP_FILE definition that does nothing - for backward source
compatibility.
1994-08-02 15:06:58 +00:00
David Greenman
09ce307fbe Completed (hopefully) the kernel support for old style "fastlinks". 1994-08-02 13:51:05 +00:00
David Greenman
3c4dd3568f Added $Id$ 1994-08-02 07:55:43 +00:00
David Greenman
b53902964f fixed bug where large amounts of unidirectional UDP traffic would fill
the interface output queue and further udp packets would be fragmented
and only partially sent - keeping the output queue full and jamming the
network, but not actually getting any real work done (because you can't
send just 'part' of a udp packet - if you fragment it, you must send
the whole thing). The fix involves adding a check to make sure that the
output queue has sufficient space for all of the fragments.
1994-08-01 12:01:45 +00:00