Commit graph

21970 commits

Author SHA1 Message Date
Matt Jacob 81372ff0ed Steal 512KB more from system memory for heap instead of 256KB. We died
without the extra space. What a pity.
2000-09-18 08:19:04 +00:00
Matt Jacob 8c8c7c234b Run SECONDARY_LOAD_ADDRESS down to 0x2000a000- this still leaves
(barely) enough room for boot1 and gives us 8KB more to play with
in the loader.
2000-09-18 08:17:57 +00:00
Matt Jacob 4b3f16308b Make the stack 12K- we seem to need a bit more.
Rename 'stack' to 'stackbase' as this variable
more correctly denotes what it is.
2000-09-18 08:16:38 +00:00
Matthew N. Dodd 098b8a1eb0 Initialize 'hints_loaded' to 0.
This allows static hints to work properly.
2000-09-17 23:57:52 +00:00
Cameron Grant 341f16cc60 dynamify- probe codec for supported channels
add channel mappings for mixer
2000-09-17 23:51:00 +00:00
Cameron Grant 7dfc932548 fix warnings 2000-09-17 23:46:32 +00:00
David Greenman 7d854d93af Added a couple more missing FXP_SPLVAR()'s. 2000-09-17 23:23:22 +00:00
David Greenman b184b38e2b As a minor optimization, do suspended checking more like it was originally
in the PR - before the while loop.
2000-09-17 23:04:57 +00:00
David Greenman 04ea20fcf4 Added missing FXP_SPLVAR() to fxp_intr(). 2000-09-17 22:59:58 +00:00
David Greenman 2053b07d7e Attempt to replicate the new fxp SMP locking in the changes committed
in the previous (APM suspend/resume) commit.
2000-09-17 22:20:33 +00:00
David Greenman 7dced78a28 Added support for APM suspend/resume.
PR:		18756
Submitted by:	mike ryan <msr@elision.org>, with modifications by me.
2000-09-17 22:12:12 +00:00
Chuck Paterson 87807fded9 Add include of proc.h to make compile without SMP defined. This
change is likely interm, the include happens automagically
when SMP is defined.

Obtained from:	Jason Evans.
2000-09-17 22:01:21 +00:00
Dag-Erling Smørgrav 8461bdba85 Silence a warning. 2000-09-17 19:41:26 +00:00
Bruce Evans 21144e2fde Brought back schedsofttty() (SMPng casualty). Didn't bring back the
other schedsoft*() functions since they have never been used.

Removed confused comment about not needing these functions.  The
functions delay scheduling of SWIs until the next hardclock tick.
For devices that only deliver a few characters per interrupt, this
reduces the number of calls to the scheduler by a large factor (about
115 for each sio port at 115200 bps).
2000-09-17 16:37:30 +00:00
MIHIRA Sanpei Yoshiro 5935ef4c44 add PIOCSRESOURCE(IOC_GET_RESOURCE_RANGE)
Now /usr/sbin/pccardd read free resource(io,irq) range with
this ioctl.

Original Idea from:	PAO3
2000-09-17 15:37:53 +00:00
Bruce Evans df6bd679e7 Declare sigexit() as non-returning.
Fixed some style bugs.
2000-09-17 15:23:10 +00:00
Bruce Evans 33510ef17a Unpessimized CURSIG(). The fast path through CURSIG() was broken in
the 128-bit sigset_t changes by moving conditionally (rarely) executed
code to the beginning where it is always executed, and since this code
now involves 3 128-bit operations, the pessimization was relatively
large.  This change speeds up lmbench's pipe latency benchmark by
3.5%.

Fixed style bugs in CURSIG().
2000-09-17 15:12:04 +00:00
Bruce Evans fbbeeb6cd6 Uninlined CURSIG() and unpolluted <sys/signalvar.h>. CURSIG() had become
very bloated, first with 128-bit sigset_t's, then with locking in the
SMP case, then with locking in all cases.  The space bloat was probably
also time bloat, partly because the fast path through CURSIG() was
pessimized by the sigset_t changes.  This change speeds up lmbench's
pipe-based latency benchmark by 4% on a Celeron.  <sys/signalvar.h>
had become very polluted to support the bloat.
2000-09-17 14:28:33 +00:00
Kenjiro Cho e645a1ca27 change the evaluation order of the rsvp socket in rsvp_input()
in favor of the new-style per-vif socket.

this does not affect the behavior of the ISI rsvpd but allows
another rsvp implementation (e.g., KOM rsvp) to take advantage
of the new style for particular sockets while using the old style
for others.

in the future, rsvp supporn should be replaced by more generic
router-alert support.

PR:		kern/20984
Submitted by:	Martin Karsten <Martin.Karsten@KOM.tu-darmstadt.de>
Reviewed by:	kjc
2000-09-17 13:50:12 +00:00
Poul-Henning Kamp e4bdf25dc8 Properly jail UDP sockets. This is quite a bit more tricky than TCP.
This fixes a !root userland panic, and some cases where the wrong
interface was chosen for a jailed UDP socket.

PR:		20167, 19839, 20946
2000-09-17 13:35:42 +00:00
Poul-Henning Kamp 24b261c720 Reverse last commit, a better fix has been found. 2000-09-17 13:34:18 +00:00
Chuck Paterson 0f4dc94cfc Add locking to make able to run without the Giant lock being held. This
is enabling as all entries are still called with Giant being held.
Maintaining compatability with NetBSD makes what should be very simple
kinda ugly.

Reviewed by:	Jason Evans
2000-09-17 13:26:25 +00:00
Bruce Evans 621dbe43df Added used include of <sys/mutex.h> (don't depend on pollution in
<sys/signalvar.h>).
2000-09-17 12:20:49 +00:00
Poul-Henning Kamp e2cabba9d7 Make sure UDP sockets are explicitly bind(2)'ed [sic] before we connect(2)
them.

PR:     20946
Isolated by:    Aaron Gifford <agifford@infowest.com>
2000-09-17 11:34:33 +00:00
Boris Popov 3ff1a2f43e Add new flag PDIRUNLOCK to the component.cn_flags which should be set by
filesystem lookup() routine if it unlocks parent directory. This flag should
be carefully tracked by filesystems if they want to work properly with nullfs
and other stacked filesystems.

VFS takes advantage of this flag to perform symantically correct usage
of vrele() instead of vput() if parent directory already unlocked.

If filesystem fails to track this flag then previous codepath in VFS left
unchanged.

Convert UFS code to set PDIRUNLOCK flag if necessary. Other filesystmes will
be changed after some period of testing.

Reviewed in general by:	mckusick, dillon, adrian
Obtained from:	NetBSD
2000-09-17 07:26:42 +00:00
Don Lewis 69f15de9b1 Silence lame compiler warning.
Submitted by:	bde
2000-09-16 23:59:37 +00:00
Poul-Henning Kamp a2cdd9d93e Move SPECNAMELEN from <sys/conf.h> to <sys/param.h> 2000-09-16 21:57:16 +00:00
Justin T. Gibbs f175cbb663 Today is just not my day. Really get the right file. 2000-09-16 21:55:31 +00:00
Daniel C. Sobral 818c39998e Use _ instead or - where proper, according to the style I have been
using.

Overload "?" so it will also show loader.4th commands.
2000-09-16 21:04:49 +00:00
Justin T. Gibbs ae7c64e466 Pull the correct file over to freefall. 2000-09-16 20:59:12 +00:00
Poul-Henning Kamp 3a1c157588 I have no idea how a '#' became a '*'...
Spotted by:	Bernd Walter <ticso@cicely5.cicely.de>
2000-09-16 20:45:11 +00:00
Daniel C. Sobral df3c7d3993 Solve a name clash.
Add something to help debugging.
2000-09-16 20:20:44 +00:00
Justin T. Gibbs cb4d93cdbf Add PCIM_CMD_MWRICEN, the bit in the command register in PCI space
that enables memory write and invalidate cycles on a bus master.
2000-09-16 20:06:13 +00:00
Justin T. Gibbs 32da3127a3 Move aicasm to its own subdirectory. 2000-09-16 20:02:39 +00:00
Justin T. Gibbs 717d424718 Move aicasm to its own subdirectory.
Separate our platform independent hooks from core driver functionality
shared between platforms (FreeBSD and Linux at this time).

Add sequencer workarounds for several chip->chipset interactions.

Correct external SCB corruption problem on aic7895 based cards (3940AUW).

Lots of cleanups resulting from the port to another OS.
2000-09-16 20:02:28 +00:00
Justin T. Gibbs 0562034333 Update for aic7xxx driver reorganization. 2000-09-16 20:00:13 +00:00
Daniel C. Sobral c888e59b84 The default search path for kernel and modules was bogus. It makes no
sense for /boot/kernel to come last.
2000-09-16 19:59:41 +00:00
Daniel C. Sobral 54329571cf The module_path set by default was bogus. It had /boot/kernel last,
which makes little sense.
2000-09-16 19:56:23 +00:00
Daniel C. Sobral 7b9e034490 Both boot and boot-conf were using a different algorithm from the one
used by start to find the kernel. Fix this.

Also, boot would proceed immediately in the absence of a path as
argument. Check first if a kernel has already been loaded, and, if
not, fall back to load kernel&modules behavior.

Some further factorizing. I deem this code to be mostly readable by
now! :-)

Many thanks to: Makoto MATSUSHITA <matusita@jp.FreeBSD.org>
2000-09-16 19:49:52 +00:00
Poul-Henning Kamp 2404c3f99d Make LINT link.
cy driver is broken post SMPng.
2000-09-16 19:04:24 +00:00
Poul-Henning Kamp fa0c13fabb Add proper prototype for dkcksum(); 2000-09-16 18:57:51 +00:00
Poul-Henning Kamp a164712178 Remove SMP hack. 2000-09-16 18:57:18 +00:00
Poul-Henning Kamp c866ec47e3 Make LINT compile. 2000-09-16 18:55:05 +00:00
Poul-Henning Kamp 3045b21279 Make LINT compile. 2000-09-16 18:48:11 +00:00
Poul-Henning Kamp 574abd2926 Make LINT compile again. 2000-09-16 18:46:05 +00:00
Poul-Henning Kamp 1534d6d899 Remove a few leftover CLBYTES related bits.
This was in the good old days to patch up a logical pagesize on hardware
with stupid VM pagesize.  Vaxen I belive.
2000-09-16 18:28:24 +00:00
Poul-Henning Kamp ae2276e657 Remove a pointless casting of a gid_t to a gid_t. 2000-09-16 18:20:27 +00:00
Brian Somers 30fa52a6fa Call bpfattach() correctly from if_ppp.c
Submitted by:	Andy Adams <ala@merit.edu>
PR:		18506
2000-09-16 14:17:15 +00:00
Poul-Henning Kamp fc87418be0 Turn dkcksum() into an __inline function.
Change its type to u_int_16_t.
2000-09-16 13:43:00 +00:00
Poul-Henning Kamp c80d29139c Add canonical checks to devfs_setattr(). 2000-09-16 12:06:58 +00:00