Commit Graph

11242 Commits

Author SHA1 Message Date
Bruce Evans
bc9e7c3b42 Fixed double counting of runtime after a process exits. The last
timeslice of the exiting process was counted for both the exiting
process and the next process to run if the next process runs
immediately.

Broken in:	mostly in kern_clock.c rev.1.70 (1998/05/28)
1998-07-27 19:16:21 +00:00
Jonathan Lemon
a69546de82 Extend cpl workaround so that it applies when we are returning to
user-mode as well as vm86 mode.
1998-07-27 16:51:33 +00:00
Jonathan Lemon
f80a0a922c Re-arrange the page layout used by vm86_bioscall so that we can
potentially re-use the stack page.

Cosmetic cleanup of the code to de-obfuscate it and make it easier
to follow.  There should be no functional changes in this commit.
1998-07-27 16:45:05 +00:00
Bruce Evans
15b29aabe4 Made lazy syncing of timestamps for special files non-optional. 1998-07-27 15:37:00 +00:00
KATO Takenori
a0f5e52d43 Sync with sys/i386/isa/diskslice_machdep.c revision 1.30. 1998-07-27 09:49:23 +00:00
Doug Rabson
3260ced375 Macros for accessing alpha sparse device ports and memory. 1998-07-27 09:40:35 +00:00
Doug Rabson
86dd710847 Support the disabled keyword for isa devices. 1998-07-27 09:38:26 +00:00
David Greenman
01ddfa33e6 Only call m_reclaim() if M_WAIT since calling it from an interrupt can
cause problems.
PR:	7403
1998-07-27 03:59:48 +00:00
Alexander Langer
3f47ee5c4d A better solution to the rm_at_exit problem: Register the exit function
during first mount.  Unregister the exit function at last unmount.

Concept by:	sef
Reviewed by:	sef
Implemented by:	alex
1998-07-27 01:07:01 +00:00
Doug Rabson
976a1a2b57 Fix a stupid bug preventing gdb from seeing the right value of 'v0'. 1998-07-26 18:47:47 +00:00
Doug Rabson
56e7ede1c4 Notify pmap when a page is freed on the alpha to allow it to clean up
its emulated modified/referenced bits.
1998-07-26 18:15:20 +00:00
Doug Rabson
c99399cccd Fix a stupid bug in the new pmap_remove().
Arrange for soft modified/referenced flags to be cleared when a page is freed.
1998-07-26 18:13:28 +00:00
Bruce Evans
f69c53b019 Don't pass the label to diskerr(), since the label is being constructed
and may be invalid.  In particular, d_secpercyl may be 0, and diskerr()
divides by it.
1998-07-25 16:35:06 +00:00
Alexander Langer
ca2be56ff9 Override the default VFS LKM dispatch functions so that a module
unload function can be provided (this is necessary to unregister
the at_exit handler).
1998-07-25 15:52:44 +00:00
Alexander Langer
38bfd69bda Allow VFS LKMs to override the default module dispatch functions if
VFS_LKM_NO_DEFAULT_DISPATCH is defined.
1998-07-25 15:52:10 +00:00
Doug Rabson
4adac93d90 Make __weak_reference work on the alpha. 1998-07-25 14:37:37 +00:00
Matt Jacob
d8bb36bb8a minphys means something different in FreeBSD than in NetBSD- not needed here and in the way 1998-07-25 04:48:25 +00:00
Doug Rabson
72c36e1413 Fix pmap_enter_quick() to stop it hanging when the level 2 pte exists
but isn't valid.

Reimplement pmap_remove() to be much more efficient at removing large
stretches of addresses.

As part of reimplementing pmap_remove() fix pmap_protect() so that it stands
a hope of working.
1998-07-24 09:43:27 +00:00
Julian Elischer
9294f850e8 Don't return an error if the magic number is not present.
Failing to probe is not an error, just a negative result.
1998-07-22 19:37:02 +00:00
David Greenman
f3679e351a Improved pager input failure message. 1998-07-22 09:38:04 +00:00
Doug Rabson
e8669bcd43 Add a call to a platform-specific irq remapping function for alpha platforms
which don't record the correct irqs in PCI config space.

Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
1998-07-22 08:40:46 +00:00
Doug Rabson
df9efa9a8b On the alpha, ports may be allocated above 64k.
Change the port address argument to pci_map_port to pci_port_t* which is
defined as u_int on the alpha, u_short on i386.  This is a stopgap with a
hopefully limited lifetime.

Discussed with: Stefan Esser <se@freebsd.org>
1998-07-22 08:39:08 +00:00
Doug Rabson
0bf030847d Add some very simple support for a compiled in (from config(8)) resource
database.
1998-07-22 08:35:52 +00:00
Doug Rabson
b9ca15f807 Add ISA support.
Remove mcclock hack which was only needed because ISA wasn't done yet.
1998-07-22 08:34:19 +00:00
Doug Rabson
ec0bebbce7 Add some convenience functions for accessing device memory.
Add a DDB command to allow access to inb/inw/inl from the debugger.
1998-07-22 08:33:30 +00:00
Doug Rabson
53db2ca0b6 Merge KN20AA support from Andrew Gallatin.
Add support for device memory access.
Add support for ISA irqs.
1998-07-22 08:32:17 +00:00
Doug Rabson
31de2a7751 Move the mcclock from root to isa. 1998-07-22 08:30:25 +00:00
Doug Rabson
905f48d3af Add a simple ISA bus. No support for memory mapped devices or DMA as yet. 1998-07-22 08:29:26 +00:00
Doug Rabson
dd5f72e1fc Add sio support. 1998-07-22 08:28:24 +00:00
Doug Rabson
68f6efc27f Add platform functions for manipulating PCI irqs.
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
1998-07-22 08:27:56 +00:00
Doug Rabson
41fc8fe7b9 Add declaration of {aquire,release}_timer2(). 1998-07-22 08:26:23 +00:00
Doug Rabson
bb4752537d Add macros and chipset support for accessing device i/o memory on the alpha. 1998-07-22 08:25:39 +00:00
Doug Rabson
515cfabc48 Add KN20AA support.
Add entries for syscons.
1998-07-22 08:24:39 +00:00
Doug Rabson
237064085b Tweaks to stop config(8) warnings and to avoid an unfinished experiment of
mine.

Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
1998-07-22 08:23:26 +00:00
Doug Rabson
eb7262ea39 Support header for sysbeep() implementation. 1998-07-22 08:21:36 +00:00
Doug Rabson
e0fbfbb576 Make ptrace single steps over branches work as intended. 1998-07-22 08:20:57 +00:00
Doug Rabson
cbda18ce6e Add support for sio's soft interrupts. 1998-07-22 08:20:15 +00:00
Doug Rabson
e17e8bbe72 Support the DEC_KN20AA platform (AlphaStation 500 or 600).
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
1998-07-22 08:19:39 +00:00
Doug Rabson
bba326e18d Merge changes from Andrew Gallatin to support different CIA based platforms.
Add code to identify and attach either serial or VGA console as appropriate.
1998-07-22 08:18:34 +00:00
Doug Rabson
b50b3bda1d Add code from i386 to support sysbeep(). Doesn't work quite right but it
makes a noise.
1998-07-22 08:16:34 +00:00
Doug Rabson
e344c90c80 Change siocn{put,get}c to siogdb{put,get}c to allow a serial console on one
line with serial debugging on another line.
1998-07-22 08:15:34 +00:00
Poul-Henning Kamp
db7ac2451b There is a comment in vm_param.h which doesn't belong to the
code still left in there.  The macros it describes disapeared some-
time since 4.4BSD lite.

PR:		7246
Reviewed by:	phk
Submitted by:	Stefan Eggers <seggers@semyam.dinoco.de>
1998-07-22 06:21:55 +00:00
KATO Takenori
991b3bc3b3 Sync with sys/i386/i386/userconfig.c revision 1.105. 1998-07-21 12:07:49 +00:00
KATO Takenori
fd3ae2d6b0 Sync with sys/i386/conf/files.i386 revision 1.199. 1998-07-21 12:07:02 +00:00
KATO Takenori
649db6f0df Sync with sys/i386/isa/diskslice_machdep.c revision 1.29. 1998-07-21 12:06:04 +00:00
Mike Smith
b16d163da1 Add the 'cs' driver for Crystal Semiconductor CS89x0 devices. This
supports PnP and if_media.  I've been running a slightly older version
here for several weeks now.
Submitted by:	Maxim Bolotin <max@rsu.ru>
1998-07-20 20:00:43 +00:00
Mike Smith
2512c3e67b Add support for PCNet PCI chips that only work when we talk to them as ISA
devices. Specifically fix the case for the Hitachi version as used in
their VisionBook models.

Submitted by:	Ted Faber <faber@isi.edu>
1998-07-20 17:33:01 +00:00
Bruce Evans
7cb743ae28 Initialize more defaults for the in-core label for the whole disk.
Callers only need to initialize d_secperunit now, but should
initialize d_type (to reduce the IDE/SCSI confusion), d_typename
(put the disk model in it) and geometry info (if it isn't completely
ficticious).  Callers will soon need to initialize d_secsize.
1998-07-20 14:35:27 +00:00
Bruce Evans
f21b93a0f7 Cleaned up rev.1.39 - the shadowing variable should have just gone away. 1998-07-20 13:51:11 +00:00
Bruce Evans
92d1f65ed2 Moved allocation of the slices struct to the right place. Initialize
everything in it (the devsw pointers were not initialized early or at
all for the !DEVFS case, but this was harmless on i386's).
1998-07-20 13:39:45 +00:00