Commit Graph

1196 Commits

Author SHA1 Message Date
thorpej 673ddc8f39 Style nits in last. 1999-03-01 22:24:53 +00:00
scottr d32ed292af defopt BUFCACHE and BUFPAGES. 1999-02-27 06:39:34 +00:00
is f71d843a65 synchronize types, and s/curproc/p/ in one forgotten place 1999-02-26 22:37:57 +00:00
is 84971e3a6f mac68k part of pr 6152 fix, and missing bits for the others 1999-02-25 23:13:39 +00:00
is 75469a775e hp300 part of the fix for pr 6152 1999-02-25 22:34:41 +00:00
lukem dcab0210a0 convert from NOxxx= to MKxxx=no.
include <bsd.own.mk> if testing a MKxxx variable.
1999-02-13 02:54:17 +00:00
kleink f69591d12d Use of casts as lvalues is a GNU C extension; rearrange slightly. 1999-02-10 17:03:26 +00:00
he f96a373094 Avoid "unterminated character constant" error from mkdep (really cc).
I guess in some sense it's stupid to use cc to depend .S files.
1999-02-08 10:45:11 +00:00
thorpej ad268a2ff8 STANDALONE -> _STANDALONE. 1999-02-06 21:19:24 +00:00
thorpej 3bf59302e9 Workaround for disks that respond slowly to selection, provided by
David Jones in PR #3769.

Every day, I hate this driver more and more.
1999-02-06 03:30:32 +00:00
thorpej e6c245c5df Implement pmap_prefer() for systems w/ an HP MMU (which has a virtually-tagged
cache).  This helps to avoid cache alias problems, and can improve performance
in the case where physical pages have multiple mappings (since the pages will
not have to be marked cache-inhibited).
1999-02-02 21:06:55 +00:00
thorpej f108f0164e Nuke the "pmapvacflush" debugging code. 1999-02-02 20:52:21 +00:00
mycroft 3e6e3322d8 Format consistency. 1999-01-24 12:56:51 +00:00
mycroft 082204fa3d Standardize format. 1999-01-22 14:12:07 +00:00
thorpej 2fb041ce0a No need for <sys/mtio.h> 1999-01-19 18:18:41 +00:00
chuck 32244a5a1a MNN is now the default 1999-01-16 20:13:17 +00:00
thorpej 8922647c58 Some minor, mostly costmetic, changes to CPPFLAGS/CFLAGS. 1999-01-15 23:37:05 +00:00
thorpej c84a74b16b Don't define "mc68020". Nothing uses it. 1999-01-15 23:21:25 +00:00
bouyer dc306354b0 Move the bswap functions from libutil to libc (this bups the
minor of libc and the major of libutil). For little-endian architectures
merge the bnswap() assembly versions with nto* and hton* using symbols
aliasing. Use symbol renaming for the bswap function in this case to avoid
namespace pollution.
Declare bswap* in machine/bswap.h, not machine/endian.h. For little-endian
machines, common code for inline macros go in machine/byte_swap.h
Sync libkern with libc.
Adjust #include in kernel sources for machine/bswap.h.
1999-01-15 13:31:15 +00:00
thorpej e598335d1c Garbage-collect `mbutl'. 1999-01-09 22:10:12 +00:00
thorpej e279508ccb In pmap_zero_page() and pmap_copy_page(), set copyback caching of the
pages on the 68040 and 68060, to avoid cache consistency problems (since
regular mappings are copyback cached as well), per Ignatios Souvatzis.
1999-01-09 18:40:12 +00:00
augustss fc5f9ee81d Add -Wpointer-arith warning since `void *' arithmetic is not ANSI C. 1999-01-08 19:26:12 +00:00
thorpej 289b2e154f In pmap_zero_page() and pmap_copy_page(), don't bother invalidating the
mappings for CADDR1 and CADDR2, unless we're built w/ -DDEBUG.  Since
these addresses are only used in these routines, we can "lazily invalidate"
them by just using them again.  This makes these routines a teeny bit faster,
as it saves 1 or 2 TBIS's per call.

Suggested by Leo Weppelman <leo@netbsd.org>.
1999-01-08 05:15:43 +00:00
thorpej 795321b030 Make pmap_collect() actually do something useful. If called with a user
pmap, forget all of the mappings for the user address space for that pmap.
This causes the PT pages to be freed so that they can be reclaimed by the
VM system. [*]

[*] Actually, in the current implementation, it merely causes the wiring
count on the PT pages to drop to 0, which allows them to be reclaimed by
the pagedaemon.  Handling of PT pages needs to be completely rewritten.
1998-12-21 09:02:43 +00:00
thorpej d6bc370c2d Use a memory pool for pmap structures. 1998-12-21 08:51:39 +00:00
thorpej 43b82d559c Improve pmap_copy_page() and pmap_zero_page(). Don't use pmap_enter()/
pmap_remove() for the temporary addresses.  This is completely unnecessary
as the temps are used ONLY for these routines, and we have better control
over the mapping by manipulating the PTE ourself.  On VAC systems, cache-
inhibit the mapping to prevent wasting a cache load.

These routines are now significantly faster.

Add a DIAGNOSTIC check in pmap_enter() for kernel pmap and (CADDR1 or CADDR2);
nothing should be adding mappings there via that interface.
1998-12-21 06:58:39 +00:00
thorpej 51d0670371 Clean up a whole load of comments, making many of them truthful, and
indicating which functions are INTERFACE vs. which aren't.
1998-12-20 01:15:52 +00:00
thorpej 80e50e26a8 Unclutter a bit. Use a PMAP_DPRINTF() macro where possible, getting rid
of loads of #ifdef DEBUG ... stuff.
1998-12-19 23:51:47 +00:00
thorpej 6659e7174e In pmap_testbit(), if we end up traversing the PV list, cache a hit on the
bit to speed future tests.
1998-12-19 23:21:51 +00:00
thorpej 97f422e85b Make pmap_changebit() take bits to set and a PTE mask, rather than a bit
and a boolean.
1998-12-19 23:01:47 +00:00
thorpej d36a60d9cb G/c PMAPSTATS. It's unneeded clutter right now. 1998-12-19 21:11:14 +00:00
thorpej 459052a0fb Wire down some SCSI disk units. 1998-12-19 20:05:19 +00:00
itohy 36416d8500 Added options COMPAT_LINUX and EXEC_ELF32 as comments
to GENERIC configuration files.
1998-12-15 19:50:00 +00:00
itohy c05dadc113 Added COMPAT_LINUX support. 1998-12-15 19:36:36 +00:00
thorpej ebd6430388 Add some brackes to make EGCS happy. 1998-12-11 18:46:18 +00:00
hwr 59acb69be1 Add (commented out) 'gre' pseudo device line. Feedback is welcome.
Inspired by Klaus Klein.
1998-11-25 20:30:44 +00:00
kleink 94696a9ac3 Add support for the BUFCACHE config option. 1998-11-24 13:19:48 +00:00
mrg db3051d720 fix problems in many d_mmap routines:
- returned EOPNOTSUPP rather than -1.
	- no check for negative offset.
many of these fix potential security problems in these drivers.


XXX XXX XXX
the d_mmap cdev routine should be changed to have a prototype like:
	paddr_t (*d_mmap) __P((dev_t, off_t, int));

by someone!
1998-11-19 15:38:20 +00:00
oster cf5710c839 Added a (commented out) configuration line for the RAIDframe device driver. 1998-11-16 16:44:55 +00:00
oster c74d32c5fc Updating of bdev's and cdev's to support RAIDframe. 1998-11-13 04:47:03 +00:00
thorpej 49c62c4336 Changes to support fork_kthread():
- cpu_set_kpc() now takes void *arg third argument, passed to the
  entry point.
- cpu_fork() allows parent to be non-curproc iff parent is proc0.
  When forking non-curproc, assume its state has already been saved.
- Adjust various pieces of machine-dependent code to account of all of this.
1998-11-11 06:41:23 +00:00
frueauf a894e3a493 Add braces to make egcs happy. 1998-11-09 15:53:51 +00:00
tron b296275bb4 Defopt SYSVMSG, SYSVSEM and SYSVSHM. 1998-10-19 22:09:13 +00:00
thorpej 49dd2f5055 Need <sys/device.h> 1998-10-18 18:39:43 +00:00
thorpej adc5eb3a57 Fix a thinko in previous. 1998-10-18 17:15:04 +00:00
drochner 2468738337 change handling of libkern:
-sys/lib/libkern builds as library per default (as it was documented all
 the time)
-ports able to LKM set "KERN_AS=obj" explicitely in their Makefiles
 (for now; should depend on actual "option LKM" or -better- functions
 included for LKM use should be pulled in by a stub)
-always link libcompat before libkern - libkern stuff can be referred to
 by libcompat, but not the other way
1998-10-15 18:37:13 +00:00
chuck 8bef431273 remove unused share map code from UVM:
- update calls to uvm_unmap_remove/uvm_unmap (mainonly boolean arg
        has been removed)
1998-10-11 23:20:59 +00:00
thorpej 908cdfe4e3 Add scsibus entry points to the cdevsw[]. 1998-10-10 02:00:49 +00:00
thorpej d681cf055a configure() prototype is in <sys/device.h> 1998-10-06 20:50:15 +00:00
thorpej c01f29bdf1 Sigh, we are going to have to burn a vector for Just Sigreturn again,
for now, until we have a more generic {get,set}context().  Update the
comment for trap #3 accordingly.
1998-10-04 23:38:55 +00:00