]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years ago- Move opt_fb.h earlier so it covers included headers.
John Baldwin [Fri, 8 Nov 2002 21:10:53 +0000 (21:10 +0000)]
- Move opt_fb.h earlier so it covers included headers.
- Update cdevsw.  Block majors are out, kqfilters are in.
- No need to bzero softc's that we get from new-bus.  They come that way to
  begin with.

21 years agoMove include of opt_fb.h earlier so that FB_INSTALL_CDEV is defined in
John Baldwin [Fri, 8 Nov 2002 21:04:44 +0000 (21:04 +0000)]
Move include of opt_fb.h earlier so that FB_INSTALL_CDEV is defined in
included headers.

21 years agoFix printf warnings with %j and uintmax_t.
John Baldwin [Fri, 8 Nov 2002 20:59:23 +0000 (20:59 +0000)]
Fix printf warnings with %j and uintmax_t.

21 years agoFix warnings when compiled with SIMOS defined.
John Baldwin [Fri, 8 Nov 2002 20:52:21 +0000 (20:52 +0000)]
Fix warnings when compiled with SIMOS defined.

21 years agoAssign value of NULL to imgp->execlabel when imgp is initialized
Robert Watson [Fri, 8 Nov 2002 20:49:50 +0000 (20:49 +0000)]
Assign value of NULL to imgp->execlabel when imgp is initialized
in the ELF code.  Missed in earlier merge from the MAC tree.

Approved by: re
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoCast pointers in mem*_io() compat macros to uintptr_t so that they can be
John Baldwin [Fri, 8 Nov 2002 20:49:43 +0000 (20:49 +0000)]
Cast pointers in mem*_io() compat macros to uintptr_t so that they can be
used as bus handles by the bus_space functions implementing these macros.

21 years agoMake 3dfx i386-only. The memrange API it uses may be defined in an MI
John Baldwin [Fri, 8 Nov 2002 20:04:38 +0000 (20:04 +0000)]
Make 3dfx i386-only.  The memrange API it uses may be defined in an MI
header, but it is only implemented on i386.

21 years agoMove rc(4) over to MI notes and enable it as a MI module.
John Baldwin [Fri, 8 Nov 2002 19:34:26 +0000 (19:34 +0000)]
Move rc(4) over to MI notes and enable it as a MI module.

21 years agoFix two typos.
Bill Fenner [Fri, 8 Nov 2002 19:10:05 +0000 (19:10 +0000)]
Fix two typos.

21 years agoTo reduce per-return overhead of userret(), call into
Robert Watson [Fri, 8 Nov 2002 19:00:17 +0000 (19:00 +0000)]
To reduce per-return overhead of userret(), call into
mac_thread_userret() only if PS_MACPEND is set in the process AST mask.
This avoids the cost of the entry point in the common case, but
requires policies interested in the userret event to set the flag
(protected by the scheduler lock) if they do want the event.  Since
all the policies that we're working with which use mac_thread_userret()
use the entry point only selectively to perform operations deferred
for locking reasons, this maintains the desired semantics.

Approved by: re
Requested by: bde
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoConnect the ia64 versions of the release documentation to the RELNOTESng
Bruce A. Mah [Fri, 8 Nov 2002 18:36:04 +0000 (18:36 +0000)]
Connect the ia64 versions of the release documentation to the RELNOTESng
build.

This will give ia64 folks the chance to help flesh these out, and
also ensure that (when the time comes) ia64 releases with documentation
won't fall over due to missing RELNOTESng files.

21 years agoUse critical_enter/exit instead of disable/enable_intr to make this
John Baldwin [Fri, 8 Nov 2002 18:32:57 +0000 (18:32 +0000)]
Use critical_enter/exit instead of disable/enable_intr to make this
driver more portable.

21 years agoAdd some text warning about the state (incomplete) of this document.
Bruce A. Mah [Fri, 8 Nov 2002 18:28:00 +0000 (18:28 +0000)]
Add some text warning about the state (incomplete) of this document.

21 years ago- Change the ATM stack functions to use intptr_t instead of int for opaque
John Baldwin [Fri, 8 Nov 2002 18:27:30 +0000 (18:27 +0000)]
- Change the ATM stack functions to use intptr_t instead of int for opaque
  arguments.
- Fix a few other places that assumed that sizeof(int) == sizeof(void *).

Reviewed by: mdodd

21 years agoInclude the pc98 installation document in the RELNOTESng snapshot
Bruce A. Mah [Fri, 8 Nov 2002 18:15:02 +0000 (18:15 +0000)]
Include the pc98 installation document in the RELNOTESng snapshot
page.

21 years ago- Set RS (Report Status) bit on all descriptors of a packet instead of just the last...
Prafulla Deuskar [Fri, 8 Nov 2002 18:14:17 +0000 (18:14 +0000)]
- Set RS (Report Status) bit on all descriptors of a packet instead of just the last one.
- Set RDTR to zero by default instead of 28.
- Fixed a problem with TX hangs with jumbo frames when number of fragments in the mbuf chain
is large.
- Added support for 82540EP based cards.

MFC after: 3 days

21 years agoUpdate MAC modules for changes in arguments for exec MAC policy
Robert Watson [Fri, 8 Nov 2002 18:04:36 +0000 (18:04 +0000)]
Update MAC modules for changes in arguments for exec MAC policy
entry points to include an explicit execlabel.

Approved by: re
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoAdd an explicit execlabel argument to exec-related MAC policy entry
Robert Watson [Fri, 8 Nov 2002 18:04:00 +0000 (18:04 +0000)]
Add an explicit execlabel argument to exec-related MAC policy entry
points, rather than relying on policies to grub around in the
image activator instance structure.

Approved by: re
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoInfrastructure for ia64 release notes and installation guide.
Bruce A. Mah [Fri, 8 Nov 2002 18:01:31 +0000 (18:01 +0000)]
Infrastructure for ia64 release notes and installation guide.

21 years agoDiff reduction to NetBSD:
Gordon Tetlow [Fri, 8 Nov 2002 17:06:15 +0000 (17:06 +0000)]
Diff reduction to NetBSD:

Fix a minor grammer nit.
Get rid of _echo that uses logger and just call logger in the 3 routines
that need it.
Add a comment referencing rc.conf(5) in the case of an incorrectly set
variable.

This pulls rc.subr up to 1.51 of rc.subr. I skipped 1.49 (systrace support)
since I don't have systrace.

Obtained from: NetBSD

21 years agoConvert some broken cases where Error is called, but we try to continue,
Juli Mallett [Fri, 8 Nov 2002 16:59:11 +0000 (16:59 +0000)]
Convert some broken cases where Error is called, but we try to continue,
to Fatal errors, because the logic that we use to try to continue is far
too broken, and makes things look and act weird, because we end up pointing
past the end of a buffer boundry into freed memory in the caller, as we
don't come close to setting the lengthPtr to a sane value.

Reviewed by: make@

(This only changes failure cases which would have died horrid deaths to
 explicit clean death failure cases.)

21 years agoRemove harmless but irritating printf.
Poul-Henning Kamp [Fri, 8 Nov 2002 15:35:16 +0000 (15:35 +0000)]
Remove harmless but irritating printf.

21 years agoAlways recalculate the SRM checksum if the label is at 64 bytes offset.
Poul-Henning Kamp [Fri, 8 Nov 2002 15:31:34 +0000 (15:31 +0000)]
Always recalculate the SRM checksum if the label is at 64 bytes offset.

Tested by: jhb

21 years ago- Recognize FAT partition on MBR and PC98 disks.
Yoshihiro Takahashi [Fri, 8 Nov 2002 15:25:15 +0000 (15:25 +0000)]
- Recognize FAT partition on MBR and PC98 disks.
- Fix to convert to the name of partition.

21 years ago- Update the SIMOS fp-stealing code to KSE-II and beyond.
John Baldwin [Fri, 8 Nov 2002 15:24:32 +0000 (15:24 +0000)]
- Update the SIMOS fp-stealing code to KSE-II and beyond.
- Wrap a variable only used in !SIMOS case in #ifndef SIMOS.

21 years agoRemove #ifdef DEBUG signal debugging code brought over from NetBSD with the
John Baldwin [Fri, 8 Nov 2002 15:16:38 +0000 (15:16 +0000)]
Remove #ifdef DEBUG signal debugging code brought over from NetBSD with the
original signal handling code.  It doesn't compile and it seems the rest of
the infrastructure was never brought over from NetBSD.

21 years agoMake xrpu(4) i386-only. Consumers of i386_btop() are not MI.
John Baldwin [Fri, 8 Nov 2002 15:05:16 +0000 (15:05 +0000)]
Make xrpu(4) i386-only.  Consumers of i386_btop() are not MI.

21 years agoUpdate acl.3 to xref getfacl(1) and setfacl(1), the recommended tools for
Robert Watson [Fri, 8 Nov 2002 15:01:28 +0000 (15:01 +0000)]
Update acl.3 to xref getfacl(1) and setfacl(1), the recommended tools for
manipulating file ACLs.  Update the status of the implementation a bit,
update the copyright, etc.

Obtained from: TrustedBSD Project

21 years agoWrap a device_printf() that violates bus space abstractions to figure out
John Baldwin [Fri, 8 Nov 2002 15:01:02 +0000 (15:01 +0000)]
Wrap a device_printf() that violates bus space abstractions to figure out
if it's IO port resource is IO or memory mapped for the sake of a printf
using i386-specific values in #ifdef __i386__.

21 years agoUse %z to print a size_t value.
John Baldwin [Fri, 8 Nov 2002 14:58:35 +0000 (14:58 +0000)]
Use %z to print a size_t value.

21 years agoUse %z to print size_t values.
John Baldwin [Fri, 8 Nov 2002 14:57:18 +0000 (14:57 +0000)]
Use %z to print size_t values.

21 years agoCast a ptrdiff_t to an int to printf.
John Baldwin [Fri, 8 Nov 2002 14:52:26 +0000 (14:52 +0000)]
Cast a ptrdiff_t to an int to printf.

21 years agoUse %z to print a size_t value.
John Baldwin [Fri, 8 Nov 2002 14:50:44 +0000 (14:50 +0000)]
Use %z to print a size_t value.

21 years agoMFi386: revision 1.4.
Yoshihiro Takahashi [Fri, 8 Nov 2002 14:49:14 +0000 (14:49 +0000)]
MFi386: revision 1.4.

21 years agoDon't set a variable to a bogus value right before setting it to the
John Baldwin [Fri, 8 Nov 2002 14:00:44 +0000 (14:00 +0000)]
Don't set a variable to a bogus value right before setting it to the
correct value in the next statement.

21 years agoDon't set a value to a variable that we don't use.
John Baldwin [Fri, 8 Nov 2002 14:00:09 +0000 (14:00 +0000)]
Don't set a value to a variable that we don't use.

21 years agoMake changes such that the kernel will now fit on floppy #1.
Matt Jacob [Fri, 8 Nov 2002 13:58:48 +0000 (13:58 +0000)]
Make changes such that the kernel will now fit on floppy #1.

This involves moving all networking drivers to floppy 2, all RAID
nfsclient && cd9660 to floppy 3 (and adding ispfw).

COMPAT_FREEBSD4 had to come back, maxusers is down to 2.

AHC is out until somebody can make it a module for alpha.

TurboLaser is out because it rarely has a floppy and probably
no longer works *anyway*.

There's probably some cleanup that can be done because I was told
that things in drivers.conf don't have to be removed via dokern.sh,
but because I formally tested the release build with this version
and don't want to spend another half day on it, it's going in.

I'm sure we'll have to do some further changes prior to release.

NB: Not tested end to end yet because I've been told that sysinstall
doesn't yet work.

21 years ago- Merge Write_FreeBSD() into Write_Disk().
John Baldwin [Fri, 8 Nov 2002 13:58:00 +0000 (13:58 +0000)]
- Merge Write_FreeBSD() into Write_Disk().
- Miscellaneous cleanups.

21 years agoDON'T EVER PUT THIS BACK!
Ruslan Ermilov [Fri, 8 Nov 2002 12:47:51 +0000 (12:47 +0000)]
DON'T EVER PUT THIS BACK!

Pointy hat to: obrien

21 years agoRemove physmem from here, too, as it is defined in vm_init.c since
Thomas Moestl [Fri, 8 Nov 2002 11:10:50 +0000 (11:10 +0000)]
Remove physmem from here, too, as it is defined in vm_init.c since
r1.35 (forgotten in my last commit due to a botched patch).

Pointy hat to: tmm

21 years agoReviewed by: n_hibma
Michael Reifenberger [Fri, 8 Nov 2002 07:57:42 +0000 (07:57 +0000)]
Reviewed by: n_hibma
MFC after: 2 weeks
fix XPT_CALC_GEOMETRY to fill in some reasonable values.
Inspired by aic/sbp.

21 years agoTry to make this document reflect reality better:
Bruce A. Mah [Fri, 8 Nov 2002 04:52:04 +0000 (04:52 +0000)]
Try to make this document reflect reality better:

o All of the Q&A entries in this section were i386-specific, so make
these conditional on building for that platform.  If
non-i386-platforms need similar Q&A lists, they can add their own to
this file.

o Coalesce the two Q&A sections into one.

o Correct some device names (wd(4) devices have not existed for quite
some time).

o Get rid of a number of obsolete entries (suggested by jhb).

21 years agoAdd fwe.4.
Hidetoshi Shimokawa [Fri, 8 Nov 2002 03:24:32 +0000 (03:24 +0000)]
Add fwe.4.

21 years agoadjust critical section to only wrap around vm86_bioscall().
David Xu [Fri, 8 Nov 2002 03:07:52 +0000 (03:07 +0000)]
adjust critical section to only wrap around vm86_bioscall().

21 years agoDetect and exit on write errors.
Tim J. Robbins [Fri, 8 Nov 2002 02:51:31 +0000 (02:51 +0000)]
Detect and exit on write errors.

21 years agouse critical_enter/exit to add a critical section around BIOS call,
David Xu [Fri, 8 Nov 2002 01:09:16 +0000 (01:09 +0000)]
use critical_enter/exit to add a critical section around BIOS call,
this unbreaks WITNESS.

Pointed out by: jhb

21 years agoMove the definitions of the hw.physmem, hw.usermem and hw.availpages
Thomas Moestl [Thu, 7 Nov 2002 23:57:17 +0000 (23:57 +0000)]
Move the definitions of the hw.physmem, hw.usermem and hw.availpages
sysctls to MI code; this reduces code duplication and makes all of them
available on sparc64, and the latter two on powerpc.
The semantics by the i386 and pc98 hw.availpages is slightly changed:
previously, holes between ranges of available pages would be included,
while they are excluded now. The new behaviour should be more correct
and brings i386 in line with the other architectures.

Move physmem to vm/vm_init.c, where this variable is used in MI code.

21 years agoBetter printf() formats.
Maxime Henrion [Thu, 7 Nov 2002 23:16:22 +0000 (23:16 +0000)]
Better printf() formats.

21 years agoSome more printf() format fixes.
Maxime Henrion [Thu, 7 Nov 2002 23:03:04 +0000 (23:03 +0000)]
Some more printf() format fixes.

21 years agoAdd a cast to quiet a warning.
John Baldwin [Thu, 7 Nov 2002 22:49:15 +0000 (22:49 +0000)]
Add a cast to quiet a warning.

21 years agoCorrectly print vm_offset_t types.
Maxime Henrion [Thu, 7 Nov 2002 22:49:07 +0000 (22:49 +0000)]
Correctly print vm_offset_t types.

21 years agoProperly parenthesize the DBREG_DRX macro's variables to allow for
Alfred Perlstein [Thu, 7 Nov 2002 22:42:31 +0000 (22:42 +0000)]
Properly parenthesize the DBREG_DRX macro's variables to allow for
DBREG_DRX(&dbregs, n) usage.

21 years ago- Use %j to print intmax_t values.
John Baldwin [Thu, 7 Nov 2002 22:41:08 +0000 (22:41 +0000)]
- Use %j to print intmax_t values.
- Cast more daddr_t values to intmax_t when printing to quiet warnings.

21 years agoUse %z to quiet a warning.
John Baldwin [Thu, 7 Nov 2002 22:38:04 +0000 (22:38 +0000)]
Use %z to quiet a warning.

21 years agoCast a pointer to a uintptr_t to quiet a warning.
John Baldwin [Thu, 7 Nov 2002 22:35:45 +0000 (22:35 +0000)]
Cast a pointer to a uintptr_t to quiet a warning.

21 years agoThird argument to copyinstr() is a pointer to a size_t, not a pointer to a
John Baldwin [Thu, 7 Nov 2002 22:33:28 +0000 (22:33 +0000)]
Third argument to copyinstr() is a pointer to a size_t, not a pointer to a
u_int.

21 years ago- Use %p to print pointers instead of casting pointers to an int and using
John Baldwin [Thu, 7 Nov 2002 22:25:13 +0000 (22:25 +0000)]
- Use %p to print pointers instead of casting pointers to an int and using
  0x%x.
- Add a cast to quiet a warning.

21 years agoUse [u]intmax_t and %j instead of long long and %ll to better fix warnings
John Baldwin [Thu, 7 Nov 2002 22:23:46 +0000 (22:23 +0000)]
Use [u]intmax_t and %j instead of long long and %ll to better fix warnings
I fixed earlier.

Requested by: mux, jake

21 years agoMake 'rclk' a uintptr_t to quiet warnings.
John Baldwin [Thu, 7 Nov 2002 22:22:10 +0000 (22:22 +0000)]
Make 'rclk' a uintptr_t to quiet warnings.

21 years agoUse %z and %j to quiet warnings.
John Baldwin [Thu, 7 Nov 2002 22:21:38 +0000 (22:21 +0000)]
Use %z and %j to quiet warnings.

21 years agoCast a pointer to a uintptr_t instead of a u_int.
John Baldwin [Thu, 7 Nov 2002 22:21:10 +0000 (22:21 +0000)]
Cast a pointer to a uintptr_t instead of a u_int.

21 years agoFix the rest of the 64-bit warnings for the sr(4) driver.
John Baldwin [Thu, 7 Nov 2002 22:20:22 +0000 (22:20 +0000)]
Fix the rest of the 64-bit warnings for the sr(4) driver.

21 years agoDon't exit with success after timeout during authentication.
Bill Fenner [Thu, 7 Nov 2002 22:16:45 +0000 (22:16 +0000)]
Don't exit with success after timeout during authentication.

21 years agoMark ed(4) as i386/pc98-only, tracking src/sys/conf/NOTES rev. 1.1109.
Bruce A. Mah [Thu, 7 Nov 2002 22:11:00 +0000 (22:11 +0000)]
Mark ed(4) as i386/pc98-only, tracking src/sys/conf/NOTES rev. 1.1109.

21 years ago- Fix a bunch of casts to long which were truncating off_t's.
Maxime Henrion [Thu, 7 Nov 2002 21:56:05 +0000 (21:56 +0000)]
- Fix a bunch of casts to long which were truncating off_t's.
- Remove the comments which were justifying this by the fact
that we don't have %q in the kernel, this was probably right
back in time, but we now have %q, and we even have better to
print those types (%j).

21 years agoExplicitly cast daddr_t's to long long's when printing them since daddr_t
John Baldwin [Thu, 7 Nov 2002 21:52:51 +0000 (21:52 +0000)]
Explicitly cast daddr_t's to long long's when printing them since daddr_t
is not long long on all archs.  (They happen to be long's on 64-bit arch's
and gcc considers that significant enough to warn about it.)  These should
probably be uintmax_t but I didn't feel like adding all the extra includes.

21 years agoA bunch of style(9) fixes.
Maxime Henrion [Thu, 7 Nov 2002 21:45:28 +0000 (21:45 +0000)]
A bunch of style(9) fixes.

Obtained from: bde

21 years agoCast a ptrdiff_t value to an int to quiet a warning since we don't support
John Baldwin [Thu, 7 Nov 2002 21:34:21 +0000 (21:34 +0000)]
Cast a ptrdiff_t value to an int to quiet a warning since we don't support
%t in the kernel printf yet.

21 years agoQuiet some DIAGNOSTIC printf warnings with band-aid casts.
John Baldwin [Thu, 7 Nov 2002 21:25:49 +0000 (21:25 +0000)]
Quiet some DIAGNOSTIC printf warnings with band-aid casts.

21 years agoPrint pointers with %p instead of casting them to an unsigned int and
John Baldwin [Thu, 7 Nov 2002 21:19:47 +0000 (21:19 +0000)]
Print pointers with %p instead of casting them to an unsigned int and
then printing them with 0x%x which doesn't work when sizeof(void *) >
sizeof(int).

21 years agoMove sr(4) over to i386-only as it is yet another user of kvtop().
John Baldwin [Thu, 7 Nov 2002 21:15:29 +0000 (21:15 +0000)]
Move sr(4) over to i386-only as it is yet another user of kvtop().

21 years ago- Use a better definition for MNAMELEN which doesn't require
Maxime Henrion [Thu, 7 Nov 2002 21:15:02 +0000 (21:15 +0000)]
- Use a better definition for MNAMELEN which doesn't require
  to have one #ifdef per architecture.
- Change a space to a tab after a nearby #define.

Obtained from: bde

21 years agoFix lots of places that assume sizeof(int) == sizeof(void *). Most of
John Baldwin [Thu, 7 Nov 2002 21:13:15 +0000 (21:13 +0000)]
Fix lots of places that assume sizeof(int) == sizeof(void *).  Most of
these were cosmetic.  However, at least two of these changes fix bugs
where a (fake) pointer to a RX descriptor was added to the physical
memory start address with the cast in the wrong place resulting in
rxdesc[pmem_start] instead of (uintptr_t)rxdesc + pmem_start.

21 years agoDo a bit more work in the aio code to simulate the credential environment
Robert Watson [Thu, 7 Nov 2002 20:46:37 +0000 (20:46 +0000)]
Do a bit more work in the aio code to simulate the credential environment
of the original AIO request: save and restore the active thread credential
as well as using the file credential, since MAC (and some other bits of
the system) rely on the thread credential instead of/as well as the
file credential.  In brief: cache td->td_ucred when the AIO operation
is queued, temporarily set and restore the kernel thread credential,
and release the credential when done.  Similar to ktrace credential
management.

Reviewed by: alc
Approved by: re
Obtained from: TrustedBSD Project
Sponsored by: DARPA, Network Associates Laboratories

21 years agoLicense and blurb update authorized by Network Associates.
Robert Watson [Thu, 7 Nov 2002 20:37:04 +0000 (20:37 +0000)]
License and blurb update authorized by Network Associates.

21 years agoCast a pointer to uintptr_t instead ot u_int.
John Baldwin [Thu, 7 Nov 2002 19:56:21 +0000 (19:56 +0000)]
Cast a pointer to uintptr_t instead ot u_int.

21 years agoFactor out the code that determines whether a message must be skipped
Thomas Quinot [Thu, 7 Nov 2002 19:53:29 +0000 (19:53 +0000)]
Factor out the code that determines whether a message must be skipped
as a consequence of a host or program name specification into a common
function, skip_pmessage.

Reviewed by: roberto

21 years agoSimplify and optimize pmap_object_init_pt(). More specifically,
Alan Cox [Thu, 7 Nov 2002 18:33:55 +0000 (18:33 +0000)]
Simplify and optimize pmap_object_init_pt().   More specifically,
take advantage of the fact that the vm object's list of pages is
now ordered to reduce the overhead of finding the desired set of
pages to be mapped.  (See revision 1.215 of vm/vm_page.c.)

21 years agod_read_t functions return an int, not a ssize_t. (Considering the fact
John Baldwin [Thu, 7 Nov 2002 17:26:38 +0000 (17:26 +0000)]
d_read_t functions return an int, not a ssize_t.  (Considering the fact
that read(2) returns a ssize_t perhaps this is a bug in d_read_t?)

21 years agoWarning fixes for sizeof(int) != sizeof(void *).
John Baldwin [Thu, 7 Nov 2002 17:25:33 +0000 (17:25 +0000)]
Warning fixes for sizeof(int) != sizeof(void *).

21 years ago- Make 'irq' intptr_t instead of uintptr_t so it handles a value of -1
John Baldwin [Thu, 7 Nov 2002 17:20:58 +0000 (17:20 +0000)]
- Make 'irq' intptr_t instead of uintptr_t so it handles a value of -1
  properly.
- Add a cast to quiet a printf warning.

21 years agoFix to support pc98.
Yoshihiro Takahashi [Thu, 7 Nov 2002 16:42:37 +0000 (16:42 +0000)]
Fix to support pc98.
It is mostly merged from MBR specific part.

Reviewed by: phk

21 years agoAdd a Firewire section to the hardware notes, with initial content
Bruce A. Mah [Thu, 7 Nov 2002 16:30:17 +0000 (16:30 +0000)]
Add a Firewire section to the hardware notes, with initial content
derived from the fwohci(4) and sbp(4) manpages.

21 years agoNew release notes: getconf(1), nextboot(8),
Bruce A. Mah [Thu, 7 Nov 2002 16:23:30 +0000 (16:23 +0000)]
New release notes:  getconf(1), nextboot(8),

Updated release notes:  ntp-4.1.1b.

MFCs noted:  firewire(4), atapicam(4), uudecode(1)/uuencode(1) base64
links.

21 years agoMove firewire back to being MI.
John Baldwin [Thu, 7 Nov 2002 16:19:43 +0000 (16:19 +0000)]
Move firewire back to being MI.

21 years agoAdd two new workaround for firmware anomalies:
Thomas Moestl [Thu, 7 Nov 2002 16:07:46 +0000 (16:07 +0000)]
Add two new workaround for firmware anomalies:
1. At least some Netra t1 models have PCI buses with no associated
   interrupt map, but obviously expect the PCI swizzle to be done with
   the interrupt number from the higher level as intpin. In this case,
   the mapping also needs to continue at parent bus nodes.
   To handle that, add a quirk table based on the "name" property of
   the root node to avoid breaking other boxen. This property is now
   retrieved and printed at boot.
2. On SPARCengine Ultra AX machines, interrupt numbers are not mapped
   at all, and full interrupt numbers (not just INOs) are given in
   the interrupt properties. This is more or less cosmetical; the
   PCI interrupt numbers would be wrong, but the psycho resource
   allocation method would pass the right numbers on anyway.

Tested by: mux (1), Maxim Mazurok <maxim@km.ua> (2)

21 years agoProperly fix the occassional random crash issue that revision 1.142
Ian Dowse [Thu, 7 Nov 2002 16:04:07 +0000 (16:04 +0000)]
Properly fix the occassional random crash issue that revision 1.142
just limited to the DEVICE_POLLING case. This removes the FXP_RFA_RNRMARK
hack, and replaces it with a softc flag that is used to record when
the handling of a no-resource condition was deferred due to running
out of DEVICE_POLLING cycles. This was tested on -stable, but the
code is essentially the same as in -current. It should only affect
the case where DEVICE_POLLING is defined.

The details of the mechanism behind the crashes are still uncertain
but the most likely cause seems to be some kind of hardware confusion
when the no-resource recovery code is accidentally invoked while
the receiver is still active. This could have happened if the
hardware left the 0x4000 bit of the RFA status word set. The comments
in the commit log for revision 1.142 stating that the driver could
clash with the hardware writing to this status word were not correct.

Tested by: Guy Helmer <ghelmer@palisadesys.com>

21 years agoFix one misspelling and two punctuation nits.
Bruce A. Mah [Thu, 7 Nov 2002 16:00:48 +0000 (16:00 +0000)]
Fix one misspelling and two punctuation nits.

21 years agoMerge over ar(4) and db(4) to files.pc98.
John Baldwin [Thu, 7 Nov 2002 15:52:34 +0000 (15:52 +0000)]
Merge over ar(4) and db(4) to files.pc98.

21 years agoCreate a MACHINE_ARCH == i386 section that first lists modules common
John Baldwin [Thu, 7 Nov 2002 15:01:12 +0000 (15:01 +0000)]
Create a MACHINE_ARCH == i386 section that first lists modules common
to both i386 and pc98 and then uses MACHINE subsections to list modules
specific to i386/i386 and i386/pc98.

21 years agoFixed pc98 support.
Yoshihiro Takahashi [Thu, 7 Nov 2002 14:54:53 +0000 (14:54 +0000)]
Fixed pc98 support.
(merged i386 changes from chunk.c 1.41 and disk.c 1.100)

21 years agoGet this closer to working. The Write_Disk() function's for loop needed
John Baldwin [Thu, 7 Nov 2002 14:39:21 +0000 (14:39 +0000)]
Get this closer to working.  The Write_Disk() function's for loop needed
to use the same start condition as the i386 version.  However, since
Alpha's only have one fake "slice" from sysinstall's perspective we don't
need to use a loop, but can just write out the BSD label in the first
fake "slice".

21 years agoTurn on GEOM. It works on Alpha and we are going to need it for installs.
John Baldwin [Thu, 7 Nov 2002 14:36:56 +0000 (14:36 +0000)]
Turn on GEOM.  It works on Alpha and we are going to need it for installs.

21 years ago- Fix printf warnings on 64bit plathome.
Hidetoshi Shimokawa [Thu, 7 Nov 2002 02:13:40 +0000 (02:13 +0000)]
- Fix printf warnings on 64bit plathome.
- Disable fw_vmaccess() because it's not used and unuseful 64bit plathome.

21 years ago1.Fix smp race between kernel vm86 BIOS calling and userland vm86 mode code,
David Xu [Thu, 7 Nov 2002 01:34:23 +0000 (01:34 +0000)]
1.Fix smp race between kernel vm86 BIOS calling and userland vm86 mode code,
  remove global variable in_vm86call, set vm86 calling flag in PCB flags.

2.Fix vm86 BIOS calling preempted problem by changing vm86_lock mutex type
  from MTX_DEF to MTX_SPIN. vm86pcb is not remembered in thread struct,
  when the thread calling vm86 BIOS is preempted by interrupt thread,
  and later switching back to the thread would cause incorrect context be
  loaded into CPU registers, this leads to kernel crash.

21 years agos/HZ/Hz/
Maxime Henrion [Wed, 6 Nov 2002 23:31:01 +0000 (23:31 +0000)]
s/HZ/Hz/

21 years agoAdjust some casts to quiet warnings.
John Baldwin [Wed, 6 Nov 2002 22:58:55 +0000 (22:58 +0000)]
Adjust some casts to quiet warnings.

Reviewed by: mdodd

21 years agolnc(4) uses kvtop() and is thus i386-only for now.
John Baldwin [Wed, 6 Nov 2002 22:56:57 +0000 (22:56 +0000)]
lnc(4) uses kvtop() and is thus i386-only for now.

21 years agoStyle sync with rest of FreeBSD.
David E. O'Brien [Wed, 6 Nov 2002 22:54:58 +0000 (22:54 +0000)]
Style sync with rest of FreeBSD.

21 years agoUse long long to print out a bus_addr_t again.
John Baldwin [Wed, 6 Nov 2002 22:42:45 +0000 (22:42 +0000)]
Use long long to print out a bus_addr_t again.