]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
21 years agoDon't complain about not knowing the remote file size when working in
des [Sun, 27 Oct 2002 17:33:08 +0000 (17:33 +0000)]
Don't complain about not knowing the remote file size when working in
quiet mode.

PR: bin/37079
Submitted by: Nicolas Rachinsky <list@rachinsky.de>

21 years agoDon peril sensitive sun glasses and change the default system call vector
jake [Sun, 27 Oct 2002 17:22:43 +0000 (17:22 +0000)]
Don peril sensitive sun glasses and change the default system call vector
for sparc64 from trap #9 to trap #65.  This is one of the ABI "blessed"
system call vectors and is different from any other system that we might
want to emulate, making the emulation easier by reducing the number of
code paths that need to be shared.  Compatibility with old applications
is provided with COMPAT_FREEBSD4.
Add defines for a few special traps that we may need to implement for
compatibility with 32bit applications, and add comments on which vectors
are used for what in other systems, and which are available.
Pass magic flags to trap() for deprecated or unimplemented system call
vectors so they will deliver SIGSYS instead of SIGILL.

This piggy backs nicely with the recent sigaction(2) system call number
change, and provided the rules are followed for upgrading past it, this
change should not be noticed.

21 years agoSlight amendment to rev 1.34: instead of considering any short read an
des [Sun, 27 Oct 2002 17:20:49 +0000 (17:20 +0000)]
Slight amendment to rev 1.34: instead of considering any short read an
error, only report an error if no data was read at all (unless len was
0 to start with).  Otherwise, the final read of practically any transfer
will end in a fatal error.

21 years agoDon't roll our own clean target, the default one
mux [Sun, 27 Oct 2002 17:06:03 +0000 (17:06 +0000)]
Don't roll our own clean target, the default one
works better.

21 years agoMerged various changes from src/sbin/fdisk/fdisk.c revisions up to 1.66.
nyan [Sun, 27 Oct 2002 16:43:59 +0000 (16:43 +0000)]
Merged various changes from src/sbin/fdisk/fdisk.c revisions up to 1.66.

21 years agoIntroduce _fetch_writev(), which is the conn_t version of writev(2). In
des [Sun, 27 Oct 2002 16:11:21 +0000 (16:11 +0000)]
Introduce _fetch_writev(), which is the conn_t version of writev(2).  In
the SSL case, it is no different from the old _fetch_write(), but in the
non-SSL case it uses writev(2) to send the entire vector as a single
packet (provided it can fit in one packet).  Implement _fetch_write()
and _fetch_putln() in terms of _fetch_writev().

This should improve performance in the non-SSL case (by reducing protocol
overhead) and solve the problem where too-smart-for-their-own-good
firewalls reject FTP packets that do not end in CRLF.

PR: bin/44123
Submitted by: fenner

21 years agoCentrally manage enforcement of {reboot,swapon,sysctl} using the
rwatson [Sun, 27 Oct 2002 15:50:49 +0000 (15:50 +0000)]
Centrally manage enforcement of {reboot,swapon,sysctl} using the
mac_enforce_system toggle, rather than several separate toggles.

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

21 years agoEliminate two cases of undefined behaviour: total in _fetch_write() was
des [Sun, 27 Oct 2002 15:43:40 +0000 (15:43 +0000)]
Eliminate two cases of undefined behaviour: total in _fetch_write() was
not initialized before use, and _http_growbuf() did not return a value
on success.

Reported by: Peter Edwards <pmedwards@eircom.net>
MFC after: 2 weeks

21 years agoWhitespace and indentation cleanup.
des [Sun, 27 Oct 2002 15:32:51 +0000 (15:32 +0000)]
Whitespace and indentation cleanup.

21 years agoForgot to disable alarms after fetchXGet() in previous commit.
des [Sun, 27 Oct 2002 15:32:06 +0000 (15:32 +0000)]
Forgot to disable alarms after fetchXGet() in previous commit.

21 years agoAdd an ETA timer that kicks in after 30 seconds.
des [Sun, 27 Oct 2002 15:15:13 +0000 (15:15 +0000)]
Add an ETA timer that kicks in after 30 seconds.

Re-add alarm(2) calls around the calls to fetchStat(3) and fetchXGet(3),
since these calls can still time out on DNS lookups or TCP connect(2).

Remove the alarm(2) calls in the main loop, since all methods properly
handle transfer timeouts (as opposed to connection timeouts).

Set the sigalrm flag if a timeout occurs in the main loop.

Move the signal: label up a little so we still set the atime and mtime
when the transfer times out or is interrupted, so that restarted transfers
will work as expected (as long as the file still exists).

MFC after: 2 weeks

21 years agoBack out the previous commit, and fix the bug rather than try to hide its
des [Sun, 27 Oct 2002 15:08:21 +0000 (15:08 +0000)]
Back out the previous commit, and fix the bug rather than try to hide its
symptoms: make timeouts and short transfers fatal, and set errno to an
appropriate value (ETIMEDOUT for a timeout, EPIPE for a short transfer).

MFC after: 2 weeks

21 years agoFix a bunch of warnings on 64 bit platforms in the
mux [Sun, 27 Oct 2002 12:27:04 +0000 (12:27 +0000)]
Fix a bunch of warnings on 64 bit platforms in the
CISS_DEBUG case by appropriately using %z and %j.

21 years ago- Comment a line which sets CISS_DEBUG by default.
mux [Sun, 27 Oct 2002 12:09:51 +0000 (12:09 +0000)]
- Comment a line which sets CISS_DEBUG by default.
- Use -DCISS_DEBUG rather than -DCISS_DEBUG=0.

21 years agoFix a warning when CISS_DEBUG is defined.
mux [Sun, 27 Oct 2002 12:05:11 +0000 (12:05 +0000)]
Fix a warning when CISS_DEBUG is defined.

21 years agoCorrect visibility for v*wscanf(), wcstoll() and wcstoull(). These functions
tjr [Sun, 27 Oct 2002 11:30:36 +0000 (11:30 +0000)]
Correct visibility for v*wscanf(), wcstoll() and wcstoull(). These functions
did not exist in ISO C Amd. 1. Add #ifdef __LONG_LONG_SUPPORTED and lint
comments around wcstoll() and wcstoull().

21 years agoFix the regexp evilness so that fdisk can (again?) find the device
phk [Sun, 27 Oct 2002 10:52:54 +0000 (10:52 +0000)]
Fix the regexp evilness so that fdisk can (again?) find the device
root is on from the root mount path.

Spotted by: imp

21 years agoStyle sweep.
tjr [Sun, 27 Oct 2002 10:41:21 +0000 (10:41 +0000)]
Style sweep.

21 years agoDon't attempt to find the geometry with disklabel based ioctl, it just
phk [Sun, 27 Oct 2002 10:33:38 +0000 (10:33 +0000)]
Don't attempt to find the geometry with disklabel based ioctl, it just
issues a useless warning now.

21 years agoDon't truncate on large disks.
phk [Sun, 27 Oct 2002 10:17:38 +0000 (10:17 +0000)]
Don't truncate on large disks.

21 years agoTranslation of DOS and QIC/SCSI tape section
gioria [Sun, 27 Oct 2002 10:11:27 +0000 (10:11 +0000)]
Translation of DOS and QIC/SCSI tape section

21 years agoMFbed:
ue [Sun, 27 Oct 2002 09:58:43 +0000 (09:58 +0000)]
MFbed:
hardware/alpha/proc-alpha.sgml: 1.44  -> 1.45
hardware/common/dev.sgml: 1.110 -> 1.111
relnotes/common/new.sgml: 1.437 -> 1.439

21 years agoMFen 1.5
gioria [Sun, 27 Oct 2002 08:47:38 +0000 (08:47 +0000)]
MFen 1.5

21 years agoImplement mac_check_system_sysctl(), a MAC Framework entry point to
rwatson [Sun, 27 Oct 2002 07:12:34 +0000 (07:12 +0000)]
Implement mac_check_system_sysctl(), a MAC Framework entry point to
permit MAC policies to augment the security protections on sysctl()
operations.  This is not really a wonderful entry point, as we
only have access to the MIB of the target sysctl entry, rather than
the more useful entry name, but this is sufficient for policies
like Biba that wish to use their notions of privilege or integrity
to prevent inappropriate sysctl modification.  Affects MAC kernels
only.  Since SYSCTL_LOCK isn't in sysctl.h, just kern_sysctl.c,
we can't assert the SYSCTL subsystem lockin the MAC Framework.

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

21 years agoHook up mac_check_system_reboot(), a MAC Framework entry point that
rwatson [Sun, 27 Oct 2002 07:03:29 +0000 (07:03 +0000)]
Hook up mac_check_system_reboot(), a MAC Framework entry point that
permits MAC modules to augment system security decisions regarding
the reboot() system call, if MAC is compiled into the kernel.

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

21 years agoMerge from MAC tree: rename mac_check_vnode_swapon() to
rwatson [Sun, 27 Oct 2002 06:54:06 +0000 (06:54 +0000)]
Merge from MAC tree: rename mac_check_vnode_swapon() to
mac_check_system_swapon(), to reflect the fact that the primary
object of this change is the running kernel as a whole, rather
than just the vnode.  We'll drop additional checks of this
class into the same check namespace, including reboot(),
sysctl(), et al.

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

21 years agoMACHINE_ARCH is wrong here. It should be MACHINE
imp [Sun, 27 Oct 2002 06:31:37 +0000 (06:31 +0000)]
MACHINE_ARCH is wrong here.  It should be MACHINE

Submitted by: nyan@ (Yoshihiro-san)

21 years agoMention the need to have COMPAT_FREEBSD4 in the kernel you use for the
imp [Sun, 27 Oct 2002 04:48:31 +0000 (04:48 +0000)]
Mention the need to have COMPAT_FREEBSD4 in the kernel you use for the
upgrade.

21 years agoRemove unnecessary compatibility macro. (We were only printing, not parsing,
wollman [Sun, 27 Oct 2002 04:14:08 +0000 (04:14 +0000)]
Remove unnecessary compatibility macro.  (We were only printing, not parsing,
intmax_t.)

21 years agoAdd new 1003.1-2001/TC1/D6 parameters. (On final recirculation ballot now,
wollman [Sun, 27 Oct 2002 04:10:34 +0000 (04:10 +0000)]
Add new 1003.1-2001/TC1/D6 parameters.  (On final recirculation ballot now,
so this should be officially TC1 before the New Year.)

Add TrustedBSD pathconf parameters.

Add compilation support for -stable (to be merged momentarily).

21 years agoAdd NO_COMPAT_FREEBSD4.
peter [Sun, 27 Oct 2002 03:41:35 +0000 (03:41 +0000)]
Add NO_COMPAT_FREEBSD4.

Submitted by: Sergey Mokryshev <mokr@mokr.net>

21 years agoDISTDIR is used by 'make release', lets see if this is enough to turn
peter [Sun, 27 Oct 2002 03:38:57 +0000 (03:38 +0000)]
DISTDIR is used by 'make release', lets see if this is enough to turn
off the execution test.

21 years agoo Add functionality to add a GPT partition,
marcel [Sun, 27 Oct 2002 03:23:05 +0000 (03:23 +0000)]
o  Add functionality to add a GPT partition,
o  Use DCE compliant UUID functions and provide local
   implementations if they don't exist,
o  Move dumping of the map to show.c and print the
   partition type,
o  Some cleanups and rearrangements.

The default GPT partition type is UFS. When no starting block
or size are specified, the tool will create a partition in the
first free space it find (or that fits, depending on the size).

21 years agoDon't dump core if none of the directories in /etc/manpath.config exist.
kris [Sun, 27 Oct 2002 02:12:52 +0000 (02:12 +0000)]
Don't dump core if none of the directories in /etc/manpath.config exist.

Submitted by: terry

21 years agoAllow users to read the Early Adopter's Guide in sysinstall.
bmah [Sun, 27 Oct 2002 01:49:45 +0000 (01:49 +0000)]
Allow users to read the Early Adopter's Guide in sysinstall.

21 years agoNote the scary messages that you see when mounting / and other things
imp [Sun, 27 Oct 2002 01:35:36 +0000 (01:35 +0000)]
Note the scary messages that you see when mounting / and other things
are expected and normal when you've booted a 5.0 kernel with a 4.x
userland.

21 years agoUn-remove sysinstall for ia64.
peter [Sun, 27 Oct 2002 00:22:05 +0000 (00:22 +0000)]
Un-remove sysinstall for ia64.

21 years agoAdd back ia64 support that was removed in the last few revisions.
peter [Sun, 27 Oct 2002 00:21:02 +0000 (00:21 +0000)]
Add back ia64 support that was removed in the last few revisions.
I've cloned write_ia64_disk.c from write_i386_disk.c.

21 years agoConvert speed into KB/s. This was missed by the previous commit to this file.
njl [Sat, 26 Oct 2002 23:32:35 +0000 (23:32 +0000)]
Convert speed into KB/s.  This was missed by the previous commit to this file.

MFC after: 3 weeks

21 years agoState the stupidly obvious:
imp [Sat, 26 Oct 2002 22:55:43 +0000 (22:55 +0000)]
State the stupidly obvious:
If /tmp is on /, then blowing away its contents, as appropriate,
may get you enough space to do the installworld.

21 years agoNote: 5.0 requires more space. If you don't have 30MB free on /, then
imp [Sat, 26 Oct 2002 22:50:31 +0000 (22:50 +0000)]
Note: 5.0 requires more space.  If you don't have 30MB free on /, then
you are going to get hurt badly if you try to do an update from
sources.  Make a note of this.  While 'experts' could install it in
less space, I think 30MB is a good number.

21 years agoMake geom_mbr.c optional on PC98, use GEOM_MBR option to include it.
phk [Sat, 26 Oct 2002 20:17:59 +0000 (20:17 +0000)]
Make geom_mbr.c optional on PC98, use GEOM_MBR option to include it.

Disable check for supposedly magic "IPL1" string for PC98 labels, its
thaumaturgical power is in doubt.

21 years agoquery ip6.arpa then ip6.int for IPv6 reverse lookup. follows RFC3152.
ume [Sat, 26 Oct 2002 19:00:14 +0000 (19:00 +0000)]
query ip6.arpa then ip6.int for IPv6 reverse lookup.  follows RFC3152.

MFC after: 5 days

21 years agoFix a style nit.
mux [Sat, 26 Oct 2002 18:19:46 +0000 (18:19 +0000)]
Fix a style nit.

21 years agoIn VOP_LOOKUP, don't deny DELETE and RENAME operations
mux [Sat, 26 Oct 2002 18:16:19 +0000 (18:16 +0000)]
In VOP_LOOKUP, don't deny DELETE and RENAME operations
when ISLASTCN is not set.  The actual file which is being
looked up may live in a different filesystem.

21 years agoAdd a missing destroy_dev().
mux [Sat, 26 Oct 2002 18:09:30 +0000 (18:09 +0000)]
Add a missing destroy_dev().

21 years agoAllow deprecated or unimplemented system call vectors to deliver SIGSYS,
jake [Sat, 26 Oct 2002 17:38:20 +0000 (17:38 +0000)]
Allow deprecated or unimplemented system call vectors to deliver SIGSYS,
as suggested by the sparc v9 ABI.

21 years agoRemove an unused macro.
jake [Sat, 26 Oct 2002 17:36:37 +0000 (17:36 +0000)]
Remove an unused macro.

21 years agoModified release notes: Add a note to geom(4) entry about the demise
bmah [Sat, 26 Oct 2002 17:27:35 +0000 (17:27 +0000)]
Modified release notes:  Add a note to geom(4) entry about the demise
of compatability slices.

21 years agoGet rid of the boilerplate upgrading information in the release notes,
bmah [Sat, 26 Oct 2002 17:13:01 +0000 (17:13 +0000)]
Get rid of the boilerplate upgrading information in the release notes,
and point to the Early Adopter's Guide instead, at least for the next
release or two.

21 years agoPoint to the Early Adopter's Guide from the README.
bmah [Sat, 26 Oct 2002 17:00:39 +0000 (17:00 +0000)]
Point to the Early Adopter's Guide from the README.

21 years agoMFi386: revisions 1.544 and 1.545.
nyan [Sat, 26 Oct 2002 15:44:06 +0000 (15:44 +0000)]
MFi386: revisions 1.544 and 1.545.

21 years agoAdd & hookup manpage for pthread_attr_get_np(3).
phantom [Sat, 26 Oct 2002 15:04:29 +0000 (15:04 +0000)]
Add & hookup manpage for pthread_attr_get_np(3).

MFC after: 3 days

21 years agoSlightly change the semantics of vnode labels for MAC: rather than
rwatson [Sat, 26 Oct 2002 14:38:24 +0000 (14:38 +0000)]
Slightly change the semantics of vnode labels for MAC: rather than
"refreshing" the label on the vnode before use, just get the label
right from inception.  For single-label file systems, set the label
in the generic VFS getnewvnode() code; for multi-label file systems,
leave the labeling up to the file system.  With UFS1/2, this means
reading the extended attribute during vfs_vget() as the inode is
pulled off disk, rather than hitting the extended attributes
frequently during operations later, improving performance.  This
also corrects sematics for shared vnode locks, which were not
previously present in the system.  This chances the cache
coherrency properties WRT out-of-band access to label data, but in
an acceptable form.  With UFS1, there is a small race condition
during automatic extended attribute start -- this is not present
with UFS2, and occurs because EAs aren't available at vnode
inception.  We'll introduce a work around for this shortly.

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

21 years agoPrototype for pthread_attr_get_np().
phantom [Sat, 26 Oct 2002 13:58:12 +0000 (13:58 +0000)]
Prototype for pthread_attr_get_np().

21 years agoHook uthread_attr_get_np.c to build
phantom [Sat, 26 Oct 2002 13:55:35 +0000 (13:55 +0000)]
Hook uthread_attr_get_np.c to build

21 years agoAdd pthread_attr_get_np() function. This is FreeBSD non-portable POSIX threads
phantom [Sat, 26 Oct 2002 13:53:22 +0000 (13:53 +0000)]
Add pthread_attr_get_np() function. This is FreeBSD non-portable POSIX threads
extenston function. It supposed to provide facility to get already created
thread's attributes. Looks like it's last thing we need to make JDK's Hotspot
building without requirement to have source tree.

Reviewed by: deischen
MFC after: 3 days

21 years agoBe more agresive on arguments' checking.
phantom [Sat, 26 Oct 2002 13:47:06 +0000 (13:47 +0000)]
Be more agresive on arguments' checking.

OK'ed by: deischen
MFC after: 3 days

21 years agoMFen 1.111
gioria [Sat, 26 Oct 2002 12:36:38 +0000 (12:36 +0000)]
MFen 1.111

21 years agoCall bpf_mtap() on outgoing packets.
tmm [Sat, 26 Oct 2002 12:32:07 +0000 (12:32 +0000)]
Call bpf_mtap() on outgoing packets.

21 years agoAdd tests for btowc(), mbrlen(), mbsrtowcs() and wcsrtombs().
tjr [Sat, 26 Oct 2002 07:56:47 +0000 (07:56 +0000)]
Add tests for btowc(), mbrlen(), mbsrtowcs() and wcsrtombs().

21 years agoA simple make world is too dangerous for most upgrades. Instead, copy
imp [Sat, 26 Oct 2002 06:25:11 +0000 (06:25 +0000)]
A simple make world is too dangerous for most upgrades.  Instead, copy
the updating from stable section, and remove the kludge-o-rounds that
are in there.

Sometimes make world is safe, sometimes not.  Let's document the safe
way to deal.

21 years agoNote GEOM removal of the compatibility slice, and what you, the
imp [Sat, 26 Oct 2002 06:21:00 +0000 (06:21 +0000)]
Note GEOM removal of the compatibility slice, and what you, the
upgrader needs to do about it.

Also, greatly expand the the upgrading from -stable part of this file.
It appears that you need new boot blocks to ensure that the right
kernel is booted single user.

21 years agoAdd COMPAT_FREEBSD4 here too. It has COMPAT_43 as well.
peter [Sat, 26 Oct 2002 04:58:36 +0000 (04:58 +0000)]
Add COMPAT_FREEBSD4 here too.  It has COMPAT_43 as well.

21 years agoAdd COMPAT_FREEBSD4 for the transition. COMPAT_43 was already present.
peter [Sat, 26 Oct 2002 04:57:19 +0000 (04:57 +0000)]
Add COMPAT_FREEBSD4 for the transition.  COMPAT_43 was already present.

21 years agoHook the early adopter's guide into the release build. Its basename
bmah [Sat, 26 Oct 2002 04:48:55 +0000 (04:48 +0000)]
Hook the early adopter's guide into the release build.  Its basename
is "EARLY", which is kind of an odd name, but I couldn't think of
anything better that fit the traditional 8.3 naming convention for
release documentation files.

21 years agoiBack out david's last commit. the suspension code needs to be called
julian [Sat, 26 Oct 2002 04:44:17 +0000 (04:44 +0000)]
iBack out david's last commit. the suspension code needs to be called
for non KSE processes too.

21 years agoAdd regression tests for mbrtowc(3) and wcrtomb(3).
tjr [Sat, 26 Oct 2002 04:40:16 +0000 (04:40 +0000)]
Add regression tests for mbrtowc(3) and wcrtomb(3).

21 years agoMove suspension checking code from userret() into thread_userret().
davidxu [Sat, 26 Oct 2002 02:56:51 +0000 (02:56 +0000)]
Move suspension checking code from userret() into thread_userret().

21 years agoSigh. Fix past-o in install directory.
bmah [Sat, 26 Oct 2002 01:37:09 +0000 (01:37 +0000)]
Sigh.  Fix past-o in install directory.

21 years agoBackout revision 1.48.
davidxu [Sat, 26 Oct 2002 01:26:36 +0000 (01:26 +0000)]
Backout revision 1.48.

21 years agoFix "err" usage.
fenner [Sat, 26 Oct 2002 00:18:06 +0000 (00:18 +0000)]
Fix "err" usage.

PR: 44284
Submitted by: Sergey Mokryshev <mokr@mokr.net>

21 years agoFix two typos. Using "-eq" instead of "=" in a [ caused an error when
fenner [Fri, 25 Oct 2002 23:02:54 +0000 (23:02 +0000)]
Fix two typos.  Using "-eq" instead of "=" in a [ caused an error when
an entry in /etc/sysctl.conf didn't exist.  Fixing this exposed a minor
typo "exits" vs. "exist".

However, there doesn't appear to be any provision to run this with
the "lastload" argument, meaning that the error will never appear.

21 years agoDocument the `maclabel' command.
chris [Fri, 25 Oct 2002 22:40:37 +0000 (22:40 +0000)]
Document the `maclabel' command.

Obtained from: TrustedBSD Project
Sponsored by: Network Associates Laboratories

21 years agoo m_adj rx buffer so IP header is 32-bit aligned
sam [Fri, 25 Oct 2002 22:06:03 +0000 (22:06 +0000)]
o m_adj rx buffer so IP header is 32-bit aligned
o replace EPIC_MGETCLUSER with m_getcl

MFC after: 1 week

21 years agoAdd lt_LT.ISO8859-13
ache [Fri, 25 Oct 2002 22:02:10 +0000 (22:02 +0000)]
Add lt_LT.ISO8859-13

PR:             44268
Submitted by:   Kestutis Paulikas <kestas@elen.ktu.lt>

21 years agoAdd lt_LT.ISO8859-13
ache [Fri, 25 Oct 2002 22:00:34 +0000 (22:00 +0000)]
Add lt_LT.ISO8859-13

21 years agoHook up early-adopter article to RELNOTESng build.
bmah [Fri, 25 Oct 2002 21:58:06 +0000 (21:58 +0000)]
Hook up early-adopter article to RELNOTESng build.

21 years agoGah. Ditch elements I inserted while thinking about LaTeX and
bmah [Fri, 25 Oct 2002 21:56:05 +0000 (21:56 +0000)]
Gah.  Ditch elements I inserted while thinking about LaTeX and
substitute normal DocBook.  Pure coincidence that this thing built.

While here, pull in some more entities so that we can build in the
RELNOTESng environment.

21 years agoMFP4: Add the "Early Adopter's Guide to FreeBSD 5.0-RELEASE"
bmah [Fri, 25 Oct 2002 21:29:00 +0000 (21:29 +0000)]
MFP4:  Add the "Early Adopter's Guide to FreeBSD 5.0-RELEASE"
from the releng Perforce repository.  (Not hooked up to the RELNOTESng
build yet.)

Makefile#1
article.sgml#6

Reviewed by: re

21 years agoCast si_drv2 to intptr_t instead of int to quiet warnings.
jhb [Fri, 25 Oct 2002 20:49:24 +0000 (20:49 +0000)]
Cast si_drv2 to intptr_t instead of int to quiet warnings.

Tested on: alpha
Reported by: marcel (on an ia64)

21 years agoComment describing the semantics of mac_late.
rwatson [Fri, 25 Oct 2002 20:45:27 +0000 (20:45 +0000)]
Comment describing the semantics of mac_late.
Trim trailing whitespace.

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

21 years agoUse the newer "+" modifier on output contraints when a register or
jhb [Fri, 25 Oct 2002 20:22:12 +0000 (20:22 +0000)]
Use the newer "+" modifier on output contraints when a register or
memory datum is used for both input and output instead of using
matching constraints.

21 years agoReduce the GEOM verbosity under bootverbose to something more sufferable.
phk [Fri, 25 Oct 2002 20:09:45 +0000 (20:09 +0000)]
Reduce the GEOM verbosity under bootverbose to something more sufferable.
This is not quite the set of information I would want, but the tree where
I have the "correct" version is messed up with conflicts.

Sponsored by: DARPA & NAI Labs.

21 years agoDon't copy td_md. Instead, let the MD code handle it just like it
jhb [Fri, 25 Oct 2002 20:07:52 +0000 (20:07 +0000)]
Don't copy td_md.  Instead, let the MD code handle it just like it
handles the MD fields of all the other MD portions of proc-related
structures.

Tested on: i386, alpha, sparc64

21 years agoNote that the sched_lock protects md_ldt of struct mdproc.
jhb [Fri, 25 Oct 2002 20:06:16 +0000 (20:06 +0000)]
Note that the sched_lock protects md_ldt of struct mdproc.

21 years ago- Rename the DDB specific %z printf format to %y.
mux [Fri, 25 Oct 2002 19:41:32 +0000 (19:41 +0000)]
- Rename the DDB specific %z printf format to %y.
- Make DDB use %y instead of %z.
- Teach GCC about %y.
- Implement support for the C99 %z format modifier.

Approved by: re@
Reviewed by: peter
Tested on: i386, sparc64

21 years agoProvide a bit of anti-foot-shooting protection. Make sure that in the
peter [Fri, 25 Oct 2002 19:15:29 +0000 (19:15 +0000)]
Provide a bit of anti-foot-shooting protection.  Make sure that in the
non-cross cases without DESTDIR, that the bin/sh that we're about to
install works.  Otherwise, a 'make installworld' without having already
rebooted with a post-signal-fix kernel is a rather big disaster when
important things like /bin/sh coredump.

21 years agoFinish fixing the 5.x FPU code for dealing with signal handlers.
peter [Fri, 25 Oct 2002 19:12:16 +0000 (19:12 +0000)]
Finish fixing the 5.x FPU code for dealing with signal handlers.

Obtained from:  bde

21 years agoNo need for this config file; games are gone.
markm [Fri, 25 Oct 2002 19:11:19 +0000 (19:11 +0000)]
No need for this config file; games are gone.

21 years agoSplit 4.x and 5.x signal handling so that we can keep 4.x signal
peter [Fri, 25 Oct 2002 19:10:58 +0000 (19:10 +0000)]
Split 4.x and 5.x signal handling so that we can keep 4.x signal
handling clean and functional as 5.x evolves.  This allows some of the
nasty bandaids in the 5.x codepaths to be unwound.

Encapsulate 4.x signal handling under COMPAT_FREEBSD4 (there is an
anti-foot-shooting measure in place, 5.x folks need this for a while) and
finish encapsulating the older stuff under COMPAT_43.  Since the ancient
stuff is required on alpha (longjmp(3) passes a 'struct osigcontext *'
to the current sigreturn(2), instead of the 'ucontext_t *' that sigreturn
is supposed to take), add a compile time check to prevent foot shooting
there too.  Add uniform COMPAT_43 stubs for ia64/sparc64/powerpc.

Tested on: i386, alpha, ia64.  Compiled on sparc64 (a few days ago).
Approved by: re

21 years agoChange method to determine base memory size.
iwasaki [Fri, 25 Oct 2002 18:46:36 +0000 (18:46 +0000)]
Change method to determine base memory size.
Try INT 15H/E820H first, then fall back to the old compatibility
method (INT 12H).
This is a workaround for newer machines which have broken INT 12H BIOS
service implementation.

Reviewed by: -current ML
MFC after: 3 days

21 years ago#include <geom/geom.h> to get proper prototypes. Contrary to my fears we
phk [Fri, 25 Oct 2002 18:44:42 +0000 (18:44 +0000)]
#include <geom/geom.h> to get proper prototypes.  Contrary to my fears we
seem to have all the prerequisites already.

Call g_waitidle() as the first thing in vfs_mountroot() so that we have
it out of the way before we even decide if we should call .._ask() or
.._try().

Call the g_dev_print() function to provide better guidance for the
root-mount prompt.

21 years agoAdd a g_dev_print() function which prints all the /dev entries GEOM
phk [Fri, 25 Oct 2002 18:42:42 +0000 (18:42 +0000)]
Add a g_dev_print() function which prints all the /dev entries GEOM
know about.

21 years agoInitialize tick_MHz and related variables much earlier. After the last
tmm [Fri, 25 Oct 2002 17:42:14 +0000 (17:42 +0000)]
Initialize tick_MHz and related variables much earlier. After the last
revision of tick.c, this was done at SI_SUB_CLOCKS, which is too late
because tick_MHz is required for DELAY() to work.

Reviewed by: jake

21 years agoFix iommu_dvmamap_sync() to use the right address when flushing the
tmm [Fri, 25 Oct 2002 17:31:31 +0000 (17:31 +0000)]
Fix iommu_dvmamap_sync() to use the right address when flushing the
streaming cache. This bug could have the potential to cause data
corruption on systems with Psycho U2P bridges (Sabre bridges have no
streaming cache).
However, due to the usual driver architecture, it is believed that
corruption did occur only in rare cases (if at all).

21 years agoDon't check IFF_RUNNING in previous change.
simokawa [Fri, 25 Oct 2002 17:31:03 +0000 (17:31 +0000)]
Don't check IFF_RUNNING in previous change.
The flag is sometimes unset if the interface has IPv6 link-local
address only.

21 years ago- scopeid is u_int32_t
ume [Fri, 25 Oct 2002 17:07:02 +0000 (17:07 +0000)]
- scopeid is u_int32_t
- strtoul pedant.  pointed out by deraadt

Obtained from: KAME
MFC after: 1 week

21 years agoExplicitly list architectures supporting sysinstall
phk [Fri, 25 Oct 2002 17:04:49 +0000 (17:04 +0000)]
Explicitly list architectures supporting sysinstall

21 years agoLoose the g_dev_clone() noise.
phk [Fri, 25 Oct 2002 17:00:15 +0000 (17:00 +0000)]
Loose the g_dev_clone() noise.