]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoAdd a missing dereference level. This caused nfsm_postop_attr_xx()
Ian Dowse [Tue, 25 Sep 2001 00:00:33 +0000 (00:00 +0000)]
Add a missing dereference level. This caused nfsm_postop_attr_xx()
to try and extract node attributes from an RPC reply even if none
were present.

Reviewed by: peter

23 years agoUse b6 instead of b1 - b1 is supposed to be preserved and b6 is scratch.
Doug Rabson [Mon, 24 Sep 2001 22:50:20 +0000 (22:50 +0000)]
Use b6 instead of b1 - b1 is supposed to be preserved and b6 is scratch.

23 years agoMake the Alternate {I,D} TLB vector code actually work for virtual
Doug Rabson [Mon, 24 Sep 2001 22:49:20 +0000 (22:49 +0000)]
Make the Alternate {I,D} TLB vector code actually work for virtual
addresses greater than 256M (the page size for region 6 and 7).

23 years agoghostscript6 has been replaced by ghostscript-gnu in the Ports
Bruce A. Mah [Mon, 24 Sep 2001 20:15:04 +0000 (20:15 +0000)]
ghostscript6 has been replaced by ghostscript-gnu in the Ports
Collection, so we need to change the ports we need for a release.

23 years agoDon't try to access external files from SKI unless we are actually running
Doug Rabson [Mon, 24 Sep 2001 20:01:29 +0000 (20:01 +0000)]
Don't try to access external files from SKI unless we are actually running
in SKI.

23 years agoIncrease the number of bootstrap PVs.
Doug Rabson [Mon, 24 Sep 2001 20:00:20 +0000 (20:00 +0000)]
Increase the number of bootstrap PVs.

23 years agoInclude <machine/pte.h> instead of <machine/pmap.h>
Doug Rabson [Mon, 24 Sep 2001 19:58:15 +0000 (19:58 +0000)]
Include <machine/pte.h> instead of <machine/pmap.h>

23 years agoWe need different call stubs for static and stacked calling conventions.
Doug Rabson [Mon, 24 Sep 2001 19:41:16 +0000 (19:41 +0000)]
We need different call stubs for static and stacked calling conventions.

23 years agoAdd commands to dump the itrs and dtrs.
Doug Rabson [Mon, 24 Sep 2001 19:39:34 +0000 (19:39 +0000)]
Add commands to dump the itrs and dtrs.

23 years agoReturn the mapkey which EFI gave us when we read the memory map - we need
Doug Rabson [Mon, 24 Sep 2001 19:37:44 +0000 (19:37 +0000)]
Return the mapkey which EFI gave us when we read the memory map - we need
it to call ExitBootServices.

23 years agoTidy up a little - don't try to print anything or enable interrupts after
Doug Rabson [Mon, 24 Sep 2001 19:36:45 +0000 (19:36 +0000)]
Tidy up a little - don't try to print anything or enable interrupts after
we start changing translation registers. Also, call ExitBootServices
before we jump into the kernel.

23 years agoPick up pal.s from the kernel sources.
Doug Rabson [Mon, 24 Sep 2001 19:31:44 +0000 (19:31 +0000)]
Pick up pal.s from the kernel sources.

23 years agoFactor out PTE and related definitions from pmap.h - they are useful in
Doug Rabson [Mon, 24 Sep 2001 19:27:38 +0000 (19:27 +0000)]
Factor out PTE and related definitions from pmap.h - they are useful in
the loader.

23 years agoThe addition of i_dirhash to struct inode pushed RELENG_4's
Ian Dowse [Mon, 24 Sep 2001 18:29:20 +0000 (18:29 +0000)]
The addition of i_dirhash to struct inode pushed RELENG_4's
sizeof(struct inode) into a new malloc bucket on the i386. This
didn't happen in -current due to the removal of i_lock, but it does
no harm to apply the workaround to -current first.

Reduce the size of the i_spare[] array in struct inode from 4 to
3 entries, and change ext2fs to use i_din.di_spare[1] so that it
does not need i_spare[3].

Reviewed by: bde
MFC after: 3 days

23 years agoBack out rev 1.274. The previous behavior was documented in rc.8 and
David E. O'Brien [Mon, 24 Sep 2001 18:17:11 +0000 (18:17 +0000)]
Back out rev 1.274.  The previous behavior was documented in rc.8 and
the original commit of local_startup depended on the scripts being
executable; so there is too much precedence to change it now.  About all
anyone could agree on is that rev 1.274 broke POLA and before rev 1.274
also broke POLA.

23 years agoClarify that -h implies -c.
David E. O'Brien [Mon, 24 Sep 2001 17:42:37 +0000 (17:42 +0000)]
Clarify that -h implies -c.

23 years agoFix a few comment typos from the last commit.
Doug Rabson [Mon, 24 Sep 2001 17:38:58 +0000 (17:38 +0000)]
Fix a few comment typos from the last commit.

23 years agoAdd some code which can be used to change to/from physical mode when
Doug Rabson [Mon, 24 Sep 2001 17:07:23 +0000 (17:07 +0000)]
Add some code which can be used to change to/from physical mode when
calling various firmware functions.

23 years agoOnly compile with strtofflags.c from libc if bootstrapping.
Ruslan Ermilov [Mon, 24 Sep 2001 15:38:38 +0000 (15:38 +0000)]
Only compile with strtofflags.c from libc if bootstrapping.

23 years agoo Reduce userland inclusion of kernel headers -- remove unneeded include
Robert Watson [Mon, 24 Sep 2001 15:00:16 +0000 (15:00 +0000)]
o Reduce userland inclusion of kernel headers -- remove unneeded include
  of <sys/mbuf.h>.

Reviewed by: jlemon

23 years agoo Reduce userland inclusion of kernel headers -- remove unneeded include
Robert Watson [Mon, 24 Sep 2001 14:43:35 +0000 (14:43 +0000)]
o Reduce userland inclusion of kernel headers -- remove unneeded include
  of <sys/mbuf.h>.

Reviewed by: jlemon

23 years agoFix typo.
Hiroki Sato [Mon, 24 Sep 2001 12:56:52 +0000 (12:56 +0000)]
Fix typo.

23 years agoMerge the following from the English version:
Hiroki Sato [Mon, 24 Sep 2001 12:38:06 +0000 (12:38 +0000)]
Merge the following from the English version:

1.139 -> 1.145 relnotes/common/new.sgml

23 years agoAdd a couple of functions to create persistent variables that will
Murray Stokely [Mon, 24 Sep 2001 10:16:23 +0000 (10:16 +0000)]
Add a couple of functions to create persistent variables that will
survive a sysinstall Ctrl-C -> 'Restart'.  This fixes another annoying
bug where restarting sysinstall will try to reload kernel modules and
do other external things that have already been done.  For now, use
these persistent variables to keep track of module, usbd, and pccardd
initialization.

Bug found by: rwatson
MFC after:    1 week

23 years agoFixed the buildworld breakage in cross-tools caused by
Ruslan Ermilov [Mon, 24 Sep 2001 09:25:05 +0000 (09:25 +0000)]
Fixed the buildworld breakage in cross-tools caused by
misuse of /usr/src/include headers.  This REALLY fixes
the 20010919 src/UPDATING entry.

With this patch the 4.2-RELEASE box was able to survive
the 5.0-CURRENT "make world".

Beat over the head with this patch: obrien

23 years agoNote that it might not be a good idea to mount /var 'noatime'.
Murray Stokely [Mon, 24 Sep 2001 07:35:37 +0000 (07:35 +0000)]
Note that it might not be a good idea to mount /var 'noatime'.

PR: docs/30675
Submitted by: Eugene M. Kim <gene@nttmcl.com>, and
                Giorgos Keramidas <charon@labs.gr>

23 years ago- Added #include <sys/systm.h>
Yoshihiro Takahashi [Mon, 24 Sep 2001 06:41:06 +0000 (06:41 +0000)]
- Added #include <sys/systm.h>
- Cosmetic change

23 years agoCorrectly document the return value of mb_fixhdr.
Murray Stokely [Mon, 24 Sep 2001 06:30:52 +0000 (06:30 +0000)]
Correctly document the return value of mb_fixhdr.

PR: docs/30697
Submitted by: Jonathan Hanna <jhanna@home.com>
Reviewed by: bp

23 years agoFix a null pointer dereference introduced in the last commit, plus
Luigi Rizzo [Mon, 24 Sep 2001 05:24:19 +0000 (05:24 +0000)]
Fix a null pointer dereference introduced in the last commit, plus
remove a useless assignment and move a comment.

Submitted by: Thomas Moestl

23 years ago+ Fix misplacement of `txp'
David E. O'Brien [Mon, 24 Sep 2001 03:23:48 +0000 (03:23 +0000)]
+ Fix misplacement of `txp'
+ Document our -CURRENT debugging bits

23 years ago+ explicitly use the shell to run ${local_startup} so one does not have to
David E. O'Brien [Mon, 24 Sep 2001 03:03:51 +0000 (03:03 +0000)]
+ explicitly use the shell to run ${local_startup} so one does not have to
  ensure the executable bit is set on the scripts.
+ Fix VCS ID's

23 years agoCatch up to recent removal of curpcb from globals.h.
Mark Peek [Mon, 24 Sep 2001 02:58:49 +0000 (02:58 +0000)]
Catch up to recent removal of curpcb from globals.h.

23 years ago- Correct capitalization of a function name.
Murray Stokely [Mon, 24 Sep 2001 02:51:43 +0000 (02:51 +0000)]
- Correct capitalization of a function name.
- Add a missing word to form a complete sentence.

PR: docs/30626
Submitted by: Dan Lukes <dan@obluda.cz>, and
swear@blarg.net (Gary W. Swearingen)

23 years agoTurn on WARNS?= 2 for libdevstat.
Kenneth D. Merry [Sun, 23 Sep 2001 23:03:23 +0000 (23:03 +0000)]
Turn on WARNS?= 2 for libdevstat.

Submitted by: "Sergey A. Osokin" <osa@freebsd.org.ru>

23 years agoThe README file in the floppies/ directory of a distribution should
Bruce A. Mah [Sun, 23 Sep 2001 22:32:09 +0000 (22:32 +0000)]
The README file in the floppies/ directory of a distribution should
come from the installation document (of the release documentation)
not the readme.  The installation document is the one that has the
content of the old FLOPPIES.TXT.

Pointy hat to: bmah

23 years agoDon't print running stats unless we're the foreground process.
Dag-Erling Smørgrav [Sun, 23 Sep 2001 22:03:39 +0000 (22:03 +0000)]
Don't print running stats unless we're the foreground process.

PR: 30764

23 years agoSend a .4 second BREAK rather than a .25 second break so that
Brian Somers [Sun, 23 Sep 2001 20:03:40 +0000 (20:03 +0000)]
Send a .4 second BREAK rather than a .25 second break so that
we're at least consistent with what tcsendbreak(3) is documented
to do.

MFC after: 2 weeks

Note, the MFC will be to sys/dev/dgb/dgm.c on the RELENG_4 branch

23 years agoAdd commands to dump the configuration tables and the SAL System Table.
Doug Rabson [Sun, 23 Sep 2001 10:28:01 +0000 (10:28 +0000)]
Add commands to dump the configuration tables and the SAL System Table.

23 years agoAdd definitions of SAL System Table.
Doug Rabson [Sun, 23 Sep 2001 10:26:20 +0000 (10:26 +0000)]
Add definitions of SAL System Table.

23 years agoReinitialize the keyboard after the ACPI resume event.
Kazutaka YOKOTA [Sun, 23 Sep 2001 08:42:06 +0000 (08:42 +0000)]
Reinitialize the keyboard after the ACPI resume event.

I am not sure if this is absolutely necessary on all systems.  Yet,
there certainly are motherboards and notebook systems which require
this, although there are other systems which just don't.  I hope we
shall know when to do this on which systems, as the development of our
ACPI subsystem progresses...  (I know we didn't need this for the APM
resume.)

23 years ago- Update pnp(4) to match the reality in -CURRENT.
Kazutaka YOKOTA [Sun, 23 Sep 2001 08:41:12 +0000 (08:41 +0000)]
- Update pnp(4) to match the reality in -CURRENT.
- Add pnpbios(4).

23 years agoFix a comment line which wrongly states this man page is for
Kazutaka YOKOTA [Sun, 23 Sep 2001 06:56:09 +0000 (06:56 +0000)]
Fix a comment line which wrongly states this man page is for
the `asc' scanner device :-)

MFC after: 3 days

23 years agoWhile I am THE Warner Losh (I know of no others that share my name and
Warner Losh [Sun, 23 Sep 2001 06:44:07 +0000 (06:44 +0000)]
While I am THE Warner Losh (I know of no others that share my name and
I've looked), there's no need to refer to myself here as the Warner Losh.

Noticed by: mike

23 years agoFix grammar of the newest UPDATING entry. Also, minor rewording of
Warner Losh [Sun, 23 Sep 2001 06:36:41 +0000 (06:36 +0000)]
Fix grammar of the newest UPDATING entry.  Also, minor rewording of
grant to use this document to be less restrictive than before.  The
document now can be used without restriction.

23 years agoOverzealous search/replace: PROC_LOCK_ASSERT(p, not PROC_LOCK_ASSERT(td,
Peter Wemm [Sun, 23 Sep 2001 06:16:01 +0000 (06:16 +0000)]
Overzealous search/replace: PROC_LOCK_ASSERT(p, not PROC_LOCK_ASSERT(td,

Pointy-hat to:  peter

23 years agoImplement TCP/IP checksum off-loading on receive. Announce
Alan Cox [Sun, 23 Sep 2001 05:13:12 +0000 (05:13 +0000)]
Implement TCP/IP checksum off-loading on receive.  Announce
rxcsum capabilities.

Reviewed by: wpaul

23 years agoDon't add space after comma unless there was already one (prevents a
Mike Heffner [Sun, 23 Sep 2001 04:07:26 +0000 (04:07 +0000)]
Don't add space after comma unless there was already one (prevents a
rare overflow). Optimize loop slightly. Don't exit if realloc can't
shrink a buffer (just continue and use the larger buffer).

Obtained from: OpenBSD
MFC after: 2 weeks

23 years agoNote that strdup(3) will set errno if memory allocation fails. This
Mike Barcroft [Sun, 23 Sep 2001 00:37:42 +0000 (00:37 +0000)]
Note that strdup(3) will set errno if memory allocation fails.  This
is also required by the forthcoming POSIX.1-200x standard.

Obtained from: malloc.3

23 years agoblah, remove debugging code that crept in with last commit.
David E. O'Brien [Sun, 23 Sep 2001 00:11:48 +0000 (00:11 +0000)]
blah, remove debugging code that crept in with last commit.

23 years agoSilence more warnings.
Murray Stokely [Sat, 22 Sep 2001 23:17:37 +0000 (23:17 +0000)]
Silence more warnings.

23 years agoSilence warnings on alpha :
Murray Stokely [Sat, 22 Sep 2001 22:34:14 +0000 (22:34 +0000)]
Silence warnings on alpha :

  Use '%p' when printing out the address of a function.
  sizeof(int) != sizeof(long)

23 years agoSilence warnings.
Murray Stokely [Sat, 22 Sep 2001 22:25:35 +0000 (22:25 +0000)]
Silence warnings.

23 years agoSilence warnings :
Murray Stokely [Sat, 22 Sep 2001 22:21:01 +0000 (22:21 +0000)]
Silence warnings :

   Use static as necessary.
   Use __unused as necessary.
   sizeof(int) != sizeof(void *)

23 years agoSince we no longer inline any debugging code in the mutex operations, move
John Baldwin [Sat, 22 Sep 2001 21:19:55 +0000 (21:19 +0000)]
Since we no longer inline any debugging code in the mutex operations, move
all the debugging code into the function versions of the mutex operations
in kern_mutex.c.  This reduced the __mtx_* macros to simply wrappers of
the _{get,rel}_lock_* macros, so the __mtx_* macros were also abolished in
favor of just calling the _{get,rel}_lock_* macros.  The tangled hairy mass
of macros calling macros is at least a bit more sane now.

23 years agoAdd CONSPEED option.
Doug Rabson [Sat, 22 Sep 2001 19:53:36 +0000 (19:53 +0000)]
Add CONSPEED option.

23 years agoDon't activate the ssc console unless we are running in SKI.
Doug Rabson [Sat, 22 Sep 2001 19:52:02 +0000 (19:52 +0000)]
Don't activate the ssc console unless we are running in SKI.

23 years agoAdd implementations of readx() and writex().
Doug Rabson [Sat, 22 Sep 2001 19:51:18 +0000 (19:51 +0000)]
Add implementations of readx() and writex().

23 years agoAdd declaration of ia64_running_in_simulator().
Doug Rabson [Sat, 22 Sep 2001 19:50:54 +0000 (19:50 +0000)]
Add declaration of ia64_running_in_simulator().

23 years ago* Turn off memory descriptor debugging - its served its purpose.
Doug Rabson [Sat, 22 Sep 2001 19:50:12 +0000 (19:50 +0000)]
* Turn off memory descriptor debugging - its served its purpose.
* Don't get confused when memory regions don't lie on page boundaries -
  remember our page size is typically larger than the firmware's page size.
* Add a function ia64_running_in_simulator() which is intended to detect
  whether the kernel is running in SKI or on real hardware.

23 years agoRemove a redundant stop.
Doug Rabson [Sat, 22 Sep 2001 19:46:41 +0000 (19:46 +0000)]
Remove a redundant stop.

23 years agoTweak so I can use a serial console on ia64.
Doug Rabson [Sat, 22 Sep 2001 19:44:37 +0000 (19:44 +0000)]
Tweak so I can use a serial console on ia64.

23 years agoAdd EFI network support.
Doug Rabson [Sat, 22 Sep 2001 19:12:30 +0000 (19:12 +0000)]
Add EFI network support.

23 years ago* Flesh out elf_exec and bootinfo.
Doug Rabson [Sat, 22 Sep 2001 19:10:56 +0000 (19:10 +0000)]
* Flesh out elf_exec and bootinfo.
* Add EFI network support.

23 years agoUpdate NFS_ROOT comments to reflect the NFSCLIENT option
Chris D. Faulhaber [Sat, 22 Sep 2001 19:02:18 +0000 (19:02 +0000)]
Update NFS_ROOT comments to reflect the NFSCLIENT option
instead of the depricated NFS option.

Reviewed by: peter

23 years agoAdd the 'h' flag to operate on symlinks rather than what they point to.
David E. O'Brien [Sat, 22 Sep 2001 18:45:03 +0000 (18:45 +0000)]
Add the 'h' flag to operate on symlinks rather than what they point to.

23 years agoAdd getsecs() for the libstand network code.
Doug Rabson [Sat, 22 Sep 2001 18:33:09 +0000 (18:33 +0000)]
Add getsecs() for the libstand network code.

23 years agoAdd a twiddle meter when reading from files. Gives me something to look
Doug Rabson [Sat, 22 Sep 2001 18:31:02 +0000 (18:31 +0000)]
Add a twiddle meter when reading from files. Gives me something to look
at when a kernel is loading from a floppy.

23 years agoFix a whole bunch of dependancy bugs and make it actually work when the
Doug Rabson [Sat, 22 Sep 2001 18:27:01 +0000 (18:27 +0000)]
Fix a whole bunch of dependancy bugs and make it actually work when the
size is not a multiple of eight.

23 years agoFix byte swapping - it was totally broken.
Doug Rabson [Sat, 22 Sep 2001 18:22:53 +0000 (18:22 +0000)]
Fix byte swapping - it was totally broken.

23 years agoAdd a function to free all of sysinstall's internal variables from the
Murray Stokely [Sat, 22 Sep 2001 18:10:56 +0000 (18:10 +0000)]
Add a function to free all of sysinstall's internal variables from the
environment.  This fixes an annoying bug where hitting Ctrl-C and
telling sysinstall to 'restart' will do no such thing since many of
the options are still set and so you won't be prompted for them
again.

MFC after: 1 week

23 years ago Use the isDebug() function rather than making up a new environment
Murray Stokely [Sat, 22 Sep 2001 18:07:47 +0000 (18:07 +0000)]
  Use the isDebug() function rather than making up a new environment
variable to check for debug functionality.  Previously, you had to set
both 'debug' and 'SYSINSTALL_DEBUG' to get a log of sysinstall's
activities.  Now, only 'debug' is necessary.

23 years agoGive the pps driver an additional 8 inputs if we can persuade the
Poul-Henning Kamp [Sat, 22 Sep 2001 16:34:59 +0000 (16:34 +0000)]
Give the pps driver an additional 8 inputs if we can persuade the
ppc to go into EPP mode.  These 8 inputs are timestamped in polled
loop so their resolution will be nanoseconds but their granularity
will only be 1/hz.

23 years agoRemove IFCAP_HWCSUM, it is merely a luxury #define, not a real flag.
Jeroen Ruigrok van der Werven [Sat, 22 Sep 2001 15:59:09 +0000 (15:59 +0000)]
Remove IFCAP_HWCSUM, it is merely a luxury #define, not a real flag.
Remove explicit mention of IP stack, since it might not be accurate for all
interfaces.
Change if_enable to if_capenable, as it is spelled.

Submitted by: jlemon

23 years agoCommit descriptions for the interface capabilities support.
Jeroen Ruigrok van der Werven [Sat, 22 Sep 2001 12:34:19 +0000 (12:34 +0000)]
Commit descriptions for the interface capabilities support.

23 years agoFixing a bug that causes very short uptimes (less than one minute) to
Crist J. Clark [Sat, 22 Sep 2001 09:33:00 +0000 (09:33 +0000)]
Fixing a bug that causes very short uptimes (less than one minute) to
be 30 seconds off. (Thanks to Vladimir B.Grebenschikov
<vova@express.ru> for the PR and ru for a more elegant fix.)

PR: bin/30680
Approved by: ru

23 years agoDocument COMPAT4X.
Ruslan Ermilov [Sat, 22 Sep 2001 08:21:56 +0000 (08:21 +0000)]
Document COMPAT4X.

23 years agoReset ``schg'' flag before removing stale libs.
Ruslan Ermilov [Sat, 22 Sep 2001 08:11:24 +0000 (08:11 +0000)]
Reset ``schg'' flag before removing stale libs.

23 years agoNote in the BUGS section that the driver will not work on controllers that
Scott Long [Sat, 22 Sep 2001 04:26:35 +0000 (04:26 +0000)]
Note in the BUGS section that the driver will not work on controllers that
have version 1.x firmware.  This might also need to go into the release
documentation, as many people seem to have been bitten by this.

MFC after: 3 days

23 years agoo vpaccess() -> vn_access() -- Peter reminds me that there is already
Robert Watson [Sat, 22 Sep 2001 03:07:41 +0000 (03:07 +0000)]
o vpaccess() -> vn_access()  -- Peter reminds me that there is already
  a convention for vnop helper routines of this sort.

Submitted by: Mr Wemm <peter>

23 years agoUse __FILE__ and __LINE__ explicitly since we know we will be using them
John Baldwin [Fri, 21 Sep 2001 23:15:10 +0000 (23:15 +0000)]
Use __FILE__ and __LINE__ explicitly since we know we will be using them
when calling _mtx_assert() to prevent mtx_assert() from requiring
sys/lock.h as well as sys/mutex.h.

23 years agoFix statclock() on non-boot processors: curproc -> curthread.
John Baldwin [Fri, 21 Sep 2001 23:10:11 +0000 (23:10 +0000)]
Fix statclock() on non-boot processors: curproc -> curthread.

23 years agoo Update unistd.h with the prototype for the new eaccess(2) system call.
Robert Watson [Fri, 21 Sep 2001 22:50:39 +0000 (22:50 +0000)]
o Update unistd.h with the prototype for the new eaccess(2) system call.

Obtained from: TrustedBSD Project

23 years agoUse the passed in thread to selrecord() instead of curthread.
John Baldwin [Fri, 21 Sep 2001 22:46:54 +0000 (22:46 +0000)]
Use the passed in thread to selrecord() instead of curthread.

23 years agoUse the passed in thread to selrecord() instead of curthread.
John Baldwin [Fri, 21 Sep 2001 22:26:51 +0000 (22:26 +0000)]
Use the passed in thread to selrecord() instead of curthread.

23 years agoUse the passed in thread pointer instead of curthread in calls to
John Baldwin [Fri, 21 Sep 2001 22:22:25 +0000 (22:22 +0000)]
Use the passed in thread pointer instead of curthread in calls to
selrecord() in ptcpoll().  The pre-KSE code used the passed in proc pointer
rather than curproc, and an earlier seltrue() call uses the passed in
thread and not curthread.

23 years agoNew release notes: KSE milestone 2, groups(1) and whoami(1) subsumed
Bruce A. Mah [Fri, 21 Sep 2001 22:19:45 +0000 (22:19 +0000)]
New release notes:  KSE milestone 2, groups(1) and whoami(1) subsumed
by id(1).

23 years agoNew release notes: Load addresses of kernels exported, KVA_SPACE,
Bruce A. Mah [Fri, 21 Sep 2001 22:06:54 +0000 (22:06 +0000)]
New release notes:  Load addresses of kernels exported, KVA_SPACE,
client/server NFS split, find(1) primaries now take time units.

23 years agoThe P_SELECT flag was moved from p->p_flag to td->td_flags, but p_flag
John Baldwin [Fri, 21 Sep 2001 22:06:22 +0000 (22:06 +0000)]
The P_SELECT flag was moved from p->p_flag to td->td_flags, but p_flag
was locked by the proc lock and td_flags is locked by the sched_lock.
The places that read, set, and cleared TDF_SELECT weren't updated, so they
read and modified td_flags w/o holding the sched_lock, meaning that they
could corrupt the per-thread flags field.  As an immediate band-aid,
grab sched_lock while reading and manipulating td_flags in relation to
TDF_SELECT.  This will probably be cleaned up some later on.

23 years agoRemove unneeded proc variables and fix comments.
John Baldwin [Fri, 21 Sep 2001 21:54:45 +0000 (21:54 +0000)]
Remove unneeded proc variables and fix comments.

23 years agoo Modify access(2) man page to describe eaccess(2), and add a symlink
Robert Watson [Fri, 21 Sep 2001 21:35:22 +0000 (21:35 +0000)]
o Modify access(2) man page to describe eaccess(2), and add a symlink
  so man eaccess will return the access(2) man page.

Obtained from: TrustedBSD Project

23 years agoo Part two of eaccess(2) commit, rebuilt system call code.
Robert Watson [Fri, 21 Sep 2001 21:34:06 +0000 (21:34 +0000)]
o Part two of eaccess(2) commit, rebuilt system call code.

Obtained from: TrustedBSD Project

23 years agoo Introduce eaccess(2), a version of access(2) that uses the effective
Robert Watson [Fri, 21 Sep 2001 21:33:22 +0000 (21:33 +0000)]
o Introduce eaccess(2), a version of access(2) that uses the effective
  credentials rather than the real credentials.  This is useful for
  implementing GUI's which need to modify icons based on access rights,
  but where use of open(2) is too expensive, use of stat(2) doesn't
  reflect the file system's real protection model, and use of
  access() suffers from real/effective credential confusion.  This
  implementation provides the same semantics as the call of the same
  name on SCO OpenServer.  Note: using this call improperly can
  leave you subject to some of the same races present in the
  access(2) call.
o To implement this, break out the basic logic of access(2) into
  vpaccess(), which accepts a passed credential to perform the
  invocation of VOP_ACCESS().  Add eaccess(2) to invoke vpaccess(),
  and modify access(2) to use vpaccess().

Obtained from: TrustedBSD Project

23 years agoo Fix two eaccess() checks -- in one case, the wrong test file
Robert Watson [Fri, 21 Sep 2001 21:28:43 +0000 (21:28 +0000)]
o Fix two eaccess() checks -- in one case, the wrong test file
  was used, resulting in a regression failure, and in the other,
  the test on an error return was inverted.

Obtained from: TrustedBSD Project

23 years agoNew release notes: cvs(1) -T, cvs(1) diff -j.
Bruce A. Mah [Fri, 21 Sep 2001 21:11:18 +0000 (21:11 +0000)]
New release notes:  cvs(1) -T, cvs(1) diff -j.

MFCs noted:  cat(1) UNIX-domain sockets, dirpref.

While I'm here, group CVS-related items together in their own section
under userland contributed software.

23 years agoIntroduce new syscons(4) kernel options:
Maxim Sobolev [Fri, 21 Sep 2001 20:41:24 +0000 (20:41 +0000)]
Introduce new syscons(4) kernel options:
- SC_CUT_SPACES2TABS - when copying text into the cut buffer convert leading
  spaces into the tabs;
- SC_CUT_SEPCHARS="XYZ" - treat supplied characters as possible words
  separators when the driver searches for words boundaries when doing cut
  operation.

Also unify cut code a bit to decrease amount of duplicated code. This fixes
line cut mode, so that it is no longer pads line with useless spaces.

Approved by: ru

23 years agoRemove a bogus comment. "atomic" doesn't mean that the operation is done
John Baldwin [Fri, 21 Sep 2001 19:26:57 +0000 (19:26 +0000)]
Remove a bogus comment.  "atomic" doesn't mean that the operation is done
as a physical atomic operation.  That would require the code to use the
atomic API, which it does not.  Instead, the operation is made psuedo
atomic (hence the quotes) by use of the lock to protect clearing all of the
flags in question.

23 years agoGC some #if 0'd code.
John Baldwin [Fri, 21 Sep 2001 19:21:18 +0000 (19:21 +0000)]
GC some #if 0'd code.

23 years agoWhitespace and spelling fixes.
John Baldwin [Fri, 21 Sep 2001 19:16:12 +0000 (19:16 +0000)]
Whitespace and spelling fixes.

23 years agooops, forgot to commit man page for cvs changes
Matthew Dillon [Fri, 21 Sep 2001 19:05:36 +0000 (19:05 +0000)]
oops, forgot to commit man page for cvs changes

23 years agoKeep up with sys/net/rtsock.c,v 1.58.
Ruslan Ermilov [Fri, 21 Sep 2001 16:14:53 +0000 (16:14 +0000)]
Keep up with sys/net/rtsock.c,v 1.58.

Prodded by: brian