]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoAdd A section on the retransmit timer sysctls.
Matthew Dillon [Sun, 25 Aug 2002 01:51:57 +0000 (01:51 +0000)]
Add A section on the retransmit timer sysctls.

MFC after: 3 days

22 years agoFixed editing errors in rev.1.4 which manifested as printf format errors
Bruce Evans [Sun, 25 Aug 2002 01:32:22 +0000 (01:32 +0000)]
Fixed editing errors in rev.1.4 which manifested as printf format errors
at compile time and probably as panics at runtime.

22 years agoFixed printf format errors and style bugs in previous commit.
Bruce Evans [Sun, 25 Aug 2002 01:08:27 +0000 (01:08 +0000)]
Fixed printf format errors and style bugs in previous commit.

22 years agoReorg just a tad to better express our bzip intentions.
David E. O'Brien [Sun, 25 Aug 2002 01:01:08 +0000 (01:01 +0000)]
Reorg just a tad to better express our bzip intentions.

22 years agoFind remaining hardcoded gzip bits and change to bzip2 bits.
David E. O'Brien [Sun, 25 Aug 2002 01:00:16 +0000 (01:00 +0000)]
Find remaining hardcoded gzip bits and change to bzip2 bits.
Mark with XXX so someone that cares about being able to handle either
bziped or gziped packages knows where the remaining nits are.

22 years ago o Retire vm_page_zero_fill() and vm_page_zero_fill_area(). Ever since
Alan Cox [Sun, 25 Aug 2002 00:22:31 +0000 (00:22 +0000)]
 o Retire vm_page_zero_fill() and vm_page_zero_fill_area().  Ever since
   pmap_zero_page() and pmap_zero_page_area() were modified to accept
   a struct vm_page * instead of a physical address, vm_page_zero_fill()
   and vm_page_zero_fill_area() have served no purpose.

22 years agoFixed white space around operators, casts and reserved words.
Jake Burkholder [Sat, 24 Aug 2002 22:55:16 +0000 (22:55 +0000)]
Fixed white space around operators, casts and reserved words.

Reviewed by: md5

22 years agoUse 'p' as the partition specifier instead of 's'. We continue to use
Marcel Moolenaar [Sat, 24 Aug 2002 22:42:16 +0000 (22:42 +0000)]
Use 'p' as the partition specifier instead of 's'. We continue to use
's' for compatibility partitions (ie partitions with a BSD disklabel).
Partition numbers continue to start with 1.
Example /etc/fstab:
# Device        Mountpoint      FStype          Options ...
/dev/da0p1      /efi            msdos           rw ...
/dev/da0p2      /               ufs             rw ...
/dev/da0p3      none            swap            sw ...

22 years agoReally add new fxp(4) adapter names this time.
Bruce A. Mah [Sat, 24 Aug 2002 22:26:52 +0000 (22:26 +0000)]
Really add new fxp(4) adapter names this time.

Forgotten part of a patch from:

PR: 41759
Submitted by: David Siebörger <drs@rucus.ru.ac.za>

22 years agoreturn x; -> return (x);
Jake Burkholder [Sat, 24 Aug 2002 22:01:40 +0000 (22:01 +0000)]
return x; -> return (x);
return(x); -> return (x);

Reviewed by: md5

22 years agoPreserve suid/gid permissions when extracting packages remotelly
Mario Sergio Fujikawa Ferreira [Sat, 24 Aug 2002 20:04:17 +0000 (20:04 +0000)]
Preserve suid/gid permissions when extracting packages remotelly
fetched with -r option

Reviewed by: will
Approved by: will

22 years agoBe sure to unregister from sndstat on unregister. Gets rid of phantom
Nick Sayer [Sat, 24 Aug 2002 19:13:11 +0000 (19:13 +0000)]
Be sure to unregister from sndstat on unregister. Gets rid of phantom
sndstat output after removing uaudio.

22 years ago o Use vm_object_lock() in place of directly locking Giant.
Alan Cox [Sat, 24 Aug 2002 18:44:52 +0000 (18:44 +0000)]
 o Use vm_object_lock() in place of directly locking Giant.

Reviewed by: md5

22 years ago o Use vm_object_lock() in place of Giant when manipulating a vm object
Alan Cox [Sat, 24 Aug 2002 17:52:08 +0000 (17:52 +0000)]
 o Use vm_object_lock() in place of Giant when manipulating a vm object
   in vm_map_insert().

22 years agoFix a bug in __ivaliduser_sa() which caused some rsh/rlogin attempts
John Polstra [Sat, 24 Aug 2002 17:37:42 +0000 (17:37 +0000)]
Fix a bug in __ivaliduser_sa() which caused some rsh/rlogin attempts
to fail needlessly if a reverse DNS lookup of the IP address didn't
come up with a hostname.  As a comment in the code clearly stated,
the "damn hostname" was looked up only for the purpose of netgroup
matching.  But if that lookup failed, the function bailed out
immediately even though in many cases netgroup matching would not
be used.

This change marks the hostname as unknown but continues.  Where
netgroup matching is performed, an unknown hostname is handled
conservatively.  I.e., for "+@netgroup" (accept) entries an unknown
hostname never matches, and for "-@netgroup" (reject) entries an
unknown hostname always matches.

In the lines affected (only), I also fixed a few bogus casts.  There
are others, and in fact this entire file would be a good candidate
for a cleanup sweep.

Reviewed by: imp (wearing his flourescent yellow Security Team cap)
MFC after: 2 days

22 years agoCorrect bug in t_bw_rtttime rollover, #undef USERTT
Matthew Dillon [Sat, 24 Aug 2002 17:22:44 +0000 (17:22 +0000)]
Correct bug in t_bw_rtttime rollover, #undef USERTT

22 years agoUpdate the list of chips and cards supported by this driver.
Bruce A. Mah [Sat, 24 Aug 2002 17:00:17 +0000 (17:00 +0000)]
Update the list of chips and cards supported by this driver.

Based heavily on patches contained in...

PR: 41759
Submitted by: David "Siebörger" <drs@rucus.ru.ac.za>

22 years agoThere is a built-in command called "builtin"; spell its name correctly
Tim J. Robbins [Sat, 24 Aug 2002 12:41:23 +0000 (12:41 +0000)]
There is a built-in command called "builtin"; spell its name correctly
after rev. 1.77 called it "built-in".

22 years agoUse "kbytes" instead of "blocks" in messages as this is really
Jens Schweikhardt [Sat, 24 Aug 2002 11:03:48 +0000 (11:03 +0000)]
Use "kbytes" instead of "blocks" in messages as this is really
what is calculated.

PR: 41936
Submitted by: Lukas Ertl <le@univie.ac.at>
MFC after: 1 week

22 years agoFix typos: s/O2Mirco/O2Micro/
Marc Fonvieille [Sat, 24 Aug 2002 07:46:08 +0000 (07:46 +0000)]
Fix typos: s/O2Mirco/O2Micro/

Approved by: imp

22 years agoDon't show the process ID of background jobs that have terminated,
Tim J. Robbins [Sat, 24 Aug 2002 07:19:01 +0000 (07:19 +0000)]
Don't show the process ID of background jobs that have terminated,
for consistency with ksh.

22 years ago&x is not a format string
Kris Kennaway [Sat, 24 Aug 2002 07:15:55 +0000 (07:15 +0000)]
&x is not a format string

22 years ago o Resurrect vm_object_lock() and vm_object_unlock() from revision 1.19.
Alan Cox [Sat, 24 Aug 2002 07:15:14 +0000 (07:15 +0000)]
 o Resurrect vm_object_lock() and vm_object_unlock() from revision 1.19.
   (For now, they simply acquire and release Giant.)

22 years agoWork around a GCC optimization bug on ia64: In link_elf_symbol_values(),
Marcel Moolenaar [Sat, 24 Aug 2002 05:01:33 +0000 (05:01 +0000)]
Work around a GCC optimization bug on ia64: In link_elf_symbol_values(),
a pointer to a symbol is given and we have to find the containing symbol
table. We do this by bounds checking. For some strange reason (ie I
haven't found the root cause) the first test succeeded for said symbol,
implying that the symbol came from the .dynsym table. In reality however
the symbol actually resided in the .symtab table. Needless to say that
all that was returned was junk.

The upper bounds check was: (symptr - baseptr) < symtab_size
This has been rewritten to: symptr < (baseptr + symtab_size)

As a side-effect, slightly more optimal (and still correct :-) code can
be generated on ia64.

22 years agocheck packet length before fetching ESP crypto checksum.
Hajimu UMEMOTO [Sat, 24 Aug 2002 04:48:13 +0000 (04:48 +0000)]
check packet length before fetching ESP crypto checksum.

Obtained from: KAME
MFC after: 2 days

22 years agoMake this compile again (libgmd now is a port, and remove some
Luigi Rizzo [Sat, 24 Aug 2002 03:54:38 +0000 (03:54 +0000)]
Make this compile again (libgmd now is a port, and remove some
binaries to save space).

22 years agoMention the special pathname @__CWD__@
Luigi Rizzo [Sat, 24 Aug 2002 03:33:31 +0000 (03:33 +0000)]
Mention the special pathname @__CWD__@

Modify entry for libgmd (used by ssh) now that it is not part of the
base system anymore.

22 years agoUse `uint32_t' instead of `unsigned long', since the code assumes 32-bit
Jacques Vidrine [Sat, 24 Aug 2002 02:53:23 +0000 (02:53 +0000)]
Use `uint32_t' instead of `unsigned long', since the code assumes 32-bit
arithmetic.

Reviewed by: make test

The fact that bdes(1) didn't work was
Reported by: Fred Clift <fclift@verio.net>

22 years agoAdd some \n's to printf()s
Peter Wemm [Sat, 24 Aug 2002 02:51:28 +0000 (02:51 +0000)]
Add some \n's to printf()s

22 years agoFix unit wiring. Also, change the variable "hit" to "wired" so that it
Peter Wemm [Sat, 24 Aug 2002 02:47:01 +0000 (02:47 +0000)]
Fix unit wiring.  Also, change the variable "hit" to "wired" so that it
is more obvious.

Obtained from: gibbs/ken

22 years agoNote that <sys/types.h> in no longer a prerequisite for <utime.h> and
Mike Barcroft [Sat, 24 Aug 2002 00:39:43 +0000 (00:39 +0000)]
Note that <sys/types.h> in no longer a prerequisite for <utime.h> and
<sys/mman.h>.

22 years agoAdd time_t typedef for POSIX compatibility, move an include.
Mike Barcroft [Sat, 24 Aug 2002 00:11:52 +0000 (00:11 +0000)]
Add time_t typedef for POSIX compatibility, move an include.

22 years agostyle: put return types on a line by themselves.
Alfred Perlstein [Sat, 24 Aug 2002 00:02:03 +0000 (00:02 +0000)]
style: put return types on a line by themselves.

22 years agostyle:
Alfred Perlstein [Fri, 23 Aug 2002 23:49:02 +0000 (23:49 +0000)]
style:
 put return values on a line by themselves.
 fix some paste issues where whitespace was used instead of tabs.

22 years agoPut return values from functions on a line by themselves.
Alfred Perlstein [Fri, 23 Aug 2002 23:19:25 +0000 (23:19 +0000)]
Put return values from functions on a line by themselves.

Ok'd previously by: wpaul

22 years agoMove the TAILQ_INIT(&td->td_selq) before the retry: label. Otherwise in
Peter Wemm [Fri, 23 Aug 2002 22:43:28 +0000 (22:43 +0000)]
Move the TAILQ_INIT(&td->td_selq) before the retry: label.  Otherwise in
some circumstances when we get a select collision, we can end up with
cases where we do not clear some sip->si_thread on the way out, leading to
page faults in selwakeup().  This should solve the problem where postfix
can crash the kernel during select collisions.

Reviewed by: alfred

22 years agoNew release note: MAC (mostly a placeholder for now).
Bruce A. Mah [Fri, 23 Aug 2002 22:32:57 +0000 (22:32 +0000)]
New release note:  MAC (mostly a placeholder for now).

MFCs noted:  smbfs-1.4.5 (small grammatical fix).

Markup fix:  Use manpage entity for uvisor(4).

22 years agos/_BSD_VA_LIST_/__va_list/. The former type doesn't exist anymore.
Marcel Moolenaar [Fri, 23 Aug 2002 22:19:57 +0000 (22:19 +0000)]
s/_BSD_VA_LIST_/__va_list/. The former type doesn't exist anymore.

22 years agoOk, somebody please shoot me. The asm I wrote for the ranged IPI shootdown
Peter Wemm [Fri, 23 Aug 2002 21:45:59 +0000 (21:45 +0000)]
Ok, somebody please shoot me.  The asm I wrote for the ranged IPI shootdown
was wrong.  It only ever invalidated one page due to me getting the loop
terminator wrong.  This explains the DISABLE_PG_G effect on SMP.

22 years agoMake sure channel buffer start is associated with channel, otherwise
Orion Hodson [Fri, 23 Aug 2002 20:54:32 +0000 (20:54 +0000)]
Make sure channel buffer start is associated with channel, otherwise
getptr is broken.  Noise reported by Thomas Draney <tmdraney@yahoo.com> who
also tested the patch.

22 years agoAdd the complex state TDS_SUSP_SLP.
Julian Elischer [Fri, 23 Aug 2002 20:13:22 +0000 (20:13 +0000)]
Add the complex state TDS_SUSP_SLP.

This state is to allow some experimentation and not YET used..
The theory is that a thread that is about to sleep is placed on the sleep
queue and then discovers it should suspend, and is placed on suspend queue.
(these are separate queues and it can be on both). It will not become runnable
until it has been removed from BOTH queues. i.e. a wakeup event
has occured AND the process has been unsuspended. If it were not on the sleep
queue when suspended, then the (possibly only) wakeup event might arrive and
not find any process to wake up. this would result in the thread
sleeping 'forever' when the suspension is lifted. This state will
transition to one of TDS_SLP or TDS_SUSPENDED, depending upon which
constraint is lifted first.

22 years agoDon't use "NULL" when "0" is really meant.
Archie Cobbs [Fri, 23 Aug 2002 20:07:19 +0000 (20:07 +0000)]
Don't use "NULL" when "0" is really meant.
But in this case, "-1" is really meant.

Reviewed by: darrenr

22 years agoo Fix namespace issues in <sys/mman.h>.
Mike Barcroft [Fri, 23 Aug 2002 20:04:49 +0000 (20:04 +0000)]
o Fix namespace issues in <sys/mman.h>.
o Move mode_t details from <sys/types.h> into <sys/_types.h>.
o Add primitives for sharing the mode_t and off_t typedefs.
o Add typedefs mode_t, off_t, and size_t to <sys/mman.h>.

PR: 21644

22 years agoAdd Netgear GA621, uvisor(4) devices.
Bruce A. Mah [Fri, 23 Aug 2002 19:22:22 +0000 (19:22 +0000)]
Add Netgear GA621, uvisor(4) devices.

22 years agoNew hardware list additions: mpt and pst devices.
Bruce A. Mah [Fri, 23 Aug 2002 18:01:10 +0000 (18:01 +0000)]
New hardware list additions:  mpt and pst devices.

22 years agoNew release notes: mpt driver, pst driver.
Bruce A. Mah [Fri, 23 Aug 2002 18:00:13 +0000 (18:00 +0000)]
New release notes:  mpt driver, pst driver.

While I'm here, add a missing &merged; for a recent sh(1) note.

22 years agoUpdate list of fxp(4) devices.
Bruce A. Mah [Fri, 23 Aug 2002 17:26:18 +0000 (17:26 +0000)]
Update list of fxp(4) devices.

PR: 41760
Submitted by: David "Siebörger" <drs@bashir.dsl.ru.ac.za>

22 years agoNew release notes: sh(1) parser changes, via8233 sound driver, ich
Bruce A. Mah [Fri, 23 Aug 2002 17:09:27 +0000 (17:09 +0000)]
New release notes:  sh(1) parser changes, via8233 sound driver, ich
sound driver supports ich4.

Submitted by: tjr (sh), orion (via8233, ich)

22 years agoMove several MI types from <machine/_types.h> to <sys/_types.h>.
Mike Barcroft [Fri, 23 Aug 2002 16:49:06 +0000 (16:49 +0000)]
Move several MI types from <machine/_types.h> to <sys/_types.h>.
These types are unlikely to ever become very MD.  They include:
clockid_t, ct_rune_t, fflags_t, intrmask_t, mbstate_t, off_t, pid_t,
rune_t, socklen_t, timer_t, wchar_t, and wint_t.

While moving them, make a few adjustments (submitted by bde):
o __ct_rune_t needs to be precisely `int', not necessarily __int32_t,
  since the arg type of the ctype functions is int.
o __rune_t, __wchar_t and __wint_t inherit this via a typedef of
  __ct_rune_t.
o Some minor wording changes in the comment blocks for ct_rune_t and
  mbstate_t.

Submitted by: bde (partially)

22 years agoRemove stddef.h from the header list
Scott Long [Fri, 23 Aug 2002 14:10:55 +0000 (14:10 +0000)]
Remove stddef.h from the header list

Prodded by: peter

22 years agoTweak chain example to show more possibilities
Andrey A. Chernov [Fri, 23 Aug 2002 14:03:59 +0000 (14:03 +0000)]
Tweak chain example to show more possibilities

22 years agoSimilar to bsd.subdir.mk, echo the name of the dir before
Johan Karlsson [Fri, 23 Aug 2002 12:49:16 +0000 (12:49 +0000)]
Similar to bsd.subdir.mk, echo the name of the dir before
doing the cd. This is done for bootstrap-tools,
build-tools, cross-tools, and the libraries loop.

Reviewed by: ru
Approved by: sheldonh (mentor)
MFC after: 1 week

22 years agoAdd option '-W': don't log FTP sessions to wtmp.
Yaroslav Tykhiy [Fri, 23 Aug 2002 09:06:28 +0000 (09:06 +0000)]
Add option '-W': don't log FTP sessions to wtmp.

Submitted by: maxim
MFC after: 1 week

22 years agoAdd a belated entry for amdpm(4).
Murray Stokely [Fri, 23 Aug 2002 08:00:31 +0000 (08:00 +0000)]
Add a belated entry for amdpm(4).

Submitted by: marius@alchemy.franken.de
MFC After: 1 day

22 years agoDon't re-lock the sched lock if we didn't unlock it.
Julian Elischer [Fri, 23 Aug 2002 07:23:44 +0000 (07:23 +0000)]
Don't re-lock the sched lock if we didn't unlock it.

Original error by: David Xu <bsddiy@yahoo.com>
Fix by: David Xu <bsddiy@yahoo.com>
Completely failed to spot it: Julian Elischer <julian@freebsd.org>

22 years agoDo some minor cleanups found during backport to RELENG_4.
Matt Jacob [Fri, 23 Aug 2002 06:56:31 +0000 (06:56 +0000)]
Do some minor cleanups found during backport to RELENG_4.

22 years agoPick a cleaner method (and put in a separate function) for finding
Matt Jacob [Fri, 23 Aug 2002 06:56:08 +0000 (06:56 +0000)]
Pick a cleaner method (and put in a separate function) for finding
the peer device on a dual board.

22 years agos/AC97_MIX_PHONES/AC97_MIX_AUXOUT/ to match ac97r2.{2,3}.
Orion Hodson [Fri, 23 Aug 2002 06:19:28 +0000 (06:19 +0000)]
s/AC97_MIX_PHONES/AC97_MIX_AUXOUT/ to match ac97r2.{2,3}.

Attempt to determine what function of AUX_OUT is: "True line level
out", "Headphone out", or "4-Channel out" and frig OSS mixer label
accordingly.

Addresses problem raised by Randy Bush on -multimedia of not being
able to hear audio on ich2 m/b which was eventually found to be
because the mixer monitor value was 0.  On this h/w the label
"monitor" should now be presented as the marginally more intuitive
"ogain".

22 years agoRemoved unneeded include of machine/types.h (which no longer exists).
Jake Burkholder [Fri, 23 Aug 2002 05:01:19 +0000 (05:01 +0000)]
Removed unneeded include of machine/types.h (which no longer exists).

22 years agoFix warning when calling w(1) when logged in via xdm/kdm. This is
Sean Chittenden [Fri, 23 Aug 2002 04:31:58 +0000 (04:31 +0000)]
Fix warning when calling w(1) when logged in via xdm/kdm.  This is
really a problem with utmp/wtmp, but takes the same approach as who(1).

Reviewed by: knu (mentor), mini, silence on -audit
Approved by: knu (mentor), mini

22 years agoAllow chains in any form like <name1><name2> or \xf1\xf2, not binary
Andrey A. Chernov [Fri, 23 Aug 2002 04:18:26 +0000 (04:18 +0000)]
Allow chains in any form like <name1><name2> or \xf1\xf2, not binary
representation only.

22 years agoWe cannot use an alloc with only inputs and/or locals. The kernel
Marcel Moolenaar [Fri, 23 Aug 2002 03:47:50 +0000 (03:47 +0000)]
We cannot use an alloc with only inputs and/or locals. The kernel
assumes that the parameters are passed in output registers. Remove
the alloc entirely, but don't depend on the kernel not trashing
our registers.

22 years ago - Fix a mistake in my last few commits. The PDROP flag stops msleep from
Jeff Roberson [Fri, 23 Aug 2002 00:32:03 +0000 (00:32 +0000)]
 - Fix a mistake in my last few commits.  The PDROP flag stops msleep from
   re-acquiring the mutex.

Pointy hat to: me
Noticed by: tegge

22 years agoWrap the header to prevent multiple inclusion, and mark the DECLS section.
Juli Mallett [Thu, 22 Aug 2002 23:35:35 +0000 (23:35 +0000)]
Wrap the header to prevent multiple inclusion, and mark the DECLS section.

Reminded by: Rachel Hestilow <hestilow@ximian.com>

22 years agoFix a bug where large msdos partitions were not handled correctly, and fix
Tom Rhodes [Thu, 22 Aug 2002 22:17:11 +0000 (22:17 +0000)]
Fix a bug where large msdos partitions were not handled correctly, and fix
a few fsck_msdosfs related 'issues'

PR: 28536, 30168
Submitted by: Jiangyi Liu <jyliu@163.net> && NetBSD
Approved by: rwatson (mentor)

22 years agos/sus/sys/ in the a.out kernel case.
Peter Wemm [Thu, 22 Aug 2002 22:01:53 +0000 (22:01 +0000)]
s/sus/sys/ in the a.out kernel case.

Submitted by: julian

22 years agoslight cleanup of single-threading code for KSE processes
Julian Elischer [Thu, 22 Aug 2002 21:45:58 +0000 (21:45 +0000)]
slight cleanup of single-threading code for KSE processes

22 years agoReplace (ab)uses of "NULL" where "0" is really meant.
Archie Cobbs [Thu, 22 Aug 2002 21:24:01 +0000 (21:24 +0000)]
Replace (ab)uses of "NULL" where "0" is really meant.

22 years agoUse machine/limits.h rather than userland limits.h (which may come from
Peter Wemm [Thu, 22 Aug 2002 20:44:17 +0000 (20:44 +0000)]
Use machine/limits.h rather than userland limits.h (which may come from
the userland source tree, or even worse: /usr/include)

22 years agoInstead of grabbing the userland a.out.h/link.h (or worse, from
Peter Wemm [Thu, 22 Aug 2002 20:43:07 +0000 (20:43 +0000)]
Instead of grabbing the userland a.out.h/link.h (or worse, from
/usr/include!), use sys/nlist_aout.h, machine/reloc.h, sys/imgact_aout.h
and sys/link_aout.h.

22 years agoInstead of nlist.h and link.h, use sys/nlist_aout.h and sys/link_elf.h
Peter Wemm [Thu, 22 Aug 2002 20:39:30 +0000 (20:39 +0000)]
Instead of nlist.h and link.h, use sys/nlist_aout.h and sys/link_elf.h
This avoids reaching out into userland sources (or worse: /usr/include!)
for building the kernel.

22 years agoUse sys/nlist_aout.h for shared-with-kernel definitions.
Peter Wemm [Thu, 22 Aug 2002 20:37:57 +0000 (20:37 +0000)]
Use sys/nlist_aout.h for shared-with-kernel definitions.

22 years agoUse sys/link_elf.h or sys/link_aout.h based on compile environment
Peter Wemm [Thu, 22 Aug 2002 20:37:30 +0000 (20:37 +0000)]
Use sys/link_elf.h or sys/link_aout.h based on compile environment

22 years agorepo copy nlist.h to sys/nlist_aout.h, and strip out the userland
Peter Wemm [Thu, 22 Aug 2002 20:36:20 +0000 (20:36 +0000)]
repo copy nlist.h to sys/nlist_aout.h, and strip out the userland
declarations from the shared-with-kernel side of it.

22 years agoRepo copy link.h to sys/link_elf.h and sys/link_aout.h since they are
Peter Wemm [Thu, 22 Aug 2002 20:35:23 +0000 (20:35 +0000)]
Repo copy link.h to sys/link_elf.h and sys/link_aout.h since they are
shared with the kernel.  This should make it easier to #include them both
at once as well as stop grabbing stuff from /usr/include.

22 years agoNew release note: whereis -a
Johan Karlsson [Thu, 22 Aug 2002 20:35:07 +0000 (20:35 +0000)]
New release note: whereis -a

Reviewed by: bmah

22 years agoInclude sys/libkern.h for the kernel prototypes of these libkern functions
John Baldwin [Thu, 22 Aug 2002 20:08:07 +0000 (20:08 +0000)]
Include sys/libkern.h for the kernel prototypes of these libkern functions
instead of including the userland string.h header.

Prompted by: breakage with old string.h from recent machine/types.h fix
Approved by: peter (in principle)

22 years agoMany of the alpha low level console drivers still have bogus compile time
Peter Wemm [Thu, 22 Aug 2002 19:52:16 +0000 (19:52 +0000)]
Many of the alpha low level console drivers still have bogus compile time
dependencies on syscons.  Bandaid for now.

22 years ago1) add mvterm
Jens Schweikhardt [Thu, 22 Aug 2002 19:48:13 +0000 (19:48 +0000)]
1) add mvterm
2) add 132 column entries for VESA
3) fix kterm entry tc

PR: 1) misc/39628, 2) conf/13918, 3) misc/22660
Submitted by: 1) Daniel Rudy <dcrudy@pacbell.net>
Submitted by: 2) Martin Kammerhofer <dada@sbox.tu-graz.ac.at>
Submitted by: 3) Koji Mori <mori@tri.asanuma.co.jp>
MFC after: 2 weeks

22 years agoConvert NEXUS_ACCESSOR to use the __BUS_ACCESSOR
Maxime Henrion [Thu, 22 Aug 2002 19:47:10 +0000 (19:47 +0000)]
Convert NEXUS_ACCESSOR to use the __BUS_ACCESSOR
macro instead of reimplementing it.

Approved by: peter

22 years agoDefine ISP_DMA_ADDR_T to be a bus_addr_t, not a u_int32_t.
Matt Jacob [Thu, 22 Aug 2002 16:14:19 +0000 (16:14 +0000)]
Define ISP_DMA_ADDR_T to be a bus_addr_t, not a u_int32_t.

This is in preparation to completing A64 PCI support.

22 years agounused variable removal (pointed out by bde)
Matt Jacob [Thu, 22 Aug 2002 16:13:31 +0000 (16:13 +0000)]
unused variable removal (pointed out by bde)

22 years agoPrepare for importing newer version of ACPI CA.
Mitsuru IWASAKI [Thu, 22 Aug 2002 15:21:15 +0000 (15:21 +0000)]
Prepare for importing newer version of ACPI CA.

22 years agoSpell proprly properly:
Robert Watson [Thu, 22 Aug 2002 14:36:03 +0000 (14:36 +0000)]
Spell proprly properly:

  failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()
  failed to set signal flags proprly for ast()

22 years agoMerged from sys/dev/syscons/scvtb.c revision 1.9.
Yoshihiro Takahashi [Thu, 22 Aug 2002 13:24:00 +0000 (13:24 +0000)]
Merged from sys/dev/syscons/scvtb.c revision 1.9.

22 years agoMerged from sys/dev/syscons/syscons.c revision 1.388.
Yoshihiro Takahashi [Thu, 22 Aug 2002 13:23:15 +0000 (13:23 +0000)]
Merged from sys/dev/syscons/syscons.c revision 1.388.

22 years agoInclude <sys/systm.h> for the declarations of many things instead of
Bruce Evans [Thu, 22 Aug 2002 12:47:22 +0000 (12:47 +0000)]
Include <sys/systm.h> for the declarations of many things instead of
depending on namespace pollution in <sys/mumble.h>.

22 years agoAdd an entry for Avlab's low profile 4 serial port card.
John Hay [Thu, 22 Aug 2002 10:40:00 +0000 (10:40 +0000)]
Add an entry for Avlab's low profile 4 serial port card.

PR: 41146
Submitted by: Shinsuke Matsui <smatsui@internet-inc.co.jp>
Obtained from: NetBSD

22 years agoAdd an entry for VScom PCI-200L 2 port serial card.
John Hay [Thu, 22 Aug 2002 10:37:55 +0000 (10:37 +0000)]
Add an entry for VScom PCI-200L 2 port serial card.

PR: 41142
Submitted by: Shinsuke Matsui <smatsui@internet-inc.co.jp>
Obtained from: OpenBSD

22 years ago o Remove the AIOCBLIST_ASYNCFREE flag and related code. It's never set.
Alan Cox [Thu, 22 Aug 2002 08:50:15 +0000 (08:50 +0000)]
 o Remove the AIOCBLIST_ASYNCFREE flag and related code.  It's never set.

Submitted by: Romer Gil <rgil@cs.rice.edu>

22 years ago - Closer inspection revealed a possible deadlock situation in vn_lock() that
Jeff Roberson [Thu, 22 Aug 2002 07:57:43 +0000 (07:57 +0000)]
 - Closer inspection revealed a possible deadlock situation in vn_lock() that
   was introduced by my last commit but not caught by stress testing.  Fix
   that and slightly restructure the code so that it is more readable.

22 years ago - Make vn_lock() vget() and VOP_LOCK() all behave the same way WRT
Jeff Roberson [Thu, 22 Aug 2002 07:44:45 +0000 (07:44 +0000)]
 - Make vn_lock() vget() and VOP_LOCK() all behave the same way WRT
   LK_INTERLOCK.  The interlock will never be held on return from these
   functions even when there is an error.  Errors typically only occur when
   the XLOCK is held which means this isn't the vnode we want anyway.  Almost
   all users of these interfaces expected this behavior even though it was
   not provided before.

22 years ago - Return two shared locks to exclusive locks. This was premature.
Jeff Roberson [Thu, 22 Aug 2002 07:26:18 +0000 (07:26 +0000)]
 - Return two shared locks to exclusive locks.  This was premature.
 - Document the problems that prevent us from using shared locks.

22 years ago(forced commit to get the correct log message)
Jeff Roberson [Thu, 22 Aug 2002 07:01:52 +0000 (07:01 +0000)]
(forced commit to get the correct log message)
 - Remove unsafe and dead code from getnewvnode().
 - Fixup vclean() so that it grabs the VOP_LOCK before calling VOP_GETATTR()
   and grabs interlock afterwards.

22 years ago - Fix interlock handling in vn_lock(). Previously, vn_lock() could return
Jeff Roberson [Thu, 22 Aug 2002 06:58:11 +0000 (06:58 +0000)]
 - Fix interlock handling in vn_lock().  Previously, vn_lock() could return
   with interlock held in error conditions when the caller did not specify
   LK_INTERLOCK.
 - Add several comments to vn_lock() describing the rational behind the code
   flow since it was not immediately obvious.

22 years ago - Fix interlock handling in vn_lock(). Previously, vn_lock() could return
Jeff Roberson [Thu, 22 Aug 2002 06:51:06 +0000 (06:51 +0000)]
 - Fix interlock handling in vn_lock().  Previously, vn_lock() could return
   with interlock held in error conditions when the caller did not specify
   LK_INTERLOCK.
 - Add several comments to vn_lock() describing the rational behind the code
   flow since it was not immediately obvious.

22 years agoEncrypted strings (after hex decoding) aren't null terminated, because
Nick Sayer [Thu, 22 Aug 2002 06:19:07 +0000 (06:19 +0000)]
Encrypted strings (after hex decoding) aren't null terminated, because
0 might simply be part of the ciphertext.

PR: bin/40266
Submitted by: andr@dgap.mipt.ru
MFC after: 3 days

22 years agoFix a nasty memory corruption bug caused by having a bogus pointer
Marcel Moolenaar [Thu, 22 Aug 2002 03:56:57 +0000 (03:56 +0000)]
Fix a nasty memory corruption bug caused by having a bogus pointer
for the DT_IA64_PLT_RESERVE dynamic table entry. When a shared object
does not have any PLT relocations, the linker apparently doesn't find
it necessary to actually reserve the space for the BOR (Bind On
Reference) entries as pointed to by the DTE. As a result, relocatable
data in the PLT was overwritten, causing some unexpected control flow
with annoyingly predictable outcome: coredump.
To reproduce:
% echo 'int main() { return 0; }' > foo.c
% cc -o foo foo.c -lxpg4

22 years agoUse correct document date for the -a option.
Johan Karlsson [Thu, 22 Aug 2002 01:54:18 +0000 (01:54 +0000)]
Use correct document date for the -a option.

22 years agoAdd the -a option to report all matches instead of only the
Johan Karlsson [Thu, 22 Aug 2002 01:50:51 +0000 (01:50 +0000)]
Add the -a option to report all matches instead of only the
first of each requested type.

Approved by: joerg, sheldonh (mentor)