]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoUse the MI ithread helper functions in the x86 interrupt code.
jhb [Fri, 9 Feb 2001 17:47:44 +0000 (17:47 +0000)]
Use the MI ithread helper functions in the x86 interrupt code.

23 years ago- Catch up to the new swi API changes:
jhb [Fri, 9 Feb 2001 17:46:35 +0000 (17:46 +0000)]
- Catch up to the new swi API changes:
  - Use swi_* function names.
  - Use void * to hold cookies to handlers instead of struct intrhand *.
- In sio.c, use 'driver_name' instead of "sio" as the name of the driver
  lock to minimize diffs with cy(4).

23 years ago- Move struct ithd to sys/interrupt.h.
jhb [Fri, 9 Feb 2001 17:42:43 +0000 (17:42 +0000)]
- Move struct ithd to sys/interrupt.h.
- Add a set of MI helper functions for interrupt threads:
  - ithread_create() creates a new interrupt thread
  - ithread_destroy() destroys an interrupt thread
  - ithread_add_handler() attaches a new handler to an interrupt thread
  - ithread_remove_handler() detaches a handler from an interrupt thread
- Rename sinthand_add() and sched_swi() to swi_add() and swi_sched()
  respectively so that they live in a consistent namespace.
- struct intrhand is no longer a public type.  It would be private to
  kern_intr.c but the current implementation of fast interrupts on the
  alpha requires the type to be exported.  However, all handlers should
  be treated as void * cookies in the way that new-bus treats them.  This
  includes references to software interrupt handlers.

23 years agoAdd a new SYSINIT for interrupt thread initialization and stick
jhb [Fri, 9 Feb 2001 17:38:19 +0000 (17:38 +0000)]
Add a new SYSINIT for interrupt thread initialization and stick
initialization right after it.

23 years agoAxe INTR_HEAVY, INTR_LIGHT, and INTR_THREADED. Add in a INTR_TYPE_CLK
jhb [Fri, 9 Feb 2001 17:35:22 +0000 (17:35 +0000)]
Axe INTR_HEAVY, INTR_LIGHT, and INTR_THREADED.  Add in a INTR_TYPE_CLK
priorirty for clock interrupt handlers and an INTR_ENTROPY flag to mark
interrupt sources that are good sources of entropy.

23 years agoRemove extra space after int_currency_symbol
ache [Fri, 9 Feb 2001 17:30:34 +0000 (17:30 +0000)]
Remove extra space after int_currency_symbol

23 years agoUse 4-digit year in x_fmt
ache [Fri, 9 Feb 2001 17:18:45 +0000 (17:18 +0000)]
Use 4-digit year in x_fmt

23 years agoUse " " as thousands_sep
ache [Fri, 9 Feb 2001 16:55:39 +0000 (16:55 +0000)]
Use " " as thousands_sep

23 years agoRemove extra space after int_curr_sym
ache [Fri, 9 Feb 2001 16:52:19 +0000 (16:52 +0000)]
Remove extra space after int_curr_sym

23 years agoRemove extra space after int_curr_sym
ache [Fri, 9 Feb 2001 16:50:43 +0000 (16:50 +0000)]
Remove extra space after int_curr_sym
Add "." after currency_sym
Replace decimal_point with ","
Add " " as thousands_sep

23 years agoRelease the proc lock around crfree() and uifree() in wait1(). It leads to
jhb [Fri, 9 Feb 2001 16:43:18 +0000 (16:43 +0000)]
Release the proc lock around crfree() and uifree() in wait1().  It leads to
a lock order violation, and since p is already a zombie at this point,
I'm not sure that we even need all the locking currently in wait1().

23 years agoUnbreak EISA. The PCI-EISA bridge bus is named `eisab', not `isab'.
joerg [Fri, 9 Feb 2001 16:33:53 +0000 (16:33 +0000)]
Unbreak EISA.  The PCI-EISA bridge bus is named `eisab', not `isab'.
This mistake seems to have been benign until very recently, probably
until msmith's PCI code reshuffle which cleaned up a lot of things.

Still, my AIC7770 doesn't work again, but it at least probes the
EISA bus now.

23 years agoProc locking.
jhb [Fri, 9 Feb 2001 16:27:41 +0000 (16:27 +0000)]
Proc locking.

23 years agoMove the initailization of the proc lock for proc0 very early into the MD
jhb [Fri, 9 Feb 2001 16:25:16 +0000 (16:25 +0000)]
Move the initailization of the proc lock for proc0 very early into the MD
startup code.

23 years agoWoops, remove an obsolete reference to gd_cpu_lockid.
jhb [Fri, 9 Feb 2001 16:13:57 +0000 (16:13 +0000)]
Woops, remove an obsolete reference to gd_cpu_lockid.

23 years ago- Change the 'witness_list' ddb command to 'show mutexes'. Note that this
jhb [Fri, 9 Feb 2001 15:19:41 +0000 (15:19 +0000)]
- Change the 'witness_list' ddb command to 'show mutexes'.  Note that this
  will only display sleep mutexes held by the current process.
- Clean up some nits in the witness_display() function and add a ddb
  command 'show witness' that dumps the hierarchy and order lists to the
  console.
- Use queue(3) macros where appropriate.
- Resort the spin lock order list so that "com" is before "sched_lock".
  Also, add appropriate #ifdef's around SMP and i386-specific mutexes.
- Add two new mutexes used to protect the ithread lists and tables to the
  order list.

Requested by: bde (1)

23 years agoChange the ktr ddb commands to be show commands. The commands are now as
jhb [Fri, 9 Feb 2001 15:07:30 +0000 (15:07 +0000)]
Change the ktr ddb commands to be show commands.  The commands are now as
follows:
 - show ktr_first display the first entry
 - show ktr_next display the next entry
 - show ktr display the entire buffer

The /v modifiers continue to work as described previously.

Requested by: bde

23 years ago- Point out that we don't lock anything during the idle setup because
jhb [Fri, 9 Feb 2001 14:59:43 +0000 (14:59 +0000)]
- Point out that we don't lock anything during the idle setup because
  only the boot processor should be running in the comments.
- Initialize curproc to point to each CPU's respective idleproc if their
  curproc is NULL.
- Keep track of the number of context switches performed by idleproc.

23 years agoRemove bogus #if 0'd code that dinked with the saved interrupt state in
jhb [Fri, 9 Feb 2001 14:50:52 +0000 (14:50 +0000)]
Remove bogus #if 0'd code that dinked with the saved interrupt state in
sched_lock.

23 years agoRemove unused forward_irq counters.
jhb [Fri, 9 Feb 2001 14:30:03 +0000 (14:30 +0000)]
Remove unused forward_irq counters.

23 years agoAxe gd_cpu_lockid as it is no longer used.
jhb [Fri, 9 Feb 2001 14:25:22 +0000 (14:25 +0000)]
Axe gd_cpu_lockid as it is no longer used.

23 years agoAdd is_IS locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 14:08:52 +0000 (14:08 +0000)]
Add is_IS locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd fi_FI locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 13:37:06 +0000 (13:37 +0000)]
Add fi_FI locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoFix some problems that were introduced in revision 1.97. Instead
iedowse [Fri, 9 Feb 2001 13:24:06 +0000 (13:24 +0000)]
Fix some problems that were introduced in revision 1.97. Instead
of returning an error code to the caller, NFS server op routines
must themselves build an error reply and return 0 to the caller.

This is achieved by replacing the erroneous return statements with
code that jumps forward to the op function's reply code. We need
to be careful to ensure that the 'struct mount' pointer is NULL
though, so that the final vn_finished_write() call becomes a no-op.

Reviewed by: mckusick, dillon

23 years agoFix login so that it exports environmental variables that are set by PAM
nectar [Fri, 9 Feb 2001 13:21:50 +0000 (13:21 +0000)]
Fix login so that it exports environmental variables that are set by PAM
modules (via pam_putenv).  The following variables will never be set in
this fashion:

   SHELL, HOME, LOGNAME, MAIL, CDPATH, IFS, PATH
   any variable starting with `LD_'

23 years agomdoc(7) police: prevent generation of the extra
ru [Fri, 9 Feb 2001 12:47:23 +0000 (12:47 +0000)]
mdoc(7) police: prevent generation of the extra
whitespace after ``i.e.'', use .Rv, update .Dd.

23 years agoAdd af_ZA locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 12:39:32 +0000 (12:39 +0000)]
Add af_ZA locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd fr_FR locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 12:30:53 +0000 (12:30 +0000)]
Add fr_FR locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd de_DE locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 12:21:17 +0000 (12:21 +0000)]
Add de_DE locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd da_DK locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 12:08:12 +0000 (12:08 +0000)]
Add da_DK locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd no_NO locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 11:53:55 +0000 (11:53 +0000)]
Add no_NO locale support for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoAdd sv_SE locale files for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.
asmodai [Fri, 9 Feb 2001 11:38:57 +0000 (11:38 +0000)]
Add sv_SE locale files for LC_MESSAGES, LC_MONETARY, and LC_NUMERIC.

23 years agoRepeat after me: no commits when tired.
asmodai [Fri, 9 Feb 2001 09:09:14 +0000 (09:09 +0000)]
Repeat after me: no commits when tired.

Add the correct named file for LC_NUMERIC for nl_NL locales.

23 years agomdoc(7) police: fixed the weird construct.
ru [Fri, 9 Feb 2001 09:08:53 +0000 (09:08 +0000)]
mdoc(7) police: fixed the weird construct.

23 years agoRemove some leftovers. This is obviously unused, since the #defines
peter [Fri, 9 Feb 2001 08:26:26 +0000 (08:26 +0000)]
Remove some leftovers.  This is obviously unused, since the #defines
referred to members that no longer exist.

23 years agopoll(2) array limits (take 2) - after some input from bde.
peter [Fri, 9 Feb 2001 08:10:22 +0000 (08:10 +0000)]
poll(2) array limits (take 2) - after some input from bde.

23 years agomdoc(7) police: cosmetics.
ru [Fri, 9 Feb 2001 07:54:59 +0000 (07:54 +0000)]
mdoc(7) police: cosmetics.

23 years agoChange and clean the mutex lock interface.
bmilekic [Fri, 9 Feb 2001 06:11:45 +0000 (06:11 +0000)]
Change and clean the mutex lock interface.

mtx_enter(lock, type) becomes:

mtx_lock(lock) for sleep locks (MTX_DEF-initialized locks)
mtx_lock_spin(lock) for spin locks (MTX_SPIN-initialized)

similarily, for releasing a lock, we now have:

mtx_unlock(lock) for MTX_DEF and mtx_unlock_spin(lock) for MTX_SPIN.
We change the caller interface for the two different types of locks
because the semantics are entirely different for each case, and this
makes it explicitly clear and, at the same time, it rids us of the
extra `type' argument.

The enter->lock and exit->unlock change has been made with the idea
that we're "locking data" and not "entering locked code" in mind.

Further, remove all additional "flags" previously passed to the
lock acquire/release routines with the exception of two:

MTX_QUIET and MTX_NOSWITCH

The functionality of these flags is preserved and they can be passed
to the lock/unlock routines by calling the corresponding wrappers:

mtx_{lock, unlock}_flags(lock, flag(s)) and
mtx_{lock, unlock}_spin_flags(lock, flag(s)) for MTX_DEF and MTX_SPIN
locks, respectively.

Re-inline some lock acq/rel code; in the sleep lock case, we only
inline the _obtain_lock()s in order to ensure that the inlined code
fits into a cache line. In the spin lock case, we inline recursion and
actually only perform a function call if we need to spin. This change
has been made with the idea that we generally tend to avoid spin locks
and that also the spin locks that we do have and are heavily used
(i.e. sched_lock) do recurse, and therefore in an effort to reduce
function call overhead for some architectures (such as alpha), we
inline recursion for this case.

Create a new malloc type for the witness code and retire from using
the M_DEV type. The new type is called M_WITNESS and is only declared
if WITNESS is enabled.

Begin cleaning up some machdep/mutex.h code - specifically updated the
"optimized" inlined code in alpha/mutex.h and wrote MTX_LOCK_SPIN
and MTX_UNLOCK_SPIN asm macros for the i386/mutex.h as we presently
need those.

Finally, caught up to the interface changes in all sys code.

Contributors: jake, jhb, jasone (in no particular order)

23 years agoApply patch to add support for the intergrated ethernet in the SiS630E
wpaul [Fri, 9 Feb 2001 00:45:29 +0000 (00:45 +0000)]
Apply patch to add support for the intergrated ethernet in the SiS630E
chipset. The MAC address is stored in the APC CMOS RAM and we have to
commit trememdous evil in order to read it. The code to do this is only
activated on the i386 platform. Thanks to Cameron Grant for providing
access to a test box for me to tinker with.

This will fix the problem where the sis driver ends up with a station
address of 00:00:00:00:00:00 on boards that use the 630E chipset.

23 years agoAdd includes of sys/malloc.h so this actually compiles.
jhb [Fri, 9 Feb 2001 00:16:41 +0000 (00:16 +0000)]
Add includes of sys/malloc.h so this actually compiles.

Pointy-hat to: asmodai

23 years agoApply clue to mind and correct the ISO setting to the correct one.
asmodai [Thu, 8 Feb 2001 23:06:11 +0000 (23:06 +0000)]
Apply clue to mind and correct the ISO setting to the correct one.

This adds LC_MESSAGES, LC_MONETARY, and LC_NUMERIC suppport for nl_NL
locales.

23 years agoWhoops... forgotten a few pieces in previous patch:
luigi [Thu, 8 Feb 2001 22:54:57 +0000 (22:54 +0000)]
Whoops... forgotten a few pieces in previous patch:
 * a ">" is really ">=" ;
 * do not try to fetch zero-sized blocks from the card;
 * make sure that bpf gets the packets it wants even with
   bridging active;

23 years agoAdd LC_NUMERIC, LC_MONETARY, and LC_MESSAGES definitions for
asmodai [Thu, 8 Feb 2001 22:49:38 +0000 (22:49 +0000)]
Add LC_NUMERIC, LC_MONETARY, and LC_MESSAGES definitions for
nl_NL.ISO_8859-9.

23 years agoFix comment to reflect what it should be:
asmodai [Thu, 8 Feb 2001 22:45:44 +0000 (22:45 +0000)]
Fix comment to reflect what it should be:

yesexpr/noexpr

23 years agoMake sure that we do not accept packets shorter than a full ethernet
luigi [Thu, 8 Feb 2001 22:21:30 +0000 (22:21 +0000)]
Make sure that we do not accept packets shorter than a full ethernet
header, or we would have a negative length passed to ether_input().

23 years agoFix memory leaks with dev_get_children().
imp [Thu, 8 Feb 2001 22:14:27 +0000 (22:14 +0000)]
Fix memory leaks with dev_get_children().

Submitted by: Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>

23 years agoFix a cosmetic problem with some very defensive programming: The devfs
phk [Thu, 8 Feb 2001 22:07:08 +0000 (22:07 +0000)]
Fix a cosmetic problem with some very defensive programming:  The devfs
mount would show up as "/dev/", loose that trailing slash.

23 years agoFix a memory leak.
sos [Thu, 8 Feb 2001 21:59:27 +0000 (21:59 +0000)]
Fix a memory leak.

Pointed out by: Asmodai

23 years agoFix memoryleaks with device_get_children().
asmodai [Thu, 8 Feb 2001 21:53:05 +0000 (21:53 +0000)]
Fix memoryleaks with device_get_children().

Approved by: wpaul

23 years agoFix memory leaks with dev_get_children().
imp [Thu, 8 Feb 2001 21:47:45 +0000 (21:47 +0000)]
Fix memory leaks with dev_get_children().

Submitted by: Jeroen Ruigrok/Asmodai <asmodai@wxs.nl>

23 years agoDefine HAVE_PAM_GETENVLIST for build. Now environmental variables set
nectar [Thu, 8 Feb 2001 21:16:34 +0000 (21:16 +0000)]
Define HAVE_PAM_GETENVLIST for build.  Now environmental variables set
by PAM modules will be exported (correctly).

23 years agoHardcode c_fmt in a different way since used in nl_langinfo now
ache [Thu, 8 Feb 2001 20:55:38 +0000 (20:55 +0000)]
Hardcode c_fmt in a different way since used in nl_langinfo now

23 years agoFree the memory we get from devclass_get_devices and device_get_children.
msmith [Thu, 8 Feb 2001 20:44:49 +0000 (20:44 +0000)]
Free the memory we get from devclass_get_devices and device_get_children.

Submitted by: wpaul

23 years agoShow denied secondary bind transfer attempts
brian [Thu, 8 Feb 2001 20:31:21 +0000 (20:31 +0000)]
Show denied secondary bind transfer attempts

Submitted by: inTEXT Communications <glenn@intextonline.com>
Ok'd by: imp, kris
Not objected to by: freebsd-audit

23 years agoFix typos: initalise -> initialise.
asmodai [Thu, 8 Feb 2001 20:28:57 +0000 (20:28 +0000)]
Fix typos: initalise -> initialise.

Initalise is not an english word.

23 years agoProtect freelists managed by Balloc/Bfree with a spinlock.
tegge [Thu, 8 Feb 2001 20:22:28 +0000 (20:22 +0000)]
Protect freelists managed by Balloc/Bfree with a spinlock.
Change __dtoa to not free the string it allocated the previous time it was
called.  The caller now frees the string after usage if appropiate.

PR: 15070

23 years agoFix style(9) bugs from previous commit: remove space after function
mikeh [Thu, 8 Feb 2001 20:15:59 +0000 (20:15 +0000)]
Fix style(9) bugs from previous commit: remove space after function
names and revert to char *[].

bde by: bde, imp

23 years agoWhen we receive an incoming Ethernet frame that was unicast to a
archie [Thu, 8 Feb 2001 17:56:49 +0000 (17:56 +0000)]
When we receive an incoming Ethernet frame that was unicast to a
different hardware address, we should drop it (this should only
happen in promiscuous mode). Relocate the code for this check
from before ng_ether(4) processing to after ng_ether(4) processing.
Also fix a compiler warning.

PR: kern/24465

23 years ago- By default treat supplied arguments as a shell globs to be matched against
sobomax [Thu, 8 Feb 2001 17:44:00 +0000 (17:44 +0000)]
- By default treat supplied arguments as a shell globs to be matched against
  names of installed packages;
- add new `-G' option to disable glob matching and revert to previous
  behaviour (I have no idea why this could be necessary, though);
- add a new `-x' option, which instructs pkg_info(1) to treat supplied
  arguments as a regular expressions.

For example:
$ pkg_info foo\* - displays information about all packages whose names start
   from foo
$ pkg_info -G foo\*-1.1 - displays information about package named "foo*-1.1"
$ pkg_info -x ^foo.\* - displays information about all packages whose names
start from foo

Original idea submitted by: Edwin Groothuis <mavetju@chello.nl> (bin/24695)
Reviewed by: jkh, roam
Approved by: jkh

23 years agoactivate monetdef, msgdef and numericdef
phantom [Thu, 8 Feb 2001 17:33:55 +0000 (17:33 +0000)]
activate monetdef, msgdef and numericdef

23 years agoAdd sources for following categories: LC_NUMERIC (share/numericdef),
phantom [Thu, 8 Feb 2001 17:33:19 +0000 (17:33 +0000)]
Add sources for following categories: LC_NUMERIC (share/numericdef),
LC_MONETARY (share/monetdef), LC_MESSAGES (share/msgdef). Now only
en_US.ISO_8859-1 and ru_RU.KOI8-R locales ready. I will find some time
in near future and'll try to make defintions for other locales.

23 years agoadd langinfo.h
phantom [Thu, 8 Feb 2001 17:14:28 +0000 (17:14 +0000)]
add langinfo.h

23 years agoadd lmonetary.?, lnumeric.?, ldpart.?, lmessages.?, nl_langinfo.c
phantom [Thu, 8 Feb 2001 17:13:24 +0000 (17:13 +0000)]
add lmonetary.?, lnumeric.?, ldpart.?, lmessages.?, nl_langinfo.c
remove lconv.c

23 years agoAdd SUSv2 compatible nl_langinfo() function. It still need some work, but
phantom [Thu, 8 Feb 2001 17:12:03 +0000 (17:12 +0000)]
Add SUSv2 compatible nl_langinfo() function. It still need some work, but
this is already usable one.

23 years agoCatch up to latest chanage in timelocal structures exporting.
phantom [Thu, 8 Feb 2001 17:08:13 +0000 (17:08 +0000)]
Catch up to latest chanage in timelocal structures exporting.

23 years agoExport internal data structures in same manner as numeric/monetary/messages
phantom [Thu, 8 Feb 2001 17:06:37 +0000 (17:06 +0000)]
Export internal data structures in same manner as numeric/monetary/messages
structures exported.
Protect timelocal.h from multiple inclusions.

23 years agoMake localeconv() actual function. Now it will use LC_MONETARY/LC_NUMERIC
phantom [Thu, 8 Feb 2001 17:03:01 +0000 (17:03 +0000)]
Make localeconv() actual function. Now it will use LC_MONETARY/LC_NUMERIC
information to fill return structure. Remove unused anymore stub.

23 years agoMake FreeBSD locale support complete: add support for rest locale categories
phantom [Thu, 8 Feb 2001 16:58:53 +0000 (16:58 +0000)]
Make FreeBSD locale support complete: add support for rest locale categories
LC_MONETARY, LC_NUMERIC and LC_MESSAGES. Remove stub functions since they
don't need anymore.

Reviewed by: silence on -i18n

23 years agoOptimize gencat(1) sources in same manner as its libc part some time ago:
phantom [Thu, 8 Feb 2001 16:38:16 +0000 (16:38 +0000)]
Optimize gencat(1) sources in same manner as its libc part some time ago:

. remove SYSV compatibility bits
. sort #include's
. ifdef unused code
. cleanup BDECFLAGS warnings
. fixed few typos in diagnostics messages
. style(9) fixes

as part of this cleanup I have changed all K&R declarations to ANSI style.
Main reason for doing that was great mix of declarations used here - clean
ASNI , clean K&R, ifdef'ed (both).

Reviewed by: ache (mostly)

23 years agoAdded pc98 supports.
nyan [Thu, 8 Feb 2001 13:24:30 +0000 (13:24 +0000)]
Added pc98 supports.

23 years agoDocument the cases in which setreuid changes the saved-user-ID.
nectar [Thu, 8 Feb 2001 12:43:33 +0000 (12:43 +0000)]
Document the cases in which setreuid changes the saved-user-ID.

23 years agoFixed the .St strings so that no single-character component of
ru [Thu, 8 Feb 2001 12:00:18 +0000 (12:00 +0000)]
Fixed the .St strings so that no single-character component of
the name could be wrapped, e.g. ``ISO C'' is always printed on
the same line.

Ported from: mdocNG

23 years agoFixed the .Xr macro so that it produces the right spacing
ru [Thu, 8 Feb 2001 11:22:34 +0000 (11:22 +0000)]
Fixed the .Xr macro so that it produces the right spacing
and no hyphenation is possible on a dash (`-') character.

Ported from: mdocNG

23 years agomdoc(7) police: print #ifdef/#endif literally,
ru [Thu, 8 Feb 2001 10:40:42 +0000 (10:40 +0000)]
mdoc(7) police: print #ifdef/#endif literally,
and as list items (to get the correct offset).

23 years agomdoc(7) police: C types should be declared with either .Ft or .Vt.
ru [Thu, 8 Feb 2001 10:22:35 +0000 (10:22 +0000)]
mdoc(7) police: C types should be declared with either .Ft or .Vt.

23 years agomdoc(7) police: removed hard sentence break, corrected .Nd.
ru [Thu, 8 Feb 2001 10:06:09 +0000 (10:06 +0000)]
mdoc(7) police: removed hard sentence break, corrected .Nd.

23 years agoRemove the manpage since the actual software was put in the Attic.
obrien [Thu, 8 Feb 2001 09:30:21 +0000 (09:30 +0000)]
Remove the manpage since the actual software was put in the Attic.

Noticed by: bde

23 years agoDe-__P()-ify, convert K&R style declarations to ANSI ones
mikeh [Thu, 8 Feb 2001 05:58:55 +0000 (05:58 +0000)]
De-__P()-ify, convert K&R style declarations to ANSI ones

23 years agoPrevent string overflows by using safer string functions
mikeh [Thu, 8 Feb 2001 05:33:19 +0000 (05:33 +0000)]
Prevent string overflows by using safer string functions
Use tolower() rather than bit or-ing
Sanity check user specified printf() format
Exit when too many input files are given
Remove register from variables

Reviewed by: markm, imp

23 years agoBump __FreeBSD_cc_version to 500002 to show the -pthread option change.
obrien [Thu, 8 Feb 2001 05:27:17 +0000 (05:27 +0000)]
Bump __FreeBSD_cc_version to 500002 to show the -pthread option change.

23 years agoAdd recently added PHY drivers sources.
semenu [Thu, 8 Feb 2001 04:58:17 +0000 (04:58 +0000)]
Add recently added PHY drivers sources.

Reminded by: Peter Wemm <peter@netplex.com.au>

23 years agoDon't bother with acquiring/releasing Giant around kmem_malloc() and
jhb [Thu, 8 Feb 2001 00:27:38 +0000 (00:27 +0000)]
Don't bother with acquiring/releasing Giant around kmem_malloc() and
kmem_free() for now.  Kmem_malloc() and kmem_free() now have appropriate
assertions in place, and these checks aren't feasible until more of the
networking code is locked down.  Also, the extra assertions here should
already be caught by the WITNESS code as lock order violations should
mutex operations on Giant be reintroduced here later.

23 years agoDon't enable interrupts for a kernel breakpoint or trace trap. Otherwise,
jhb [Thu, 8 Feb 2001 00:10:07 +0000 (00:10 +0000)]
Don't enable interrupts for a kernel breakpoint or trace trap.  Otherwise,
this negates the explicit disabling of interrupts when entering the
debugger in Debugger().

23 years ago- Check if the ccb_infos in the softc is NULL before trying to free it in
jhb [Thu, 8 Feb 2001 00:00:12 +0000 (00:00 +0000)]
- Check if the ccb_infos in the softc is NULL before trying to free it in
  adv_free() as the ISA probe routine doesn't malloc() ccb_infos but does
  call adv_free().
- Release the ISA-only overrun DMA tags, bufs, and maps if the probe fails.

Tested by: rwatson

23 years agoThe code I picked up from NetBSD in '97 had a nasty bug. It limited
peter [Wed, 7 Feb 2001 23:28:01 +0000 (23:28 +0000)]
The code I picked up from NetBSD in '97 had a nasty bug.  It limited
the index of the pollfd array to the number of fd's currently open, not
the maximum number of fd's.  ie: if you had 0,1,2 open, you could not
use pollfd slots higher than 20.  The specs say we only have to support
OPEN_MAX [64] entries but we allow way more than that.

23 years agoWhen SMPng was first committed, we removed 'cpl' from the interrupt
jhb [Wed, 7 Feb 2001 22:41:47 +0000 (22:41 +0000)]
When SMPng was first committed, we removed 'cpl' from the interrupt
frame.  Teach ddb about this as there is one less word for it to skip
over when finding a trapframe on the interrupt frame stack.

23 years agoOops... I knew i shouldn't use arrows in vi. Fix a typo causing errors in
semenu [Wed, 7 Feb 2001 22:34:15 +0000 (22:34 +0000)]
Oops... I knew i shouldn't use arrows in vi. Fix a typo causing errors in
compilation.

23 years agoSynch: Properly constify sccsid[].
asmodai [Wed, 7 Feb 2001 22:25:56 +0000 (22:25 +0000)]
Synch: Properly constify sccsid[].
Replace bcopy() with memmove().

23 years agoSynch: Prefer memmove() over bcopy().
asmodai [Wed, 7 Feb 2001 22:22:57 +0000 (22:22 +0000)]
Synch: Prefer memmove() over bcopy().

23 years agoSynch: Properly constify sccsid[].
asmodai [Wed, 7 Feb 2001 22:20:44 +0000 (22:20 +0000)]
Synch: Properly constify sccsid[].

23 years agoSynch: prefer memmove() over bcopy(), since the first is a C-standard
asmodai [Wed, 7 Feb 2001 22:18:58 +0000 (22:18 +0000)]
Synch: prefer memmove() over bcopy(), since the first is a C-standard
interface, whilst the latter is a BSD'ism.

23 years agoSynch: Add $FreeBSD$.
asmodai [Wed, 7 Feb 2001 21:58:16 +0000 (21:58 +0000)]
Synch: Add $FreeBSD$.

23 years agoback out rev 1.49, looks like i committed it from the wrong tree.
cg [Wed, 7 Feb 2001 20:45:40 +0000 (20:45 +0000)]
back out rev 1.49, looks like i committed it from the wrong tree.

include the changes 1.49 was supposed to have.

Noticed by: peter

23 years agoReflect recently added support for SMC9432FTX cards.
semenu [Wed, 7 Feb 2001 20:18:54 +0000 (20:18 +0000)]
Reflect recently added support for SMC9432FTX cards.

23 years agoAdd support for SMC9432FTX card, possibly othe fiber optic SMC9432 family
semenu [Wed, 7 Feb 2001 20:11:02 +0000 (20:11 +0000)]
Add support for SMC9432FTX card, possibly othe fiber optic SMC9432 family
cards will work too.

23 years agoUpdate tx(4) man page to reflect changes since 1997, especially
semenu [Wed, 7 Feb 2001 20:09:50 +0000 (20:09 +0000)]
Update  tx(4) man page to reflect changes since 1997, especially
recent changes.

23 years agoAdd QS6612, AC101 and LXT970 personal drivers and register definitions.
semenu [Wed, 7 Feb 2001 19:57:40 +0000 (19:57 +0000)]
Add QS6612, AC101 and LXT970 personal drivers and register definitions.

Obtained from: NetBSD/OpenBSD

23 years ago1. Disable hyphenation on a dash (`-') character in macro arguments.
ru [Wed, 7 Feb 2001 17:45:22 +0000 (17:45 +0000)]
1.  Disable hyphenation on a dash (`-') character in macro arguments.
2.  Disable recognition of end-of-sentence characters in text-type
    macro arguments.
3.  Fix the missing space bug when an end-of-sentence character was
    followed by another punctuation character.
4.  Fix the bug where hyphenation was left disabled after .Re.

Obtained from: mdocNG

23 years agoFix a f^Hdamn typo, which prevented to fopen() more that 17 files at once.
sobomax [Wed, 7 Feb 2001 17:34:48 +0000 (17:34 +0000)]
Fix a f^Hdamn typo, which prevented to fopen() more that 17 files at once.

Tested by: knu, sobomax and other #bsdcode'rs

23 years agomdoc(7) police: replace two -unfilled displays with -tag lists.
ru [Wed, 7 Feb 2001 16:02:21 +0000 (16:02 +0000)]
mdoc(7) police: replace two -unfilled displays with -tag lists.
This works only because of bugs in current implementation: the
first .It after ``.Bd -unfilled'' re-enables filling mode and
does not restore (disable) it back afterwards.