]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoClarify firewire setup, including minimal error recovery.
grog [Tue, 30 Dec 2003 02:04:25 +0000 (02:04 +0000)]
Clarify firewire setup, including minimal error recovery.

20 years agoDescribe kernel debug macros for Vinum. This was originally intended
grog [Tue, 30 Dec 2003 01:40:38 +0000 (01:40 +0000)]
Describe kernel debug macros for Vinum.  This was originally intended
as part of gdb(4).

Suggested by: njl

20 years agoNew man page for kernel debugging with gdb.
grog [Tue, 30 Dec 2003 01:39:08 +0000 (01:39 +0000)]
New man page for kernel debugging with gdb.

Reviewed by: jkoshy, des, gallatin, njl.

20 years agoRemove current implementation of kldstat macro.
grog [Tue, 30 Dec 2003 01:23:16 +0000 (01:23 +0000)]
Remove current implementation of kldstat macro.

Add new kldstat, kldstat-v and kernel macros.  The kldstat macro is
functionally equivalent to the previous implementation, but it looks
prettier and it matches the kldstat-v macro better.  kldstat-v gives
output similar to userland kldstat -v (note lacking space), and kernel
loads a new kernel and dump.

Submitted by: des

20 years ago- Bump up the general and status RID sizes
ambrisko [Tue, 30 Dec 2003 01:07:12 +0000 (01:07 +0000)]
- Bump up the general and status RID sizes
- Clear out an_dma_vaddr on free so we can test to see if dma is
  setup when the card is kldunloaded/kldloaded etc. only for MPI350
- Use a common detach like wi(4)
- Notify on RID read overflow and truncate this currently causes
  a panic in -stable when the stack during an ifconfig an0 is done
  with newer firmware
- Convert from UNLOCK/tsleep/LOCK to msleep.  I thought I did that
  a while ago.

20 years agoRemember to use stringcvt() to expand %foo% strings in registry values in
wpaul [Tue, 30 Dec 2003 00:30:45 +0000 (00:30 +0000)]
Remember to use stringcvt() to expand %foo% strings in registry values in
dump_defaultinfo().

Noticed by: green@freebsd.org

20 years agoRework resource allocation. Replace the "feel around like a blind man"
wpaul [Mon, 29 Dec 2003 23:51:59 +0000 (23:51 +0000)]
Rework resource allocation. Replace the "feel around like a blind man"
method with something a little more intelligent: use BUS_GET_RESOURCE_LIST()
to run through all resources allocated to us and map them as needed. This
way we know exactly what resources need to be mapped and what their RIDs
are without having to guess. This simplifies both ndis_attach() and
ndis_convert_res(), and eliminates the unfriendly "ndisX: couldn't map
<foo>" messages that are sometimes emitted during driver load.

20 years agoForgot to commit this file for last commit. :(
davidxu [Mon, 29 Dec 2003 23:33:51 +0000 (23:33 +0000)]
Forgot to commit this file for last commit. :(

20 years agoImplement sigaltstack() as per-threaded. Current only scope process thread
davidxu [Mon, 29 Dec 2003 23:21:09 +0000 (23:21 +0000)]
Implement sigaltstack() as per-threaded. Current only scope process thread
is supported, for scope system process, kernel signal bits need to be
changed.

Reviewed by: deischen
Tested on  : i386 amd64 ia64

20 years agoMerge the following changes from English version:
motoyuki [Mon, 29 Dec 2003 15:40:41 +0000 (15:40 +0000)]
Merge the following changes from English version:

early-adopter/article.sgml 1.16 -> 1.17

20 years agoDescribe kern.ipc.nsfbufsused and kern.ipc.nsfbufspeak.
maxim [Mon, 29 Dec 2003 12:29:37 +0000 (12:29 +0000)]
Describe kern.ipc.nsfbufsused and kern.ipc.nsfbufspeak.

Reviewed by: silby

20 years agoFixed some (most) style bugs in rev.1.33. Mainly 4-char indentation
bde [Mon, 29 Dec 2003 11:59:05 +0000 (11:59 +0000)]
Fixed some (most) style bugs in rev.1.33.  Mainly 4-char indentation
(msdosfs uses normal 8-char indentation almost everywhere else),
too-long lines, and minor English usage errors.  The verbose formal
comment before the new function is still abnormal.

20 years agoGarbage-collected some vestiges of objformat support (mainly ${FMT}).
bde [Mon, 29 Dec 2003 11:34:33 +0000 (11:34 +0000)]
Garbage-collected some vestiges of objformat support (mainly ${FMT}).

20 years agoFixed some style bugs (manly a few English usage errors in comments and
bde [Mon, 29 Dec 2003 11:28:21 +0000 (11:28 +0000)]
Fixed some style bugs (manly a few English usage errors in comments and
many unusual indentations for continued lines in code).

20 years agoFixed some minor style bugs in rev.1.144. All related to msdosfs_advlock()
bde [Mon, 29 Dec 2003 10:12:02 +0000 (10:12 +0000)]
Fixed some minor style bugs in rev.1.144.  All related to msdosfs_advlock()
(mainly unsorting).  There were no changes related to the dirty flag
here.  The reference NetBSD implementation put msdosfs_advlock() in a
different place.  This commit only moves its declarations and changes
some of the function body to be like the NetBSD version.

20 years agoFixed style bugs in rev.1.112. The bugs started with obscure magic
bde [Mon, 29 Dec 2003 09:50:42 +0000 (09:50 +0000)]
Fixed style bugs in rev.1.112.  The bugs started with obscure magic
numbers in comments (Apple PR numbers?) and didn't improve.

20 years agoUse double quotes instead of single quotes for echos to build options
bde [Mon, 29 Dec 2003 09:15:48 +0000 (09:15 +0000)]
Use double quotes instead of single quotes for echos to build options
headers as in most other modules Makefiles (5 yesterday, none now).
Fixed any simple nearby style bugs (not many).

20 years agoFixed some style bugs (mainly the bad examples @echo, NOMAN, and non-use
bde [Mon, 29 Dec 2003 09:03:38 +0000 (09:03 +0000)]
Fixed some style bugs (mainly the bad examples @echo, NOMAN, and non-use
of ${.TARGET}).  This was the last instances of @echo in module Makefiles
after it was removed in sio/Makefile.  NOMAN is nonsense in module
Makefiles, and was removed, but came back here and in too many other
places.

20 years agoFixed some style bugs:
bde [Mon, 29 Dec 2003 08:45:38 +0000 (08:45 +0000)]
Fixed some style bugs:
- SRCS was totally disordered.
- the echos to create options headers were hidden using '@'.
- the args of echos to create options headers were single-quoted.
- the target names for the rules to create options headers were repeated.
- the unused option COMPAT_FREEBSD4 was put in opt_compat.h.

20 years agoFixed some style bugs. SRCS and the opt_*.h rules were totally
bde [Mon, 29 Dec 2003 08:35:33 +0000 (08:35 +0000)]
Fixed some style bugs.  SRCS and the opt_*.h rules were totally
disordered.  This commit only fixes the external disorder by
rearranging whole lines.

20 years agoUse floating point instead of unsigned long longs in percentage
bde [Mon, 29 Dec 2003 08:25:32 +0000 (08:25 +0000)]
Use floating point instead of unsigned long longs in percentage
calculations.  Long longs should never be used, since they break compiling
with C90 compilers and don't necessarily work any better than longs for
avoiding overflow.

Print percentages with another digit of precision since they can be small
and this is easy to do now that the format is floating point.

Restored some more of the old -m output:
Print the percentage of allocated memory that is in use.  This is the
amount of memory in active mbufs and mbuf clusters relative to the
total amount of memory soft-allocated for mbufs and mbuf clusters.

Print the percentage of allocated memory that is wired (cached).  The
old mbuf allocator never freed memory so printing this value wasn't
useful.  A previous version of netstat for the new allocator printed
the in-use amount as a percentage of the wired amount.

Fixed some nearby style bugs (excessive parenthesization and a redundant
return).

Reviewed by: alfred

20 years agoUnbreak build of bpf-free kernels.
tjr [Mon, 29 Dec 2003 08:23:11 +0000 (08:23 +0000)]
Unbreak build of bpf-free kernels.

20 years agoSorted includes. Removed duplicates exposed by this.
bde [Mon, 29 Dec 2003 06:51:10 +0000 (06:51 +0000)]
Sorted includes.  Removed duplicates exposed by this.

20 years agoSorted includes.
bde [Mon, 29 Dec 2003 06:47:38 +0000 (06:47 +0000)]
Sorted includes.

20 years agogdbinit target: Remove comment lines from dot.gdbinit. This allows us to
grog [Mon, 29 Dec 2003 06:40:00 +0000 (06:40 +0000)]
gdbinit target: Remove comment lines from dot.gdbinit.  This allows us to
put dire warnings in the original and not find them in the installed version.

20 years agoAdd documentation for previously undocumented macros, and correct
grog [Mon, 29 Dec 2003 06:38:27 +0000 (06:38 +0000)]
Add documentation for previously undocumented macros, and correct
existing documentation.  This is not made any easier by the
brain-damaged format required for the documentation: one line, no full
stops or commas.

20 years agoAdd documentation for previously undocumented macros. This is not
grog [Mon, 29 Dec 2003 06:37:11 +0000 (06:37 +0000)]
Add documentation for previously undocumented macros.  This is not
made any easier by the brain-damaged format required for the
documentation: one line, no full stops or commas.

Remove macro xy.  I can't see any use for it any more.

20 years agoAdd comments to deter people from using this file directly. That's
grog [Mon, 29 Dec 2003 06:35:14 +0000 (06:35 +0000)]
Add comments to deter people from using this file directly.  That's
not the intention.

Rearrange "you may want to change these values" values to the top, and
reduce their number as much as possible.

tr macro: Require a parameter (because gdb is too stupid to understand
optional parameters), and create macros tr0, tr1 and trf which call it
to connect to /dev/cuaa0, /dev/cuaa1 and firewire connections
respectively.

Split kld symbol load into two separate macros: revision 1.5
simplified things for the /dev/mem case, but broke it for anything
else.  Now the simple /dev/mem version is called kldsyms, and the
version for serial debugging and processor dumps is called getsyms,
and still requires this irritating cut and paste.

Change comments on startup to make life easier for the poor
(de)bugger.

20 years agoFixed missing declaration of pluralies(). This showed up as strange
bde [Mon, 29 Dec 2003 04:41:38 +0000 (04:41 +0000)]
Fixed missing declaration of pluralies().  This showed up as strange
printf format warnings for inet6.c (pluralies() was implicit int, but
the context requires a "char *").

Added WARNS?=2 to the Makefile so that such errors don't come back.
Added NO_WERROR?= to the Makefile because I haven't checked that setting
WARNS doesn't uncover more bugs except on i386's.

20 years agoFixed style bugs created in rev.1.27 by removing "__P(" and its closing ")"
bde [Mon, 29 Dec 2003 04:34:36 +0000 (04:34 +0000)]
Fixed style bugs created in rev.1.27 by removing "__P(" and its closing ")"
without removing the space before it.

20 years agoRemove swap_pager_un_object_list; it is unused.
alc [Mon, 29 Dec 2003 04:21:44 +0000 (04:21 +0000)]
Remove swap_pager_un_object_list; it is unused.

20 years agoExplicitly state that mknod(8) can be used to undelete entries under devfs.
alfred [Mon, 29 Dec 2003 00:40:44 +0000 (00:40 +0000)]
Explicitly state that mknod(8) can be used to undelete entries under devfs.

20 years agoDocument that mknod(8) can be used to undelete entries under devfs.
alfred [Mon, 29 Dec 2003 00:37:52 +0000 (00:37 +0000)]
Document that mknod(8) can be used to undelete entries under devfs.

20 years agoDon't bother clearing PG_ZERO on the page table page in _pmap_allocpte();
alc [Sun, 28 Dec 2003 23:57:09 +0000 (23:57 +0000)]
Don't bother clearing PG_ZERO on the page table page in _pmap_allocpte();
it serves no purpose.

20 years agoHook up acpi_thermal.4 man page.
takawata [Sun, 28 Dec 2003 23:23:16 +0000 (23:23 +0000)]
Hook up acpi_thermal.4 man page.

20 years agomdoc(7): fixed document title.
ru [Sun, 28 Dec 2003 22:34:47 +0000 (22:34 +0000)]
mdoc(7): fixed document title.

20 years agoDon't attach throttling if the P_BLK is 0, even if the P_BLK_LEN is 6.
njl [Sun, 28 Dec 2003 22:15:24 +0000 (22:15 +0000)]
Don't attach throttling if the P_BLK is 0, even if the P_BLK_LEN is 6.
This is more strict but no known systems have this problem.

20 years agoImplement NdisInitUnicodeString().
wpaul [Sun, 28 Dec 2003 21:36:03 +0000 (21:36 +0000)]
Implement NdisInitUnicodeString().

20 years agoRemove the sanity test in ndis_adjust_buflen(). I'm not sure what the
wpaul [Sun, 28 Dec 2003 21:28:36 +0000 (21:28 +0000)]
Remove the sanity test in ndis_adjust_buflen(). I'm not sure what the
nb_size field in an ndis_buffer is meant to represent, but it does not
represent the original allocation size, so the sanity check doesn't
make any sense now that we're using the Windows-mandated initialization
method.

Among other things, this makes the following card work with the
NDISulator:

ndis0: <NETGEAR PA301 Phoneline10X PCI Adapter> mem 0xda004000-0xda004fff irq 10 at device 9.0 on pci0

This is that notoriously undocumented 10Mbps HomePNA Broadcom chipset
that people wanted support for many moons ago. Sadly, the only other
HomePNA NIC I have handy is a 1Mbps device, so I can't actually do
any 10Mbps performance tests, but it talks to my 1Mbps ADMtek card
just fine.

20 years agoPlug a leak of open files that happens when you exec a suid program
dwmalone [Sun, 28 Dec 2003 19:27:14 +0000 (19:27 +0000)]
Plug a leak of open files that happens when you exec a suid program
with one of std{in,out,err} open. This helps with the file descriptor
leaks reported on -current. This should probably be merged into 5.2.

Reviewed by: ru
Tested by: Bjoern A. Zeeb <bzeeb-lists@lists.zabbadoz.net>

20 years agoAdd my birthdate (gifts are welcome :-)).
ale [Sun, 28 Dec 2003 18:13:57 +0000 (18:13 +0000)]
Add my birthdate (gifts are welcome :-)).

Approved by: blackend (mentor)

20 years agoUse intmax_t to avoid wrap on large images in verbose output.
alfred [Sun, 28 Dec 2003 13:32:49 +0000 (13:32 +0000)]
Use intmax_t to avoid wrap on large images in verbose output.

20 years agoSomebody called mdoc(7) janitor.
ru [Sun, 28 Dec 2003 12:56:01 +0000 (12:56 +0000)]
Somebody called mdoc(7) janitor.

A bit of wordsmithing from a native speaker is still necessary.

20 years agoFix to support pc98 and currect typo.
nyan [Sun, 28 Dec 2003 12:27:34 +0000 (12:27 +0000)]
Fix to support pc98 and currect typo.

20 years agoAdd detach method.
nyan [Sun, 28 Dec 2003 12:25:49 +0000 (12:25 +0000)]
Add detach method.

20 years agoCorrectly retrieve sigaction flags.
davidxu [Sun, 28 Dec 2003 12:20:04 +0000 (12:20 +0000)]
Correctly retrieve sigaction flags.

20 years agov_vxproc was a bogus name for a thread (pointer).
bde [Sun, 28 Dec 2003 09:12:56 +0000 (09:12 +0000)]
v_vxproc was a bogus name for a thread (pointer).

20 years agoMove the declarations for nsfbufs and friends over to sf_buf.h from
silby [Sun, 28 Dec 2003 09:10:51 +0000 (09:10 +0000)]
Move the declarations for nsfbufs and friends over to sf_buf.h from
mbuf.h

20 years agoTeach netstat about the new sendfile statistics.
silby [Sun, 28 Dec 2003 08:59:41 +0000 (08:59 +0000)]
Teach netstat about the new sendfile statistics.

20 years agoTrack three new sendfile-related statistics:
silby [Sun, 28 Dec 2003 08:57:09 +0000 (08:57 +0000)]
Track three new sendfile-related statistics:
- The number of times sendfile had to do disk I/O
- The number of times sfbuf allocation failed
- The number of times sfbuf allocation had to wait

20 years agoupdate radiotap support to reflect recent changes:
sam [Sun, 28 Dec 2003 07:00:32 +0000 (07:00 +0000)]
update radiotap support to reflect recent changes:

o move tx taps from ath_start to ath_tx_start so lots more
  state is available to tap
o add tx flags
o add tx rate
o add tx power (constant for the moment)
o add tx antenna state

20 years agoupdate radiotap support to reflect recent changes:
sam [Sun, 28 Dec 2003 06:58:52 +0000 (06:58 +0000)]
update radiotap support to reflect recent changes:

o add xmit rate
o drop rx time
o add rx flags

20 years agoSync with netbsd:
sam [Sun, 28 Dec 2003 06:57:28 +0000 (06:57 +0000)]
Sync with netbsd:

o #ifdef _KERNEL the fallback definition for DLT_IEEE802_11_RADIO
o fix many comments
o rename antenna stuff and fix units/reference signal
o change IEEE80211_RADIOTAP_DBM_TX_POWER from unsigned 16-bit value
  to a signed 8-bit value
o change IEEE80211_RADIOTAP_FLAGS from 16 bits to 8 bits to simplify
  padding requirements
o drop IEEE80211_RADIOTAP_TIME
o change IEEE80211_RADIOTAP_ANTENNA from 16 bits to 8 bits
o drop IEEE80211_RADIOTAP_PAD
o add channel flag definitions for outside the kernel so radiotap
  doesn't depend on stuff in ieee80211*.h

Obtained from: NetBSD

20 years agoDon't bother clearing and setting PG_BUSY on page table directory pages.
alc [Sun, 28 Dec 2003 05:10:21 +0000 (05:10 +0000)]
Don't bother clearing and setting PG_BUSY on page table directory pages.

20 years agoFixed some style bugs (mainly, try to always use explicit comparisons with
bde [Sun, 28 Dec 2003 04:37:59 +0000 (04:37 +0000)]
Fixed some style bugs (mainly, try to always use explicit comparisons with
NULL when checking for null pointers).

20 years agoFixed some disordering in revs.1.194 and 1,196. Moved the exceve() syscall
bde [Sun, 28 Dec 2003 04:18:13 +0000 (04:18 +0000)]
Fixed some disordering in revs.1.194 and 1,196.  Moved the exceve() syscall
function back to near the beginning of the file.  Rev.1.194 moved it into
the middle of auxiliary functions following kern_execve().  Moved the
__mac_execve() syscall function up together with execve().  It was new in
rev1.1.196 and perfectly misplaced after execve().

20 years agoo eliminate widespread on-stack mbuf use for bpf by introducing
sam [Sun, 28 Dec 2003 03:56:00 +0000 (03:56 +0000)]
o eliminate widespread on-stack mbuf use for bpf by introducing
  a new bpf_mtap2 routine that does the right thing for an mbuf
  and a variable-length chunk of data that should be prepended.
o while we're sweeping the drivers, use u_int32_t uniformly when
  when prepending the address family (several places were assuming
  sizeof(int) was 4)
o return M_ASSERTVALID to BPF_MTAP* now that all stack-allocated
  mbufs have been eliminated; this may better be moved to the bpf
  routines

Reviewed by: arch@ and several others

20 years agoPR 28852
cracauer [Sun, 28 Dec 2003 02:05:12 +0000 (02:05 +0000)]
PR 28852

sh -e behaviour was incorrect when && and || statements where used in
"if" clauses.

This is the patch submitted by MORI Kouji <mori@tri.asanuma.co.jp>.

It fixes the issue at hand, but sh fixes like this are super-hard to
verify that they don't break anything else.  I ran some of my old test
cases and a few big GNU configure scripts that detected mistakes
before, with the previous sh, patched sh and bash.  No differences in
behaviour found.  MFC recommended after longer than usual time.

Compiles on i386 and sledge.

20 years agoMove the declaration of sfbufspeak and sfbufsused to mbuf.h,
silby [Sun, 28 Dec 2003 01:43:22 +0000 (01:43 +0000)]
Move the declaration of sfbufspeak and sfbufsused to mbuf.h,
and use imax instead of max, as sfbufspeak and sfbufsused
are signed.

Submitted by:   bde

20 years agoUse of the MTX_SYSINIT macro requires the inclusion of <sys/kernel.h>.
jkoshy [Sun, 28 Dec 2003 01:39:18 +0000 (01:39 +0000)]
Use of the MTX_SYSINIT macro requires the inclusion of <sys/kernel.h>.

20 years agoMake this manual page reflect the sources better.
jkoshy [Sun, 28 Dec 2003 01:37:48 +0000 (01:37 +0000)]
Make this manual page reflect the sources better.

- A #include of <sys/mutex.h> is no longer needed to use sx(9) (since
  2001/05/01).
- Use of the SX_SYSINIT() macro requires inclusion of '<sys/kernel.h>'

20 years agoOops, highly non-KNF indentation is normal for large expressions in
bde [Sun, 28 Dec 2003 01:30:31 +0000 (01:30 +0000)]
Oops, highly non-KNF indentation is normal for large expressions in
this program.  Gnu indentation is used for these.  Redo the fix for
the large expression at the end of the previous commit to give gnu
indentation.  The original version was gnuish but had 9 bogus extra
characters of indentation in its continuation lines, perfect tab
lossage on every line, and other bugs.

The previous commit log should have claimed to fix style bugs in the
previous-1 commit (1.5), not the forced null previous commit (1.6).

20 years agoFix signal behaviour.
cracauer [Sun, 28 Dec 2003 01:20:03 +0000 (01:20 +0000)]
Fix signal behaviour.

In my last change I made sure that the signal as reported from a truss
exit is the same as if truss wasn't between parent and trussed
program.  I was smart enough to not have it coredump on SIGQUIT but it
didn't ocur to me SIGSEGV might cause a coredump, too :-)

So get rid of SIGQUIT extra hack and limit coredumpsize to zero
instead.

Tested: still works, correct signal reported.  No more codedumps from
SIGSEGV in the trussed proces.  This file compiles cleanly on AMD64
(sledge).

PR:
Submitted by:
Reviewed by:
Approved by:
Obtained from:
MFC after:

20 years agoFix the maxpipekva warning message so that it points to the correct
silby [Sun, 28 Dec 2003 01:19:58 +0000 (01:19 +0000)]
Fix the maxpipekva warning message so that it points to the correct
sysctl, and shorten the message.

Noticed by: bde

20 years agoRemove GIANT_REQUIRED from kmem_suballoc().
alc [Sun, 28 Dec 2003 00:10:48 +0000 (00:10 +0000)]
Remove GIANT_REQUIRED from kmem_suballoc().

20 years agounlinke -> unlike
brueffer [Sat, 27 Dec 2003 21:14:32 +0000 (21:14 +0000)]
unlinke -> unlike

Submitted by: Pawel Jakub Dawidek <nick@garage.freebsd.pl>

20 years agoCorrect reference to mac_none.4
brueffer [Sat, 27 Dec 2003 20:48:22 +0000 (20:48 +0000)]
Correct reference to mac_none.4

Submitted by: Pawel Jakub Dawidek <nick@garage.freebsd.pl>

20 years agoNew sio module. A number of people have suggested this over the years
imp [Sat, 27 Dec 2003 19:50:37 +0000 (19:50 +0000)]
New sio module.  A number of people have suggested this over the years
(most recently bde), so I'll commit the module I've had knocking
around in my tree for a while.  This may have some rough edges, so if
you are able to build it on non-i386 platform (including pc98) please
let me know you succeeded.  When I get enough reports, I'll connect it
to the build.  If there are problems, feel free to fix them.

Suggested by: bde

20 years agoAdd detach methods so we can unload the sio module.
imp [Sat, 27 Dec 2003 19:47:10 +0000 (19:47 +0000)]
Add detach methods so we can unload the sio module.

20 years agoRemove GIANT_REQUIRED from exec_unmap_first_page().
alc [Sat, 27 Dec 2003 19:40:03 +0000 (19:40 +0000)]
Remove GIANT_REQUIRED from exec_unmap_first_page().

20 years agoFix slight disordering of supported cards.
imp [Sat, 27 Dec 2003 18:07:50 +0000 (18:07 +0000)]
Fix slight disordering of supported cards.

20 years agoAdd acpi_thermal manual page,
takawata [Sat, 27 Dec 2003 16:25:11 +0000 (16:25 +0000)]
Add acpi_thermal manual page,
because I asked about the unit of thermal
zone-related value too freqently.
Come on, mdoc police.

20 years agoStop "fortune -l" from looping forever.
ceri [Sat, 27 Dec 2003 16:04:41 +0000 (16:04 +0000)]
Stop "fortune -l" from looping forever.

PR: bin/58730
Submitted by: pak <pak@cns.utoronto.ca>

20 years agoo use .Nm
brueffer [Sat, 27 Dec 2003 16:00:42 +0000 (16:00 +0000)]
o use .Nm
o fix language usage

MFC after: 5 days

20 years agoAdd missing word
brueffer [Sat, 27 Dec 2003 15:46:30 +0000 (15:46 +0000)]
Add missing word

20 years agoo add missing sysctl to synopsis
brueffer [Sat, 27 Dec 2003 15:45:08 +0000 (15:45 +0000)]
o add missing sysctl to synopsis
o remove some useless words
o fix spelling

MFC after: 5 days

20 years agoo add reference to safe.4
brueffer [Sat, 27 Dec 2003 15:40:47 +0000 (15:40 +0000)]
o add reference to safe.4
o remove useless word

MFC after: 5 days

20 years agoo Grammar fix
brueffer [Sat, 27 Dec 2003 15:38:08 +0000 (15:38 +0000)]
o Grammar fix
o Add missing punctuation

20 years agoGrammar fixes
brueffer [Sat, 27 Dec 2003 15:35:09 +0000 (15:35 +0000)]
Grammar fixes

20 years agoo grammar fixes
brueffer [Sat, 27 Dec 2003 15:30:49 +0000 (15:30 +0000)]
o grammar fixes
o FireWire.4 -> firewire.4
o wordsmithing

MFC after: 5 days

20 years agoo grammar fixes
brueffer [Sat, 27 Dec 2003 15:14:58 +0000 (15:14 +0000)]
o grammar fixes
o wordsmithing
o sort references

MFC after: 5 days

20 years agoAdd reference to safe.4
brueffer [Sat, 27 Dec 2003 15:00:07 +0000 (15:00 +0000)]
Add reference to safe.4

MFC after: 5 days

20 years agoCorrect compilation with "#define really_paranoid".
ceri [Sat, 27 Dec 2003 14:58:00 +0000 (14:58 +0000)]
Correct compilation with "#define really_paranoid".

PR: bin/59674
Submitted by: Vaclav Rehak <vasek@zoom-int.cz>

20 years agoo Fix a comment: softticks lives in sys/kern/kern_timeout.c.
maxim [Sat, 27 Dec 2003 14:08:53 +0000 (14:08 +0000)]
o Fix a comment: softticks lives in sys/kern/kern_timeout.c.

PR: kern/60613
Submitted by: Gleb Smirnoff
MFC after: 3 days

20 years agoExpanded the comment about the -F flag.
bde [Sat, 27 Dec 2003 14:02:52 +0000 (14:02 +0000)]
Expanded the comment about the -F flag.

Fixed a nearby style bug (unreachable break).

20 years agoImprove continuity of text.
jkoshy [Sat, 27 Dec 2003 14:01:08 +0000 (14:01 +0000)]
Improve continuity of text.

20 years agoUse __printflike() and __dead2 instead of hard-coded gccisms.
bde [Sat, 27 Dec 2003 13:54:02 +0000 (13:54 +0000)]
Use __printflike() and __dead2 instead of hard-coded gccisms.

Declare perror().  We define and use a home made version of perror(3)
that can't simply be removed (although it has the same interface as
perror(3)) since it is very different (it prints on stdout, doesn't
always print the program name, and sometimes exits).  Declare it to
get a reminder of this brokenness when WARNS is increased enough.

20 years agoGarbage-collected hotroot, rawname() and unrawname() again. These
bde [Sat, 27 Dec 2003 13:29:49 +0000 (13:29 +0000)]
Garbage-collected hotroot, rawname() and unrawname() again.  These
became garbage when block devices were axed and were removed a few
months later, but they came back (with hotroot renamed to hot + hotroot())
when the NetBSD fsck was mismerged.

20 years agofsck_msdosfs/main.c:
bde [Sat, 27 Dec 2003 13:08:55 +0000 (13:08 +0000)]
fsck_msdosfs/main.c:
- Don't use errexit() to (mis)implement usage().  Using errexit() just
  gave the bogus exit code 8.
- Fixed 3 other style bugs in usage().

fsck/fsutil.[ch]:
- Garbage-collected errexit().  It is essentially just one of NetBSD's
  fsck_ext2fs error printing functions, but we don't have fsck_ext2fs
  and the function is unsuitable for use there too (since pfatal() is
  also used and it printf to a different stream).

20 years agoFixed style bugs in rev.1.12 (unsorting of prototypes by adding new ones
bde [Sat, 27 Dec 2003 10:30:43 +0000 (10:30 +0000)]
Fixed style bugs in rev.1.12 (unsorting of prototypes by adding new ones
at the end).

Fixed some older style bugs (missing parameter names in all the kernel
prototypes except the ones added in rev.1.12).

20 years agoClean up the style of the previous commit, and fix a few
silby [Sat, 27 Dec 2003 09:02:31 +0000 (09:02 +0000)]
Clean up the style of the previous commit, and fix a few
type mismatches as well.

Suggested by: bde

20 years agoTeach netstat to read and display the new sfbuf statistics.
silby [Sat, 27 Dec 2003 07:58:31 +0000 (07:58 +0000)]
Teach netstat to read and display the new sfbuf statistics.

20 years agoTrack current and peak sfbuf usage, export the values via sysctl.
silby [Sat, 27 Dec 2003 07:52:47 +0000 (07:52 +0000)]
Track current and peak sfbuf usage, export the values via sysctl.

20 years agoFixed style bugs in previous commit (unsorting of declarations and poor
bde [Sat, 27 Dec 2003 06:44:32 +0000 (06:44 +0000)]
Fixed style bugs in previous commit (unsorting of declarations and poor
wording in a comment).

20 years agoFixed quoting of `clean'.
bde [Sat, 27 Dec 2003 06:30:14 +0000 (06:30 +0000)]
Fixed quoting of `clean'.

Obtained from: fsck_ffs.8

20 years agoFixed some style bugs in previous commit (mainly highly non-KNF indentation).
bde [Sat, 27 Dec 2003 06:21:44 +0000 (06:21 +0000)]
Fixed some style bugs in previous commit (mainly highly non-KNF indentation).

20 years agoFixed style bugs in previous commit (unsorting of the global declarations
bde [Sat, 27 Dec 2003 05:57:20 +0000 (05:57 +0000)]
Fixed style bugs in previous commit (unsorting of the global declarations
and addition of a tab to a blank line).

20 years ago - Reduce Giant's scope in vm_fault().
alc [Fri, 26 Dec 2003 23:33:37 +0000 (23:33 +0000)]
 - Reduce Giant's scope in vm_fault().
 - Use vm_object_reference_locked() instead of vm_object_reference()
   in vm_fault().

20 years agoMinor correction to revision 1.258: Use the proc pointer that is passed to
alc [Fri, 26 Dec 2003 21:54:45 +0000 (21:54 +0000)]
Minor correction to revision 1.258: Use the proc pointer that is passed to
vm_map_growstack() in the RLIMIT_VMEM check rather than curthread.

20 years agoFix all cases of unintended word doublings (the the, and and, ...) and a
schweikh [Fri, 26 Dec 2003 21:32:47 +0000 (21:32 +0000)]
Fix all cases of unintended word doublings (the the, and and, ...) and a
few typos.