]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
20 years agoAdd missing () to function invocation.
Alexander Kabaev [Wed, 28 Jul 2004 05:37:18 +0000 (05:37 +0000)]
Add missing () to function invocation.

20 years agoDo not use casts as lvalues.
Alexander Kabaev [Wed, 28 Jul 2004 05:33:58 +0000 (05:33 +0000)]
Do not use casts as lvalues.

20 years agoBmake glue for GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 05:27:21 +0000 (05:27 +0000)]
Bmake glue for GCC 3.4.2-prerelease.

20 years agoWe generate this file from config.h automatically.
Alexander Kabaev [Wed, 28 Jul 2004 05:23:28 +0000 (05:23 +0000)]
We generate this file from config.h automatically.

20 years agoMerge rev. 1.2 (allow profiling of FORTRAN77 programs) info
Alexander Kabaev [Wed, 28 Jul 2004 05:08:31 +0000 (05:08 +0000)]
Merge rev. 1.2 (allow profiling of FORTRAN77 programs) info
GCC 3.4.3-prerelease sources.

20 years agoMerge FreeBSD changes into GCC 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 05:06:59 +0000 (05:06 +0000)]
Merge FreeBSD changes into GCC 3.4.2-prerelease:
  1.2   Printf format errors fixes.

20 years agoUse stock version of this file now.
Alexander Kabaev [Wed, 28 Jul 2004 05:03:19 +0000 (05:03 +0000)]
Use stock version of this file now.

20 years agoUpdate sparc64 MD bits to GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 05:00:13 +0000 (05:00 +0000)]
Update sparc64 MD bits to GCC 3.4.2-prerelease.

20 years agoMerge FreeBSD changes into GCC 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 04:56:54 +0000 (04:56 +0000)]
Merge FreeBSD changes into GCC 3.4.2-prerelease:
  1.2   Tweaks to make this file live beside our hacked config.

20 years agoMerge r1.10 (-mno-align-long-strings) and r1.11 (make LONG_TYPE_SIZE
Alexander Kabaev [Wed, 28 Jul 2004 04:49:58 +0000 (04:49 +0000)]
Merge r1.10 (-mno-align-long-strings) and r1.11 (make LONG_TYPE_SIZE
overridable from the command line).

20 years agoMerge r1.11 (-mno-align-long-strings).
Alexander Kabaev [Wed, 28 Jul 2004 04:47:35 +0000 (04:47 +0000)]
Merge r1.11 (-mno-align-long-strings).

20 years agoUpdate Amd64 MD bits for GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 04:44:23 +0000 (04:44 +0000)]
Update Amd64 MD bits for GCC 3.4.2-prerelease.

20 years agoUpdate i386 MD bits for GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 04:41:42 +0000 (04:41 +0000)]
Update i386 MD bits for GCC 3.4.2-prerelease.

20 years agoUpdate Alpha config MD bits for GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 04:39:15 +0000 (04:39 +0000)]
Update Alpha config MD bits for GCC 3.4.2-prerelease.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 04:35:51 +0000 (04:35 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prerelease:
  1.8 Change symbol values to match freebsd.h.

20 years agoUpdate MI config bits for GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 04:34:05 +0000 (04:34 +0000)]
Update MI config bits for GCC 3.4.2-prerelease.

20 years agoUpdate for GCC 3.4.2. Bump __FreeBSD_cc_version_ and use correct ELF
Alexander Kabaev [Wed, 28 Jul 2004 04:31:47 +0000 (04:31 +0000)]
Update for GCC 3.4.2. Bump __FreeBSD_cc_version_ and use correct ELF
interpreter on FreeBSD 5.x series.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 04:27:06 +0000 (04:27 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prerelease:
  1.2   Localize the version string.

20 years ago-O0 -O1 should optimize alignment for time, not size. Moved from toplev.c.
Alexander Kabaev [Wed, 28 Jul 2004 04:23:38 +0000 (04:23 +0000)]
-O0 -O1 should optimize alignment for time, not size. Moved from toplev.c.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prelelease:
Alexander Kabaev [Wed, 28 Jul 2004 04:20:36 +0000 (04:20 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prelelease:
  1.2  -fformat-extensions.

Remove r1.7 (FORCE_OPTIMIZATION_DOWNGRADE knob for Alpha) as obsolete.
Move r1.14 (-O0 -O1 optimize alignment for time, not size) to opts.c.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 04:05:09 +0000 (04:05 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prerelease:
  1.{2,15}      FREEBSD_NATIVE.
  1.21          "GCC_OPTIONS".
  1.{16,25,27}  Cross-compiling issues.
  1.38 Correct priority for GCC_EXEC_PREFIX.

20 years agoMerge FreeBSD changes into GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 04:01:01 +0000 (04:01 +0000)]
Merge FreeBSD changes into GCC 3.4.2-prerelease.
  1.19 Fix -mprofiler-epilogue on i386.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 03:57:21 +0000 (03:57 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prerelease:
  1.2 -fformat-extensions.
  1.3 -printf0
  1.6 teach GCC about %y for DDB.

20 years agoUse stock version of this file.
Alexander Kabaev [Wed, 28 Jul 2004 03:53:20 +0000 (03:53 +0000)]
Use stock version of this file.
Revision 1.3 to suppress warnings on K&R main is not useful anymore
since GCC suppresses such warnings unconditionally now.

20 years agoMerge FreeBSD modifications into gcc 3.4.2-prerelease:
Alexander Kabaev [Wed, 28 Jul 2004 03:46:02 +0000 (03:46 +0000)]
Merge FreeBSD modifications into gcc 3.4.2-prerelease:
  1.2   framework for our kernel printf enhancements.
  1.3   suppress prototype warnings on K&R main.

Both flag variables declared here were disconnected through the 5-current
lifetime and are retained here for backwards compatibility reasons only.

20 years agoMerge FreeBSD changes into GCC 3.4.2-prerelease.
Alexander Kabaev [Wed, 28 Jul 2004 03:40:54 +0000 (03:40 +0000)]
Merge FreeBSD changes into GCC 3.4.2-prerelease.
r1.18: Move FreeBSD specific flag variables from c-decl.c here.

20 years agoUse stock GCC versions on these files.
Alexander Kabaev [Wed, 28 Jul 2004 03:36:15 +0000 (03:36 +0000)]
Use stock GCC versions on these files.

20 years agoErroneously decreased __FreeBSD_version, which should have been increased instead:
Oliver Eikemeier [Wed, 28 Jul 2004 03:36:05 +0000 (03:36 +0000)]
Erroneously decreased __FreeBSD_version, which should have been increased instead:

  "__FreeBSD_version should only ever increment.  It is a historial record
  of events in the system.  Decrementing it is akin to trying to go back
  in time and change history."

Reminded by: kuriyama, scottl

20 years agoThis commit was generated by cvs2svn to compensate for changes in r132724,
Alexander Kabaev [Wed, 28 Jul 2004 03:12:14 +0000 (03:12 +0000)]
This commit was generated by cvs2svn to compensate for changes in r132724,
which included commits to RCS files with non-trunk default branches.

20 years agoGcc 3.4.2 20040728 libf2c bits.
Alexander Kabaev [Wed, 28 Jul 2004 03:12:14 +0000 (03:12 +0000)]
Gcc 3.4.2 20040728 libf2c bits.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r132722,
Alexander Kabaev [Wed, 28 Jul 2004 03:12:12 +0000 (03:12 +0000)]
This commit was generated by cvs2svn to compensate for changes in r132722,
which included commits to RCS files with non-trunk default branches.

20 years agoGcc 3.4.2 20040728 Objective C support bits.
Alexander Kabaev [Wed, 28 Jul 2004 03:12:12 +0000 (03:12 +0000)]
Gcc 3.4.2 20040728 Objective C support bits.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r132720,
Alexander Kabaev [Wed, 28 Jul 2004 03:12:05 +0000 (03:12 +0000)]
This commit was generated by cvs2svn to compensate for changes in r132720,
which included commits to RCS files with non-trunk default branches.

20 years agoGcc 3.4.2 20040728 C++ support bits.
Alexander Kabaev [Wed, 28 Jul 2004 03:12:05 +0000 (03:12 +0000)]
Gcc 3.4.2 20040728 C++ support bits.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r132718,
Alexander Kabaev [Wed, 28 Jul 2004 03:11:36 +0000 (03:11 +0000)]
This commit was generated by cvs2svn to compensate for changes in r132718,
which included commits to RCS files with non-trunk default branches.

20 years agoGcc 3.4.2 20040728.
Alexander Kabaev [Wed, 28 Jul 2004 03:11:36 +0000 (03:11 +0000)]
Gcc 3.4.2 20040728.

20 years agoFix a bug in the sack code that was causing data to be retransmitted
Jayanth Vijayaraghavan [Wed, 28 Jul 2004 02:15:14 +0000 (02:15 +0000)]
Fix a bug in the sack code that was causing data to be retransmitted
with the FIN bit set for all segments, if a FIN has already been sent before.
The fix will allow the FIN bit to be set for only the last segment, in case
it has to be retransmitted.

Fix another bug that would have caused snd_nxt to be pulled by len if
there was an error from ip_output. snd_nxt should not be touched
during sack retransmissions.

20 years agoback out the localkg changes until things have settled.
Oliver Eikemeier [Wed, 28 Jul 2004 00:09:19 +0000 (00:09 +0000)]
back out the localkg changes until things have settled.

Discussed with: mtm

20 years agoForced commit to note that the previous commit was:
Robert Watson [Tue, 27 Jul 2004 23:45:19 +0000 (23:45 +0000)]
Forced commit to note that the previous commit was:

Approved by: suz (core@kame.net)

20 years agoCommit a first pass at in6pcb and pcbinfo locking for IPv6,
Robert Watson [Tue, 27 Jul 2004 23:44:03 +0000 (23:44 +0000)]
Commit a first pass at in6pcb and pcbinfo locking for IPv6,
synchronizing IPv6 protocol control blocks and lists.  These changes
are modeled on the inpcb locking for IPv4, submitted by Jennifer Yang,
and committed by Jeffrey Hsu.  With these locking changes, IPv6 use of
inpcbs is now substantially more MPSAFE, and permits IPv4 inpcb locking
assertions to be run in the presence of IPv6 compiled into the kernel.

20 years agoAdd an UPDATING entry about recompiling modules implementing network
Robert Watson [Tue, 27 Jul 2004 23:32:06 +0000 (23:32 +0000)]
Add an UPDATING entry about recompiling modules implementing network
interfaces due to a change in the size of struct ifnet.

20 years agoAdd a new network interface flag, IFF_NEEDSGIANT, which will allow
Robert Watson [Tue, 27 Jul 2004 23:20:45 +0000 (23:20 +0000)]
Add a new network interface flag, IFF_NEEDSGIANT, which will allow
device drivers to declare that the ifp->if_start() method implemented
by the driver requires Giant in order to operate correctly.

Add a 'struct task' to 'struct ifnet' that can be used to execute a
deferred ifp->if_start() in the event that if_start needs to be called
in a Giant-free environment.  To do this, introduce if_start(), a
wrapper function for ifp->if_start().  If the interface can run MPSAFE,
it directly dispatches into the interface start routine.  If it can't
run MPSAFE, we're running with debug.mpsafenet != 0, and Giant isn't
currently held, the task is queued to execute in a swi holding Giant
via if_start_deferred().

Modify if_handoff() to use if_start() instead of direct dispatch.
Modify 802.11 to use if_start() instead of direct dispatch.

This is intended to provide increased compatibility for non-MPSAFE
network device drivers in the presence of Giant-free operation via
asynchronous dispatch.  However, this commit does not mark any network
interfaces as IFF_NEEDSGIANT.

20 years agoAdd workaround for brain damaged cell phones
Maksim Yevmenkin [Tue, 27 Jul 2004 22:40:42 +0000 (22:40 +0000)]
Add workaround for brain damaged cell phones

PR: bin/67906

20 years agoConvert the vfsconf list to a TAILQ.
Poul-Henning Kamp [Tue, 27 Jul 2004 22:32:01 +0000 (22:32 +0000)]
Convert the vfsconf list to a TAILQ.

Introduce vfs_byname() function to find things on it.

Staticize vfs_nmount() function under the name vfs_donmount().

Various cleanups.

20 years agoStart new sentence on new line.
Colin Percival [Tue, 27 Jul 2004 22:17:26 +0000 (22:17 +0000)]
Start new sentence on new line.

Pointed out by: simon

20 years agoUse kernel_vmount() instead of vfs_nmount().
Poul-Henning Kamp [Tue, 27 Jul 2004 21:38:42 +0000 (21:38 +0000)]
Use kernel_vmount() instead of vfs_nmount().

20 years agoAdd a BUGS entry pointing out that -mindepth and -maxdepth are global
Colin Percival [Tue, 27 Jul 2004 21:22:14 +0000 (21:22 +0000)]
Add a BUGS entry pointing out that -mindepth and -maxdepth are global
options even though they look like primaries.  (This is already documented
in the options themselves, but is sufficiently astonishing that I think it
deserves a BUGS entry as well.)

20 years agoMerge the "multibyte not supported" BUG into the pre-existing BUGS
Colin Percival [Tue, 27 Jul 2004 21:13:04 +0000 (21:13 +0000)]
Merge the "multibyte not supported" BUG into the pre-existing BUGS
section.

Move the HISTORY section to place it before BUGS rather than after BUGS,
in order to minimize the chance of this error being reproduced in the
future.  (Both mdoc(7) and 63% of manual pages have these sections listed
in this order.)

20 years agoWhen making a peer of unknown type framework tries to load module
Gleb Smirnoff [Tue, 27 Jul 2004 20:30:56 +0000 (20:30 +0000)]
When making a peer of unknown type framework tries to load module
using linker_load_module(). This works OK if NGM_MKPEER message came
from userland and we have process associated with thread. But when
NGM_MKPEER was queued because target node was busy, linker_load_module()
is called from netisr thread leading to panic.
To workaround that we do not load modules by framework, instead ng_socket
loads module (if this is required) before sending NGM_MKPEER.
However, the race condition between return from NgSendMsg() and actual
creation of node still exist and needs to be solved.

PR: kern/62789
Approved by: julian

20 years agoAfter conferring with Apple legal, update the license of this to APSL 2.0
Jordan K. Hubbard [Tue, 27 Jul 2004 20:01:43 +0000 (20:01 +0000)]
After conferring with Apple legal, update the license of this to APSL 2.0
and change the URL link to point to the appropriate license text.

20 years agoWhen node is server serve both standard RFC2516 and non-standard 3Com
Gleb Smirnoff [Tue, 27 Jul 2004 19:47:13 +0000 (19:47 +0000)]
When node is server serve both standard RFC2516 and non-standard 3Com
clients simultaneously. When node is client its mode is configured
with a control message.

sysctl net.graph.nonstandard_pppoe is deprecated but kept for
backward compatibility for some time.

Approved by: julian

20 years agoCorrect typo in prior commit: s/cd/td/
Robert Watson [Tue, 27 Jul 2004 19:44:08 +0000 (19:44 +0000)]
Correct typo in prior commit: s/cd/td/

20 years agoEven though binaries will give 0 matches, make it explicit.
Mike Makonnen [Tue, 27 Jul 2004 16:59:35 +0000 (16:59 +0000)]
Even though binaries will give 0 matches, make it explicit.

20 years agoPass a thread argument into cpu_critical_{enter,exit}() rather than
Robert Watson [Tue, 27 Jul 2004 16:41:01 +0000 (16:41 +0000)]
Pass a thread argument into cpu_critical_{enter,exit}() rather than
dereference curthread.  It is called only from critical_{enter,exit}(),
which already dereferences curthread.  This doesn't seem to affect SMP
performance in my benchmarks, but improves MySQL transaction throughput
by about 1% on UP on my Xeon.

Head nodding: jhb, bmilekic

20 years agoDon't consider TCP connections beyond LISTEN state
Yaroslav Tykhiy [Tue, 27 Jul 2004 16:35:09 +0000 (16:35 +0000)]
Don't consider TCP connections beyond LISTEN state
(i.e. with the foreign address being not wildcard) when checking
for possible port theft since such connections cannot be stolen.

The port theft check is FreeBSD-specific and isn't in the KAME tree.

PR: bin/65928 (in the audit trail)
Reviewed by: -net, -hackers (silence)
Tested by: Nick Leuta <skynick at mail.sc.ru>
MFC after: 1 month

20 years agoAdd "options ADAPTIVE_GIANT" which causes Giant to also be treated in
Robert Watson [Tue, 27 Jul 2004 16:34:48 +0000 (16:34 +0000)]
Add "options ADAPTIVE_GIANT" which causes Giant to also be treated in
an adaptive fashion when adaptive mutexes are enabled.  The theory
behind non-adaptive Giant is that Giant will be held for long periods
of time, and therefore spinning waiting on it is wasteful.  However,
in MySQL benchmarks which are relatively Giant-free, running Giant
adaptive makes an observable difference on SMP (5% transaction rate
improvement).  As such, make adaptive behavior on Giant an option so
it can be more widely benchmarked.

20 years agoFixed style of previous commit.
Ruslan Ermilov [Tue, 27 Jul 2004 12:31:38 +0000 (12:31 +0000)]
Fixed style of previous commit.

20 years agoDon't strip trailing linear whitespace from passwords.
Dag-Erling Smørgrav [Tue, 27 Jul 2004 11:34:25 +0000 (11:34 +0000)]
Don't strip trailing linear whitespace from passwords.

MFC after: 2 weeks

20 years agoStyle nits.
Dag-Erling Smørgrav [Tue, 27 Jul 2004 11:30:35 +0000 (11:30 +0000)]
Style nits.

20 years agoCorrect a couple of noticed slips in sorting order
Yaroslav Tykhiy [Tue, 27 Jul 2004 10:37:04 +0000 (10:37 +0000)]
Correct a couple of noticed slips in sorting order
and add missing markup at one spot.

20 years agoNew release notes:
Yaroslav Tykhiy [Tue, 27 Jul 2004 10:30:11 +0000 (10:30 +0000)]
New release notes:
fxp(4) allows toggling its VLAN_MTU capability;
ifconfig(8) provides controls for VLAN_MTU and VLAN_HWTAGGING.

20 years agoThis commit was generated by cvs2svn to compensate for changes in r132691,
Dag-Erling Smørgrav [Tue, 27 Jul 2004 10:15:55 +0000 (10:15 +0000)]
This commit was generated by cvs2svn to compensate for changes in r132691,
which included commits to RCS files with non-trunk default branches.

20 years agoBump the date, .Dd, since the document content has been changed
Yaroslav Tykhiy [Tue, 27 Jul 2004 09:51:49 +0000 (09:51 +0000)]
Bump the date, .Dd, since the document content has been changed
in the previous commit.

Pointed out by: ru

20 years agoProperly implement kdb_cpu_{set|clear}_singlestep to allow DDB to
Peter Grehan [Tue, 27 Jul 2004 07:06:20 +0000 (07:06 +0000)]
Properly implement kdb_cpu_{set|clear}_singlestep to allow DDB to
continue from breakpoints.

20 years agoMake sure icache is sync'd whenever memory is touched. It may
Peter Grehan [Tue, 27 Jul 2004 07:04:58 +0000 (07:04 +0000)]
Make sure icache is sync'd whenever memory is touched. It may
be more optimal to override the BKPT_WRITE macro, but DDB performance
isn't really a goal at this stage...

20 years agoAdd UTF-8-specific implementations of mbsnrtowcs() and wcsnrtombs().
Tim J. Robbins [Tue, 27 Jul 2004 06:29:48 +0000 (06:29 +0000)]
Add UTF-8-specific implementations of mbsnrtowcs() and wcsnrtombs().
These convert plain ASCII characters in-line, making them only slightly
slower than the single-byte ("NONE" encoding) version when processing
ASCII strings.

20 years agoafter reading the commit mail, I better understand what cnt_hold is for..
John-Mark Gurney [Tue, 27 Jul 2004 04:23:00 +0000 (04:23 +0000)]
after reading the commit mail, I better understand what cnt_hold is for..
improve the verbage in the manpage so others are more likely to understand
it...

Reviewed by: rwatson

20 years agoVarious improvements for ia64:
Marcel Moolenaar [Tue, 27 Jul 2004 03:54:16 +0000 (03:54 +0000)]
Various improvements for ia64:
o  s/TARGET_XFER_DIRTY/NATIVE_XFER_DIRTY/g to be conformant,
o  Handle TARGET_OBJECT_DIRTY for core files. Required on Linux,
o  Replace ia64_fbsd_supply_fpregs() and ia64_fbsd_supply_gregs()
   with supply_fpregset() and supply_gregset() resp. Needed to
   support libthread_db.
o  For target_read_partial(), pass the address of bspstore as the
   annex instead of bspstore itself. This fixes crossbuilds. The
   annex is a pointer, while bspstore is a ULONGEST.
o  In ia64_frame_prev_register(), handle dirty stacked registers.

20 years ago - Use atomic ops for updating the vmspace's refcnt and exitingcnt.
Alan Cox [Tue, 27 Jul 2004 03:53:41 +0000 (03:53 +0000)]
 - Use atomic ops for updating the vmspace's refcnt and exitingcnt.
 - Push down Giant into shmexit().  (Giant is acquired only if the vmspace
   contains shm segments.)
 - Eliminate the acquisition of Giant from proc_rwmem().
 - Reduce the scope of Giant in exit1(), uncovering the destruction of the
   address space.

20 years agoSave DAR/DSISR in DDB regsave area when stack overflow detected. It's
Peter Grehan [Tue, 27 Jul 2004 03:46:34 +0000 (03:46 +0000)]
Save DAR/DSISR in DDB regsave area when stack overflow detected. It's
hard to work out where the problem was without these.

20 years agoMove the schedlock owner state update following the context
Bosko Milekic [Tue, 27 Jul 2004 03:46:31 +0000 (03:46 +0000)]
Move the schedlock owner state update following the context
switch in fork_exit() to before anything else is done (but keep
schedlock for the deadthread check).  This means one less
nasty bug if ever in the future whatever might have been called
before the update played with schedlock or critical sections.

Discussed with: tjr

20 years agoImprove boot-time debugging with DDB by extracting the ksym start/end
Peter Grehan [Tue, 27 Jul 2004 03:41:34 +0000 (03:41 +0000)]
Improve boot-time debugging with DDB by extracting the ksym start/end
values from the loader.

20 years agoUse file2c instead of a combination of hexdump, sed and shell script to
Tim J. Robbins [Tue, 27 Jul 2004 01:33:27 +0000 (01:33 +0000)]
Use file2c instead of a combination of hexdump, sed and shell script to
generate the wakecode[] array from acpi_wakecode.bin. The old method was
not safe in multibyte locales.

20 years agoadd a command to calculate the ratios of cnt_hold and cnt_lock to locks
John-Mark Gurney [Tue, 27 Jul 2004 00:36:53 +0000 (00:36 +0000)]
add a command to calculate the ratios of cnt_hold and cnt_lock to locks
from mutex profiling...  This takes cnt_* and divides it by count, so
lower is better...

Sample command:
sh mtxpercent.sh | sort -rn +7 -8

SEE ALSO
MUTEX_PROFILING(9)

20 years agoFix typo in description of pflog_logfile.
Max Laier [Tue, 27 Jul 2004 00:28:16 +0000 (00:28 +0000)]
Fix typo in description of pflog_logfile.

Submitted by: Mike Jakubik

20 years agoFix URLs.
Hiroki Sato [Tue, 27 Jul 2004 00:02:19 +0000 (00:02 +0000)]
Fix URLs.

20 years agoFix for a SACK bug where the very last segment retransmitted
Jayanth Vijayaraghavan [Mon, 26 Jul 2004 23:41:12 +0000 (23:41 +0000)]
Fix for a SACK bug where the very last segment retransmitted
from the SACK scoreboard could result in the next (untransmitted)
segment to be skipped.

20 years agocompare pointer against NULL, not 0
John-Mark Gurney [Mon, 26 Jul 2004 21:29:56 +0000 (21:29 +0000)]
compare pointer against NULL, not 0

when inpcb is NULL, this is no longer invalid since jlemon added the
tcp_twstart function... this prevents close "failing" w/ EINVAL when it
really was successful...

Reviewed by: jeremy (NetBSD)

20 years agodocument that times are in microseconds, yes it's at the end in the NOTES
John-Mark Gurney [Mon, 26 Jul 2004 21:21:08 +0000 (21:21 +0000)]
document that times are in microseconds, yes it's at the end in the NOTES
section, but this makes it easier to find.

20 years agoThe name of the landlord in Shaw's _The Man of Destiny_ is Giuseppe (not
Dag-Erling Smørgrav [Mon, 26 Jul 2004 21:18:58 +0000 (21:18 +0000)]
The name of the landlord in Shaw's _The Man of Destiny_ is Giuseppe (not
Guiseppe) Gandi.

Submitted by: Massimiliano Stucchi <stucchi@willystudios.com>
MFC after: 3 days

20 years agoNo capital letter after :
Philippe Charnier [Mon, 26 Jul 2004 20:24:59 +0000 (20:24 +0000)]
No capital letter after :

20 years agoAdd __FBSDID. Replace local variable sin by sockin to not conflict with sin(3).
Philippe Charnier [Mon, 26 Jul 2004 20:18:11 +0000 (20:18 +0000)]
Add __FBSDID. Replace local variable sin by sockin to not conflict with sin(3).
Use warnx() instead of warn() when error message is not of any interest. Add
prototypes.

20 years agoIntroduce options list the standard way. Correct style(9) in FILES section.
Philippe Charnier [Mon, 26 Jul 2004 19:59:29 +0000 (19:59 +0000)]
Introduce options list the standard way. Correct style(9) in FILES section.

20 years agoIntroduce options list the standard way.
Philippe Charnier [Mon, 26 Jul 2004 19:53:44 +0000 (19:53 +0000)]
Introduce options list the standard way.

20 years agoRemove useless .Pp. Typo: gcos -> gecos (as spelled in passwd(5)).
Philippe Charnier [Mon, 26 Jul 2004 19:49:29 +0000 (19:49 +0000)]
Remove useless .Pp. Typo: gcos -> gecos (as spelled in passwd(5)).

20 years agoEnclose .Fa fn with ``The ... function'' at the beginning of sentences.
Philippe Charnier [Mon, 26 Jul 2004 19:35:20 +0000 (19:35 +0000)]
Enclose .Fa fn with ``The ... function'' at the beginning of sentences.

20 years agoImplement the protection check required by the pmap_extract_and_hold()
Alan Cox [Mon, 26 Jul 2004 18:10:10 +0000 (18:10 +0000)]
Implement the protection check required by the pmap_extract_and_hold()
specification.

Reviewed and tested by: grehan@

20 years agoImprove geom(8)'s 'list' command to show geoms and their providers and
Pawel Jakub Dawidek [Mon, 26 Jul 2004 17:14:47 +0000 (17:14 +0000)]
Improve geom(8)'s 'list' command to show geoms and their providers and
consumers. Teach STRIPE, CONCAT and NOP classes about this improvement.

20 years agoChange naming scheme from /dev/<name>.stripe to /dev/stripe/<name>.
Pawel Jakub Dawidek [Mon, 26 Jul 2004 16:10:27 +0000 (16:10 +0000)]
Change naming scheme from /dev/<name>.stripe to /dev/stripe/<name>.

20 years agoChange naming scheme from /dev/<name>.concat to /dev/concat/<name>.
Pawel Jakub Dawidek [Mon, 26 Jul 2004 16:08:32 +0000 (16:08 +0000)]
Change naming scheme from /dev/<name>.concat to /dev/concat/<name>.

20 years agoM_WAITOK is ok here, while I'm using M_WAITOK later in this function.
Pawel Jakub Dawidek [Mon, 26 Jul 2004 15:41:28 +0000 (15:41 +0000)]
M_WAITOK is ok here, while I'm using M_WAITOK later in this function.

20 years agoM_WAITOK is ok here, while I'm using M_WAITOK later in this function.
Pawel Jakub Dawidek [Mon, 26 Jul 2004 15:35:04 +0000 (15:35 +0000)]
M_WAITOK is ok here, while I'm using M_WAITOK later in this function.

20 years agoAssign the result of getopt() to an int rather than to a char (which is
Stefan Farfeleder [Mon, 26 Jul 2004 15:04:57 +0000 (15:04 +0000)]
Assign the result of getopt() to an int rather than to a char (which is
possibly unsigned).

20 years agoStop tinkering with the parent's VLAN_MTU capability.
Yaroslav Tykhiy [Mon, 26 Jul 2004 14:46:04 +0000 (14:46 +0000)]
Stop tinkering with the parent's VLAN_MTU capability.
Now it is user-controlled through ifconfig(8).

The former ``automagic'' way of operation created more
trouble than good.  First, VLAN_MTU consumers other than
vlan(4) had appeared, e.g., ng_vlan(4).  Second, there was
no way to disable VLAN_MTU manually if it were causing
trouble, e.g., data corruption.

Dropping the ``automagic'' should be completely invisible
to the user since
a) all the drivers supporting VLAN_MTU
have it enabled by default, and in the first place
b) there is only one driver that can really toggle VLAN_MTU
in the hardware under its control (it's fxp(4), to which
I added VLAN_MTU controls to illustrate the principle.)

20 years agoAdd two knobs to ifconfig(8), `vlanmtu' and `vlanhwtag',
Yaroslav Tykhiy [Mon, 26 Jul 2004 13:25:45 +0000 (13:25 +0000)]
Add two knobs to ifconfig(8), `vlanmtu' and `vlanhwtag',
that provide control over the respective capabilities
of an interface, reception of extended frames and hardware
VLAN multiplexor.

20 years agore-enabled Rev 1.15 (lost during KAME merge at Rev 1.25, due to a KAME bug...)
SUZUKI Shinsuke [Mon, 26 Jul 2004 08:18:37 +0000 (08:18 +0000)]
re-enabled Rev 1.15 (lost during KAME merge at Rev 1.25, due to a KAME bug...)

Obtained from: KAME

20 years agoIn revision 1.228, I accidentally broke the "total number of processes in
Colin Percival [Mon, 26 Jul 2004 07:54:39 +0000 (07:54 +0000)]
In revision 1.228, I accidentally broke the "total number of processes in
the system" resource limit code: When checking if the caller has superuser
privileges, we should be checking the *real* user, not the *effective*
user.  (In general, resource limiting is done based on the real user, in
order to avoid resource-exhaustion-by-setuid-program attacks.)

Now that a SUSER_RUID flag to suser_cred exists, use it here to return
this code to its correct behaviour.

Pointed out by: rwatson

20 years agoSave the vinum config back to disk after syncing two plexes.
Lukas Ertl [Mon, 26 Jul 2004 07:30:21 +0000 (07:30 +0000)]
Save the vinum config back to disk after syncing two plexes.

20 years agoRename suser_cred()'s PRISON_ROOT flag to SUSER_ALLOWJAIL. This is
Colin Percival [Mon, 26 Jul 2004 07:24:04 +0000 (07:24 +0000)]
Rename suser_cred()'s PRISON_ROOT flag to SUSER_ALLOWJAIL.  This is
somewhat clearer, but more importantly allows for a consistent naming
scheme for suser_cred flags.

The old name is still defined, but will be removed in a few days (unless I
hear any complaints...)

Discussed with: rwatson, scottl
Requested by: jhb

20 years agoAdd DragonFly BSD 1.0.
Sergey A. Osokin [Mon, 26 Jul 2004 06:40:48 +0000 (06:40 +0000)]
Add DragonFly BSD 1.0.

Approved by: wosch (maintainer)