]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
24 years agoBack out the previous change to install the examples - it breaks the release builds.
jkh [Tue, 11 Jan 2000 02:18:33 +0000 (02:18 +0000)]
Back out the previous change to install the examples - it breaks the release builds.

24 years ago correct the entry for the Linksys EtherFast 10/100
jmb [Tue, 11 Jan 2000 02:03:12 +0000 (02:03 +0000)]
correct the entry for the Linksys EtherFast 10/100
PC Card (PCMPC100).  the entry was one character
short...the final ")" was missing.

Pointed out by: Chris D. Faulhaber - jedgar@fxp.org - jedgar@FreeBSD.org

24 years agoAttempt to fix a problem with receiving packets on USB ethernet interfaces.
wpaul [Mon, 10 Jan 2000 23:12:54 +0000 (23:12 +0000)]
Attempt to fix a problem with receiving packets on USB ethernet interfaces.
Packets are received inside USB bulk transfer callbacks, which run at
splusb() (actually splbio()). The packet input queues are meant to be
manipulated at splimp(). However the locking apparently breaks down under
certain circumstances and the input queues can get trampled.

There's a similar problem with if_ppp, which is driven by hardware/tty
interrupts from the serial driver, but which must also manipulate the
packet input queues at splimp(). The fix there is to use a netisr, and
that's the fix I used here. (I can hear you groaning back there. Hush up.)

The usb_ethersubr module maintains a single queue of its own. When a
packet is received in the USB callback routine, it's placed on this
queue with usb_ether_input(). This routine also schedules a soft net
interrupt with schednetisr(). The ISR routine then runs later, at
splnet, outside of the USB callback/interrupt context, and passes the
packet to ether_input(), hopefully in a safe manner.

The reason this is implemented as a separate module is that there are
a limited number of NETISRs that we can use, and snarfing one up for
each driver that needs it is wasteful (there will be three once I get
the CATC driver done). It also reduces code duplication to a certain
small extent. Unfortunately, it also needs to be linked in with the
usb.ko module in order for the USB ethernet drivers to share it.

Also removed some uneeded includes from if_aue.c and if_kue.c

Fix suggested by: peter
Not rejected as a hairbrained idea by: n_hibma

24 years agoDon't fail if less then MAXDEV /dev/usb\d+ entries exist.
n_hibma [Mon, 10 Jan 2000 22:35:33 +0000 (22:35 +0000)]
Don't fail if less then MAXDEV /dev/usb\d+ entries exist.

24 years agoAdd an entry for dhclient and the USB ethernet adapters.
n_hibma [Mon, 10 Jan 2000 22:34:36 +0000 (22:34 +0000)]
Add an entry for dhclient and the USB ethernet adapters.

24 years agoAdd an extra debugging message.
n_hibma [Mon, 10 Jan 2000 22:33:43 +0000 (22:33 +0000)]
Add an extra debugging message.

24 years agoAdd Sun keyboard and NetChip
n_hibma [Mon, 10 Jan 2000 22:31:01 +0000 (22:31 +0000)]
Add Sun keyboard and NetChip

24 years agoUpdate syscalls to Linux kernel 2.3.38. These syscalls are
marcel [Mon, 10 Jan 2000 22:19:06 +0000 (22:19 +0000)]
Update syscalls to Linux kernel 2.3.38. These syscalls are
implemented as dummies to ease problem solving.

24 years agoUpdate the syscalls to Linux kernel 2.3.38.
marcel [Mon, 10 Jan 2000 22:16:35 +0000 (22:16 +0000)]
Update the syscalls to Linux kernel 2.3.38.

24 years agoFix 'make world' breakage caused by missing comment character.
archie [Mon, 10 Jan 2000 21:44:28 +0000 (21:44 +0000)]
Fix 'make world' breakage caused by missing comment character.

24 years agomalloc more space for temp file name
kris [Mon, 10 Jan 2000 20:26:24 +0000 (20:26 +0000)]
malloc more space for temp file name

Noticed by: marcel

24 years agoInclude a note below the example qmail entry that mentions that inetd is
dbaker [Mon, 10 Jan 2000 20:02:28 +0000 (20:02 +0000)]
Include a note below the example qmail entry that mentions that inetd is
no longer the correct way to have qmail handle incoming qmail smtp
connections.  Also provide a url to the correct method.

24 years agoAdd a comment to hopefully prevent more bugs like kern/8596.
archie [Mon, 10 Jan 2000 19:34:40 +0000 (19:34 +0000)]
Add a comment to hopefully prevent more bugs like kern/8596.

24 years agoMGETHDR() does not initialize m_pkthdr.rcvif, do it here.
ru [Mon, 10 Jan 2000 18:46:05 +0000 (18:46 +0000)]
MGETHDR() does not initialize m_pkthdr.rcvif, do it here.

This fixes page fault panic observed when diverting packets
with IP options (e.g. ping -R remoteIP over natd).

PR: kern/8596, kern/11199

24 years agoUse GENERIC for both pccard and non-pccard disks now that GENERIC is
imp [Mon, 10 Jan 2000 16:27:32 +0000 (16:27 +0000)]
Use GENERIC for both pccard and non-pccard disks now that GENERIC is
(or should be) everything that PCCARD is.

24 years agoSigh. RFC2038 and bind 8.2.2 have a slight variation of interpretation
peter [Mon, 10 Jan 2000 15:31:40 +0000 (15:31 +0000)]
Sigh. RFC2038 and bind 8.2.2 have a slight variation of interpretation
of the SOA 'minimum' field.  Now it's necessary to define $TTL seperately
to shut it up.  Bind does reasonable things by default but it's annoying
still.

PR: 15834
Submitted by: Daniel Lewart <d-lewart@uiuc.edu>

24 years agoFix the use of an uninitialized variable in the previous commit.
peter [Mon, 10 Jan 2000 14:20:53 +0000 (14:20 +0000)]
Fix the use of an uninitialized variable  in the previous commit.

Also, in addition to the previous log message, the last change had a fix
for the case where where f.mntfromname is a relative path like da0a.

Submitted by: bde

24 years agoAdd items for pop3 and imap
phantom [Mon, 10 Jan 2000 14:08:30 +0000 (14:08 +0000)]
Add items for pop3 and imap

PR: conf/14714
Submitted by: Ronald F. Guilmette <rfg@monkeys.com>

24 years agoReturn Linux kernel version 2.2.12 by default. This is in line
marcel [Mon, 10 Jan 2000 13:09:08 +0000 (13:09 +0000)]
Return Linux kernel version 2.2.12 by default. This is in line
with linux_base-6.1.

24 years agoAdded PCIR_BIOS (0x30).
yokota [Mon, 10 Jan 2000 12:53:19 +0000 (12:53 +0000)]
Added PCIR_BIOS (0x30).

24 years agoAdd `.Nm rrestore' to NAME section.
phantom [Mon, 10 Jan 2000 12:27:33 +0000 (12:27 +0000)]
Add `.Nm rrestore' to NAME section.

24 years agoAdd `.Nm fastboot' and `.Nm fasthalt' to NAME section.
phantom [Mon, 10 Jan 2000 12:24:47 +0000 (12:24 +0000)]
Add `.Nm fastboot' and `.Nm fasthalt' to NAME section.

24 years agoremove check now done in vn_isdisk().
phk [Mon, 10 Jan 2000 12:24:36 +0000 (12:24 +0000)]
remove check now done in vn_isdisk().

24 years agoAlso handle zero return from dscheck().
phk [Mon, 10 Jan 2000 12:21:39 +0000 (12:21 +0000)]
Also handle zero return from dscheck().

PR: 15956

24 years agoAdd `.Nm red' to NAME section.
phantom [Mon, 10 Jan 2000 12:20:30 +0000 (12:20 +0000)]
Add `.Nm red' to NAME section.

24 years agoAdd `.Nm rdump' to NAME section.
phantom [Mon, 10 Jan 2000 12:18:10 +0000 (12:18 +0000)]
Add `.Nm rdump' to NAME section.

24 years agoInstall html files to /usr/share/doc/ncurses/
phantom [Mon, 10 Jan 2000 12:12:51 +0000 (12:12 +0000)]
Install html files to /usr/share/doc/ncurses/

24 years agoCreate /usr/share/doc/ncurses (place for html files from ncurses dist)
phantom [Mon, 10 Jan 2000 12:11:51 +0000 (12:11 +0000)]
Create /usr/share/doc/ncurses (place for html files from ncurses dist)

24 years agoRemove controller miibus, there already were a device miibus.
phk [Mon, 10 Jan 2000 12:06:32 +0000 (12:06 +0000)]
Remove controller miibus, there already were a device miibus.

24 years agoGive vn_isdisk() a second argument where it can return a suitable errno.
phk [Mon, 10 Jan 2000 12:04:27 +0000 (12:04 +0000)]
Give vn_isdisk() a second argument where it can return a suitable errno.

Suggested by: bde

24 years agoInstall contents of the TESTS subdirectory to /usr/share/examples/libdialog
phantom [Mon, 10 Jan 2000 12:04:18 +0000 (12:04 +0000)]
Install contents of the TESTS subdirectory to /usr/share/examples/libdialog

24 years agoCreate /usr/share/examples/libdialog (examples of dialog(3) usage)
phantom [Mon, 10 Jan 2000 12:01:27 +0000 (12:01 +0000)]
Create /usr/share/examples/libdialog (examples of dialog(3) usage)

24 years agoRemove -lncurses and -lmytinfo from LDFLAGS
phantom [Mon, 10 Jan 2000 11:54:55 +0000 (11:54 +0000)]
Remove -lncurses and -lmytinfo from LDFLAGS

24 years agoCleanup warnings.
phantom [Mon, 10 Jan 2000 11:52:10 +0000 (11:52 +0000)]
Cleanup warnings.

24 years agoFix typo: MAINAINER -> MAINTAINER
phantom [Mon, 10 Jan 2000 11:38:51 +0000 (11:38 +0000)]
Fix typo: MAINAINER -> MAINTAINER

24 years agorevision 1.40 backed out the removal of PCATCH in a tsleep allowing
alfred [Mon, 10 Jan 2000 10:24:47 +0000 (10:24 +0000)]
revision 1.40 backed out the removal of PCATCH in a tsleep allowing
an infinite loop if a signal is delivered here.
remove it again, this still ought to be revisited as the error should
probably be returned.

24 years ago10 X's for mktemp(). This seems to be free from race conditions.
kris [Mon, 10 Jan 2000 09:33:37 +0000 (09:33 +0000)]
10 X's for mktemp(). This seems to be free from race conditions.

24 years agoCompletely reconstruct page. Make it mdoc(7) style compatible also :)
phantom [Mon, 10 Jan 2000 09:29:21 +0000 (09:29 +0000)]
Completely reconstruct page. Make it mdoc(7) style compatible also :)
Page still needs some work about english refinements and some actualization.

24 years ago10 X's for mkstemp(), and don't redefine _PATH_TMP
kris [Mon, 10 Jan 2000 09:25:32 +0000 (09:25 +0000)]
10 X's for mkstemp(), and don't redefine _PATH_TMP

24 years agoActivate doc/ subdir
phantom [Mon, 10 Jan 2000 09:20:27 +0000 (09:20 +0000)]
Activate doc/ subdir

24 years agoInstall ntp html docs as /usr/share/doc/ntp
phantom [Mon, 10 Jan 2000 09:20:03 +0000 (09:20 +0000)]
Install ntp html docs as /usr/share/doc/ntp

24 years agoCreate doc dir for ntp (/usr/share/doc/ntp)
phantom [Mon, 10 Jan 2000 09:18:02 +0000 (09:18 +0000)]
Create doc dir for ntp (/usr/share/doc/ntp)

24 years ago10 X's in mkstemp().
kris [Mon, 10 Jan 2000 09:17:46 +0000 (09:17 +0000)]
10 X's in mkstemp().

24 years agoInstall additional documentation:
phantom [Mon, 10 Jan 2000 09:05:30 +0000 (09:05 +0000)]
Install additional documentation:
contrib/bind/doc/{html,misc} as /usr/share/doc/bind/{html,misc}

Install additional manpages:
. named.conf.5
. nsupdate.8
. named-bootconf.8

24 years agoCreate dirs for bind's additional documentation:
phantom [Mon, 10 Jan 2000 08:58:00 +0000 (08:58 +0000)]
Create dirs for bind's additional documentation:
/usr/share/doc/bind/{html,misc}

24 years agoUse X x X's in mkstemp().
kris [Mon, 10 Jan 2000 08:56:01 +0000 (08:56 +0000)]
Use X x X's in mkstemp().

24 years agoinstall contrib/bc/Examples as /usr/share/examples/bc
phantom [Mon, 10 Jan 2000 08:55:02 +0000 (08:55 +0000)]
install contrib/bc/Examples as /usr/share/examples/bc

24 years agoadd /usr/share/examples/bc
phantom [Mon, 10 Jan 2000 08:54:24 +0000 (08:54 +0000)]
add /usr/share/examples/bc

24 years agoFeed mkstemp() some more X's to keep it safe.
kris [Mon, 10 Jan 2000 08:54:09 +0000 (08:54 +0000)]
Feed mkstemp() some more X's to keep it safe.

24 years agoObtain the initial key repeat rate setting via BIOS in i386 if
yokota [Mon, 10 Jan 2000 08:52:32 +0000 (08:52 +0000)]
Obtain the initial key repeat rate setting via BIOS in i386 if
possible.

24 years agoAdd some keyboard IDs.
yokota [Mon, 10 Jan 2000 08:50:43 +0000 (08:50 +0000)]
Add some keyboard IDs.

24 years agoRework the algorithm to detect MDA/CGA/EGA/VGA cards, so that
yokota [Mon, 10 Jan 2000 08:49:49 +0000 (08:49 +0000)]
Rework the algorithm to detect MDA/CGA/EGA/VGA cards, so that
the vga driver won't be fooled to believe it has a CGA card when
in fact it is a VGA card.

24 years ago. add manpage for svr4(8)
phantom [Mon, 10 Jan 2000 08:48:11 +0000 (08:48 +0000)]
. add manpage for svr4(8)
. s/freebsd/FreeBSD

24 years ago- Fix typo: CGA40 -> CGA80
yokota [Mon, 10 Jan 2000 08:47:04 +0000 (08:47 +0000)]
- Fix typo: CGA40 -> CGA80
- Remove erroneous comments.

24 years agoMissing FREE_LOCK call before handle_workitem_freeblocks.
mckusick [Mon, 10 Jan 2000 08:39:03 +0000 (08:39 +0000)]
Missing FREE_LOCK call before handle_workitem_freeblocks.

Submitted by: "Kenneth D. Merry" <ken@kdm.org>

24 years ago- Style/bde changes.
peter [Mon, 10 Jan 2000 08:21:22 +0000 (08:21 +0000)]
- Style/bde changes.
  - Don't use realpath as stat does the right thing.
  - Only check ufs filesystems in getmntpt.
  - Dont' bother checking that the ufs-mounted-on
    device is a special file.  It *must* be a special
    file, or ufs wouldn't have mounted it.

Submitted by: Paul Saab <ps@yahoo-inc.com>

24 years ago- Forgot to nuke hotroot completely.
peter [Mon, 10 Jan 2000 08:18:18 +0000 (08:18 +0000)]
- Forgot to nuke hotroot completely.

Submitted by: Paul Saab <ps@yahoo-inc.com>

24 years agoMove xe driver from dev/pccard to dev/xe. Convert driver to newbus.
imp [Mon, 10 Jan 2000 08:05:53 +0000 (08:05 +0000)]
Move xe driver from dev/pccard to dev/xe.  Convert driver to newbus.
Driver is not functional yet, but does compile.  Tests with xe cards
indicates that it doesn't panic the machine when they are present, but
fail to probe.  Interface help in the pcic/pccard layers are needed to
complete this driver.

24 years agoPanic if proc0 hasn't been created and we try to call kthread_create.
imp [Mon, 10 Jan 2000 08:00:58 +0000 (08:00 +0000)]
Panic if proc0 hasn't been created and we try to call kthread_create.
This prevents a more mysterious crash later.

XXX The long term solution is defer creation of these things until
XXX proc0 lives

24 years agomake dsp_read() assert that it has a read channel, not a write channel
cg [Mon, 10 Jan 2000 07:05:15 +0000 (07:05 +0000)]
make dsp_read() assert that it has a read channel, not a write channel

Submitted by: green

24 years agoEliminate pccard_chip_* tonight.
imp [Mon, 10 Jan 2000 06:58:17 +0000 (06:58 +0000)]
Eliminate pccard_chip_* tonight.
o ifdef out pccardchip.h (almost all of it, there are dangling bits
o Add rid/res members to pccard_function
o remove pct/pch from pccard_softc
o map memory properly in scan_cis (almost, see XXX for more work)
o manage ccr.
o remove bogus comment I added about touching the ccr being a layering
  violation for pccard.  It is properly done at that level.
o More function prototyping

24 years agoZap NO_IDEA
kris [Mon, 10 Jan 2000 06:28:04 +0000 (06:28 +0000)]
Zap NO_IDEA

24 years agoList of files to nuke prior to import.
kris [Mon, 10 Jan 2000 06:27:12 +0000 (06:27 +0000)]
List of files to nuke prior to import.

24 years agoAdapt to the new `ccp' now that the traditional-behaving /usr/bin/cpp
obrien [Mon, 10 Jan 2000 06:24:49 +0000 (06:24 +0000)]
Adapt to the new `ccp' now that the traditional-behaving /usr/bin/cpp
script is gone.

PR: 15932
Submitted by: Jos Backus <Jos.Backus@nl.origin-it.com>
Tested by: brian, Manfred Antar <mantar@pacbell.net>

24 years agoThis commit was generated by cvs2svn to compensate for changes in r55714,
kris [Mon, 10 Jan 2000 06:22:05 +0000 (06:22 +0000)]
This commit was generated by cvs2svn to compensate for changes in r55714,
which included commits to RCS files with non-trunk default branches.

24 years agoInitial import of OpenSSL 0.9.4, sans IDEA and RSA code for patent
kris [Mon, 10 Jan 2000 06:22:05 +0000 (06:22 +0000)]
Initial import of OpenSSL 0.9.4, sans IDEA and RSA code for patent
infringement reasons.

24 years agodon't complain about bad intrs unless we get 1000 of them consecutively
cg [Mon, 10 Jan 2000 06:19:20 +0000 (06:19 +0000)]
don't complain about bad intrs unless we get 1000 of them consecutively
whilst we are playing or recording.  since we should irq ~20 times/sec when
active, this should never trigger.  in theory.  if it never does trigger,
the check will be removed.

24 years agoAdd some comments about things that have become bad as I've been
imp [Mon, 10 Jan 2000 06:17:27 +0000 (06:17 +0000)]
Add some comments about things that have become bad as I've been
working on this code, or things that I think are layering violations.
Plus a minor whitespace cleanup in one place while I'm at it.

24 years ago- Revert --strip-all of /kernel on a new installed system (it would work
peter [Mon, 10 Jan 2000 06:15:43 +0000 (06:15 +0000)]
- Revert --strip-all of /kernel on a new installed system (it would work
with kld etc just fine, but tracebacks would have less information and
nm /kernel wouldn't be so good).
- Just strip the kernel on the boot disk. This does not affect kld or
module loading, there are two symbol tables in a kernel.  There is the
dynamic linking one (.dynsym+.strtab) with just global symbols and a user
symbol table (.symtab+.strtab) with all symbols.  BTW; objdump lies and
hides the second one.  There's a good half a meg or so that can be saved
from an average kernel by stripping it.

24 years agoAdd new resource flag type: RF_PCCARD_ATTR for pccard's attribute memory.
imp [Mon, 10 Jan 2000 06:13:52 +0000 (06:13 +0000)]
Add new resource flag type: RF_PCCARD_ATTR for pccard's attribute memory.
This was suggested by Doug a while ago.

24 years agoZap the IDEA stuff - it's patented internationally (at least in some
kris [Mon, 10 Jan 2000 05:36:35 +0000 (05:36 +0000)]
Zap the IDEA stuff - it's patented internationally (at least in some
places), and we don't want people to get in trouble just for having it.

24 years agoMake sched_param parameter a const to comply with POSIX and SUSv2 specs.
deischen [Mon, 10 Jan 2000 04:14:08 +0000 (04:14 +0000)]
Make sched_param parameter a const to comply with POSIX and SUSv2 specs.
This doesn't need to be applied to stable, because somehow -stable seems
to have gotten it right.

Reviewed by: jasone

24 years agoHandle the case where we truss an SUGID program -- in particular, we need
sef [Mon, 10 Jan 2000 04:09:05 +0000 (04:09 +0000)]
Handle the case where we truss an SUGID program -- in particular, we need
to wake up any processes waiting via PIOCWAIT on process exit, and truss
needs to be more aware that a process may actually disappear while it's
waiting.

Reviewed by: Paul Saab <ps@yahoo-inc.com>

24 years agogeneral tidyup
cg [Mon, 10 Jan 2000 03:22:28 +0000 (03:22 +0000)]
general tidyup

24 years agoMerge from OpenBSD:
imp [Mon, 10 Jan 2000 03:20:13 +0000 (03:20 +0000)]
Merge from OpenBSD:
o Realloc memory leak fixed which won't matter but would trigger purify
o Default to sendmail when no mailer.conf exists.

Fixed bugs in OpenBSD version:
o Add NULL termination in the right place.

Also put back the err.  free shouldn't touch errno.

Pointed out by: theo de raadt (except the NULL bug :-)

24 years agoAdd 3com OfficeConnect 572
imp [Mon, 10 Jan 2000 02:46:57 +0000 (02:46 +0000)]
Add 3com OfficeConnect 572

24 years agoCommit kernel part of patches to get the 3Com OfficeConnect 3CXSH572BT
imp [Mon, 10 Jan 2000 02:32:43 +0000 (02:32 +0000)]
Commit kernel part of patches to get the 3Com OfficeConnect 3CXSH572BT
working on -current.

Submitted by: Eric D. Futch <efutch@nyct.net>
Approved by: mdodd

24 years agoUncomment pcic device and put pccard in GENERIC. PCCARD will be removed
imp [Mon, 10 Jan 2000 02:29:23 +0000 (02:29 +0000)]
Uncomment pcic device and put pccard in GENERIC.  PCCARD will be removed
in a little while as soon as I find all the places it is used in the
tree.

24 years agoreturn the sample rate set instead of 0. oops. mpg123 should now work.
cg [Mon, 10 Jan 2000 01:59:12 +0000 (01:59 +0000)]
return the sample rate set instead of 0.  oops.  mpg123 should now work.

24 years agoAvoid setting DPARCKEN until I can figure out why it causes
gibbs [Mon, 10 Jan 2000 01:47:51 +0000 (01:47 +0000)]
Avoid setting DPARCKEN until I can figure out why it causes
spurious parity errors on some controllers.

24 years agoFix mysterious sendmail coredump on systems where malloc.conf pointed to
imp [Mon, 10 Jan 2000 01:25:53 +0000 (01:25 +0000)]
Fix mysterious sendmail coredump on systems where malloc.conf pointed to
a string containign 'J'.

o Properly terminate argv list with a NULL entry.
o Use warn() to report the exec failure because free could change errno and
  err would report the wrong reason.
o Don't terminate string to err with ':' since this results in two colons.

24 years agoSeveral performance improvements for soft updates have been added:
mckusick [Mon, 10 Jan 2000 00:24:24 +0000 (00:24 +0000)]
Several performance improvements for soft updates have been added:
1) Fastpath deletions. When a file is being deleted, check to see if it
   was so recently created that its inode has not yet been written to
   disk. If so, the delete can proceed to immediately free the inode.
2) Background writes: No file or block allocations can be done while the
   bitmap is being written to disk. To avoid these stalls, the bitmap is
   copied to another buffer which is written thus leaving the original
   available for futher allocations.
3) Link count tracking. Constantly track the difference in i_effnlink and
   i_nlink so that inodes that have had no change other than i_effnlink
   need not be written.
4) Identify buffers with rollback dependencies so that the buffer flushing
   daemon can choose to skip over them.

24 years agoAdd bwillwrite to all system calls that create things in the filesystem.
mckusick [Mon, 10 Jan 2000 00:08:53 +0000 (00:08 +0000)]
Add bwillwrite to all system calls that create things in the filesystem.
Benchmarks that create huge trees of empty files overwhelm the buffer cache.

24 years agoRemove the P_BUFEXHAUST flag from the syncer process (leaving
mckusick [Mon, 10 Jan 2000 00:07:24 +0000 (00:07 +0000)]
Remove the P_BUFEXHAUST flag from the syncer process (leaving
it only on the buf_daemon process). The problem is that when the
syncer process starts running the worklist, it wants to delete
lots of files. It does this by VFS_VGET'ing the vnodes, clearing
the blocks in them and bdwrite'ing the buffer. It can process close
to a thousand files per second which generates a large number of
dirty buffers. So, giving it special priviledge at the buffer trough
leads to trouble as the buf_daemon does occationally need a free
buffer to proceed and if the syncer has used every last one up,
we are toast.

24 years agoKeep tighter control of removal dependencies by limiting the number
mckusick [Sun, 9 Jan 2000 23:35:38 +0000 (23:35 +0000)]
Keep tighter control of removal dependencies by limiting the number
of dirrem structure rather than the collaterally created freeblks
and freefile structures. Limit the rate of buffer dirtying by the
syncer process during periods of intense file removal.

24 years ago* Add `sym' SCSI driver options.
obrien [Sun, 9 Jan 2000 23:33:38 +0000 (23:33 +0000)]
* Add `sym' SCSI driver options.
* bring in NIC comments from GENERIC
* slightly reorder a few things in an feable attempt at making the
  organization of LINT more logical.

24 years agoReorganize softdep_fsync so that it only does the inode-is-flushed
mckusick [Sun, 9 Jan 2000 23:14:57 +0000 (23:14 +0000)]
Reorganize softdep_fsync so that it only does the inode-is-flushed
check before the inode is unlocked while grabbing its parent directory.
Once it is unlocked, other operations may slip in that could make
the inode-is-flushed check fail. Allowing other writes to the inode
before returning from fsync does not break the semantics of fsync
since we have flushed everything that was dirty at the time of the
fsync call.

24 years agoGet rid of unreferenced function.
mckusick [Sun, 9 Jan 2000 22:42:42 +0000 (22:42 +0000)]
Get rid of unreferenced function.

24 years agoMake static non-exported functions from soft updates.
mckusick [Sun, 9 Jan 2000 22:40:09 +0000 (22:40 +0000)]
Make static non-exported functions from soft updates.

24 years agochange debug printout lefvels for a couple of places
mjacob [Sun, 9 Jan 2000 21:47:39 +0000 (21:47 +0000)]
change debug printout lefvels for a couple of places

24 years agoReally really remove SHA-1 support.
kris [Sun, 9 Jan 2000 21:22:48 +0000 (21:22 +0000)]
Really really remove SHA-1 support.

24 years agoRevamp the mechanism for enumerating and calling shared objects'
jdp [Sun, 9 Jan 2000 21:13:48 +0000 (21:13 +0000)]
Revamp the mechanism for enumerating and calling shared objects'
init and fini functions.  Now the code is very careful to hold no
locks when calling these functions.  Thus the dynamic linker cannot
be re-entered with a lock already held.

Remove the tolerance for recursive locking that I added in revision
1.2 of dllockinit.c.  Recursive locking shouldn't happen any more.

Mozilla and JDK users: I'd appreciate confirmation that things still
work right (or at least the same) with these changes.

24 years agoClose PR# 15986: issue an RX reset command when initializing the interface,
wpaul [Sun, 9 Jan 2000 21:12:59 +0000 (21:12 +0000)]
Close PR# 15986: issue an RX reset command when initializing the interface,
but only for those cards that don't use miibus (i.e. all the 10mbps only
cards, and the 100baseFX card).

PR: kern/15986

24 years agoRemove the warning that this interface shouldn't be used yet. Fix
jdp [Sun, 9 Jan 2000 21:01:39 +0000 (21:01 +0000)]
Remove the warning that this interface shouldn't be used yet.  Fix
a typo.  Clarify a sentence.

24 years agoThis commit was generated by cvs2svn to compensate for changes in r55682,
markm [Sun, 9 Jan 2000 20:58:00 +0000 (20:58 +0000)]
This commit was generated by cvs2svn to compensate for changes in r55682,
which included commits to RCS files with non-trunk default branches.

24 years agoImport KTH Heimdal, which will be the core of our Kerberos5.
markm [Sun, 9 Jan 2000 20:58:00 +0000 (20:58 +0000)]
Import KTH Heimdal, which will be the core of our Kerberos5.
Userland to follow.

24 years agoRemove the comment warning that the dllockinit() interface might
jdp [Sun, 9 Jan 2000 20:57:21 +0000 (20:57 +0000)]
Remove the comment warning that the dllockinit() interface might
change.  I have decided that the interface is general enough to
last.

24 years agoRemove; we don't use this file anymore.
marcel [Sun, 9 Jan 2000 20:22:35 +0000 (20:22 +0000)]
Remove; we don't use this file anymore.

24 years agotcp updates to support IPv6.
shin [Sun, 9 Jan 2000 19:17:30 +0000 (19:17 +0000)]
tcp updates to support IPv6.
also a small patch to sys/nfs/nfs_socket.c, as max_hdr size change.

Reviewed by: freebsd-arch, cvs-committers
Obtained from: KAME project

24 years agoAdd the buildkernel and installkernel targets.
marcel [Sun, 9 Jan 2000 18:17:48 +0000 (18:17 +0000)]
Add the buildkernel and installkernel targets.
While here fix some minor style bugs (whitespacing) and move the
make target from Makefile.upgrade to this file.
Simplify the make target to make it readable.