]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoDrop the references to index(3) and rindex(3), which are non-standard
Alexander Langer [Tue, 18 Jul 2000 08:05:11 +0000 (08:05 +0000)]
Drop the references to index(3) and rindex(3), which are non-standard
and people shouldn't be encouraged to use them.

Asked by: sheldonh

23 years agoClean up this new manual page. This delta includes content and
Sheldon Hearn [Tue, 18 Jul 2000 07:46:14 +0000 (07:46 +0000)]
Clean up this new manual page.  This delta includes content and
whitespace changes, which should not be a problem because this
is only the second revision of the file and translators are
unlikely to have gotten started yet.

Reviewed by: abial

23 years agoDon't take a failure to stat a file to mean that it doesn't exist, as it may
Dag-Erling Smørgrav [Tue, 18 Jul 2000 07:12:26 +0000 (07:12 +0000)]
Don't take a failure to stat a file to mean that it doesn't exist, as it may
be a dynamically created file, e.g. a tarball on servers that support sending
tarballs of directories.

23 years agoAdd in some new IN_XXX and CT_XXXX flags in preparation
Matt Jacob [Tue, 18 Jul 2000 07:06:47 +0000 (07:06 +0000)]
Add in some new IN_XXX and CT_XXXX flags in preparation
for the rototilling that !*$)~@!$_@*_(~@$*_(~@$*~@$*
Qlogic F/W changes will need.

23 years agoIf debugging set, zero out an incoming response entry
Matt Jacob [Tue, 18 Jul 2000 07:05:37 +0000 (07:05 +0000)]
If debugging set, zero out an incoming response entry
when we're done reading it (makes checking things easier).
Before calling isp_notify_ack make sure we're at RUNSTATE-
elsewise we can be responding to LIPs or SCSI bus resets
before we've finished some of the wiring.

23 years agoThe SERVICING_INTERRUPT isn't quite safe yet.
Matt Jacob [Tue, 18 Jul 2000 07:04:07 +0000 (07:04 +0000)]
The SERVICING_INTERRUPT isn't quite safe yet.

23 years agoAdd a isp_target_putback_atio- we aren't using CCINCR at this time, so
Matt Jacob [Tue, 18 Jul 2000 06:58:28 +0000 (06:58 +0000)]
Add a isp_target_putback_atio- we aren't using CCINCR at this time, so
we need a function that tells the Qlogic f/w that a target mode command
is done, so increase the resource count for that lun. Add in a timeout
function to kick the putback again if we fail to do it the first time (we
may not have the request queue space for ATIO push). Split the function
isp_handle_platform_ctio into two parts so that the timeout function for
the ATIO push or isp_handle_platform_ctio can inform CAM that the requested
CTIO(s) are now done.

Clean up (cough) residual handling. What we need for Fibre Channel
is to preserve the at_datalen field from the original incoming ATIO
so we can calculate a 'true' residual.  Unfortunately, we're not
guaranteed to get that back from CAM. We'll *try* to find it hiding
in the periph_priv field (layering violation)- but if an ATIO was
passed in from user land- forget it. This means that we'll probably
get residuals wrong for Fibre Channel commands we're completing
with an error. It's too late to 4.1 release to fix this- too bad.
Luckily the only device we'd really care about this occurring on
is a tape device and they're still so rare as FC attached devices
that this can be considered an untested combination anyway.

Remove all CCINCR usage (resource autoreplenish). When we've proved
to ourself that things are working properly, we can add it back
in.

Make sure we propage 'suggested' sense data from the incoming ATIO
into the created system ATIO- and set sense_len appropriately.
Correctly propagate tag values.

Fall back to the model of generating (well, the functions in isp_pci.c
do the work) multiple CTIOs based upon what we get from XPT. Instead
of being able to pair Qlogic generated ATIOs with CAM ATIOs, and then
to pair CAM CTIOs with Qlogic CTIOs, we have to take the CTIO passed
to us from XPT, and if it implies that we have to generate extra
Qlogic CTIOs, so be it. This means that we have to wait until the
last CTIO in a sequence we generated completes before calling xpt_done.

Executive summary- target mode actually now pretty much works well
enough to tell folks about.

23 years agoRaise debug level for some messages. Fix botched inversion
Matt Jacob [Tue, 18 Jul 2000 06:46:48 +0000 (06:46 +0000)]
Raise debug level for some messages. Fix botched inversion
about MBOX_COMMAND_ERROR vs. MBOX_COMMAND_PARAM_ERROR.

23 years agoKeep interrupts blocked for all of isp_pci_attach. Redo DMA routines
Matt Jacob [Tue, 18 Jul 2000 06:40:22 +0000 (06:40 +0000)]
Keep interrupts blocked for all of isp_pci_attach. Redo DMA routines
for target mode for cleanliness and accuracy.

23 years agoProperly detect ISA cards in EISA mode and skip them in the ISA identify
Matthew N. Dodd [Tue, 18 Jul 2000 06:37:08 +0000 (06:37 +0000)]
Properly detect ISA cards in EISA mode and skip them in the ISA identify
routine so that they will be picked up by the EISA front end.

PR: i386/2598

23 years agoPatch up some bogons in the resource_find() vs resource_find_hard()
Peter Wemm [Tue, 18 Jul 2000 06:08:27 +0000 (06:08 +0000)]
Patch up some bogons in the resource_find() vs resource_find_hard()
interfaces.  The original resource_find() returned a pointer to an internal
resource table entry.  resource_find_hard() dereferences the actual
passed in value (oops!) - effectively trashing random memory due to
the pointer being passed in with a random initial value.

Submitted by:  bde

23 years agoDon't get stuck in a loop calling exit from an atexit routine. Clean
Matt Jacob [Tue, 18 Jul 2000 04:39:36 +0000 (04:39 +0000)]
Don't get stuck in a loop calling exit from an atexit routine. Clean
up cam_fill_ctio usage to passed atio flags. Clear periph_priv area
of new ctio so if the kernel is dumb enough to look at them (this is
a SECURITY hole) the panic will be obvious instead of subtle.

23 years agoDeclare our DEC Alpha cdboot to be a fully released version 1.0.
David E. O'Brien [Tue, 18 Jul 2000 04:15:06 +0000 (04:15 +0000)]
Declare our DEC Alpha cdboot to be a fully released version 1.0.

23 years agoMove mtree to bootstrap-tools and add PATH=${TMPPATH} to IMAKEENV to pick it
Andrey A. Chernov [Tue, 18 Jul 2000 01:49:05 +0000 (01:49 +0000)]
Move mtree to bootstrap-tools and add PATH=${TMPPATH} to IMAKEENV to pick it

23 years agoChange my email address in the copyright notices for the sake of consistency
Jason Evans [Tue, 18 Jul 2000 01:38:19 +0000 (01:38 +0000)]
Change my email address in the copyright notices for the sake of consistency
(jasone@canonware.com --> jasone@freebsd.org).

23 years agoAdd mtree to cross-tools to make it updated for new -L option
Andrey A. Chernov [Tue, 18 Jul 2000 00:21:05 +0000 (00:21 +0000)]
Add mtree to cross-tools to make it updated for new -L option
It must solve make world breakage

23 years agoDocument the builtin echo command
Mark Ovens [Mon, 17 Jul 2000 23:37:55 +0000 (23:37 +0000)]
Document the builtin echo command

Reviewed by: Matthew Hunt <mph@astro.caltech.edu>

23 years agoInitialise ifnet::if_type
Brian Somers [Mon, 17 Jul 2000 23:21:42 +0000 (23:21 +0000)]
Initialise ifnet::if_type

PR: 17873
Submitted by: Kensaku Masuda <greg@greg.rim.or.jp>

23 years agoDeal correctly with statically initialized condition variables in
Jason Evans [Mon, 17 Jul 2000 22:55:05 +0000 (22:55 +0000)]
Deal correctly with statically initialized condition variables in
pthread_cond_signal(), pthread_cond_broadcast(), and pthread_cond_timedwait().

Do not dump core in pthread_cond_timedwait() (due to a NULL pointer
dereference) if attempting to wait on an uninitialized condition variable.

PR: bin/18099

23 years agoThe description of the semantics of -v and -q has long since been OBE.
Dag-Erling Smørgrav [Mon, 17 Jul 2000 22:44:36 +0000 (22:44 +0000)]
The description of the semantics of -v and -q has long since been OBE.

23 years agoWhile I'm at it, break a line that was too long, remove a pointless diagnostic
Dag-Erling Smørgrav [Mon, 17 Jul 2000 22:44:00 +0000 (22:44 +0000)]
While I'm at it, break a line that was too long, remove a pointless diagnostic
and adjust the verbosity level of another.

23 years agoReshuffle the SEE ALSO section.
Jason Evans [Mon, 17 Jul 2000 22:33:32 +0000 (22:33 +0000)]
Reshuffle the SEE ALSO section.

Prompted by: sheldonh

23 years agoDon't reply "not a plain file" when the requested file doesn't exist.
Dag-Erling Smørgrav [Mon, 17 Jul 2000 22:24:52 +0000 (22:24 +0000)]
Don't reply "not a plain file" when the requested file doesn't exist.

23 years agoDocument the fact that -r and -m are mutually exclusive.
Dag-Erling Smørgrav [Mon, 17 Jul 2000 21:51:48 +0000 (21:51 +0000)]
Document the fact that -r and -m are mutually exclusive.

23 years agoRearrange / rewrite large portions of fetch() to take advantage of new
Dag-Erling Smørgrav [Mon, 17 Jul 2000 21:49:01 +0000 (21:49 +0000)]
Rearrange / rewrite large portions of fetch() to take advantage of new
libfetch features (fetchRestartCalls, fetchXGet()).

Since it doesn't make much sense to have m_flag and r_flag set at the same
time, and it can actually cause trouble in some cases, die if they're both
set.

Set the SA_RESETHAND flag for SIGINT so that when we've caught one, we can
kill ourselves with a second SIGINT (thus notifying our parent of our tragic
fate) instead of just exiting.

These changes fix several problems that would show up when fetching ports,
as well as speeding up HTTP transfers quite a bit (at least for relatively
small files).

Most of these changes were prompted by an interaction problem with an HTTP
server called SWS-1.0, which exhibited two bugs, the first of which prevented
fetch from working around the second (the first was not sending content-type
in reply to HEAD requests, the second was sending garbage after the end of
the requested file).

23 years agoIntroduce fetchXGet*(), which combine the functionalities of fetchGet*() and
Dag-Erling Smørgrav [Mon, 17 Jul 2000 21:25:00 +0000 (21:25 +0000)]
Introduce fetchXGet*(), which combine the functionalities of fetchGet*() and
fetchStat*().  In most cases, either fetchGet*() or fetchXGet*() is a wrapper
around the other; in all cases, calling fetchGet*() is identical to calling
fetchXGet*() with the second argument set to NULL.

23 years agoDon't forget to set file descriptor to -1 after closing it, since the code
Dag-Erling Smørgrav [Mon, 17 Jul 2000 21:09:11 +0000 (21:09 +0000)]
Don't forget to set file descriptor to -1 after closing it, since the code
outside the loop inspects it to determine whether or not we succeeded in
retrieving the requested document. This fixes a bug where fetchGetHTTP()
would return a FILE with an invalid file descriptor if it hit the redirect
limit without locating the requested document.

23 years agoUse a symbolic constant instead of hardcoding 999
Dag-Erling Smørgrav [Mon, 17 Jul 2000 21:00:49 +0000 (21:00 +0000)]
Use a symbolic constant instead of hardcoding 999

23 years agoDon't forget to declare fetchRestartCalls.
Dag-Erling Smørgrav [Mon, 17 Jul 2000 20:52:06 +0000 (20:52 +0000)]
Don't forget to declare fetchRestartCalls.

23 years agoIntroduce the (undocumented) variable fetchRestartCalls, which controls whether
Dag-Erling Smørgrav [Mon, 17 Jul 2000 20:49:39 +0000 (20:49 +0000)]
Introduce the (undocumented) variable fetchRestartCalls, which controls whether
or not interrupted system calls will be restarted. This fixes a bug where
fetch(1) would hang (potentially forever) if a server stopped responding,
because the signal handler would absorb the user's efforts to interrupt the
transfer.

23 years agoRemove fetchContentType(), which disappeared in the HTTP rewrite.
Dag-Erling Smørgrav [Mon, 17 Jul 2000 20:42:13 +0000 (20:42 +0000)]
Remove fetchContentType(), which disappeared in the HTTP rewrite.

23 years agoclose PR 19544 - ipfw pipe delete causes panic when no pipes defined
Luigi Rizzo [Mon, 17 Jul 2000 20:03:27 +0000 (20:03 +0000)]
close PR 19544 - ipfw pipe delete causes panic when no pipes defined

PR: 19544

23 years ago- Don't try to free mboot.bootinst before it has been allocated. If, for
John Baldwin [Mon, 17 Jul 2000 19:51:42 +0000 (19:51 +0000)]
- Don't try to free mboot.bootinst before it has been allocated.  If, for
  some reason, mboot.bootinst is not initialized to NULL at the beginning
  of the program, then the last commit to this would try to free whatever
  bogus address is in it.
- Restore the behavior of free()'ing the mboot.bootinst buffer after we
  abuse it to determine the sector size of the disk (as clearly noted in
  the comments).  Properly fix the double free() bug by setting the pointer
  to NULL after we free it.

23 years agoOops. Remove some debugging code used to display a hexdump of the sector
John Baldwin [Mon, 17 Jul 2000 19:39:50 +0000 (19:39 +0000)]
Oops.  Remove some debugging code used to display a hexdump of the sector
we just loaded from the disk.  The code to call it was commented out, and
it easily fit into the extra sector, but remove it anyway as it is
unneeded.

23 years agoThe modem driver for Communication Device Class, ACM compliant devices.
Nick Hibma [Mon, 17 Jul 2000 19:28:29 +0000 (19:28 +0000)]
The modem driver for Communication Device Class, ACM compliant devices.
There is a number of devices that are compliant, of which the 3Com 5605 is
has been verified to work.

The driver is not perfect yet, but should be able to get you somewhere.

The driver was originally written by Lennart Augustsson, but Mike Smith
and Mike Meyer <mwm@mired.org> did the porting.

23 years agoAdd the PCI IDs for the Macronix 98727 and 98732 parts. These are
Bill Paul [Mon, 17 Jul 2000 19:27:41 +0000 (19:27 +0000)]
Add the PCI IDs for the Macronix 98727 and 98732 parts. These are
3.3volt PCI/cardbus chipsets similar to the 98715 (and they have
512-bit hash tables). Also update the man page to mention the 98727/98732
and the SOHOware SFA110A Rev B4 card with the 98715AEC-C chip.

23 years agoRemove some whitespace so the line with "brackets" changed to "angle
Ben Smithurst [Mon, 17 Jul 2000 19:05:57 +0000 (19:05 +0000)]
Remove some whitespace so the line with "brackets" changed to "angle
brackets" no longer touches the right edge of an 80 column display.

23 years agokue(4) and cue(4) do not need miibus.
Bill Paul [Mon, 17 Jul 2000 19:04:58 +0000 (19:04 +0000)]
kue(4) and cue(4) do not need miibus.

23 years agoUnbreak the build -- no manual page for this yet.
Jonathan Lemon [Mon, 17 Jul 2000 19:03:58 +0000 (19:03 +0000)]
Unbreak the build --  no manual page for this yet.

23 years agoAdd SEE ALSO section, as with the other string functions.
Alexander Langer [Mon, 17 Jul 2000 19:00:26 +0000 (19:00 +0000)]
Add SEE ALSO section, as with the other string functions.

Submitted by: dcs

23 years agoRemove __P prototypes to reduce diffs between the NetBSD and FreeBSD
Nick Hibma [Mon, 17 Jul 2000 18:41:20 +0000 (18:41 +0000)]
Remove __P prototypes to reduce diffs between the NetBSD and FreeBSD
versions.

23 years agoUse pwd instead of cwd in backticks. cwd in backticks doesn't do
Sheldon Hearn [Mon, 17 Jul 2000 17:31:59 +0000 (17:31 +0000)]
Use pwd instead of cwd in backticks.  cwd in backticks doesn't do
anything useful in a Bourne shell.

PR: 19980
Submitted by: Alex Kapranoff <alex@kapran.bitmcnit.bryansk.su>

23 years agoFix a bug which could cause programs with user threads packages to
John Polstra [Mon, 17 Jul 2000 17:18:13 +0000 (17:18 +0000)]
Fix a bug which could cause programs with user threads packages to
lock against themselves, causing infinite spinning.  Brian Feldman
found this problem when testing with Mozilla and supplied the fix,
which I have revised slightly.

Here is the failure scenario.  A thread calls dlopen() and acquires
the writer lock.  While the thread still holds the lock, a signal
is delivered and caught.  The signal handler tries to call a function
which hasn't been bound yet.  It thus enters the dynamic linker
and tries to acquire the reader lock.  Since the writer lock is
already held, it will spin forever in the signal handler.  The
thread holding the lock won't be able to progress and release the
lock.

The solution is to block almost all signals while holding the
exclusive lock.

A similar problem could conceivably occur in the opposite order.
Namely, a thread is holding the reader lock and then a signal
handler calls dlopen() or dlclose() and spins waiting for the writer
lock.  We deal with this administratively by proclaiming that signal
handlers aren't allowed to call dlopen() or dlclose().  Actually
we don't have to proclaim a thing, since signal handlers aren't
allowed to call any system functions except those which are explicitly
permitted.

Submitted by: Brian Fundakowski Feldman <green>

23 years agoHook up liloldr
Jonathan Lemon [Mon, 17 Jul 2000 17:06:27 +0000 (17:06 +0000)]
Hook up liloldr

23 years agoThis is `liloboot', which creates a file which can be treated like a
Jonathan Lemon [Mon, 17 Jul 2000 17:06:04 +0000 (17:06 +0000)]
This is `liloboot', which creates a file which can be treated like a
Linux kernel image, and is designed to be dropped into a Linux system
and booted via LILO.  Once booted, the user is greeted by the FreeBSD
loader.  This still isn't quite complete, as the the root= specification
from LILO isn't currently passed to the loader yet.

23 years agoImprove on previous commit:
Sheldon Hearn [Mon, 17 Jul 2000 13:39:48 +0000 (13:39 +0000)]
Improve on previous commit:

Don't inhibit the trailing newline for entropy-related messages.
Try harder to save the seed file on shutdown.

Reviewed by: markm

23 years agoRename MDNSECT to MD_NSECT and declare it as something that isn't
Sheldon Hearn [Mon, 17 Jul 2000 13:13:04 +0000 (13:13 +0000)]
Rename MDNSECT to MD_NSECT and declare it as something that isn't
default in NOTES.

Requested by: bde
Approved by: phk

23 years agoIf ipv6_enable is set to yes, do IPv6 setup for PCCARD ethernet
Hajimu UMEMOTO [Mon, 17 Jul 2000 12:33:57 +0000 (12:33 +0000)]
If ipv6_enable is set to yes, do IPv6 setup for PCCARD ethernet
card.  This is still at staring point and end node case only.

23 years agoAdd entropy caching. With this, some entropy is cached at shutdown
Mark Murray [Mon, 17 Jul 2000 12:28:58 +0000 (12:28 +0000)]
Add entropy caching. With this, some entropy is cached at shutdown
time, and this is used to reseed the random number generator at
boot time.

NOTE - this has no hope of working if you halt(); you need to
execute rc.shutdown to get the entropy stash.

23 years agoAdd randomness write functionality. This does absolutely nothing for
Mark Murray [Mon, 17 Jul 2000 12:23:04 +0000 (12:23 +0000)]
Add randomness write functionality. This does absolutely nothing for
entropy estimation, but causes an immediate reseed after the input
(read in sizeof(u_int64_t) chunks) is "harvested".

This will be used in the reboot "reseeder", coming in another
commit. This can be used very effectively at any time you think
your randomness is compromised; something like

# (ps -gauxwww; netstat -an; dmesg; vmstat -c10 1) > /dev/random

will give the attacker something to think about.

23 years agoCorrect the usage() message as per rev 1.25 of mtree.8 .
Sheldon Hearn [Mon, 17 Jul 2000 11:00:55 +0000 (11:00 +0000)]
Correct the usage() message as per rev 1.25 of mtree.8 .

23 years ago * Order options in the SYNOPSIS correctly.
Sheldon Hearn [Mon, 17 Jul 2000 10:59:38 +0000 (10:59 +0000)]
   * Order options in the SYNOPSIS correctly.
   * Re-order the list of options in teh DESCRIPTION as per
the SYNOPSIS.
   * Move the description of exit conditions from the
DESCRIPTION section to a new DIAGNOSTICS section.
   * Typo fix: "effect" -> "affect" when used as a verb.
   * Clear the Nm macro as appropriate.
   * Typo fix: "consider" -> "considers" for a singular subject.
   * Use Nx instead of NetBSD.

23 years agoLocale support was added by ache long ago.
Sheldon Hearn [Mon, 17 Jul 2000 10:40:48 +0000 (10:40 +0000)]
Locale support was added by ache long ago.

23 years agoSuggest looking at rc.conf(5) on how to start natd(8) during boot.
Ruslan Ermilov [Mon, 17 Jul 2000 10:06:54 +0000 (10:06 +0000)]
Suggest looking at rc.conf(5) on how to start natd(8) during boot.

Submitted by: dcs

23 years agoRegen.
Nick Hibma [Mon, 17 Jul 2000 09:56:01 +0000 (09:56 +0000)]
Regen.

23 years agoLots of new device Ids. From NetBSD.
Nick Hibma [Mon, 17 Jul 2000 09:55:34 +0000 (09:55 +0000)]
Lots of new device Ids. From NetBSD.

23 years agoAdd descriptive start/stop time information to make release. Now
Jordan K. Hubbard [Mon, 17 Jul 2000 04:50:28 +0000 (04:50 +0000)]
Add descriptive start/stop time information to make release.  Now
you can just ``grep >>> release.log'' and get all the highlights.

23 years agoFix a paste-o in the tcpoptions check (not a security problem, just a
Bill Fumerola [Mon, 17 Jul 2000 03:02:15 +0000 (03:02 +0000)]
Fix a paste-o in the tcpoptions check (not a security problem, just a
error in the usage printf())

Reviewed by: rwatson

23 years agosync with latest kame doc. wording changes and updates
Jun-ichiro itojun Hagino [Mon, 17 Jul 2000 02:37:26 +0000 (02:37 +0000)]
sync with latest kame doc.  wording changes and updates

23 years agobring in latest kame doc. talk about ah tunnel caveat.
Jun-ichiro itojun Hagino [Mon, 17 Jul 2000 02:22:18 +0000 (02:22 +0000)]
bring in latest kame doc.  talk about ah tunnel caveat.

23 years agoClean up after oneself on exit.
Matt Jacob [Mon, 17 Jul 2000 02:05:45 +0000 (02:05 +0000)]
Clean up after oneself on exit.

23 years agoPreserve CAM_DIS_DISCONNECT as passed up from SIM (like
Matt Jacob [Mon, 17 Jul 2000 01:45:51 +0000 (01:45 +0000)]
Preserve CAM_DIS_DISCONNECT as passed up from SIM (like
CAM_TAG_ACTION_VALID and CAM_DIR_MASK). Remove redundant
CAM_DEBUG line. Spiff up CAM_DEBUG printout for commands
and move the printout up to the top where we can see it,
even for the pending_ua/pending_ca cass. Add missing
newline in a CAM_DEBUG.

23 years agoWhoops- forgot to commit this other pearl from Justin- only set or
Matt Jacob [Mon, 17 Jul 2000 00:43:47 +0000 (00:43 +0000)]
Whoops- forgot to commit this other pearl from Justin- only set or
clear CAM_TAG_ACTION_VALID if this is an XPT_SCSI_IO CCB (otherwise,
the peripheral driver knows best...)..
Obtained from:gibbs@freebsd.org

23 years agoRoll back target mode f/w to 7.55- 7.65 hangs when presented with a
Matt Jacob [Mon, 17 Jul 2000 00:34:52 +0000 (00:34 +0000)]
Roll back target mode f/w to 7.55- 7.65 hangs when presented with a
non-disconnecting command. Interestingly enough, of the other flavors
of the 7.65 f/w (the dual-id and multi-id flavor)- the dual-id doesn't
hang (they're also supposed to be the same except for supporting dual
or multi-id capture!), but other things are questionable as well.

23 years agoImplement pread and pwrite.
Marcel Moolenaar [Mon, 17 Jul 2000 00:17:07 +0000 (00:17 +0000)]
Implement pread and pwrite.

PR: 17991
Submitted by: Geoffrey Speicher <geoff@caribbean.sea-incorporated.com>

23 years agoAdd prototypes for linux_pread and linux_pwrite.
Marcel Moolenaar [Mon, 17 Jul 2000 00:13:38 +0000 (00:13 +0000)]
Add prototypes for linux_pread and linux_pwrite.

PR: 17991
Submitted by: Geoffrey Speicher <geoff@caribbean.sea-incorporated.com>

23 years agoDon't try to make files immutable (ie. chflags) before setting access times.
David Malone [Sun, 16 Jul 2000 23:22:15 +0000 (23:22 +0000)]
Don't try to make files immutable (ie. chflags) before setting access times.

PR: 19973
Submitted by: Arjan de Vet <Arjan.deVet@adv.iae.nl>
Reviewed by: Matthew Jacob <mjacob@feral.com>

23 years ago#ifdef DEBUG -> #ifndef NDEBUG
Dag-Erling Smørgrav [Sun, 16 Jul 2000 23:18:44 +0000 (23:18 +0000)]
#ifdef DEBUG -> #ifndef NDEBUG

Pointed out by: ache

23 years agoImplement setfsuid and setfsgid. Implementation derived from patch
Marcel Moolenaar [Sun, 16 Jul 2000 21:23:34 +0000 (21:23 +0000)]
Implement setfsuid and setfsgid. Implementation derived from patch
in PR.

PR: 16993
Submitted by: Bjoern Groenvall <bg@sics.se>

23 years ago1. "braces" -> "brackets" when referring to [ and ].
Ben Smithurst [Sun, 16 Jul 2000 20:46:43 +0000 (20:46 +0000)]
1.  "braces" -> "brackets" when referring to [ and ].

PR: 19894
Submitted by: Tony Finch <dot@dotat.at>

2.   "brackets" -> "angle brackets" when referring to < and >.

3.  Clean up the bit about creating the usage() message.  After clarifying a
couple of points the sentence became rather long, and rather poor English, so
it was converted to a enumerated list instead.

parts 1, 2, 3:
Reviewed by: sheldonh

23 years agoClarify that CVSup doesn't understand comments in refuse files.
Ben Smithurst [Sun, 16 Jul 2000 20:34:54 +0000 (20:34 +0000)]
Clarify that CVSup doesn't understand comments in refuse files.

PR: 19949
Submitted by: Jun Ushida <ushida@msa.biglobe.ne.jp>

23 years agoAdd some detection code for the InSystem USB cable (ATAPI support not yet
Nick Hibma [Sun, 16 Jul 2000 17:40:23 +0000 (17:40 +0000)]
Add some detection code for the InSystem USB cable (ATAPI support not yet
there, so the cable does not work yet)

23 years agoPCCard entry cleanup:
Mitsuru IWASAKI [Sun, 16 Jul 2000 17:33:54 +0000 (17:33 +0000)]
PCCard entry cleanup:
 - Remove unit numbers in config lines.
 - Remove all of logger lines and add logstr lines for some cards.  This
   changes reduced file size from 84k to 45k.
 - Use '/sbin/ifconfig $device delete' instead of /etc/pccard_ether_remove
   which haven't merge from PAO yet.
 - Cosmetic changes.

23 years agoAnd another buffer overflow. Maybe next time I should read the manpage for
Nick Hibma [Sun, 16 Jul 2000 17:28:01 +0000 (17:28 +0000)]
And another buffer overflow. Maybe next time I should read the manpage for
strlen.

This one only occurs if there is exactly one element on the line without any
whitespace. This is however never a valid line, so not a big chance that
this would ever cause any problems.

23 years ago[Merge from PAO]
MIHIRA Sanpei Yoshiro [Sun, 16 Jul 2000 14:33:47 +0000 (14:33 +0000)]
[Merge from PAO]
add about 67 entry and rewrite 18 entry
change Xircom CompactCard Ethernet 10 entry
config 0x20 -> auto

Reviewed by: iwasaki
Obtained from: PAO3

23 years agoWITH_IDEA --> MAKE_IDEA fix.
Mark Murray [Sun, 16 Jul 2000 12:20:28 +0000 (12:20 +0000)]
WITH_IDEA --> MAKE_IDEA fix.

23 years agoDocument Ethernet card reqs for SRM boot. Note that 8255x (fxp)
Wilko Bulte [Sun, 16 Jul 2000 11:38:01 +0000 (11:38 +0000)]
Document Ethernet card reqs for SRM boot. Note that 8255x (fxp)
does not currently work correctly on alpha.

23 years agoBump __FreeBSD_version to indicate mtree defaults changed back to original
Andrey A. Chernov [Sun, 16 Jul 2000 10:02:34 +0000 (10:02 +0000)]
Bump __FreeBSD_version to indicate mtree defaults changed back to original

23 years agoDon't break the ability to debug the kernel when gathering entropy.
Mark Murray [Sun, 16 Jul 2000 09:25:04 +0000 (09:25 +0000)]
Don't break the ability to debug the kernel when gathering entropy.

Pointed out by: bde

23 years agoAdd -L to mtree to preserve old functionality exact
Andrey A. Chernov [Sun, 16 Jul 2000 08:58:02 +0000 (08:58 +0000)]
Add -L to mtree to preserve old functionality exact

23 years agoAdd -L to mtree.
Andrey A. Chernov [Sun, 16 Jul 2000 08:44:40 +0000 (08:44 +0000)]
Add -L to mtree.
Not shure ever it needed for MFS, but just to keep old behaviour exact, can be
removed later

23 years agoFix getopt string I forget
Andrey A. Chernov [Sun, 16 Jul 2000 08:22:38 +0000 (08:22 +0000)]
Fix getopt string I forget

23 years agoForget to add -L to one mtree call
Andrey A. Chernov [Sun, 16 Jul 2000 08:20:54 +0000 (08:20 +0000)]
Forget to add -L to one mtree call

23 years agoAdd -L to mtree
Andrey A. Chernov [Sun, 16 Jul 2000 08:19:49 +0000 (08:19 +0000)]
Add -L to mtree

23 years agoAdd -L to mtree calls since default changed back
Andrey A. Chernov [Sun, 16 Jul 2000 08:02:48 +0000 (08:02 +0000)]
Add -L to mtree calls since default changed back

23 years agoAdd -L to mtree calls since defaults changed back
Andrey A. Chernov [Sun, 16 Jul 2000 07:58:25 +0000 (07:58 +0000)]
Add -L to mtree calls since defaults changed back

23 years agos/IPSEC_IPV6FWD/IPSEC/. this avoids unexpected behavior on ipv6 fowarding.
Jun-ichiro itojun Hagino [Sun, 16 Jul 2000 07:56:54 +0000 (07:56 +0000)]
s/IPSEC_IPV6FWD/IPSEC/.  this avoids unexpected behavior on ipv6 fowarding.
(even if you ask for tunnel-mode encryption packets will go out in clear)
sync with kame.

23 years agoRevert back to original mtree default behaviour to not follow symlinks, to be
Andrey A. Chernov [Sun, 16 Jul 2000 07:55:07 +0000 (07:55 +0000)]
Revert back to original mtree default behaviour to not follow symlinks, to be
compatible with other *BSD camp.  Add -L option to follow symlinks, so remove
-P option which is now default. The next step will be to add -L to building
process.

Asked-by: bde
23 years agoAdd FreeBSD Id to make this easier to track.
David E. O'Brien [Sun, 16 Jul 2000 06:33:13 +0000 (06:33 +0000)]
Add FreeBSD Id to make this easier to track.

23 years agoSqueeze the BOOTMFS kernel image some more. The Alpha kernel was simply
David E. O'Brien [Sun, 16 Jul 2000 06:32:28 +0000 (06:32 +0000)]
Squeeze the BOOTMFS kernel image some more.  The Alpha kernel was simply
over flowing its britches.  So remove all ppbus bits except those for PLIP
(untested), and all USB bits as SRM does not know what USB is.  Also remove
/dev/random as I don't think we need it just for whacking bits onto a disk.

Approved by: JKH

23 years agoTake jdp's excellent message to -current about the internat crypto
Warner Losh [Sun, 16 Jul 2000 06:00:40 +0000 (06:00 +0000)]
Take jdp's excellent message to -current about the internat crypto
files and put it in my freebsd web area.  Rewrite the entry in
UPDATING to point to it the message and hopefully not confuse people
like the old one did.

23 years agoForced commit. This is to try and help folks that used the international
Peter Wemm [Sun, 16 Jul 2000 05:53:14 +0000 (05:53 +0000)]
Forced commit.  This is to try and help folks that used the international
crypto repo and have slightly different files but with the same version.
cvsup in 'checkout mode' has no trouble with this, but cvs can get really
silly about it.

23 years agoAdd missing $FreeBSD$ to files that are NOT still on vendor a branch.
Peter Wemm [Sun, 16 Jul 2000 05:48:49 +0000 (05:48 +0000)]
Add missing $FreeBSD$ to files that are NOT still on vendor a branch.

23 years agoGet rid of non-existent CAM 'scan' device.
Kenneth D. Merry [Sun, 16 Jul 2000 02:49:23 +0000 (02:49 +0000)]
Get rid of non-existent CAM 'scan' device.

PR: 19961
Submitted by: Cyrille Lefevre <clefevre@citeweb.net>

23 years agoimprove route/nd cache cleanup on interface removal.
Jun-ichiro itojun Hagino [Sun, 16 Jul 2000 01:46:42 +0000 (01:46 +0000)]
improve route/nd cache cleanup on interface removal.
CAVEAT: haven't really tested it yet, please report

23 years agoCatch up to the fact that we now have tcsh in the tree.
David Nugent [Sun, 16 Jul 2000 01:46:26 +0000 (01:46 +0000)]
Catch up to the fact that we now have tcsh in the tree.

23 years agoReinsert the MIT copyright, which applies to the base64 code and was
Dag-Erling Smørgrav [Sun, 16 Jul 2000 01:04:10 +0000 (01:04 +0000)]
Reinsert the MIT copyright, which applies to the base64 code and was
left out by accident during the rewrite.

23 years agoBetter signal handling.
Dag-Erling Smørgrav [Sun, 16 Jul 2000 00:47:00 +0000 (00:47 +0000)]
Better signal handling.

Submitted by:   green

23 years agoSimplify the F_GETOWN and F_SETOWN fcntl commands. The workaround
Marcel Moolenaar [Sat, 15 Jul 2000 22:33:24 +0000 (22:33 +0000)]
Simplify the F_GETOWN and F_SETOWN fcntl commands. The workaround
is not needed since the FreeBSD native implementation switched
from TIOC{G|S}PGRP to FIO{G|S}ETOWN (kern_descrip.c rev 1.55).

PR: 16946
Submitted by: Victor Salaman <salaman@teknos.com>

23 years ago"VFS entry point" -> "entry point", to be consistent with other VOP_*(9)
Ben Smithurst [Sat, 15 Jul 2000 22:28:16 +0000 (22:28 +0000)]
"VFS entry point" -> "entry point", to be consistent with other VOP_*(9)
manual pages.

PR: 18593
Submitted by: Anatoly Vorobey <mellon@pobox.com>