]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoImplement signals.
Jake Burkholder [Mon, 3 Sep 2001 22:41:40 +0000 (22:41 +0000)]
Implement signals.

23 years agoMove %ver definitions from pstate.h to ver.h. Add definitions for normal
Jake Burkholder [Mon, 3 Sep 2001 22:36:11 +0000 (22:36 +0000)]
Move %ver definitions from pstate.h to ver.h.  Add definitions for normal
kernel pstate values, which include a memory store order override.

23 years agoAdd simple macros for tracing in assembler files. There are quite
Jake Burkholder [Mon, 3 Sep 2001 22:32:49 +0000 (22:32 +0000)]
Add simple macros for tracing in assembler files.  There are quite
a few places where we cannot even call a function, and these have
proven to be very useful debugging tools for such situations.

23 years agoUse the correct copyrights. Note where most of this came from.
Jake Burkholder [Mon, 3 Sep 2001 22:27:23 +0000 (22:27 +0000)]
Use the correct copyrights.  Note where most of this came from.

Requested by: obrien

23 years agoBump UPAGES to 4. The pcb can be rather large.
Jake Burkholder [Mon, 3 Sep 2001 22:19:36 +0000 (22:19 +0000)]
Bump UPAGES to 4.  The pcb can be rather large.

23 years agomtx_savecrit is a pil level, not a pstate value, thus mtx_intr_enable
Jake Burkholder [Mon, 3 Sep 2001 22:19:04 +0000 (22:19 +0000)]
mtx_savecrit is a pil level, not a pstate value, thus mtx_intr_enable
was not doing its thing.

23 years agoAdd a flushw() macro.
Jake Burkholder [Mon, 3 Sep 2001 22:13:53 +0000 (22:13 +0000)]
Add a flushw() macro.

23 years agoAdd atomic_load and store functions without membars, fwiw.
Jake Burkholder [Mon, 3 Sep 2001 22:03:25 +0000 (22:03 +0000)]
Add atomic_load and store functions without membars, fwiw.

23 years agoThe definition for ASI_IMMU_TAG_TARGET_REG was wrong. Sort.
Jake Burkholder [Mon, 3 Sep 2001 22:02:15 +0000 (22:02 +0000)]
The definition for ASI_IMMU_TAG_TARGET_REG was wrong.  Sort.

23 years agoAllow disabling of "arp moved" messages.
Alfred Perlstein [Mon, 3 Sep 2001 21:53:15 +0000 (21:53 +0000)]
Allow disabling of "arp moved" messages.

Submitted by: Stephen Hurd <deuce@lordlegacy.org>

23 years agoI really hope this is the right answer.
Julian Elischer [Mon, 3 Sep 2001 21:07:31 +0000 (21:07 +0000)]
I really hope this is the right answer.
call ip_input directly but take the offset off the
packet first if it's an IPV4 packet encapsulated.

23 years agoCall ip_input() instead of ipip_input()
Julian Elischer [Mon, 3 Sep 2001 20:55:35 +0000 (20:55 +0000)]
Call ip_input() instead of ipip_input()
when decoding encapsulated ipv4 packets.
(allows line to compile again)

23 years agoOne caller of rip_input failed to be converted in the last commit.
Julian Elischer [Mon, 3 Sep 2001 20:40:35 +0000 (20:40 +0000)]
One caller of rip_input failed to be converted in the last commit.

23 years agoPrevious change to try to build `mkisofs' was
David E. O'Brien [Mon, 3 Sep 2001 20:30:01 +0000 (20:30 +0000)]
Previous change to try to build `mkisofs' was
Submmitted by: dirk

23 years agoTry to build `mkisofs' first, and only `pkg_add -r' if you cannot.
David E. O'Brien [Mon, 3 Sep 2001 20:29:21 +0000 (20:29 +0000)]
Try to build `mkisofs' first, and only `pkg_add -r' if you cannot.

23 years agoCause an error condition to be sensed by make(1) if we have trouble
David E. O'Brien [Mon, 3 Sep 2001 20:27:03 +0000 (20:27 +0000)]
Cause an error condition to be sensed by make(1) if we have trouble
making the ISO.

23 years agoFix my backwards logic.
David E. O'Brien [Mon, 3 Sep 2001 20:23:59 +0000 (20:23 +0000)]
Fix my backwards logic.

Submitted by: dirk

23 years agoPatches from Keiichi SHIMA <keiichi@iij.ad.jp>
Julian Elischer [Mon, 3 Sep 2001 20:03:55 +0000 (20:03 +0000)]
Patches from Keiichi SHIMA <keiichi@iij.ad.jp>
to make ip use the standard protosw structure again.

Obtained from: Well, KAME I guess.

23 years agoOpps, let one definition slip out that we define dynamically.
David E. O'Brien [Mon, 3 Sep 2001 19:19:12 +0000 (19:19 +0000)]
Opps, let one definition slip out that we define dynamically.

23 years agoThis does not describe sched_{get,set}scheduler.
Dima Dorfman [Mon, 3 Sep 2001 17:55:08 +0000 (17:55 +0000)]
This does not describe sched_{get,set}scheduler.

PR: 26001
Submitted by: OHSAWA Chitoshi <ohsawa@catv1.ccn-net.ne.jp>

23 years agoDon't capitalize jail(2) in the middle of a sentence.
Dima Dorfman [Mon, 3 Sep 2001 17:49:29 +0000 (17:49 +0000)]
Don't capitalize jail(2) in the middle of a sentence.

PR: 25876
Submitted by: Koizumi Satoru <koizumi@cms.phys.s.u-tokyo.ac.jp>

23 years agoVFS_ABORTOP hasn't existed since 1999/12/15.
Dima Dorfman [Mon, 3 Sep 2001 16:56:21 +0000 (16:56 +0000)]
VFS_ABORTOP hasn't existed since 1999/12/15.

23 years agoat_shutdown(9) hasn't existed since revision 1.61 of kern_shutdown.c.
Dima Dorfman [Mon, 3 Sep 2001 16:42:35 +0000 (16:42 +0000)]
at_shutdown(9) hasn't existed since revision 1.61 of kern_shutdown.c.

23 years agoIPFilter source code in contrib/ipfilter apparently can't make up its mind
Jonathan Lemon [Mon, 3 Sep 2001 16:37:16 +0000 (16:37 +0000)]
IPFilter source code in contrib/ipfilter apparently can't make up its mind
where the headers should live, as the code references both "ip_fil.h" and
"netinet/ip_fil.h" (among others).  As a consequence, put both
sys/contrib/ipfilter and sys/contrib/ipfilter/netinet to the include path
so either variant works.

PR: 29384
Pointed out by: Thomas.Quinot@Cuivre.FR.EU.ORG

23 years agoActually, instead of removing the comment entirely, change it to talk
Dima Dorfman [Mon, 3 Sep 2001 16:07:14 +0000 (16:07 +0000)]
Actually, instead of removing the comment entirely, change it to talk
about {vid,kbd}map, which is probably what the author meant.

Submitted by: Gary W. Swearingen <swear@aa.net>

23 years agoThis does not describe kbdcontrol or vidcontrol.
Dima Dorfman [Mon, 3 Sep 2001 16:04:59 +0000 (16:04 +0000)]
This does not describe kbdcontrol or vidcontrol.

PR: 29505

23 years agosyslogd can now be configured to bind to a specific address.
Dima Dorfman [Mon, 3 Sep 2001 15:42:10 +0000 (15:42 +0000)]
syslogd can now be configured to bind to a specific address.

23 years agoIf NO_LANGCODE_IN_DESTDIR is defined, don't put the language code in
Dima Dorfman [Mon, 3 Sep 2001 14:39:32 +0000 (14:39 +0000)]
If NO_LANGCODE_IN_DESTDIR is defined, don't put the language code in
DESTDIR.  This avoids redundant information in the path when DOCDIR
points to some directory that already implies (or specifies) a certain
language.  This is the case with the web site, where the release notes
are already installed under a language-specific directory.  This
behavior is not being made mandatory because it might still be useful
to install all the translations in one directory, such as during
testing, or in a hypothetical release notes archive.  Furthermore, it
is not being made the default because that breaks consistency with
stuff under doc/.

Reviewed by: bmah

23 years agoExtend the release note about ftpd -o to cover ftpd -O, too.
Dima Dorfman [Mon, 3 Sep 2001 14:25:59 +0000 (14:25 +0000)]
Extend the release note about ftpd -o to cover ftpd -O, too.

23 years ago`list' should be an extern'd `char **', not a local `char *' which we
Dima Dorfman [Mon, 3 Sep 2001 14:19:46 +0000 (14:19 +0000)]
`list' should be an extern'd `char **', not a local `char *' which we
never set.  Ideally, we'd get the extern from tutor.h, but that
defines a number of other variables that conflict with ours.

This fixes a segmentation fault when trying to return to the main menu.

PR: 30172

23 years agoRemove bogus implementation of _setjmp/_longjmp
Doug Rabson [Mon, 3 Sep 2001 14:19:02 +0000 (14:19 +0000)]
Remove bogus implementation of _setjmp/_longjmp

23 years agoFor new users, create the home directory before sending the welcome
Dima Dorfman [Mon, 3 Sep 2001 14:12:42 +0000 (14:12 +0000)]
For new users, create the home directory before sending the welcome
mail, if configured to do so.  Some sites have setups where the user's
mail is delivered to their home directory, so sending mail before is
exists didn't work.

PR: 29892

23 years agoAdd a working version of setjmp/longjmp.
Doug Rabson [Mon, 3 Sep 2001 13:54:50 +0000 (13:54 +0000)]
Add a working version of setjmp/longjmp.

Obtained from: Intel's EFI toolkit.

23 years agoAllow group network to read ppp.log & slip.log.
Brian Somers [Mon, 3 Sep 2001 11:35:17 +0000 (11:35 +0000)]
Allow group network to read ppp.log & slip.log.

Suggested by: Jesper Skriver <jesper@FreeBSD.org>

23 years agoAdd more information to the 'device timeout' message in the
Murray Stokely [Mon, 3 Sep 2001 10:24:43 +0000 (10:24 +0000)]
Add more information to the 'device timeout' message in the
diagnostics section.

PR: docs/30163

23 years agoTwo characters in the devices name in the SYNOPSIS were transposed.
Murray Stokely [Mon, 3 Sep 2001 10:06:26 +0000 (10:06 +0000)]
Two characters in the devices name in the SYNOPSIS were transposed.

  matcdl0c -> matcd0lc

PR:       docs/25000
Submitted by: Rich Morin <rdm@cfcl.com>

23 years agoAdd a new field, id_config_attr to the struct isa_device.
Kazutaka YOKOTA [Mon, 3 Sep 2001 09:45:41 +0000 (09:45 +0000)]
Add a new field, id_config_attr to the struct isa_device.
It will be later used to store some flags to control PnP device
configurations.

23 years agoPT_STEP in ptrace(2) man page is described as 'addr and data fields
Murray Stokely [Mon, 3 Sep 2001 09:42:40 +0000 (09:42 +0000)]
PT_STEP in ptrace(2) man page is described as 'addr and data fields
are not used'.  This is incorrect, as addr must be passed (caddr_t)1
to do anything useful.  The source for gdb and a short test program
will confirm that this man page was in error.

PR: docs/27758
Submitted by: Jiangyi Liu <jyliu@163.net>

23 years agoAdd recognition for ToPIC95B
Warner Losh [Mon, 3 Sep 2001 08:20:52 +0000 (08:20 +0000)]
Add recognition for ToPIC95B

23 years agoSince we're cross compiling from x86, ignore the x86 CPUTYPE by default.
Peter Wemm [Mon, 3 Sep 2001 07:58:32 +0000 (07:58 +0000)]
Since we're cross compiling from x86, ignore the x86 CPUTYPE by default.

23 years agoSigh. There are actually two ioctls ranges that need to be watched on the
Scott Long [Mon, 3 Sep 2001 07:50:14 +0000 (07:50 +0000)]
Sigh.  There are actually two ioctls ranges that need to be watched on the
linux side.  This will all be over soon...

23 years agoNew release note: am-utils 6.0.7.
Bruce A. Mah [Mon, 3 Sep 2001 06:30:17 +0000 (06:30 +0000)]
New release note:  am-utils 6.0.7.

23 years agoCheck for malloc failure in a couple of cases
Kris Kennaway [Mon, 3 Sep 2001 05:57:06 +0000 (05:57 +0000)]
Check for malloc failure in a couple of cases

MFC after: 2 weeks

23 years agoFile positions are off_t nowdays, not long, so:
Andrey A. Chernov [Mon, 3 Sep 2001 05:06:48 +0000 (05:06 +0000)]
File positions are off_t nowdays, not long, so:
    atoi -> strtoll
    fseek -> fseeko

    NOTE: that fseek not works for >long offsets per POSIX:

    [EOVERFLOW] For fseek( ), the resulting file offset would be a value which
    cannot be represented correctly in an object of type long.

Fix minor cast too.

23 years agoAttach to a CD device even when the SCSI status is 'busy'.
Kenneth D. Merry [Mon, 3 Sep 2001 05:04:47 +0000 (05:04 +0000)]
Attach to a CD device even when the SCSI status is 'busy'.

Reported by: Thomas Quinot <thomas@cuivre.fr.eu.org>
MFC after: 3 weeks

23 years agoDont conflict with sysctl debug.mddebug
Peter Wemm [Mon, 3 Sep 2001 04:49:19 +0000 (04:49 +0000)]
Dont conflict with sysctl debug.mddebug

23 years agoUnindent a if (1) { that was left behind in the last commit.
Peter Wemm [Mon, 3 Sep 2001 04:39:38 +0000 (04:39 +0000)]
Unindent a if (1) { that was left behind in the last commit.
(commits were seperated to not obscure the real change)

23 years agoArgh. Make the ia64 kernel work in all situations. For some reason,
Peter Wemm [Mon, 3 Sep 2001 04:37:55 +0000 (04:37 +0000)]
Argh. Make the ia64 kernel work in all situations.  For some reason,
and I still dont know why, this was not failing on the non-kse kernel.
It certainly should have since things were using linker_kernel_file
unconditionally.  This has highlighted a different problem though that
means that trying to do a kldload on a non-dynamic kernel will implode.

23 years agoFile positions are off_t nowdays, not long, so:
Andrey A. Chernov [Mon, 3 Sep 2001 04:30:46 +0000 (04:30 +0000)]
File positions are off_t nowdays, not long, so:
    fseek -> fseeko
    ftell -> ftello
    fseek(x, 0L, 0) -> rewind(x)

    NOTE: that fseek/ftell not works for >long offsets per POSIX:

    [EOVERFLOW] For fseek( ), the resulting file offset would be a value which
    cannot be represented correctly in an object of type long.

    [EOVERFLOW] For ftell ( ), the current file offset cannot be represented
    correctly in an object of type long.

23 years agoFix the argument specifier for the PnP BIOS function 2
Kazutaka YOKOTA [Mon, 3 Sep 2001 03:43:39 +0000 (03:43 +0000)]
Fix the argument specifier for the PnP BIOS function 2
(PNP_SET_DEVNODE). The second argument is not a segment:offset
pointer, but a 16 bit short.

MFC after: 4 weeks

23 years agoMatch the declaration in net/netisr.h.
David E. O'Brien [Mon, 3 Sep 2001 03:24:31 +0000 (03:24 +0000)]
Match the declaration in net/netisr.h.

Submitted by: gcc 3.0.1

23 years agoBecause we now store SCCLUN capabilities in firmware attributes, get
Matt Jacob [Mon, 3 Sep 2001 03:12:10 +0000 (03:12 +0000)]
Because we now store SCCLUN capabilities in firmware attributes, get
rid of the silly test of isp_maxluns > 16 and use the attibutes directly.

MFC after: 4 weeks

23 years agoClarify issues about whether we have SCCLUN (65535 luns) or non-SCCLUN (16
Matt Jacob [Mon, 3 Sep 2001 03:09:48 +0000 (03:09 +0000)]
Clarify issues about whether we have SCCLUN (65535 luns) or non-SCCLUN (16
luns) firmware for the Fibre Channel cards.

We used to assume that if we didn't download firmware, we couldn't know
what the firmware capability with respect to SCCLUNs is- and it's important
because the lun field changes in the request queue entry based upon which
firmware it is.

At any rate, we *do* get back firmware attributes in mailbox register 6
when we do ABOUT FIRMWARE for all 2200/2300 cards- and for 2100 cards
with at least 1.17.0 firmware. So- we now assume non-SCCLUN behaviour
for 2100 cards with firmware < 1.17.0- and we check the firmware attributes
for other cards (loaded firmware or not).

This also allows us to get rid of the crappy test of isp_maxluns > 16-
we simply can check firmware attributes for SCCLUN behaviour.

This required an 'oops' fix to the outgoing mailbox count field for
ABOUT FIRMWARE for FC cards.

Also- while here, hardwire firmware revisions for loaded code for SBus
cards. Apparently the 1.35 or 1.37 f/w we've been loading into isp1000
just doesn't report firmware revisions out to mailbox regs 1, 2 and 3
like everyone else. Grumble. Not that this fix hardly matters for FreeBSD.

MFC after: 4 weeks

23 years agoAdd some more firmware revision macros. Add firmware attributes field
Matt Jacob [Mon, 3 Sep 2001 03:03:32 +0000 (03:03 +0000)]
Add some more firmware revision macros. Add firmware attributes field
to fcparam structure.
MFC after: 4 weeks

23 years agoWhen __SOPT is cleared, clear __SOFF too.
Andrey A. Chernov [Mon, 3 Sep 2001 02:35:10 +0000 (02:35 +0000)]
When __SOPT is cleared, clear __SOFF too.

NOTE: original stdio bug.

23 years agoRe-arrange my funopen(3) fix to minimize differences with original stdio code,
Andrey A. Chernov [Mon, 3 Sep 2001 02:24:37 +0000 (02:24 +0000)]
Re-arrange my funopen(3) fix to minimize differences with original stdio code,
no functional changes.

Add fp->_offset optimization in _SAPP+_SOPT case

23 years agoReturn updated blocksize from setblocksize function.
Orion Hodson [Mon, 3 Sep 2001 02:14:55 +0000 (02:14 +0000)]
Return updated blocksize from setblocksize function.

23 years agoFix return value of cs4281chan_setblocksize.
Orion Hodson [Mon, 3 Sep 2001 01:05:04 +0000 (01:05 +0000)]
Fix return value of cs4281chan_setblocksize.

23 years agoFix return value of svchan_setblocksize.
Orion Hodson [Mon, 3 Sep 2001 01:03:50 +0000 (01:03 +0000)]
Fix return value of svchan_setblocksize.

Expand probing range.

23 years agoFix return value of setblocksize functions. Recording is interspersed
Orion Hodson [Mon, 3 Sep 2001 00:45:00 +0000 (00:45 +0000)]
Fix return value of setblocksize functions.  Recording is interspersed
with silent intervals otherwise.

23 years ago+ Blah, there was nothing wrong in rev 1.1 talking about the i386/NOTES.
David E. O'Brien [Sun, 2 Sep 2001 23:48:37 +0000 (23:48 +0000)]
+ Blah, there was nothing wrong in rev 1.1 talking about the i386/NOTES.
  I should have diff'ed the header with the Alpha GENERIC.
+ fix style nit
+ turn on NO_MODULES for now.

23 years agoMatch the style of very other platform we have.
David E. O'Brien [Sun, 2 Sep 2001 23:37:45 +0000 (23:37 +0000)]
Match the style of very other platform we have.

23 years agoo Add sample syslogd_flags for "-ss" which causes syslogd not to bind
Robert Watson [Sun, 2 Sep 2001 23:34:19 +0000 (23:34 +0000)]
o Add sample syslogd_flags for "-ss" which causes syslogd not to bind
  an inet socket.

23 years agoWork around a bug where the driver's copy of the disklabel gets corrupted
Scott Long [Sun, 2 Sep 2001 23:16:40 +0000 (23:16 +0000)]
Work around a bug where the driver's copy of the disklabel gets corrupted
and causes diskerr() to panic.

23 years agoRestrict the range of linux ioctls we will accept.
Scott Long [Sun, 2 Sep 2001 23:15:26 +0000 (23:15 +0000)]
Restrict the range of linux ioctls we will accept.

23 years agoo Sync up prototypes for cap_size() and cap_copy_ext() with
Robert Watson [Sun, 2 Sep 2001 23:13:49 +0000 (23:13 +0000)]
o Sync up prototypes for cap_size() and cap_copy_ext() with
  sys/capability.h--this compiled fine on i386 where (int) and (ssize_t)
  are the same, but broke on Alpha where they differ.

Submitted by: Mike Barcroft <mike@FreeBSD.org>
Obtained from: TrustedBSD Project

23 years agoInternal seeks are overoptimized. They should remember fp->_offset only for
Andrey A. Chernov [Sun, 2 Sep 2001 21:22:00 +0000 (21:22 +0000)]
Internal seeks are overoptimized. They should remember fp->_offset only for
plain regular files, i.e. files with __SOPT flag set. Fix it, so ftell(stdout)
always returns the same as lseek(1, 0, 1) now.

NOTE: this bug was in original stdio code

23 years agoAdd quirk entry for FujiFilm camera.
Nick Hibma [Sun, 2 Sep 2001 20:44:49 +0000 (20:44 +0000)]
Add quirk entry for FujiFilm camera.

Submitted by: Guido van Rooij <guido@gvr.org>

23 years agoUse the native mount and nfs_prot headers.
David E. O'Brien [Sun, 2 Sep 2001 20:37:36 +0000 (20:37 +0000)]
Use the native mount and nfs_prot headers.
Also generate nfs_prot_x.c in ANSI-C mode.

23 years agoadd another prototype and a couple of stopgaps for the in_protosw variant.
Julian Elischer [Sun, 2 Sep 2001 20:36:19 +0000 (20:36 +0000)]
add another prototype and a couple of stopgaps for the in_protosw variant.

23 years agoTODO is no longer part of Amd, and we do not need, nor use clnt_sperrno.c.
David E. O'Brien [Sun, 2 Sep 2001 20:15:46 +0000 (20:15 +0000)]
TODO is no longer part of Amd, and we do not need, nor use clnt_sperrno.c.

23 years agoTeak -- we were being a little over agressive.
David E. O'Brien [Sun, 2 Sep 2001 20:15:13 +0000 (20:15 +0000)]
Teak -- we were being a little over agressive.

23 years agoThis commit was generated by cvs2svn to compensate for changes in r82819,
David E. O'Brien [Sun, 2 Sep 2001 20:14:08 +0000 (20:14 +0000)]
This commit was generated by cvs2svn to compensate for changes in r82819,
which included commits to RCS files with non-trunk default branches.

23 years agoVirgin import of AMD (am-utils) v6.0.7
David E. O'Brien [Sun, 2 Sep 2001 20:14:08 +0000 (20:14 +0000)]
Virgin import of AMD (am-utils) v6.0.7

23 years agoFix bug in off_t overflow checking: if fp->_offset overflows, just remove
Andrey A. Chernov [Sun, 2 Sep 2001 19:52:09 +0000 (19:52 +0000)]
Fix bug in off_t overflow checking: if fp->_offset overflows, just remove
__SOFF flag (i.e. we don't have offset) instead of returning EOVERFLOW.
It allows again continious reading from non-stop stream.

23 years agoUpdate for version 6.0.7.
David E. O'Brien [Sun, 2 Sep 2001 19:50:47 +0000 (19:50 +0000)]
Update for version 6.0.7.

23 years agoWe can use the stock defintion of "MOUNT_NFS3" now.
David E. O'Brien [Sun, 2 Sep 2001 19:47:53 +0000 (19:47 +0000)]
We can use the stock defintion of "MOUNT_NFS3" now.

23 years agoReduce diff from vendor source.
David E. O'Brien [Sun, 2 Sep 2001 19:35:47 +0000 (19:35 +0000)]
Reduce diff from vendor source.

23 years agoUpdate for 6.0.7.
David E. O'Brien [Sun, 2 Sep 2001 19:32:45 +0000 (19:32 +0000)]
Update for 6.0.7.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 19:31:00 +0000 (19:31 +0000)]
We are returning to the stock (6.0.7) file now.
+ The bufov vulnerability was fixed in the vendor sources.
+ The vendor sources are now POSIX [me harder] compliant.

23 years agoRemove MLINKS seek.2 to lseek.2, we don't have seek.2 syscall
Andrey A. Chernov [Sun, 2 Sep 2001 19:28:41 +0000 (19:28 +0000)]
Remove MLINKS seek.2 to lseek.2, we don't have seek.2 syscall

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 19:28:19 +0000 (19:28 +0000)]
We are returning to the stock (6.0.7) file now.
+ The bufov vulnerability was fixed in the vendor sources.
+ The vendor sources are now POSIX [me harder] compliant.
+ The db vs. dbm files issue was fixed in the vendor sources.

23 years agoTypo seek(2) -> lseek(2)
Andrey A. Chernov [Sun, 2 Sep 2001 19:24:07 +0000 (19:24 +0000)]
Typo seek(2) -> lseek(2)

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 19:14:18 +0000 (19:14 +0000)]
We are returning to the stock (6.0.7) file now.
The security issues were fixed in the vendor sources.

23 years agoMove all stdio internal flags processing and setting out of __sread(),
Andrey A. Chernov [Sun, 2 Sep 2001 19:10:10 +0000 (19:10 +0000)]
Move all stdio internal flags processing and setting out of __sread(),
__swrite() and __sseek() to higher level. According to funopen(3) they all
are just wrappers to something like standard read(2), write(2) and
lseek(2), i.e. must not touch stdio internals because they are replaceable
with any other functions knows nothing about stdio internals. See example
of funopen(3) usage in sendmail sources f.e.

NOTE: this is original stdio bug, not result of my range checkin added.

23 years agoUse sh to execute scripts and just check that the script is readable.
Jordan K. Hubbard [Sun, 2 Sep 2001 19:05:27 +0000 (19:05 +0000)]
Use sh to execute scripts and just check that the script is readable.
I guess I pooched the permissions on the scripts before committing them
since they're not executable and now it's too late to change (I think -
I suppose you could chmod the ,v files and it might inheirit but I don't
feel like asking the repomeisters to try that).

Noticed by: Dirk Froemberg <dirk@freebsd.org>

23 years agoMerge rev 1.2 (use our native headers for NFSv3), 1.9 (fs header movment)
David E. O'Brien [Sun, 2 Sep 2001 19:04:09 +0000 (19:04 +0000)]
Merge rev 1.2 (use our native headers for NFSv3), 1.9 (fs header movment)
into Amd 6.0.7.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:46:35 +0000 (18:46 +0000)]
We are returning to the stock (6.0.7) file now.
+ The bufov vulnerability was fixed in the vendor sources.
+ The vendor sources are now POSIX [me harder] compliant.

23 years agoMerge our rev 1.2,3 (use or native files for NFSv3 support)
David E. O'Brien [Sun, 2 Sep 2001 18:39:17 +0000 (18:39 +0000)]
Merge our rev 1.2,3 (use or native files for NFSv3 support)
changes into Amd 6.0.7.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:26:32 +0000 (18:26 +0000)]
We are returning to the stock (6.0.7) file now.
I don't recall why the rev 1.2 hack is needed.  It looks like Amd was
already using our local headers.  I no longer have a FreeBSD 2.x box to
test this on, and this will never be MFC'ed to RELENG_2_2.  So lets just
decrease the maintenance effort.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:22:46 +0000 (18:22 +0000)]
We are returning to the stock (6.0.7) file now.
+ The rev 1.4 addition was taken from a snapshot previous to 6.0.7, so
  it is included in 6.0.7.
+ The vendor sources are now POSIX [me harder] compliant.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:20:22 +0000 (18:20 +0000)]
We are returning to the stock (6.0.7) file now.
My "nfsv2" mount option hack was added to the vendor sources.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:17:08 +0000 (18:17 +0000)]
We are returning to the stock (6.0.7) file now.
The vendor sources are now POSIX [me harder] compliant.

23 years agoWe are returning to the stock (6.0.7) file now.
David E. O'Brien [Sun, 2 Sep 2001 18:15:24 +0000 (18:15 +0000)]
We are returning to the stock (6.0.7) file now.
The bufov vulnerability was fixed in the vendor sources.

23 years agoMerge rev 1.7 (mount directory change) into Am-utils 6.0.7.
David E. O'Brien [Sun, 2 Sep 2001 18:13:17 +0000 (18:13 +0000)]
Merge rev 1.7 (mount directory change) into Am-utils 6.0.7.

23 years agoExtend the functionality offered by the -o option into a new option
Sheldon Hearn [Sun, 2 Sep 2001 17:24:19 +0000 (17:24 +0000)]
Extend the functionality offered by the -o option into a new option
-O, which limits the impact of the write-only restriction to guest
users.

*) The existing manual page's SYNOPSIS and option listing in the
   DESCRIPTION are already horribly disordered.  No attempt has been
   made to fix this.

*) The existing source's getopt() optstring and option handling switch
   are already horribly disordered.  No attempt has been made to fix
   this.

Discussed with: nik, -audit

23 years agoVirgin import of AMD (am-utils) v6.0.7
David E. O'Brien [Sun, 2 Sep 2001 17:19:13 +0000 (17:19 +0000)]
Virgin import of AMD (am-utils) v6.0.7

23 years agoThis commit was generated by cvs2svn to compensate for changes in r82794,
David E. O'Brien [Sun, 2 Sep 2001 17:19:13 +0000 (17:19 +0000)]
This commit was generated by cvs2svn to compensate for changes in r82794,
which included commits to RCS files with non-trunk default branches.

23 years agoFile positions are off_t nowdays, not long, so:
Andrey A. Chernov [Sun, 2 Sep 2001 14:40:51 +0000 (14:40 +0000)]
File positions are off_t nowdays, not long, so:
    fseek -> fseeko
    ftell -> ftello

    NOTE: fseek/ftell not works for >long offsets per POSIX:

    [EOVERFLOW] For fseek( ), the resulting file offset would be a value which
    cannot be represented correctly in an object of type long.

    [EOVERFLOW] For ftell ( ), the current file offset cannot be represented
    correctly in an object of type long.

23 years agolong -> off_t
Andrey A. Chernov [Sun, 2 Sep 2001 14:18:28 +0000 (14:18 +0000)]
long -> off_t
long -> time_t
%ld -> %qd
fseek -> fseeko

NOTE: that fseek not works for >long offsets per POSIX:

[EOVERFLOW] For fseek( ), the resulting file offset would be a value which
cannot be represented correctly in an object of type long.