]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
22 years agoAdd play (but not record) support for the Sis 7012.
alfred [Tue, 19 Feb 2002 00:59:23 +0000 (00:59 +0000)]
Add play (but not record) support for the Sis 7012.

Submitted by: Mike Meyer <mwm@mired.org>

22 years agoFixed divots that I created when I moved prototypes of group_from_gid
imp [Tue, 19 Feb 2002 00:05:59 +0000 (00:05 +0000)]
Fixed divots that I created when I moved prototypes of group_from_gid
and user_from_uid to grp.h and pwd.h.  Update the man pages.

Submitted by: David Malone
Pointy hat to: imp

22 years agoAdd emacs21 instead of emacs20 to the first CDROM.
murray [Mon, 18 Feb 2002 23:04:03 +0000 (23:04 +0000)]
Add emacs21 instead of emacs20 to the first CDROM.

Reviewed by:     nobutaka (editors/emacs21 maintainer)

22 years agoWhen the local link address is changed, send out gratuitous ARPs
luigi [Mon, 18 Feb 2002 22:50:13 +0000 (22:50 +0000)]
When the local link address is changed, send out gratuitous ARPs
to notify other nodes about the address change. Otherwise, they
might try and keep using the old address until their arp table
entry times out and the address is refreshed.

Maybe this ought to be done for INET6 addresses as well but i have
no idea how to do it. It should be pretty straightforward though.

MFC-after: 10 days

22 years agoSome release notes should have [MERGED] markers on them but didn't:
bmah [Mon, 18 Feb 2002 22:27:35 +0000 (22:27 +0000)]
Some release notes should have [MERGED] markers on them but didn't:
kern.msgbuf sysctl and dmesg(8) not SGID, umodem driver, uscanner
driver, fpa(4) works on alpha, arp(8) -s pub/only, TCP_COMPAT_42
removal, bktr(4) update to 2.18, snd module dependencies work,
chown(8) -R does the right thing with symlinks, killall(1) works
correctly with SUID root processes, strnstr(3)/strcasestr(3),
style.perl(7).

Remove some somewhat-lacking-in content release notes:  PCI subsystem
cleanup, netgraph(4) updates.

I don't mind reinstating the latter group of release notes, if there's
something less vague that can be said about them.

22 years agoParanoia: if the process is setugid, set all sensitive files mode 0.
des [Mon, 18 Feb 2002 21:41:11 +0000 (21:41 +0000)]
Paranoia: if the process is setugid, set all sensitive files mode 0.

22 years agoWhitespace and wordwrap fixes, no content or markup changes.
bmah [Mon, 18 Feb 2002 21:22:08 +0000 (21:22 +0000)]
Whitespace and wordwrap fixes, no content or markup changes.

22 years agoNew release notes: bridge(4) enhancements, dump(8) progress information
bmah [Mon, 18 Feb 2002 21:09:27 +0000 (21:09 +0000)]
New release notes:  bridge(4) enhancements, dump(8) progress information
in proctitle, ldd(1) -a, rtld(1) object dependency tracing, usbdevs(8) -a.

MFCs noted:  ldd(1) can be used on shlibs, BIND 8.3.1-REL, bzip-1.0.2.

22 years agoo Move NTOHL() and associated macros into <sys/param.h>. These are
mike [Mon, 18 Feb 2002 20:35:27 +0000 (20:35 +0000)]
o Move NTOHL() and associated macros into <sys/param.h>.  These are
  deprecated in favor of the POSIX-defined lowercase variants.
o Change all occurrences of NTOHL() and associated marcros in the
  source tree to use the lowercase function variants.
o Add missing license bits to sparc64's <machine/endian.h>.
  Approved by: jake
o Clean up <machine/endian.h> files.
o Remove unused __uint16_swap_uint32() from i386's <machine/endian.h>.
o Remove prototypes for non-existent bswapXX() functions.
o Include <machine/endian.h> in <arpa/inet.h> to define the
  POSIX-required ntohl() family of functions.
o Do similar things to expose the ntohl() family in libstand, <netinet/in.h>,
  and <sys/param.h>.
o Prepend underscores to the ntohl() family to help deal with
  complexities associated with having MD (asm and inline) versions, and
  having to prevent exposure of these functions in other headers that
  happen to make use of endian-specific defines.
o Create weak aliases to the canonical function name to help deal with
  third-party software forgetting to include an appropriate header.
o Remove some now unneeded pollution from <sys/types.h>.
o Add missing <arpa/inet.h> includes in userland.

Tested on: alpha, i386
Reviewed by: bde, jake, tmm

22 years agoLoad the current timecounter into tc. The timecounter global can change
dillon [Mon, 18 Feb 2002 19:49:30 +0000 (19:49 +0000)]
Load the current timecounter into tc.  The timecounter global can change
at any time and we do not want to call one timercounter's function with
another timecounter's structural pointer.

MFC after: 3 days

22 years agoAdd some more partition types.
joe [Mon, 18 Feb 2002 19:32:38 +0000 (19:32 +0000)]
Add some more partition types.

PR: i386/14793
MFC after: 3 days

22 years agoAdd net/samba to the list of important packages for Disc #1.
murray [Mon, 18 Feb 2002 19:04:23 +0000 (19:04 +0000)]
Add net/samba to the list of important packages for Disc #1.

Discussed with: steve, portmgr, re

22 years agoAdd kern_giant_ucred to instrument Giant around ucred related operations
dillon [Mon, 18 Feb 2002 17:51:47 +0000 (17:51 +0000)]
Add kern_giant_ucred to instrument Giant around ucred related operations
such a getgid(), setgid(), etc...

22 years agoFix typo.
sobomax [Mon, 18 Feb 2002 16:22:23 +0000 (16:22 +0000)]
Fix typo.

Approved by: phk

22 years agoMake v_addpollinfo() visible and non-inline.
phk [Mon, 18 Feb 2002 16:18:02 +0000 (16:18 +0000)]
Make v_addpollinfo() visible and non-inline.
Have callers only call it as needed.
Add necessary call in ufs_kqfilter().

Test-case found by: Andrew Gallatin <gallatin@cs.duke.edu>

22 years agoRehash of 1.43: simply remove the comment, since it's highly redundant
rwatson [Mon, 18 Feb 2002 16:02:24 +0000 (16:02 +0000)]
Rehash of 1.43: simply remove the comment, since it's highly redundant
and only partially correct.

22 years agoCorrect path to pucdata.c
jedgar [Mon, 18 Feb 2002 15:46:10 +0000 (15:46 +0000)]
Correct path to pucdata.c

Reviewed by: jhay

22 years agoFix kgdb in the face of the last round of KSE commits.
gallatin [Mon, 18 Feb 2002 14:13:59 +0000 (14:13 +0000)]
Fix kgdb in the face of the last round of KSE commits.
The alpha world may actually build now..

22 years agoAdd bus_space_unmap() and bus_space_free() functions to release
nyan [Mon, 18 Feb 2002 13:44:46 +0000 (13:44 +0000)]
Add bus_space_unmap() and bus_space_free() functions to release
a bus_space_handle allocated by bus_space_subregion().

22 years agoAdd stubs for bus_space_unmap() and bus_space_free(). They are needed to
nyan [Mon, 18 Feb 2002 13:43:19 +0000 (13:43 +0000)]
Add stubs for bus_space_unmap() and bus_space_free(). They are needed to
release a bus_space_handle allocated by bus_space_subregion().

22 years agoAdd NO_6_BYTE quirk entry for the LaCie Ltd. 105311 80 Gig USB2 drive.
bsd [Mon, 18 Feb 2002 13:35:30 +0000 (13:35 +0000)]
Add NO_6_BYTE quirk entry for the LaCie Ltd. 105311 80 Gig USB2 drive.

Submitted by: Brian Schellenberger <bts@babbleon.org>

22 years agoAdd the braces missed by revision 1.131.
iedowse [Mon, 18 Feb 2002 12:46:18 +0000 (12:46 +0000)]
Add the braces missed by revision 1.131.

Pointy hat to: rwatson

22 years agoAdd support for the Highpoint HPT372 based cards (rocketraid 133).
sos [Mon, 18 Feb 2002 11:57:56 +0000 (11:57 +0000)]
Add support for the Highpoint HPT372 based cards (rocketraid 133).

HW Sponsored by: Mike Tancsa

22 years agoFix the problem that some (in my book broken) disks reports
sos [Mon, 18 Feb 2002 11:52:51 +0000 (11:52 +0000)]
Fix the problem that some (in my book broken) disks reports
to be able to use 48bit addressing mode, but says the 48bit
size of the disk is 0, which according to spec means it can
address zero sectors in 48bit mode, why then say it supports
48bit mode at all..

22 years agoThe DEVFS paper presented at BSDcon-euro 2001 and BSDcon-2002.
phk [Mon, 18 Feb 2002 09:48:59 +0000 (09:48 +0000)]
The DEVFS paper presented at BSDcon-euro 2001 and BSDcon-2002.

22 years agoAdd some verbiage to the comments at the top of the PR template:
des [Mon, 18 Feb 2002 09:13:44 +0000 (09:13 +0000)]
Add some verbiage to the comments at the top of the PR template:

 - point at the FDP article rather than GNU's send-pr documentation
 - warn the user that PRs are public information and will be published in
   mailing lists and on the web
 - suggest that the user contact security-officer@ directly if the report
   concerns sensitive security issues.

22 years agoTake the common case of gettimeofday(&tv, NULL) out from under Giant.
phk [Mon, 18 Feb 2002 08:40:28 +0000 (08:40 +0000)]
Take the common case of gettimeofday(&tv, NULL) out from under Giant.

22 years agoRemove yet a redundant VN_KNOTE() macro.
phk [Mon, 18 Feb 2002 08:24:48 +0000 (08:24 +0000)]
Remove yet a redundant VN_KNOTE() macro.

22 years agoAdd notes about sendmail update from gshapiro.
imp [Mon, 18 Feb 2002 06:29:17 +0000 (06:29 +0000)]
Add notes about sendmail update from gshapiro.

22 years agostyle(9)
obrien [Mon, 18 Feb 2002 06:24:55 +0000 (06:24 +0000)]
style(9)

22 years agoRemove now-obsolete __STDC__ ifdefs, remove redundant NULL definition and
imp [Mon, 18 Feb 2002 06:08:23 +0000 (06:08 +0000)]
Remove now-obsolete __STDC__ ifdefs, remove redundant NULL definition and
include stdlib.h for NULL.

22 years agoThe ICANON flag is an lflag, not an iflag.
dillon [Mon, 18 Feb 2002 06:07:11 +0000 (06:07 +0000)]
The ICANON flag is an lflag, not an iflag.

Submitted by: Neelkanth Natu <neelnatu@yahoo.com>
MFC after: 3 days

22 years agoFix last second botch of adding $FreeBSD$.
imp [Mon, 18 Feb 2002 05:40:54 +0000 (05:40 +0000)]
Fix last second botch of adding $FreeBSD$.

22 years agoRemove __P() before committing new prototypes with KSE stuff
julian [Mon, 18 Feb 2002 05:26:57 +0000 (05:26 +0000)]
Remove __P() before committing new prototypes with KSE stuff
in a couple of weeks.

22 years agoRemove __P
imp [Mon, 18 Feb 2002 05:15:19 +0000 (05:15 +0000)]
Remove __P
Remove __STDC__ (which means we now use stdarg rather than vararg)
Remove register
Remove main prototype

22 years agoUse a more robust scheme for determining how many blocks to skip
iedowse [Mon, 18 Feb 2002 02:29:47 +0000 (02:29 +0000)]
Use a more robust scheme for determining how many blocks to skip
after an EOT-terminated volume. We keep track of the current record
number, and synchronise it with the c_tapea field each time we read
a header. Avoid the use of c_firstrec because some bugs in dump can
cause it to be set incorrectly.

Move the initialisation of some variables to avoid compiler warnings.

22 years agold-elf.so.1 assumed a few too many things about the ordering of sections
peter [Mon, 18 Feb 2002 02:24:10 +0000 (02:24 +0000)]
ld-elf.so.1 assumed a few too many things about the ordering of sections
produced by ld(8) (ie: that _DYNAMIC immediately follows the _GOT).
The new binutils import changed that, and the intial GOT relocation
broke.  Use a custom linker script to provide a real end-of-GOT symbol.

Update ld.so to deal with the new (faster) PLT format that gcc-3.1 and
binutils can produce.

This is probably incomplete, but appears to be working again.

Obtained from:  NetBSD
(And a fix to a silly mistake that I made by:  gallatin)

22 years agoFix mail.local build for non-sendmail.org code
gshapiro [Mon, 18 Feb 2002 02:18:42 +0000 (02:18 +0000)]
Fix mail.local build for non-sendmail.org code

22 years agoRemove a blank line.
luigi [Mon, 18 Feb 2002 01:59:40 +0000 (01:59 +0000)]
Remove a blank line.

22 years agoManpage update: comment existing functionalities and give
luigi [Mon, 18 Feb 2002 01:57:56 +0000 (01:57 +0000)]
Manpage update: comment existing functionalities and give
more detailed examples on how to use them.
Undocument deprecated functionalities which are going to be
removed soon.

22 years agoTwo license updates.
rwatson [Mon, 18 Feb 2002 01:28:01 +0000 (01:28 +0000)]
Two license updates.

22 years agoWhen we reach the end of the dump in findinode(), ask for another
iedowse [Mon, 18 Feb 2002 00:54:18 +0000 (00:54 +0000)]
When we reach the end of the dump in findinode(), ask for another
volume if we missed some earlier tapes (the user can still enter
'none' later if the tapes are unavailable). Previously with 'x'
restores, we might not ask for all tapes if the tapes are supplied
in reverse order.

Clarify the message that describes what volume should be mounted
first; reverse order is only efficient when extracting a few files.

22 years agoAdd a 'strvalid()' call to libkern. Given a character pointer, and
rwatson [Mon, 18 Feb 2002 00:37:03 +0000 (00:37 +0000)]
Add a 'strvalid()' call to libkern.  Given a character pointer, and
buffer length, determine if the pointer is to a valid string.  Currently,
the only check is whether a '\0' appears in the buffer.  This is useful
when pulling in a structure from userland that may contain one or more
strings, and validity testing must be performed on elements of the
structure.  When copying normal string arguments, copyinstr() is
expected to be used.

22 years agoWhen vn_open() is failing because it cannot allocate a vm object, call
rwatson [Mon, 18 Feb 2002 00:26:10 +0000 (00:26 +0000)]
When vn_open() is failing because it cannot allocate a vm object, call
VOP_CLOSE() on the vnode, so that VOP_OPEN() and VOP_CLOSE() calls
are symmetric in all failure cases.  This prevents an 'open' reference
from being leaked in that unlikely failure scenario.

22 years agostyle(9) prefers formatted comments in '/*' ... '*/' as opposed to
rwatson [Mon, 18 Feb 2002 00:23:44 +0000 (00:23 +0000)]
style(9) prefers formatted comments in '/*' ... '*/' as opposed to
#if 0'd.

22 years agoPer discussion at BSDCon, note that the vop_getattr locking protocol
rwatson [Mon, 18 Feb 2002 00:22:57 +0000 (00:22 +0000)]
Per discussion at BSDCon, note that the vop_getattr locking protocol
should require a shared lock, rather than an exclusive lock, which can
improve performance.  No actual code change here, since a number of
VFS locking fixes are in the works.

22 years agoLanguage clarification: s/an or/a/
rwatson [Mon, 18 Feb 2002 00:21:05 +0000 (00:21 +0000)]
Language clarification: s/an or/a/

22 years agoClarify word choice: s/those users/users/, since there is no previous
rwatson [Mon, 18 Feb 2002 00:20:12 +0000 (00:20 +0000)]
Clarify word choice: s/those users/users/, since there is no previous
instance of 'users' to refer to.

22 years agoMore for f/w crash dumps (bug fixing and adding ioctl entry points
mjacob [Mon, 18 Feb 2002 00:00:34 +0000 (00:00 +0000)]
More for f/w crash dumps (bug fixing and adding ioctl entry points
and hints to enable for specific units)

MFC after: 1 week

22 years agoUpdate sendmail release note to reflect 8.12.2 import.
gshapiro [Sun, 17 Feb 2002 23:58:31 +0000 (23:58 +0000)]
Update sendmail release note to reflect 8.12.2 import.
Change SUID -> set-user-ID

22 years agoTurn FEATURE(relay_based_on_MX) off by default. It should not be used unless
gshapiro [Sun, 17 Feb 2002 23:38:16 +0000 (23:38 +0000)]
Turn FEATURE(relay_based_on_MX) off by default.  It should not be used unless
absolutely necessary

Requested by: peter
PR: conf/33855
MFC after: 1 week

22 years agoThis commit was generated by cvs2svn to compensate for changes in r90809,
gshapiro [Sun, 17 Feb 2002 22:51:21 +0000 (22:51 +0000)]
This commit was generated by cvs2svn to compensate for changes in r90809,
which included commits to RCS files with non-trunk default branches.

22 years agoAdd infrastructure for sendmail 8.12. If users are not starting a daemon
gshapiro [Sun, 17 Feb 2002 22:19:14 +0000 (22:19 +0000)]
Add infrastructure for sendmail 8.12.  If users are not starting a daemon
at boot (sendmail_enable=NO), a localhost-only daemon may started
(sendmail_submit_enable) as it is needed to accept mail from command line
submissions.  If this isn't desired, see etc/mail/README for more hints.

Optionally (sendmail_msp_queue_enable) start a queue runner for the
submission queue in case a daemon isn't available to accept command line
submitted mail at submission time.

Note that the syslog labels for all of these sendmail processes have been
uniquified for easier log parsing.

22 years agoAdd information about how the new sendmail set-group-ID mail submission
gshapiro [Sun, 17 Feb 2002 22:14:29 +0000 (22:14 +0000)]
Add information about how the new sendmail set-group-ID mail submission
works and ways to work around common problems people might have.
Include information on reverting to a set-user-ID root sendmail binary in
case anyone really needs to do this.

22 years agoDon't build a submit.cf file if SENDMAIL_SET_USER_ID is set
gshapiro [Sun, 17 Feb 2002 22:12:57 +0000 (22:12 +0000)]
Don't build a submit.cf file if SENDMAIL_SET_USER_ID is set

22 years agoAdd /var/spool/clientmqueue for 8.12's non-set-user-ID root mail submission
gshapiro [Sun, 17 Feb 2002 22:11:41 +0000 (22:11 +0000)]
Add /var/spool/clientmqueue for 8.12's non-set-user-ID root mail submission

22 years agoAdd new include/libmilter directory for libmilter (sendmail mail filter API)
gshapiro [Sun, 17 Feb 2002 22:10:56 +0000 (22:10 +0000)]
Add new include/libmilter directory for libmilter (sendmail mail filter API)
include files

22 years agoDocument new build knob, SENDMAIL_SET_USER_ID, which installs sendmail as a
gshapiro [Sun, 17 Feb 2002 22:10:06 +0000 (22:10 +0000)]
Document new build knob, SENDMAIL_SET_USER_ID, which installs sendmail as a
set-user-ID root binary instead of the new method (set-group-ID smmsp).

22 years agoAdd new build knob, SENDMAIL_SET_USER_ID, which installs sendmail as a
gshapiro [Sun, 17 Feb 2002 22:09:15 +0000 (22:09 +0000)]
Add new build knob, SENDMAIL_SET_USER_ID, which installs sendmail as a
set-user-ID root binary instead of the new method (set-group-ID smmsp).
Therefore, we shouldn't install /etc/mail/submit.cf if it is set.

22 years agoUpdate for sendmail 8.12 which has a new OSTYPE(freebsd5)
gshapiro [Sun, 17 Feb 2002 22:08:18 +0000 (22:08 +0000)]
Update for sendmail 8.12 which has a new OSTYPE(freebsd5)
Fix access_db usage for 8.12

22 years agoUpdate import instructions for sendmail 8.12.2
gshapiro [Sun, 17 Feb 2002 22:06:57 +0000 (22:06 +0000)]
Update import instructions for sendmail 8.12.2

22 years agoUpdate build infrastructure for sendmail 8.12.
gshapiro [Sun, 17 Feb 2002 22:06:20 +0000 (22:06 +0000)]
Update build infrastructure for sendmail 8.12.
Add new build knob, SENDMAIL_SET_USER_ID, which installs sendmail as a
set-user-ID root binary instead of the new method (set-group-ID smmsp).

22 years agoUpdate build infrastructure for sendmail 8.12.
gshapiro [Sun, 17 Feb 2002 22:05:07 +0000 (22:05 +0000)]
Update build infrastructure for sendmail 8.12.

22 years agoAdd editmap, a new utility which comes with sendmail 8.12 for editing maps in
gshapiro [Sun, 17 Feb 2002 22:02:59 +0000 (22:02 +0000)]
Add editmap, a new utility which comes with sendmail 8.12 for editing maps in
place.

22 years agoAdd two new libraries which are part of sendmail 8.12. libsm is used only
gshapiro [Sun, 17 Feb 2002 22:01:40 +0000 (22:01 +0000)]
Add two new libraries which are part of sendmail 8.12.  libsm is used only
for building sendmail and the associated utilities.  libmilter is a new
mail filtering API for sendmail.

22 years agoResolve conflicts from sendmail 8.12.2 import
gshapiro [Sun, 17 Feb 2002 21:58:34 +0000 (21:58 +0000)]
Resolve conflicts from sendmail 8.12.2 import

22 years agoThis commit was generated by cvs2svn to compensate for changes in r90792,
gshapiro [Sun, 17 Feb 2002 21:56:45 +0000 (21:56 +0000)]
This commit was generated by cvs2svn to compensate for changes in r90792,
which included commits to RCS files with non-trunk default branches.

22 years agoImport sendmail 8.12.2
gshapiro [Sun, 17 Feb 2002 21:56:45 +0000 (21:56 +0000)]
Import sendmail 8.12.2

22 years agoMove the stuff related to select and poll out of struct vnode.
phk [Sun, 17 Feb 2002 21:15:36 +0000 (21:15 +0000)]
Move the stuff related to select and poll out of struct vnode.
The use of the zone allocator may or may not be overkill.
There is an XXX: over in ufs/ufs/ufs_vnops.c that jlemon may need
to revisit.

This shaves about 60 bytes of struct vnode which on my laptop means
600k less RAM used for vnodes.

22 years agoCollect the VN_KNOTE() macro definitions on vnode.h
phk [Sun, 17 Feb 2002 21:07:57 +0000 (21:07 +0000)]
Collect the VN_KNOTE() macro definitions on vnode.h

22 years agoAlways build all modules for LINT
phk [Sun, 17 Feb 2002 21:00:20 +0000 (21:00 +0000)]
Always build all modules for LINT

22 years agoEnsure err is a useful value to prevent using fputs() with a NULL stream.
jedgar [Sun, 17 Feb 2002 20:53:56 +0000 (20:53 +0000)]
Ensure err is a useful value to prevent using fputs() with a NULL stream.
MFC after: 1 week

22 years agov_lease is unused, zap it.
phk [Sun, 17 Feb 2002 20:41:27 +0000 (20:41 +0000)]
v_lease is unused, zap it.

22 years agoRemove cache_purgeleafdirs(), it has been #if 0 for quite some time.
phk [Sun, 17 Feb 2002 20:40:29 +0000 (20:40 +0000)]
Remove cache_purgeleafdirs(), it has been #if 0 for quite some time.

22 years agoDon't even think about using v_id for magic tricks, v_id is giving
phk [Sun, 17 Feb 2002 20:39:42 +0000 (20:39 +0000)]
Don't even think about using v_id for magic tricks, v_id is giving
us enough trouble as it is for SMPng.

22 years agoAdd Proxim RangeLAN-DS.
imp [Sun, 17 Feb 2002 20:05:39 +0000 (20:05 +0000)]
Add Proxim RangeLAN-DS.

Submitted by: Matt Peterson <matt@peterson.org>
PR: 35057

Also update my note for the 3crwe737A after talking to Alan Clegg at BSDcon.

22 years agoo style and consistency fixes:
jedgar [Sun, 17 Feb 2002 20:05:20 +0000 (20:05 +0000)]
o style and consistency fixes:
  - if (!var) -> if (var == NULL)
  - return val; -> return (val);
o update copyright

22 years agoOops, forgot to commit this (s/ucontext_t/struct __ucontext/).
deischen [Sun, 17 Feb 2002 19:21:42 +0000 (19:21 +0000)]
Oops, forgot to commit this (s/ucontext_t/struct __ucontext/).

22 years agoo __P removal
imp [Sun, 17 Feb 2002 19:09:20 +0000 (19:09 +0000)]
o __P removal
o use Ansi-style function definitions

22 years agoRegenerate these files after change to syscalls.master.
deischen [Sun, 17 Feb 2002 17:42:47 +0000 (17:42 +0000)]
Regenerate these files after change to syscalls.master.

22 years agoFix prototype to sigreturn to use struct __ucontext instead of ucontext_t.
deischen [Sun, 17 Feb 2002 17:41:28 +0000 (17:41 +0000)]
Fix prototype to sigreturn to use struct __ucontext instead of ucontext_t.

22 years agoUse struct __ucontext in prototypes and associated functions instead of
deischen [Sun, 17 Feb 2002 17:40:34 +0000 (17:40 +0000)]
Use struct __ucontext in prototypes and associated functions instead of
ucontext_t.  Forward declare struct __ucontext in <sys/signal.h> and
remove reliance on <sys/ucontext.h> being included.

While I'm here, also hide osigcontext types from userland; suggested
by bde.

Namespace pollution noticed by: Kevin Day <toasty@shell.dragondata.com>

22 years agoError handling fixes for inflate.
jedgar [Sun, 17 Feb 2002 17:35:18 +0000 (17:35 +0000)]
Error handling fixes for inflate.

22 years agoDon't rely on <sys/signal.h> to include <sys/ucontext.h>
deischen [Sun, 17 Feb 2002 17:21:27 +0000 (17:21 +0000)]
Don't rely on <sys/signal.h> to include <sys/ucontext.h>

22 years agoDon't rely on <sys/signal.h> to include <sys/ucontext.h>.
deischen [Sun, 17 Feb 2002 17:19:14 +0000 (17:19 +0000)]
Don't rely on <sys/signal.h> to include <sys/ucontext.h>.

22 years agoCorrect typo.
nyan [Sun, 17 Feb 2002 14:16:17 +0000 (14:16 +0000)]
Correct typo.

22 years agoFix a formatting error.
joe [Sun, 17 Feb 2002 12:41:50 +0000 (12:41 +0000)]
Fix a formatting error.

22 years agoKNF style the code, ready for an MFC.
joe [Sun, 17 Feb 2002 12:29:39 +0000 (12:29 +0000)]
KNF style the code, ready for an MFC.

22 years agoKNF style the code, ready for an MFC.
joe [Sun, 17 Feb 2002 11:58:58 +0000 (11:58 +0000)]
KNF style the code, ready for an MFC.

22 years agoMove the bus_space_subregion function from the puc driver to the bus_space
nyan [Sun, 17 Feb 2002 09:41:23 +0000 (09:41 +0000)]
Move the bus_space_subregion function from the puc driver to the bus_space
sutff.

Reviewed by: jhay

22 years ago- Split the routine to initialize a bus_space_handle into the separate
nyan [Sun, 17 Feb 2002 09:16:45 +0000 (09:16 +0000)]
- Split the routine to initialize a bus_space_handle into the separate
  function.
- Only access a bus_space_handle if the resource type is SYS_RES_MEMORY or
  SYS_RES_IOPORT.
- Add the bus_space_subregion supports.

22 years agoMFi386: revision 1.167
nyan [Sun, 17 Feb 2002 08:56:12 +0000 (08:56 +0000)]
MFi386: revision 1.167

22 years agoMFi386: revision 1.392
nyan [Sun, 17 Feb 2002 08:55:42 +0000 (08:55 +0000)]
MFi386: revision 1.392

22 years agoPurely whitespace changes
julian [Sun, 17 Feb 2002 08:14:15 +0000 (08:14 +0000)]
Purely whitespace changes

22 years agoreplace the embedded cr_mtx in the ucred structure with cr_mtxp (a mutex
dillon [Sun, 17 Feb 2002 07:30:34 +0000 (07:30 +0000)]
replace the embedded cr_mtx in the ucred structure with cr_mtxp (a mutex
pointer), and use the mutex pool routines.  This greatly reduces the size
of the ucred structure.

22 years agoAdd support such that if LD_TRACE_LOADED_OBJECTS_ALL is defined to a
obrien [Sun, 17 Feb 2002 07:04:32 +0000 (07:04 +0000)]
Add support such that if LD_TRACE_LOADED_OBJECTS_ALL is defined to a
non-empty string in the environment; we indicate which objects caused
each object to be loaded.

PR: 30908
Submitted-by: Mike Meyer <mwm@mired.org>

22 years agoSupport for f/w crash dumps (2200 && 23XX).
mjacob [Sun, 17 Feb 2002 06:38:22 +0000 (06:38 +0000)]
Support for f/w crash dumps (2200 && 23XX).

If you want QLogic to look at a potential f/w problem for FC cards, you really
have to provide them info in the format they expect. This involves dumping
a lot of hardware registers (> 300 16 bit registers) and a lot of SRAM
(> 128KB minimum). Thus all of this code is #ifdef protected which will
become an option so that the memory allocation of where to dump the crash
image is pretty expensive. It's worth it if you have a reproducible problem
because they have some tools that can tell them, given the f/w version,
the precise state of everything.

MFC after: 1 week

22 years agoHints for WWN are now WWNN and/or WWPN.
mjacob [Sun, 17 Feb 2002 06:34:21 +0000 (06:34 +0000)]
Hints for WWN are now WWNN and/or WWPN.

MFC after: 1 week

22 years agoAdd in support firmware crash dumps. Change CFG options to split
mjacob [Sun, 17 Feb 2002 06:32:58 +0000 (06:32 +0000)]
Add in support firmware crash dumps. Change CFG options to split
WWN into WWNN and WWPN.

MFC after: 1 week

22 years agoBulk changes made during the BSDcon kernel summit and travel afterwards.
imp [Sun, 17 Feb 2002 03:11:11 +0000 (03:11 +0000)]
Bulk changes made during the BSDcon kernel summit and travel afterwards.
Appologies for making this one bulk commit, but I have tested all these
changes together and don't want to break anything by trying to disentangle
it.

o Make debugging a sysctl/tunable
o Remove flags word from yenta chip info, it is unused
o Make 16-bit card I/O range and 32-bit card I/O range tunables
o Start the rename of pccbb to cbb to match NetBSD by misc renames.
o Kill the now bogus list of softcs to create kthread.  Instead, just
  create the kthread in the attach routine.
o Remove sc_ from some structure names.  It isn't needed.
o Refine chipset lookup code.
o Match generic PCI <-> CardBus bridges.  We specifically don't generically
  match PCI PCMCIA bridges because they are not, with one exception, yenta
  devices.
o Add some comments about the why we need to have a function table ala
  OLDCARD
o The PCI interrupt routing by using the ExCA registers is needed for
  for all bridges, per the spec, not just TI ones.
o Collapse TOPIC95 and TOPIC95B.
o Using the ToPIC 97 and 100 datasheets, try to support these bridges better,
  but more work is needed.
o Generally clarify some XXX comments and add them in a few places where
  things didn't look right to me.
o Move interrupt generating register access until after we establish an ISR.
o Add support for YV and XV cards.  X and Y are numbers to be determined
  later (but maybe never).
o factor powerup code for 16-bit and 32-bit cards.
o When a card supports more than one voltage, prefer the lowest supported
  volage.  Windows does this, and MS's design guides imply this is the
  right thing to do.
o Document race between kthread_exit(0) and kldunload's unmapping of pages
  that John Baldwin and I discovered.
o Debounce the CSC interrupt a little better.
o When a 16-bit card is inserted when we don't have a pccard child,
  warn about it better.  Ditto for 32-bit card.
o Ack ALL the interrupt bits that we get, not just 0x1.
o maybe a couple minor style nits corrected.