]> CyberLeo.Net >> Repos - FreeBSD/FreeBSD.git/log
FreeBSD/FreeBSD.git
23 years agoUpdate import information
Gregory Neil Shapiro [Tue, 21 Aug 2001 01:33:05 +0000 (01:33 +0000)]
Update import information

23 years agoResolve conflicts from sendmail 8.11.6 import
Gregory Neil Shapiro [Tue, 21 Aug 2001 01:32:34 +0000 (01:32 +0000)]
Resolve conflicts from sendmail 8.11.6 import

23 years agoThis commit was generated by cvs2svn to compensate for changes in r82017,
Gregory Neil Shapiro [Tue, 21 Aug 2001 01:31:44 +0000 (01:31 +0000)]
This commit was generated by cvs2svn to compensate for changes in r82017,
which included commits to RCS files with non-trunk default branches.

23 years agoImport sendmail 8.11.6
Gregory Neil Shapiro [Tue, 21 Aug 2001 01:31:44 +0000 (01:31 +0000)]
Import sendmail 8.11.6

23 years agoUse register g6 to point to a small stack for svaing alternate globals
Jake Burkholder [Tue, 21 Aug 2001 00:07:37 +0000 (00:07 +0000)]
Use register g6 to point to a small stack for svaing alternate globals
during trap handlers.
Implement ptrace_set_pc FWIW.
Initialize the pcb window scratch area in setregs(), and setup user
registers as specified by the SCD.

Submitted by: tmm

23 years agomove forward the ability to MAKEDEV zs devices for TurboLaser
Matt Jacob [Tue, 21 Aug 2001 00:03:28 +0000 (00:03 +0000)]
move forward the ability to MAKEDEV zs devices for TurboLaser

23 years agoHandle the pcb window scratch area in cpu_fork.
Jake Burkholder [Tue, 21 Aug 2001 00:02:54 +0000 (00:02 +0000)]
Handle the pcb window scratch area in cpu_fork.
Implement cpu_exit.

Submitted by: tmm

23 years agoSave and restore %fprs and %y, which are unused by kernel code, but
Jake Burkholder [Tue, 21 Aug 2001 00:01:28 +0000 (00:01 +0000)]
Save and restore %fprs and %y, which are unused by kernel code, but
may be used by 32bit userland code.
Implement cpu_throw().

Submitted by: tmm

23 years agoDisable interrupts when calling openfirmware.
Jake Burkholder [Tue, 21 Aug 2001 00:00:18 +0000 (00:00 +0000)]
Disable interrupts when calling openfirmware.

23 years agoRename fp_init_pcb to fp_init_proc. Set the FEF bit in fprs register;
Jake Burkholder [Mon, 20 Aug 2001 23:56:19 +0000 (23:56 +0000)]
Rename fp_init_pcb to fp_init_proc.  Set the FEF bit in fprs register;
according the SCD it should be set if no user trap handler in set.

Submitted by: tmm

23 years agoAdd definitions for new assembler code.
Jake Burkholder [Mon, 20 Aug 2001 23:53:11 +0000 (23:53 +0000)]
Add definitions for new assembler code.

23 years agoCatch up with new trap entry point names.
Jake Burkholder [Mon, 20 Aug 2001 23:51:40 +0000 (23:51 +0000)]
Catch up with new trap entry point names.

23 years agoAdd variables needed by hardware watchpoint support.
Jake Burkholder [Mon, 20 Aug 2001 23:50:48 +0000 (23:50 +0000)]
Add variables needed by hardware watchpoint support.

Submitted by: tmm

23 years agoAdd code for supporting hardware watch points.
Jake Burkholder [Mon, 20 Aug 2001 23:50:08 +0000 (23:50 +0000)]
Add code for supporting hardware watch points.

Submitted by: tmm

23 years agoAdd a system call trap type and syscall() call request handler.
Jake Burkholder [Mon, 20 Aug 2001 23:43:43 +0000 (23:43 +0000)]
Add a system call trap type and syscall() call request handler.
Also add support for hardware watch point traps.

Submitted by: tmm

23 years agoAdd support for splitting the register windows on entry to the
Jake Burkholder [Mon, 20 Aug 2001 23:40:31 +0000 (23:40 +0000)]
Add support for splitting the register windows on entry to the
kernel from usermode.  The remaining user windows are spilled
to the pcb as necessary.  The user land window fault handlers
fill directly from the pcb on return.
Add system call entry points.

Submitted by: tmm

23 years agodb_expr_t is signed.
Jake Burkholder [Mon, 20 Aug 2001 23:35:15 +0000 (23:35 +0000)]
db_expr_t is signed.

23 years agoAdd definitions for bits in condition code register and the load store
Jake Burkholder [Mon, 20 Aug 2001 23:34:46 +0000 (23:34 +0000)]
Add definitions for bits in condition code register and the load store
unit control registers.  Move tstate definitions to their own file.

Submitted by: tmm

23 years agoAdd a definition for the load store unit control register.
Jake Burkholder [Mon, 20 Aug 2001 23:31:41 +0000 (23:31 +0000)]
Add a definition for the load store unit control register.

23 years agoMake the copyright consistent.
Brian Somers [Mon, 20 Aug 2001 22:57:33 +0000 (22:57 +0000)]
Make the copyright consistent.

Previously approved by: Charles Mott <cmott@scientech.com>

23 years agoSync globals.h up with the other platforms. There is still some cruft in
David E. O'Brien [Mon, 20 Aug 2001 21:29:16 +0000 (21:29 +0000)]
Sync globals.h up with the other platforms.  There is still some cruft in
here, but now all the platforms have the same cruft.  Consistantly spell
the `struct globaldata *' "globalp".

Reviewed by: peter

23 years ago+ do not hard code where a port may be installed to
David E. O'Brien [Mon, 20 Aug 2001 21:22:33 +0000 (21:22 +0000)]
+ do not hard code where a port may be installed to
+ uses `mkisofs' rather than `mkhybrid' as the new mkisofs is the merger
  of the two
+ checks for `setcdboot's existsance and tries to pkg_add it if needed
+ removes English contraction

Approved by: jkh

23 years ago+ back out my `iso' target
David E. O'Brien [Mon, 20 Aug 2001 21:22:01 +0000 (21:22 +0000)]
+ back out my `iso' target
+ allows one to set _R so one can more easily make an iso after the fact,
  in the directory `make release' was issued in
+ changes name from miniboot to miniinst since we make other bootable ISOs
  also (and we might confuse someone)

Approved by: jkh

23 years agoNew release note: pppd(8) no longer world-executable/readable.
Bruce A. Mah [Mon, 20 Aug 2001 20:45:55 +0000 (20:45 +0000)]
New release note:  pppd(8) no longer world-executable/readable.

Submitted by: kris

23 years agoAvoid sleeping while holding a mutex in dounmount(). This problem
Ian Dowse [Mon, 20 Aug 2001 19:16:31 +0000 (19:16 +0000)]
Avoid sleeping while holding a mutex in dounmount(). This problem
has existed for a long time, but I made it worse a few months ago
by by adding calls to VFS_ROOT() and checkdirs() in revision 1.179.

Also, remove the LK_REENABLE flag in the lockmgr() call; this flag
has been ignored by the lockmgr code for 4 years. This was the only
remaining mention of it apart from its definition.

Reviewed by: jhb

23 years agoHandle snprintf() returning < -1.
Brian Somers [Mon, 20 Aug 2001 18:13:50 +0000 (18:13 +0000)]
Handle snprintf() returning < -1.

23 years agoFix fencepost error causing creation of 0-length mbufs when the boundary
Bill Fenner [Mon, 20 Aug 2001 17:58:46 +0000 (17:58 +0000)]
Fix fencepost error causing creation of 0-length mbufs when the boundary
 between header and data fell on the boundary between two mbufs.

23 years agoTypo and markup fixes.
Bruce A. Mah [Mon, 20 Aug 2001 17:31:47 +0000 (17:31 +0000)]
Typo and markup fixes.

Submitted by: Alex Kapranoff <kapr@acm.org>

23 years agoClean up some ways in which we set defaults for SCSI cards
Matt Jacob [Mon, 20 Aug 2001 17:28:32 +0000 (17:28 +0000)]
Clean up some ways in which we set defaults for SCSI cards
that do not have valid NVRAM. In particular, we were leaving
a retry count set (to retry selection timeouts) when thats
not really what we want. Do some constant string additions
so that LOGDEBUG0 info is useful across all cards.

MFC after: 2 weeks

23 years agoAdd MBOX_GET_PCI_PARAMS alias.
Matt Jacob [Mon, 20 Aug 2001 17:26:45 +0000 (17:26 +0000)]
Add MBOX_GET_PCI_PARAMS alias.

MFC after: 2 weeks

23 years agoConditionalize VM_SWZONE_SIZE_MAX and VM_BCACHE_SIZE_MAX so MD sections
Matthew Dillon [Mon, 20 Aug 2001 16:29:13 +0000 (16:29 +0000)]
Conditionalize VM_SWZONE_SIZE_MAX and VM_BCACHE_SIZE_MAX so MD sections
that don't define these constants don't break.

23 years agoHandle snprintf() returning < 0 (not just -1)
Brian Somers [Mon, 20 Aug 2001 15:44:17 +0000 (15:44 +0000)]
Handle snprintf() returning < 0 (not just -1)

MFC after: 2 weeks

23 years agoBetter snprintf() handling.
Brian Somers [Mon, 20 Aug 2001 15:43:14 +0000 (15:43 +0000)]
Better snprintf() handling.

Prompted by: bde

23 years agoEven more libss removal.
Ruslan Ermilov [Mon, 20 Aug 2001 15:41:15 +0000 (15:41 +0000)]
Even more libss removal.

23 years agoHandle snprintf() returning < 0 (not just -1)
Brian Somers [Mon, 20 Aug 2001 15:09:34 +0000 (15:09 +0000)]
Handle snprintf() returning < 0 (not just -1)

MFC after: 2 weeks

23 years agoHandle ftp_copycommand() and ftp_copyresult() potentially returning < -1.
Brian Somers [Mon, 20 Aug 2001 15:01:06 +0000 (15:01 +0000)]
Handle ftp_copycommand() and ftp_copyresult() potentially returning < -1.

23 years agoHandle snprintf() returning < 0 (not just -1)
Brian Somers [Mon, 20 Aug 2001 14:53:05 +0000 (14:53 +0000)]
Handle snprintf() returning < 0 (not just -1)

MFC after: 2 weeks

23 years agoHandle snprintf() returning -1.
Brian Somers [Mon, 20 Aug 2001 14:46:40 +0000 (14:46 +0000)]
Handle snprintf() returning -1.

MFC after: 2 weeks

23 years agoHandle snrintf overflows.
Brian Somers [Mon, 20 Aug 2001 13:44:13 +0000 (13:44 +0000)]
Handle snrintf overflows.

Spotted by: bde

23 years agoHandle snprintf() returning -1.
Brian Somers [Mon, 20 Aug 2001 13:24:39 +0000 (13:24 +0000)]
Handle snprintf() returning -1.

MFC after: 2 weeks

23 years agoHandle snprintf() returning -1.
Brian Somers [Mon, 20 Aug 2001 12:56:45 +0000 (12:56 +0000)]
Handle snprintf() returning -1.

MFC after: 2 weeks

23 years agoMark some functions as __printflike() and/or taking const char * arguments
Kris Kennaway [Mon, 20 Aug 2001 12:53:36 +0000 (12:53 +0000)]
Mark some functions as __printflike() and/or taking const char * arguments
instead of char *.

MFC after: 2 weeks

23 years agoNote that submitted patches are assumed to be under the same license as
Nik Clayton [Mon, 20 Aug 2001 12:52:59 +0000 (12:52 +0000)]
Note that submitted patches are assumed to be under the same license as
the file they patch.

23 years agoFix overflow problem when giving a username that is too long.
Mark Murray [Mon, 20 Aug 2001 12:52:49 +0000 (12:52 +0000)]
Fix overflow problem when giving a username that is too long.
Also minor declaration/header fixes while auditing the code.

23 years agoHandle snprintf() returning -1.
Brian Somers [Mon, 20 Aug 2001 12:50:21 +0000 (12:50 +0000)]
Handle snprintf() returning -1.

MFC after: 2 weeks

23 years agoVery minor stylistic nit.
Mark Murray [Mon, 20 Aug 2001 12:46:11 +0000 (12:46 +0000)]
Very minor stylistic nit.

Discussed with: ru

23 years agoAdd 'try_mapped_pass' standard option.
Mark Murray [Mon, 20 Aug 2001 12:43:19 +0000 (12:43 +0000)]
Add 'try_mapped_pass' standard option.

Asked for by: lukeh@PADL.COM

23 years agoHandle snprintf() returning -1
Brian Somers [Mon, 20 Aug 2001 12:41:36 +0000 (12:41 +0000)]
Handle snprintf() returning -1

MFC after: 2 weeks

23 years agoMore libss removal.
Mark Murray [Mon, 20 Aug 2001 12:39:23 +0000 (12:39 +0000)]
More libss removal.

23 years agoDiff reduce all the crypto telnet Makefiles.
Mark Murray [Mon, 20 Aug 2001 12:32:45 +0000 (12:32 +0000)]
Diff reduce all the crypto telnet Makefiles.

23 years agoHandle snrintf() returning -1.
Brian Somers [Mon, 20 Aug 2001 12:31:13 +0000 (12:31 +0000)]
Handle snrintf() returning -1.

MFC after: 2 weeks

23 years agoCode merge and diff reduce with "base" telnet. This is the "later"
Mark Murray [Mon, 20 Aug 2001 12:28:40 +0000 (12:28 +0000)]
Code merge and diff reduce with "base" telnet. This is the "later"
telnet, so it was treated as the reference code, except where later
commits were made to "base" telnet.

23 years agoCode merge and diff reduction between this and crypto telnet.
Mark Murray [Mon, 20 Aug 2001 12:21:31 +0000 (12:21 +0000)]
Code merge and diff reduction between this and crypto telnet.

Also remove useless AUTHENTICATION code. We have never compiled this
here, and it is doubtful that it even works without crypto.

23 years agoFeature merging and diff reduction between this code and crypto telnet.
Mark Murray [Mon, 20 Aug 2001 12:12:27 +0000 (12:12 +0000)]
Feature merging and diff reduction between this code and crypto telnet.
Also remove conditional (AUTHENTICATION) code as we have never compiled
it here, and it is doubtful that it even works in this scenario.

23 years agoHandle snprintf() returning -1
Brian Somers [Mon, 20 Aug 2001 12:06:42 +0000 (12:06 +0000)]
Handle snprintf() returning -1

MFC after: 2 weeks

23 years agoNo functional changes, but:
Mark Murray [Mon, 20 Aug 2001 12:04:53 +0000 (12:04 +0000)]
No functional changes, but:

o unifdef AUTHENTICATE. We have never compiled this code, and its
  doubtful it will even work in this case.

o Style changes (some ansification, some comment updating)

o Diff reduction and code style merging with crypto telnet.

23 years agoDon't misuse the return from snprintf.
Brian Somers [Mon, 20 Aug 2001 11:58:05 +0000 (11:58 +0000)]
Don't misuse the return from snprintf.

MFC after: 2 weeks

23 years agoProperly change old interface addresses so that their destination is
Brian Somers [Mon, 20 Aug 2001 11:06:21 +0000 (11:06 +0000)]
Properly change old interface addresses so that their destination is
255.255.255.255 if our new destination address conflicts.

23 years agoAdd some comments to explain what's going on here
Brian Somers [Mon, 20 Aug 2001 10:25:23 +0000 (10:25 +0000)]
Add some comments to explain what's going on here

23 years agoDon't overflow a buffer from command line arguments.
Kris Kennaway [Mon, 20 Aug 2001 09:43:04 +0000 (09:43 +0000)]
Don't overflow a buffer from command line arguments.

MFC after: 2 weeks

23 years agomdoc(7) police: markup nits.
Ruslan Ermilov [Mon, 20 Aug 2001 08:55:07 +0000 (08:55 +0000)]
mdoc(7) police: markup nits.

23 years agoMerge vendor branch onto mainline (try to fix HAVE_KERBEROS)
Peter Wemm [Mon, 20 Aug 2001 08:47:21 +0000 (08:47 +0000)]
Merge vendor branch onto mainline (try to fix HAVE_KERBEROS)

23 years agomdoc(7) police: simplify literal display block.
Ruslan Ermilov [Mon, 20 Aug 2001 08:29:55 +0000 (08:29 +0000)]
mdoc(7) police: simplify literal display block.

23 years agomdoc(7) police: fixed markup, removed the extra word.
Ruslan Ermilov [Mon, 20 Aug 2001 08:28:18 +0000 (08:28 +0000)]
mdoc(7) police: fixed markup, removed the extra word.

23 years agoPart II of libss retirement: zap mk_cmds too.
Kris Kennaway [Mon, 20 Aug 2001 06:23:35 +0000 (06:23 +0000)]
Part II of libss retirement: zap mk_cmds too.

Pointy hat to: kris

23 years agoThe timed protocol is not implemented in a compatible way by all vendors;
Kris Kennaway [Mon, 20 Aug 2001 06:16:04 +0000 (06:16 +0000)]
The timed protocol is not implemented in a compatible way by all vendors;
the size of the tsp_name field is OS-dependent.  4.3BSD used a 32-byte
field, FreeBSD uses MAXHOSTNAMELEN and RedHat apparently uses a 64-byte
field.  As a result, sanity checking code added a few months ago to detect
short packets will fail when interoperating with one of these other vendors.

Change the short packet detection code to expect a minimum packet size
corresponding to the 4.3BSD implementation, which should be a safe minimum
size.

Submitted by: Stephen Whiteley <stevew@best.com> (based on)
PR: misc/29867

23 years agoFix an SGML glitch.
Bruce A. Mah [Mon, 20 Aug 2001 04:56:37 +0000 (04:56 +0000)]
Fix an SGML glitch.

23 years agoFix grammar.
Dima Dorfman [Mon, 20 Aug 2001 02:16:41 +0000 (02:16 +0000)]
Fix grammar.

23 years agoRespect the -N flag when changing directory attributes in setdirmode).
Dima Dorfman [Mon, 20 Aug 2001 02:15:22 +0000 (02:15 +0000)]
Respect the -N flag when changing directory attributes in setdirmode).

PR: 29671
Submitted by: Sascha Blank <sblank@addcom.de>

23 years agoUse `an' where appropriate.
Dima Dorfman [Mon, 20 Aug 2001 02:08:26 +0000 (02:08 +0000)]
Use `an' where appropriate.

23 years agoPass the username we just made to pwd_mkdb so that the latter doesn't
Dima Dorfman [Mon, 20 Aug 2001 01:20:05 +0000 (01:20 +0000)]
Pass the username we just made to pwd_mkdb so that the latter doesn't
have to regenerate the entire database.

PR: 25798
Submitted by: Domas Mituzas <domas.mituzas@delfi.lt>

23 years agoSync the default module search path with the one in
Dima Dorfman [Mon, 20 Aug 2001 01:12:28 +0000 (01:12 +0000)]
Sync the default module search path with the one in
sys/boot/common/module.c.

PR: 21405
Submitted by: Makoto MATSUSHITA <matusita@jp.FreeBSD.org>

23 years agoShoot (and kill) the second person with a high-powered rifle.
Dima Dorfman [Mon, 20 Aug 2001 01:02:05 +0000 (01:02 +0000)]
Shoot (and kill) the second person with a high-powered rifle.

23 years agoWhen attempting to change the default route, don't write the gateway
Brian Somers [Mon, 20 Aug 2001 00:46:33 +0000 (00:46 +0000)]
When attempting to change the default route, don't write the gateway
and mask to the routing socket, otherwise the update fails.

Warning provided by: markm

The code here was broken for FreeBSD when IPv6 support was added, but
was fixed for OpenBSD.  OpenBSD expects the gateway and mask to be
supplied and fails the update otherwise.

23 years agoLimit the amount of KVM reserved for the buffer cache and for swap-meta
Matthew Dillon [Mon, 20 Aug 2001 00:41:12 +0000 (00:41 +0000)]
Limit the amount of KVM reserved for the buffer cache and for swap-meta
information.  The default limits only effect machines with > 1GB of ram
and can be overriden with two new kernel conf variables VM_SWZONE_SIZE_MAX
and VM_BCACHE_SIZE_MAX, or with loader variables kern.maxswzone and
kern.maxbcache.  This has the effect of leaving more KVM available for
sizing NMBCLUSTERS and 'maxusers' and should avoid tripups where a sysad
adds memory to a machine and then sees the kernel panic on boot due to
running out of KVM.

Also change the default swap-meta auto-sizing calculation to allocate half
of what it was previously allocating.  The prior defaults were way too high.
Note that we cannot afford to run out of swap-meta structures so we still
stay somewhat conservative here.

23 years agoThere is no "(integer)" type; it's called "(int)".
Dima Dorfman [Mon, 20 Aug 2001 00:29:23 +0000 (00:29 +0000)]
There is no "(integer)" type; it's called "(int)".

PR: 20749
Submitted by: Ralf S. Engelschall <rse@engelschall.com>

23 years agoAdd new entires for firewall_flags, amd_map_program, and
Dima Dorfman [Mon, 20 Aug 2001 00:28:30 +0000 (00:28 +0000)]
Add new entires for firewall_flags, amd_map_program, and
enable_quotas.  Plus other assorted fixes.

PR: 20479
Submitted by: Ralf S. Engelschall <rse@engelschall.com>

23 years agoMark up IPFIREWALL in .Dv.
Dima Dorfman [Mon, 20 Aug 2001 00:10:42 +0000 (00:10 +0000)]
Mark up IPFIREWALL in .Dv.

23 years agoCapitalize the first word in sentences, and put periods at the end of
Dima Dorfman [Mon, 20 Aug 2001 00:03:01 +0000 (00:03 +0000)]
Capitalize the first word in sentences, and put periods at the end of
sentences.

23 years agoExpand a little on the note about apm_battery. Note that apmd can
Dima Dorfman [Sun, 19 Aug 2001 23:36:49 +0000 (23:36 +0000)]
Expand a little on the note about apm_battery.  Note that apmd can
execute commands when the battery level changes, and point users at
the examples in /etc/apmd.conf since this stuff isn't (yet?)
documented in the manual page.  Also note its MFC.

Submitted by: nsayer

23 years agoAdd some casts.
Brian Somers [Sun, 19 Aug 2001 22:23:28 +0000 (22:23 +0000)]
Add some casts.

Not spotted by: FreeBSD's gcc version 2.95.3 20010315 (release)
Spotted by: OpenBSD's gcc version 2.95.3 20010125 (prerelease)

23 years agoAdd a missing % in a format string.
Brian Somers [Sun, 19 Aug 2001 22:23:23 +0000 (22:23 +0000)]
Add a missing % in a format string.

Not spotted by: FreeBSD's gcc version 2.95.3 20010315 (release)
Spotted by: OpenBSD's gcc version 2.95.3 20010125 (prerelease)

23 years agoAvoid a compiler warning
Brian Somers [Sun, 19 Aug 2001 22:23:22 +0000 (22:23 +0000)]
Avoid a compiler warning

Not spotted by: FreeBSD's gcc version 2.95.3 20010315 (release)
Spotted by: OpenBSD's gcc version 2.95.3 20010125 (prerelease)

23 years agoZap LIBSS
Kris Kennaway [Sun, 19 Aug 2001 21:34:20 +0000 (21:34 +0000)]
Zap LIBSS

23 years agoSay goodbye to libss, which somehow managed to crouch hidden in the tree
Kris Kennaway [Sun, 19 Aug 2001 21:32:52 +0000 (21:32 +0000)]
Say goodbye to libss, which somehow managed to crouch hidden in the tree
for long after it was used.

23 years agoFix logic errors in pthread_cond_wait() and pthread_cond_timedwait() that
Jason Evans [Sun, 19 Aug 2001 20:05:42 +0000 (20:05 +0000)]
Fix logic errors in pthread_cond_wait() and pthread_cond_timedwait() that
could cause deadlock after interruption due to a signal.

Reviewed by: deischen

23 years agoFix ISA machines. Also, make it possible for the new pccardd to run
Warner Losh [Sun, 19 Aug 2001 19:16:26 +0000 (19:16 +0000)]
Fix ISA machines.  Also, make it possible for the new pccardd to run
on older kernels correctly.  Terminate the loop when we find a
suitable irq.  Also, only try to select from the pool.  Cleaned up the
two cases (IRQ picked by the user and ?) into one.

MFC upon re approval.

23 years agoRemove the building of a kerberosIV'ed cvs. `make release' has been broken
David E. O'Brien [Sun, 19 Aug 2001 15:46:10 +0000 (15:46 +0000)]
Remove the building of a kerberosIV'ed cvs.  `make release' has been broken
for the past 9 days!

23 years agoAdd about rewind+errno, describe ESPIPE, minor formatting.
Andrey A. Chernov [Sun, 19 Aug 2001 08:24:50 +0000 (08:24 +0000)]
Add about rewind+errno, describe ESPIPE, minor formatting.

23 years agoSilence non-constant format string warnings by marking functions
Kris Kennaway [Sun, 19 Aug 2001 08:19:37 +0000 (08:19 +0000)]
Silence non-constant format string warnings by marking functions
as __printflike()/__printf0like(), adding const, or adding missing "%s"
format strings, as appropriate.

MFC after: 2 weeks

23 years agoForced commit to fix last commit message:
Warner Losh [Sun, 19 Aug 2001 06:31:53 +0000 (06:31 +0000)]
Forced commit to fix last commit message:
The card can be removed in pcic's ISR.  We don't call the
card's ISR when no card exists.

23 years agoImprove interlocking for card removal. We now can remove the card in
Warner Losh [Sun, 19 Aug 2001 05:01:18 +0000 (05:01 +0000)]
Improve interlocking for card removal.  We now can remove the card in
the ISR.  We keep track of the card state and don't call the IRS when
the card isn't inserted.  This helps quite a bit with card ejection
problems that Ian was seeing.

Submitted by: Ian Dowse
MFC upon: re approvel.

23 years agoConstify m_copydata.
Julian Elischer [Sun, 19 Aug 2001 04:35:28 +0000 (04:35 +0000)]
Constify m_copydata.
this allows constants to be copied without giving errors.
(removes a couple of warning messages in the kernel.)

Note: uipc_mbuf.c is a forced commit as it was accidentally included in another commit.

MFC after: 2 weeks

23 years agoForgot to remove this un-needed test. (M_WAITOK won't fail)
Julian Elischer [Sun, 19 Aug 2001 04:30:13 +0000 (04:30 +0000)]
Forgot to remove this un-needed test. (M_WAITOK won't fail)
I vaguely remember someone once proving it COULD return NULL..
was that changed?

Reminded by: BDE

MFC after: 2 weeks

23 years agoBetter handling of the return from snprintf
Brian Somers [Sat, 18 Aug 2001 22:46:06 +0000 (22:46 +0000)]
Better handling of the return from snprintf

23 years agoBetter handling for the return of snprintf().
Brian Somers [Sat, 18 Aug 2001 22:43:11 +0000 (22:43 +0000)]
Better handling for the return of snprintf().

23 years agoIf AF_INET6 isn't available in the kernel, revert to creating AF_INET
Brian Somers [Sat, 18 Aug 2001 19:16:56 +0000 (19:16 +0000)]
If AF_INET6 isn't available in the kernel, revert to creating AF_INET
listening sockets in ``set server''.

23 years agoDon't needlessly duplicate what's basically the same copyright.
Jake Burkholder [Sat, 18 Aug 2001 19:09:55 +0000 (19:09 +0000)]
Don't needlessly duplicate what's basically the same copyright.

23 years agoBack out the previous fix to deal with kernels that don't support IPv6,
Brian Somers [Sat, 18 Aug 2001 19:07:13 +0000 (19:07 +0000)]
Back out the previous fix to deal with kernels that don't support IPv6,
and implement a far more subtle and correct fix.

The reason behind the infinite loop was that ppp was trying to make up
initial IPv6 numbers and wasn't giving up when it failed unexpectedly to
assign the addresses it just fabricated to it's interface (thinking that
the reason was because another interface was using the same address).
It now attempts this up to 100 times before just failing and trying to
muddle along (in reality, this should never happen more than a couple
of times unless our random number generator doesn't work).

Also, when IPv6 is not available, don't even try to assign the IPv6
interface address in the first place...

23 years agoImplement cpu_wait().
Jake Burkholder [Sat, 18 Aug 2001 18:11:48 +0000 (18:11 +0000)]
Implement cpu_wait().