Commit Graph

95294 Commits

Author SHA1 Message Date
Mark Murray
0f66784721 Turn malloc+snprintf into asnprintf.
Submitted by:	David Hill <hill at phobia dot em ess>
2003-10-14 20:42:13 +00:00
Dag-Erling Smørgrav
e89ab6e083 Add the "TB ---" prefix to error and warning messages so that early
failures (e.g. failure to lock the sandbox) don't result in empty failure
reports.
2003-10-14 20:01:58 +00:00
Sam Leffler
929b31ddab Lock ip forwarding route cache. While we're at it, remove the global
variable ipforward_rt by introducing an ip_forward_cacheinval() call
to use to invalidate the cache.

Supported by:	FreeBSD Foundation
2003-10-14 19:19:12 +00:00
Sam Leffler
ba00f0096d MFp4: correct locking issues in nd6_lookup
Supported by:	FreeBSD Foundation
2003-10-14 18:49:08 +00:00
Sam Leffler
888c2a3c4e remove dangling ';'s` that were harmless
Supported by:	FreeBSD Foundation
2003-10-14 18:45:50 +00:00
Sam Leffler
bd19669855 Change default for kern.polling.idle_poll back to 1. This was set to 0
because Luigi observed livelock but in recent testing it did not occur
so I'm re-enabling it by default.

Reviewed by:	luigi
2003-10-14 18:39:36 +00:00
Peter Wemm
c513c3b569 Fix just about as many bugs in my last commit here as there were lines that
I changed.  That is never a good sign.
1) only map 1 page at address zero, not 4096 pages
2) page 1 starts at address 4096 (PAGE_SIZE) not 4095 (PAGE_MASK).  I
   don't even want to think what the pte's looked like.
3) subtract the r/o page group start address from the end before
   converting it to a count.  Otherwise an extra page is mapped.

If you were affected by this, the symptoms of this was a hang at boot
after the spinner.  Sorry folks. :-(

"You broke my laptop!" by:	sam
2003-10-14 17:02:18 +00:00
Søren Schmidt
a827a5ff6a Reintroduce the "recovered from lost interrupt" code, but in a new
(hopefully) panic safe way.

Why the interrupts are lost is still a mystery, to me at least.
2003-10-14 16:53:13 +00:00
Hajimu UMEMOTO
3011d4b3e2 Fix alignment problem on 64 bit arch.
I only tested if it doesn't break anything on i368.  Since I
have no 64 bit machine, I cannot test it, actually.

Reported by:	jmallett
2003-10-14 13:37:37 +00:00
Hideyuki KURASHINA
8f04170699 Catch up with the English version:
1.592 -> 1.638 relnotes/common/new.sgml
2003-10-14 08:27:39 +00:00
Poul-Henning Kamp
b84044731d Made use of 'error' argument, which was unused (by mistake) before.
Submitted by:    Pawel Jakub Dawidek <nick@garage.freebsd.pl>
2003-10-14 08:09:43 +00:00
Warner Losh
d29516dd82 With DIAGNOSTICS, sometimes we get weird crashes when some driver
accesses softc after it is freed.  Use a different malloc type for
softc than the rest of the bus code to make it more clear when these
things happen that it is the driver that's at fault, not the bus code.

Suggested by: sam and/or phk (I think)
2003-10-14 06:22:07 +00:00
Alan Cox
7fb578933f MFia64
Move uma_small_alloc() and uma_small_free() to uma_machdep.c.
2003-10-14 05:51:31 +00:00
Hideyuki KURASHINA
abb86bf8a1 Close open parenthesis. 2003-10-14 05:36:36 +00:00
Warner Losh
11a3320287 Fix a race condition in owi. If you eject a card, then the wi_inquire
timeout would continue to happen: boom!  Fix this[*] by timing out earlier.

[*] almost fixes the race on unload: wi_inquire could be running when
untimeout is called, and there's no way to know when it has actually
returned.  This race is very rare and hard to lose.

Submitted by: scottl
2003-10-14 05:25:31 +00:00
Warner Losh
4d721ca3ae Use more modern architectures than VAX for little endian machines 2003-10-14 02:59:13 +00:00
Warner Losh
010ebba66c Replace the VAX and MIPS alternate formats with the architectures that
FreeBSD supports.  None of them support an alternate formats, except
the alpha (which prints extra register information).

# if we get a mips port, we can put the mips case back to document the
# actual behavior.
2003-10-14 02:47:23 +00:00
Jeff Roberson
85b9831dfa - Add a mising vn_finished_write()
Pointy hat:     jeff
Found by:       robert
Obtained from:  kirk
2003-10-14 00:38:34 +00:00
Alan Cox
44b0fd1cf6 Add rfork_thread(3). 2003-10-13 20:32:33 +00:00
Poul-Henning Kamp
b6badb5a54 Insert an overview of the plans here, in case I get run over by a bus. 2003-10-13 20:14:02 +00:00
Hajimu UMEMOTO
66c7fe4056 use BF_ecb_encrypt().
Obtained from:	KAME
2003-10-13 19:26:08 +00:00
Hajimu UMEMOTO
b42ac57f4f - support AES counter mode for ESP.
- use size_t as return type of schedlen(), as there's no error
  check needed.
- clear key schedule buffer before freeing.

Obtained from:	KAME
2003-10-13 14:57:41 +00:00
David Xu
3a2e2a0ec8 Don't clear signal mask in execsig(). RELENG_4 does not clear it and POSIX
asks to inherit signal mask for execv.
2003-10-13 14:03:08 +00:00
Dag-Erling Smørgrav
aeba777e45 There is no network2 any more.
Pointy hat to:	mtm
2003-10-13 10:36:26 +00:00
Doug Barton
809a0e6290 * Remove the OS specific bits, since NetBSD isn't interested in
porting this stuff back.

* Test /etc/motd for writability before trying to update it. This is
especially useful when /etc/ is mounted ro, like on a diskless boot.
(Thanks to phk for the idea on this one.)

* Make the "updating" message reflect what actually happens.
2003-10-13 08:44:07 +00:00
Doug Barton
76f9428757 Remove trailing whitespace 2003-10-13 08:20:55 +00:00
Doug Barton
9f29b7170b Update references to rc.diskless[12] 2003-10-13 08:18:05 +00:00
Tim J. Robbins
1a7c9b7f5e Fix two buffer overflows caused by off-by-one errors: avoid writing a null
character 1 byte past the end of cmdline[] when libedit is being used for
input, and avoid writing a null pointer 1 element past the end of margv[].

Reviewed by:		gad
2003-10-13 07:24:22 +00:00
Bruce Evans
1f85f71570 Include <nlist.h> for nlist-related declarations instead of depending on
namespace pollution in <kvm.h>.
2003-10-13 05:30:20 +00:00
Sam Leffler
7d4bf9ccb1 remove dangling mtx_unlock orphaned by rev 1.21 change 2003-10-13 04:57:31 +00:00
Hajimu UMEMOTO
79203b9869 support AES XCBC MAC for AH.
Obtained from:	KAME
2003-10-13 04:56:04 +00:00
Hajimu UMEMOTO
c65ee7c758 - support AES XCBC MAC for AH
- correct SADB_X_AALG_RIPEMD160HMAC to 8

Obtained from:	KAME
2003-10-13 04:54:51 +00:00
Bruce Evans
7f75c38230 Fixed some style bugs in the removal of __P(()). Blind removal of
spaces before __P(()) outdented continuation lines to column 0.
2003-10-13 04:44:55 +00:00
Sam Leffler
91101a2af3 Reduce per-packet overhead when using WEP by using an advancing IV
seeded with arc4random rather than calling arc4random for each
packet.  Note this is the same algorithm used to select the IV when
doing WEP on the host.
2003-10-13 03:42:53 +00:00
Sam Leffler
167ecdca8c Must reset the pointer to the 802.11 header after prepending
for WEP in case the prepend addes a new mbuf.  This fixes WEP.
2003-10-13 03:41:32 +00:00
Sam Leffler
2075afba8d MFp4:
o don't grab the mutex at the top of ath_detach; it does nothing
  useful
o deal with entry to ath_ioctl during detach to disable promiscuous
  mode as a result of calling bpfdetach2: cannot call ath_init when
  the device is marked invalid as the code isn't prepared to deal
  with it (in particular by that time the hal reference may have
  been yanked)
2003-10-13 03:39:38 +00:00
Sam Leffler
310e4a4ae6 MFp4:
change ath_rate_ctl_reset to handle transition from station
mode to adhoc mode; was not resetting the initial xmit rate
causing outbound frames to be dicarded
2003-10-13 03:37:38 +00:00
Alan Cox
50ac3f9939 - Modify pmap_is_current() to return FALSE when a pmap's page table is in
use because a kernel thread is borrowing it.  The borrowed page table
   can change spontaneously, making any dependence on its continued use
   subject to a race condition.
 - _pmap_unwire_pte_hold() cannot use pmap_is_current(): If a change is
   made to a page table page mapping for a borrowed page table, the TLB
   must be updated.

In collaboration with:	tegge
2003-10-13 03:28:31 +00:00
Maksim Yevmenkin
0986ab12e4 Update Bluetooth code.
Reviewed by: M. Warner Losh <imp@bsdimp.com>; John Hay <jhay@freebsd.org>
Approved by: M. Warner Losh <imp@bsdimp.com> (mentor)
2003-10-12 22:04:24 +00:00
Murray Stokely
907d866750 Update the number of ports. 2003-10-12 21:35:50 +00:00
Jeff Roberson
736c97c7b3 - In SCHED_CURR() add holding Giant to the list of criteria that will keep
you on the current queue.  In the future, it would be nice if priority
   propagation could deterministicly pluck a thread off of the next queue
   and put it on the current queue.  Until then this hack stops us from
   holding up our entire current queue, including interrupt handlers, while
   a thread on the next queue is blocked while holding Giant.
 - Inherit our pctcpu information from our parent.
2003-10-12 21:07:31 +00:00
Hajimu UMEMOTO
8f21478b6a simplify and update rijndael code.
Obtained from:	KAME
2003-10-12 21:05:05 +00:00
Jeff Roberson
a34419fee5 - Implement a mtx_ownedby() macro which can be used to determine if a
particular thread owns a mutex.  This cannot be done without races
   unless the thread is curthread.
2003-10-12 21:02:55 +00:00
Robert Watson
d32864c331 Comment spelling fix. 2003-10-12 20:51:26 +00:00
Hiroki Sato
094a5db41b Use Japanese localized version of trademarks.ent. 2003-10-12 20:11:06 +00:00
Hajimu UMEMOTO
d5d49fe472 include opencrypto/rmd160.h 2003-10-12 18:33:30 +00:00
Alan Cox
d58e70a08d In vfs_bio_clrbuf(), ignore the state of the object lock if the page is the
"bogus" page.

Found by:	tegge
2003-10-12 18:26:48 +00:00
Hajimu UMEMOTO
3dbacd2651 use opencrypto for RMD160.
Requested by:	sam
2003-10-12 18:25:38 +00:00
Daniel Eischen
077af0a4b4 Don't forget to initialize the fake tcb when the kcb is allocated. 2003-10-12 16:50:45 +00:00
Bernd Walter
8b43efbd44 Cache dev_t values in the right structure.
Tested by:	Jay Cornwall <jay@evilrealms.net>
2003-10-12 15:51:40 +00:00