ru [Sun, 19 Nov 2006 12:30:31 +0000 (12:30 +0000)]
MFC:
Remove the -C option as it does more harm than good. To be fully
compatible, it would have to (at least):
- support the "compat-compat" -T option,
- *not* support the -l, -O, and -v options,
- default to soft updates being disabled.
Worse, the compatibility mode made it impossible to mount_mfs(8)
a file system from fstab(5) with soft updates disabled (-S). [1]
Now, the only difference when called as "mount_mfs" or "mfs" (as
opposed to "mdmfs") is that the file mode of the mount point is
set by default to 01777. All options available to mdmfs(8) are
also available to mount_mfs(8); the -C option is still recognized
but ignored for backward compatibility.
bz [Sun, 19 Nov 2006 10:07:08 +0000 (10:07 +0000)]
Fix PMTU discovery in IPsec case by using an MTU hint in ICMP unreachable
fragmentation needed other then 0 when we cannot get a security policy.
This changes the code path to match what we have had in HEAD since
rev. 1.312.
PR: kern/91412
Submitted by: Tom Judge <tom tomjudge.com>
ume [Sun, 19 Nov 2006 07:27:42 +0000 (07:27 +0000)]
MFC: Teach an IPV6CP to pppd(8).
The eui64.[ch] and ipv6cp.[ch] were taken from ppp-2.3.11.
However, our stock pppd(8) doesn't provide option_t nor some
utility functions. So, I made some hacks to adjust to our
stock pppd(8).
The sys_bsd.c part was taken from NetBSD with some
modifications to adjust to our stock pppd(8).
kib [Sat, 18 Nov 2006 20:59:50 +0000 (20:59 +0000)]
On RELENG_6, vn_open() will drop Giant aquired for calls into
non-MP-safe filesystem if fdx == -1, assuming the in-kernel callers are
not MP-safe. On HEAD, vn_open() after rev. 1.240 of sys/kern/vfs_vnops.c
would keep the Giant locked when the MPSAFE flag is given.
The difference in the API caused the auditctl() syscall code to try to
unlock non-locked Giant for non-MP-safe filesystems, for instance, UFS
when QUOTA is compiled in.
This is direct commit to RELENG_6 to accomodate for RELENG_6' vn_open().
Reported by: Steve Wills (steve at stevenwills com)
Reviewed by: rwatson
Approved by: re (bmah)
X-MFC to RELENG_6_2 after: 3 days
obrien [Fri, 17 Nov 2006 00:32:53 +0000 (00:32 +0000)]
Fix ARM build:
1. NO_FORTH needs to be defined
2. bootpd.c rev 1.24: make tv_sec a time_t
3. authpf.c rev 1.7: handle timeval.tv_sec change from long to time_t
flz [Tue, 14 Nov 2006 12:54:39 +0000 (12:54 +0000)]
MFC:
Add support for Epson USB Scanners (3590 [1] and 4990 [2]).
Add entries in the uscanner.4 man page (along with missing 3500).
Add quirk for Logitech iFeel MouseMan USB mouse. [3]
Add support for Familiar Linux powered iPaq handhelds to cdce(4). [4]
Add a quirk for the Belkin FC6550-AVR UPS. [5]
Add support for Netgear FA101 ethernet USB NIC. [6]
Add support for Psion Gold Port Ethernet USB NIC. [7]
Add support for Epson Stylus CX4200 in uscanner(4). [8]
Add support for Novatech NV902 wireless NIC in ural(4). [9]
Add SCSI da(4) quirk for the iRiver H10. [10]
Add a SCSI da(4) quirk for Rekam USB Camera. [11]
Add a USB quirk for Motorola A41x/V32x USB phones. [12]
Add a USB quirk for CMOTECH CDMA USB modem. [13]
Add a USB umass(4) quirk for Panasonic KXL-840AN CD-R drive. [14]
- Consistently use if_printf() only in interface methods: if_start(),
if_watchdog, etc., or in functions used only in these methods.
In all other functions in the driver use device_printf().
- Use __func__ instead of typing function name.
- Wrap code optimized for architectures without alignment constraints
in #ifdef __NO_STRICT_ALIGNMENT rather than #ifdef __i386__. This
means that amd64 now also uses the optimized code.
While at it, fix a nearby style(9) bug.
- In dc_setcfg() suppress printing a warning when forcing the receiver
and transceiver to idle state times out for chips where the status
bits in question just never change (observed in detail with DM9102A)
and therefore the warning would be highly likely false positive.
- In dc_ifmedia_sts() add a missing DC_UNLOCK().
delphij [Mon, 13 Nov 2006 00:53:04 +0000 (00:53 +0000)]
Since SA-06:24.libarchive affects -STABLE after the latest release
that this errata documents, the change should be documented in
relnotes only, but not this errata.
Discussed with: security-officer (simon), re (bmah)
Approved by: re (bmah)
The ata(4) and isp(4) items are admittedly pretty vague, but I figured
it was important to at least mention that they had been resynced with
CURRENT. Attempts at better text welcomed.
rink [Sat, 11 Nov 2006 10:47:46 +0000 (10:47 +0000)]
MFC: added PCI ID's for:
- 0x1065: Intel 82562ET/EZ/GT/GZ PRO/100 VE Ethernet [1], as found on
Tyan GS14 barebones.
- 0x1094: Intel Pro/100 946GZ (ICH7) Network Connection [2], as found on
Intel 946GZis motherboards.
csjp [Wed, 8 Nov 2006 16:28:34 +0000 (16:28 +0000)]
Log:
Fix possible leak when bridge is in monitor mode. Use m_freem() which will
free the entire chain, instead of using m_free() which will free just the
mbuf that was passed.
dougb [Wed, 8 Nov 2006 01:32:16 +0000 (01:32 +0000)]
MFC the upgrade to BIND 9.3.2-P2. This is mostly a noop in the base,
since if you're compiling the base with this upgraded, you've already
got your openssl upgraded, but it's a good idea to include this for
completeness sake.
Do not translate the IN_ACCESS inode flag into the IN_MODIFIED while filesystem
is suspending/suspended. Doing so may result in deadlock. Instead, set the
(new) IN_LAZYACCESS flag, that becomes IN_MODIFIED when suspend is lifted.
Change the locking protocol in order to set the IN_ACCESS and timestamps
without upgrading shared vnode lock to exclusive (see comments in the
inode.h). Before that, inode was modified while holding only shared
lock.
Tested on RELENG_6 by: Peter Holm
Approved by: re (kensmith)
thompsa [Tue, 7 Nov 2006 16:29:40 +0000 (16:29 +0000)]
MFC r1.84
When the packet is for the bridge then note which interface to send the reply
to, previously it was always broadcast to all interfaces (a bug). This is
useful when the bridge is the default gateway and vlans are used to isolate
each client, the reply is now kept private to the vlan which the client
resides.
marius [Sun, 5 Nov 2006 13:32:58 +0000 (13:32 +0000)]
MFC: 1.16
In the replacement text of the __bswapN_const() macros encapsulate the
argument in parentheses so these macros are safe to use and invocations
with an expression as the argument like __bswap32_const(42 << 23 | 13)
work as expected. Additionally, mask all the individually shifted bytes
as appropriate so the bytes which exceed the width of the respective
__bswapN_const() macro in invocations like __bswap16_const(0xdead600d)
are ignored like it's the case with the corresponding __bswapN_var()
function.
hrs [Sun, 5 Nov 2006 08:50:38 +0000 (08:50 +0000)]
MFC of the following:
- sys/modules/md/Makefile rev 1.14:
Rename GEOM class kernel module g_md.ko to geom_md.ko for
consistency with the rest. The g_md.ko is now a hardlink of
geom_md.ko for backward compatibility.
- sys/sys/mdioctl.h rev 1.21:
Removed (now unused) #define.
- sys/boot/forth/loader.conf rev 1.102:
Add more GEOM classes to defaults/loader.conf. Add references to
manual pages to those already present.
- sbin/mdconfig/mdconfig.c rev 1.45:
Simplify mdmaybeload() function.
Reviewed by: ru and pjd
Approved by: re (kensmith)
kib [Fri, 3 Nov 2006 09:15:55 +0000 (09:15 +0000)]
MFC rev. 1.427:
The attempt to rename "." with MAC framework compiled in would cause attempt
to twice unlock the vnode. Check that ni_vp and ni_dvp are different before
doing second unlock.
trhodes [Fri, 3 Nov 2006 09:00:18 +0000 (09:00 +0000)]
Paritially sync with CURRENT, kills hard sentence breaks, fixes some markup.
MFC: 1.197
Add a note about rule syntax compared to the shell used so users do not get
frustraited when:
ipfw add 201 deny ip from any to table(2) in via xl1
returns "Badly placed ( )'s"
trhodes [Fri, 3 Nov 2006 08:38:26 +0000 (08:38 +0000)]
MFC:
o Note the value from getenv() should not be modified by applications[1];
o getenv.3: Put "is" on a line with other words. [2]
o getobjformat.3: "takes precedence over" is not an envrionment variable. [2]
mohans [Thu, 2 Nov 2006 19:48:17 +0000 (19:48 +0000)]
MFC: Make EWOULDBLOCK a recoverable error so that the request is
retransmitted. This bug results in data corruption. Writes are
silently dropped on EWOULDBLOCK (caused because socket send buffer is
full and sockbuf timer fires - with NFS/TCP).
Reviewed by: ups@
Approved by: re
jhb [Mon, 30 Oct 2006 18:03:04 +0000 (18:03 +0000)]
MFC: Change the x86 interrupt code to suspend/resume interrupt controllers
(PICs) rather than interrupt sources. As a result, always reset the 8259As
on resume and reset the local APIC of the BSP on resume.
keramida [Mon, 30 Oct 2006 16:09:24 +0000 (16:09 +0000)]
MFC revision 1.35 and bump date.
: The 'device fd' line became obsolete some years ago, when the transition
: to device hints was committed. The GENERIC kernel configuration lost
: its 'device fd' line at revision 1.260, committed by peter@ on
: 2000/06/13 22:28:46 UTC.
:
: Delete the reference to 'device fd', to avoid kernel build breakage like
: the one described in the PR audit trail.
:
: PR: docs/91448
: Submitted by: Eugene Grosbein <eugen@grosbein.pp.ru>
: MFC after: 1 week
Resolve the devfs deadlock caused by LOR between devfs_mount->dm_lock
and vnode lock in devfs_allocv. Add hold counters for both devfs_mount
and devfs_dirent, and DE_DOOMED flag for devfs_dirent.