Oleg Bulyzhin [Thu, 24 Aug 2006 14:41:16 +0000 (14:41 +0000)]
Properly lock ifmedia callbacks. This should prevent concurrent access to PHY.
Following issues should be resolved:
- random watchdog timeouts (caused by concurrent phy access)
- some link state issues
- non working TX if media type was set explicitly
John Baldwin [Wed, 23 Aug 2006 19:16:17 +0000 (19:16 +0000)]
Remove special handling for PC == 0. With this, kgdb can now properly
unwind across a page fault due to a null function pointer. It does a
better job than ddb now in fact.
Thomas Quinot [Wed, 23 Aug 2006 15:59:43 +0000 (15:59 +0000)]
Set alarm timer for grace period from the grace_period variable, instead
of hard-coding a value of 10 seconds. Command line flag -g is thus now
correctly taken into account.
Thomas Quinot [Tue, 22 Aug 2006 23:49:36 +0000 (23:49 +0000)]
(media_status): Factor common code between IFM_ETHER and IFM_ATM cases.
(print_media_word, print_media_word_ifconfig): Remove unnecessary
goto following test for null desc.
Florent Thoumie [Tue, 22 Aug 2006 11:17:29 +0000 (11:17 +0000)]
Backout this commit since it breaks startup and some scripts in
certain conditions. I haven't been able to find a better solution yet:
- Set a two read-only variables (${prefix} and ${etcdir}). This is
especially useful when using /etc/rc.d scripts with third-party
software installed from ports.
- Fix rc.d/sshd to work with openssh from ports using ${etcdir}
instead of hardcoded /etc.
- Reflect prefix/etcdir changes in rc.subr.8.
Marcel Moolenaar [Tue, 22 Aug 2006 03:05:51 +0000 (03:05 +0000)]
Fix misalignment bugs caused by invalid type casts of pointers
returned by md_reserve(). Space reserved by mb_reserve() is
byte aligned and need to be used in conjunction with le16enc()
and le32enc().
Pyun YongHyeon [Tue, 22 Aug 2006 02:32:48 +0000 (02:32 +0000)]
It seems that em(4) misses Tx completion interrupts under certain
conditions. The cause of missing Tx completion interrupts comes from
Tx interrupt moderation mechanism(delayed interrupts) or chipset bug.
If Tx interrupt moderation mechanism is the cause of false watchdog
timeout error we should have to fix all device drivers that have Tx
interrupt moderation capability. We may need more investigation
for this issue. Anyway, the fix is the same for both cases.
This should fix occasional watchdog timeout errors seen on a few
systems.
Reported by: -net, Patrick M. Hausen < hausen AT punkt DOT de >
Tested by: Patrick M. Hausen < hausen AT punkt DOT de >
Jung-uk Kim [Mon, 21 Aug 2006 19:10:58 +0000 (19:10 +0000)]
Use aperture base address from north bridge. Some BIOS does not encode
misc. control registers correctly and it is inconsistent with north bridge.
In fact, there are too many broken BIOS implementations out there and we
cannot fix every possible combination but at least it is consistent with
what we advertise with ioctl(2).
Marcel Moolenaar [Mon, 21 Aug 2006 18:25:43 +0000 (18:25 +0000)]
Build gperf(1) for/on PowerPC. There's no apparent reason not to
build it and nothing can be found in the commit logs. We won't
know if there are runtime problems, until we build it...
While checking for update of snapshot file in the ffs_copyonwrite,
first filter out metadata update. Otherwise, devfs vnode could be
erronously interpreted as ufs one, causing further check of i_flags
to use random memory.
PR: kern/100365
Debugged and fix described by: tegge
Approved by: pjd (mentor)
MFC after: 2 weeks
Brian Somers [Mon, 21 Aug 2006 16:31:31 +0000 (16:31 +0000)]
Revert the addition of -p. It's flawed in that dhclient should not run
on an interface without carrier. devd should be used instead to handle
link up/down events.
Olivier Houchard [Mon, 21 Aug 2006 14:42:47 +0000 (14:42 +0000)]
If __ARMEB__ is defined, we're already using the network byte order, so
there's no need to to anything in the hton* functions, beside returning
the parameter.
Spotted out by: Oleksandr Tymoshenko <gonzo@freebsd.org>
Kenneth D. Merry [Mon, 21 Aug 2006 13:24:50 +0000 (13:24 +0000)]
Implement 'camcontrol reportluns'. This allows users to send the SCSI
REPORT LUNS command to a device.
camcontrol.[c8]: Implement reportluns. This tries to print the LUNs
out in a reasonable format. Only the periph
addressing method has been tested, since very little
hardware that I know of supports the other methods.
scsi_all.[ch]: Revamp the report luns CDB structure and helper
functions. This constitutes a little bit of an API
change, but since the old CDB length was 10 bytes,
and the REPORT LUNS CDB length is actually 12 bytes,
it's clear that no one was using this API in the
first place.
Yaroslav Tykhiy [Mon, 21 Aug 2006 03:09:12 +0000 (03:09 +0000)]
For the sake of clarity, explicitly tell that comma and space
characters can be included in a custom column title string --
that's why it may appear at the end of a keyword list argument
only.
Matt Jacob [Mon, 21 Aug 2006 00:46:10 +0000 (00:46 +0000)]
Fix RELENG_4 code version- isp_roles wasn't getting initialized so
it ended up defaulting to ISP_ROLE_NONE. My testing hadn't caught it
because I was deliberatly setting role via ioctl.
Thanks to user Toni for lending me an alpha to test this on.
David Malone [Sun, 20 Aug 2006 20:10:36 +0000 (20:10 +0000)]
Regigle parens to try and get the intended affect. This should fix people
having trouble with the "me6" keyword. Also, we were using inet_pton on
the wrong variable in one place.
Reviewed by: mlaier (previous version of patch)
Obtained from: Sascha Blank (inet_pton change)
MFC after: 1 week
Sync the MI parts for amd64 with i386 and remove the corresponding special
handling for amd64 in the common code. The MD parts for amd64 are still
outstanding, but at least this fixes some panics on amd64.
Warner Losh [Sun, 20 Aug 2006 06:32:40 +0000 (06:32 +0000)]
Prefer strlcpy to strncpy. In one case, this saves us from re-zeroing
data that's alreday 0. In another, it saves us from zeroing data that
will be overwritten again.
Yaroslav Tykhiy [Sat, 19 Aug 2006 08:24:01 +0000 (08:24 +0000)]
Install pkill(1), aka pgrep(1), to /bin so that rc scripts
can use this small and nifty utility. Create compatibility
symlinks from /usr/bin for the time being to avoid breaking
custom scripts relying on the hardcoded path to the utility.
If pkill(1) takes root, its source should be repocopied some
day to src/bin.
Idea by: des
Discussed with: brooks (in cvs-src and cvs-all)
Yaroslav Tykhiy [Fri, 18 Aug 2006 14:12:38 +0000 (14:12 +0000)]
Document a detail in the present syntax of "ps -o foo=bar". Now
the header title string "bar" extends to the end of the argument
even if it contains commas or spaces, unlike in RELENG_4 or earlier.
The modern syntax agrees with SUSv3.
Yaroslav Tykhiy [Fri, 18 Aug 2006 12:10:18 +0000 (12:10 +0000)]
The ps(1) command is unfriendly to scripts by default because
it limits the width of its output to the value of $COLUMNS, or
what TIOCGWINSZ reports, or 79 columns. We should specify -ww
to ps(1) so that it removes the limit and prints lines in full.
Otherwise very long command pathnames could be mishandled, e.g.,
by _find_processes().
Follow the lead of mxgb(4) and build the module for ixgb(4).
Originally it wasn't enabled since the hardware wasn't commonplace,
but as 10GE hardware is becoming more widely used, building the module
by default should be beneficial.
Julian Elischer [Thu, 17 Aug 2006 22:49:50 +0000 (22:49 +0000)]
Allow ipfw to forward to a destination that is specified by a table.
for example:
fwd tablearg ip from any to table(1)
where table 1 has entries of the form:
1.1.1.0/24 10.2.3.4
208.23.2.0/24 router2
This allows trivial implementation of a secondary routing table implemented
in the firewall layer.
I expect more work (under discussion with Glebius) to follow this to clean
up some of the messy parts of ipfw related to tables.