mm [Sat, 16 Mar 2013 08:16:11 +0000 (08:16 +0000)]
MFC r247187,247265,247348,247398,247540,247585,247852,248265,248267
Merge various ZFS improvements and bugfixes
MFC r247187:
Import vendor change to avoid "unitialized variable" warnings.
Illumos ZFS issues:
3522 zfs module should not allow uninitialized variables
MFC r247265:
Merge the ZFS I/O deadman thread from vendor (illumos).
This feature panics the system on hanging ZFS I/O, helps debugging
and resumes failed service.
The panic behavior can be controlled with the loader-only tunables:
vfs.zfs.deadman_enabled (enable or disable panic on stalled ZFS I/O)
vfs.zfs.deadman_synctime (expiration time for stalled ZFS I/O)
By default, ZFS I/O deadman is enabled by default on amd64 and i386
excluding virtual guest machines.
MFC r247348:
Be more verbose on ZFS deadman I/O panic
Patch suggested upstream.
MFC r247398:
Import metaslab_sync() speedup from vendor (illumos).
Illumos ZFS issues:
3552 condensing one space map burns 3 seconds of CPU in spa_sync() thread
3564 spa_sync() spends 5-10% of its time in metaslab_sync() (when not
condensing)
3578 transferring the freed map to the defer map should be constant time
3579 ztest trips assertion in metaslab_weight()
MFC r247540:
Fix the zfs_ioctl compat layer to support zfs_cmd size change introduced
in r247265 (ZFS deadman thread). Both new utilities now support the old
kernel and new kernel properly detects old utilities.
For future backwards compatibility, the vfs.zfs.version.ioctl read-only
sysctl has been introduced. With this sysctl zfs utilities will be able
to detect the ioctl interface version of the currently loaded zfs module.
MFC r247585:
Merge new read-only zfs properties from vendor (illumos)
Illumos ZFS issues:
3588 provide zfs properties for logical (uncompressed) space used and
referenced
MFC r247852:
Import ZFS bpobj bugfix from vendor.
Illumos ZFS issues:
3603 panic from bpobj_enqueue_subobj()
3604 zdb should print bpobjs more verbosely
MFC r248265:
Update zfs.8 manpage date (missing in r247585)
MFC r248267:
Import minor ZFS changes from vendor
Illumos ZFS issues:
3604 zdb should print bpobjs more verbosely (fix zdb hang)
3606 zpool status -x shouldn't warn about old on-disk format
brooks [Fri, 15 Mar 2013 22:31:51 +0000 (22:31 +0000)]
MFC r245887
Only install manpages and html documentation in the ncurses/*w (wchar)
builds so that it is only installed once. This is consistent with the
existing decision to only install headers in the that case.
brooks [Fri, 15 Mar 2013 22:20:45 +0000 (22:20 +0000)]
MFC r245882,245898
Reorder so that NO_MAN is declared before bsd.own.mk is included and thus
has an effect (not installed a g++.1 manpage over the g++.1(.gz) link
created in ../cc).
bsd.own.mk needs to be included before Makefil.inc so MK_ARM_EABI is defined
brooks [Fri, 15 Mar 2013 20:52:00 +0000 (20:52 +0000)]
MFC r245571,245580:
In preparation for logging metadata about each filesystem object,
refactor the link section of distrib-dirs to alwasy install to a full
path (the link contents remain relative as they should).
Eliminate the use of the "rm -r[f] <foo>; ln -s <bar> <foo>" pattern in
favor of "ln -sfh <bar> <foo>". None of these links could be directories
on a system installed in the last decade.
brooks [Fri, 15 Mar 2013 20:26:51 +0000 (20:26 +0000)]
MFC r241311:
Change 32-bit library builds by removing use of make -E and passing AS,
CC, CXX and LD on the commandline of the sub-make instead of in the
environment of the sub-make.
brooks [Fri, 15 Mar 2013 19:27:27 +0000 (19:27 +0000)]
MFC all features as of r248336:
The user visable feature is from r246083:
Add -l option to cat(1). This option causes cat(1) to use fcntl(2) to
set an exclusive advisory lock on stdout. This will be used to guarantee
orderly writing to METALOG.
brooks [Fri, 15 Mar 2013 19:16:35 +0000 (19:16 +0000)]
MFC r245265,245271,245309:
Rather than using zic to both compile and install zoneinfo files,
generate the files during the build and install them with install(1).
This was the one place in installworld where files (vs links) were
installed by a tool other than install.
r241684:
Allow LOCAL_DIRS to install files in directories not found in the
system mtree files via a LOCAL_MTREE variable which contains a list of
mtree files to be applyed along with the base mtree files to the tmp
root and DESTDIR.
r245440:
Add an option DB_FROM_SRC to use src/etc's user/group databases when
installing. This allows things like running installworld for 10-CURRENT
on a 9.0-RELEASE system without adding extra users and groups to the
passwd and group files.
To prevent potentially risky uid/gid mismatches on systems with
non-standard local values, require that DESTDIR be set if DB_FROM_SRC is
set.
Sponsored by: DARPA, AFRL
Reviewed by: peter
r245565:
Rework the mtree portion of etc/Makefile's distrib-dirs target to run
mtree in a shell loop so there is only one mtree commandline. Move the
implementation of LOCAL_MTREE into etc/Makefile.
brooks [Fri, 15 Mar 2013 15:19:33 +0000 (15:19 +0000)]
MFC all changes to install(1) through r246784. Notable functional
changes are:
r245617:
Introduce six new options from NetBSD:
* -M <metalog> Log metadata in mtree format.
* -D <destdir> Log paths relative to <destdir>.
* -h <hash> Log digest of type <hash>.
* -T <tags> Specify which mtree tags to log.
* -l <linkflag> Create hard or symbolic links (allows logging).
* -U Install without root privileges (owner, group, mode,
and flags can be logged via -M
NOTE: In the interest of compatibility with NetBSD and because it is the
obvious letter, the nearly useless -M option (disable mmap) has been
repurposed.
r245312:
Implement the -N <dbdir> option which allows an alternate passwd and
group file to be used. This is useful for installing on systems where
a user or group does not currently exist.
jfv [Thu, 14 Mar 2013 22:55:59 +0000 (22:55 +0000)]
MFC of the E1000 drivers including revisions:
------------------------------------------------------------------------
r238765 | luigi | 2012-07-25 04:28:15 -0700 (Wed, 25 Jul 2012) | 7 lines
Use legacy interrupts as a default. This gives up to 10% speedup
when used in qemu (and this driver is for non-PCIe cards,
so probably its largest use is in virtualized environments).
------------------------------------------------------------------------
r238770 | luigi | 2012-07-25 05:51:33 -0700 (Wed, 25 Jul 2012) | 4 lines
remove some extra testing code that slipped into the previous commit
------------------------------------------------------------------------
r238953 | jfv | 2012-07-31 11:44:10 -0700 (Tue, 31 Jul 2012) | 6 lines
Clean up some unused leftover code from em
Make IRQ style a tuneable
Fix lock handling in the interrupt handler
------------------------------------------------------------------------
r238981 | sbruno | 2012-08-01 17:00:34 -0700 (Wed, 01 Aug 2012) | 9 lines
CPU_NEXT() already handles wrapping around to the beginning. Also, in a
system with sparse CPU IDs, you can have a valid CPU ID > mp_ncpus (e.g. if
you have two CPUs 0 and 4, with mp_maxid == 4 and mp_ncpus == 2).
------------------------------------------------------------------------
r239105 | jfv | 2012-08-06 13:44:05 -0700 (Mon, 06 Aug 2012) | 5 lines
Correct the mq_start routine to avoid out-of-order
packet delivery, always enqueue when possible. Also
correct the DEPLETED test as multiple bits might be
set. Thanks to Randall Stewart for the changes!
------------------------------------------------------------------------
r239109 | jfv | 2012-08-06 15:43:49 -0700 (Mon, 06 Aug 2012) | 6 lines
Make the polling interface in igb able to handle
multiqueue, and correct the rxdone handling. Update
the polling man page to include igb as well.
------------------------------------------------------------------------
r239304 | jfv | 2012-08-15 10:12:40 -0700 (Wed, 15 Aug 2012) | 10 lines
Customer report of a panic on boot due to the old
"m_getjcl:invalid cluster type" that occurred some
time back with the igb driver. This happens often when
booting over the net. I believe the NIC hardware is left
in a warm state when handed over to the driver, and a stray
RX interrupt happens earlier than the code is prepared for
it to happen. This change was verified to fix the problem,
its kind of a bandaid... but it is similar to what was done
in the igb code.
------------------------------------------------------------------------
r240693 | gavin | 2012-09-19 05:27:23 -0700 (Wed, 19 Sep 2012) | 5 lines
Switch some PCI register reads from using magic numbers to using the names
defined in pcireg.h
------------------------------------------------------------------------
r241856 | eadler | 2012-10-21 20:41:14 -0700 (Sun, 21 Oct 2012) | 7 lines
Now that device disabling is generic, remove extraneous code from the
device drivers that used to provide this feature.
------------------------------------------------------------------------
r241885 | eadler | 2012-10-22 06:06:09 -0700 (Mon, 22 Oct 2012) | 7 lines
This isn't functionally identical. In some cases a hint to disable
unit 0 would in fact disable all units. This reverts r241856
------------------------------------------------------------------------
r243570 | glebius | 2012-11-26 12:03:57 -0800 (Mon, 26 Nov 2012) | 14 lines
drbr_enqueue() awlays consumes mbuf, no matter did it
fail or not. The mbuf pointer is no longer valid, so
can't be reused after.
Fix igb_mq_start() where mbuf pointer was used after
drbr_enqueue().
This eventually leads us to all invocations of
igb_mq_start_locked() called with third argument as NULL.
This allows us to simplify this function.
------------------------------------------------------------------------
r245334 | smh | 2013-01-12 08:05:55 -0800 (Sat, 12 Jan 2013) | 9 lines
Fixed mbuf free when receive structures fail to allocate.
This prevents quad igb card on high core machines, without any nmbcluster or
igb queue tuning wedging the boot process if all nics are configured.
------------------------------------------------------------------------
r246128 | sbz | 2013-01-30 10:01:20 -0800 (Wed, 30 Jan 2013) | 5 lines
Use DEVMETHOD_END macro defined in sys/bus.h instead of {0, 0} sentinel on device_method_t arrays
------------------------------------------------------------------------
r246482 | rrs | 2013-02-07 07:20:54 -0800 (Thu, 07 Feb 2013) | 30 lines
This fixes a out-of-order problem with several
of the newer drivers. The basic problem was
that the driver was pulling the mbuf off the
drbr ring and then when sending with xmit(), encounting
a full transmit ring. Thus the lower layer
xmit() function would return an error, and the
drivers would then append the data back on to the ring.
For TCP this is a horrible scenario sure to bring
on a fast-retransmit.
The fix is to use drbr_peek() to pull the data pointer
but not remove it from the ring. If it fails then
we either call the new drbr_putback or drbr_advance
method. Advance moves it forward (we do this sometimes
when the xmit() function frees the mbuf). When
we succeed we always call advance. The
putback will always copy the mbuf back to the top
of the ring. Note that the putback *cannot* be used
with a drbr_dequeue() only with drbr_peek(). We most
of the time, in putback, would not need to copy it
back since most likey the mbuf is still the same, but
sometimes xmit() functions will change the mbuf via
a pullup or other call. So the optimial case for
the single consumer is to always copy it back. If
we ever do a multiple_consumer (for lagg?) we
will need a test and atomic in the put back possibly
a seperate putback_mc() in the ring buf.
------------------------------------------------------------------------
r247064 | jfv | 2013-02-20 16:25:45 -0800 (Wed, 20 Feb 2013) | 19 lines
Refresh on the shared code for the E1000 drivers.
- bear with me, there are lots of white space changes, I would not
do them, but I am a mere consumer of this stuff and if these drivers
are to stay in shape they need to be taken.
em driver changes: support for the new i217/i218 interfaces
igb driver changes:
- TX mq start has a quick turnaround to the stack
- Link/media handling improvement
- When link status changes happen the current flow control state
will now be displayed.
- A few white space/style changes.
lem driver changes:
- the shared code uncovered a bogus write to the RLPML register
(which does not exist in this hardware) in the vlan code,this
is removed.
------------------------------------------------------------------------
Note: 246482 includes only the changes in the ixgbe driver
and the buf_ring supporting code in if_var.h and buf_ring.h
igb and em will be seperate commits, and non-Intel drivers
can commit the changes at will.
dim [Thu, 14 Mar 2013 21:18:19 +0000 (21:18 +0000)]
MFC r247960:
Make ctfconvert work correctly on clang-compiled object files. Clang
puts the full original source filename in the STT_FILE entry of the ELF
symbol table, while gcc saves only the basename.
Since the DWARF DW_AT_name attribute contains the full source filename,
both for clang and gcc, ctfconvert takes just the basename of it, for
matching with the STT_FILE entry. So when attempting to match with such
an entry, use its basename, if necessary.
Reported by: avg
MFC r247962:
Fix error in r247960: actually assign the basename to match.iim_file.
gibbs [Wed, 13 Mar 2013 17:37:13 +0000 (17:37 +0000)]
MFC r247048.
Orphaned processes that are being traced are killed by the
kernel. Properly restore, continue, and detach from processes
being DTraced when DTrace exits with an error so the program
being inspected is not terminated.
cddl/contrib/opensolaris/cmd/dtrace/dtrace.c:
In fatal(), the generic error handler, close the DTrace
handle as is done in the "probe/script" error handler
dfatal(). fatal() can be invoked after DTrace attaches
to processes (e.g. a script specified by command line
argument can't be found) and closing the handle will
release them.
gibbs [Wed, 13 Mar 2013 17:34:42 +0000 (17:34 +0000)]
MFC kernel fixes to userland dtrace support.
r247049
-------
Avoid panic when tearing down the DTrace pid provider for a
process that has crashed.
sys/cddl/contrib/opensolaris/uts/common/dtrace/fasttrap.c:
In fasttrap_pid_disable(), we cannot PHOLD the proc
structure for a process that no longer exists, but
we still have other, fasttrap specific, state that
must be cleaned up for probes that existed in the
dead process. Instead of returning early if the
process related to our probes isn't found,
conditionalize the locking and carry on with a NULL
proc pointer. The rest of the fasttrap code already
understands that a NULL proc is possible and does
the right things in this case.
r247820
-------
Fix assertion failure when using userland DTrace probes from
the pid provider on a kernel compiled with INVARIANTS.
sys/cddl/contrib/opensolaris/uts/intel/dtrace/fasttrap_isa.c:
In fasttrap_probe_pid(), attempts to write to the
address space of the thread that fired the probe
must be performed with the process of the thread
held. Use _PHOLD() to ensure this is the case.
In fasttrap_probe_pid(), use proc_write_regs() instead
of calling set_regs() directly. proc_write_regs()
performs invariant checks to verify the calling
environment of set_regs(). PROC_LOCK()/UNLOCK() around
the call to proc_write_regs() so that it's invariants
are satisfied.
nwhitehorn [Wed, 13 Mar 2013 13:45:47 +0000 (13:45 +0000)]
MFC r245680:
Make "Finish" the default choice in the partition editor. This lets you
successfully complete an installation with all defaults by pressing
Enter repeatedly until your machine reboots.
nwhitehorn [Wed, 13 Mar 2013 13:41:28 +0000 (13:41 +0000)]
MFC miscellaneous bug fixes (r228042,228192,230522,232531,233473,233904,
235228,239500,241902,242188,243832,243854,244858,245333). This is mostly
documentation fixes and updates to the default FTP list that got left
behind in -CURRENT.
lstewart [Mon, 11 Mar 2013 06:09:08 +0000 (06:09 +0000)]
MFC r247906:
The hashmask returned by hashinit() is a valid index in the returned hash array.
Fix a siftr(4) potential memory leak and INVARIANTS triggered kernel panic in
hashdestroy() by ensuring the last array index in the flow counter hash table is
flushed of entries.
dim [Sun, 10 Mar 2013 15:02:30 +0000 (15:02 +0000)]
MFC r247953:
Make c99(1) invoke /usr/bin/cc with argv[0] set to "/usr/bin/cc" instead
of just "cc", since there is no reason to cause additional path searches
in this case.
dim [Sun, 10 Mar 2013 15:01:06 +0000 (15:01 +0000)]
MFC r247951:
Make c89(1) invoke /usr/bin/cc with argv[0] also set to /usr/bin/cc,
similar to what c99(1) does, to prevent "c89: illegal option -- 1"
messages, when clang is /usr/bin/cc.
marius [Sat, 9 Mar 2013 02:36:32 +0000 (02:36 +0000)]
MFC: r227309 (partial)
Mark all SYSCTL_NODEs static that have no corresponding SYSCTL_DECLs.
The SYSCTL_NODE macro defines a list that stores all child-elements of
that node. If there's no SYSCTL_DECL macro anywhere else, there's no
reason why it shouldn't be static.
rstone [Fri, 8 Mar 2013 21:07:01 +0000 (21:07 +0000)]
MFC r244631
Correct a series of errors in the hand-rolled locking for drace_debug.c:
- Use spinlock_enter()/spinlock_exit() to prevent a thread holding a
debug lock from being preempted to prevent other threads waiting
on that lock from starvation.
- Handle the possibility of CPU migration in between the fetch of curcpu
and the call to spinlock_enter() by saving curcpu in a local variable.
- Use memory barriers to prevent reordering of loads and stores of the
data protected by the lock outside of the critical section
- Eliminate false sharing of the locks by moving them into the structures
that they protect and aligning them to a cacheline boundary.
- Record the owning thread in the lock to make debugging future problems
easier.
rstone [Fri, 8 Mar 2013 19:36:36 +0000 (19:36 +0000)]
MFC r241009
Ensure that all cases that enqueue a netgraph item for delivery by a
ngthread properly set the item's depth to 1. In particular, prior to this
change if ng_snd_item failed to acquire a lock on a node, the item's depth
would not be set at all. This fix ensures that the error code from rcvmsg/
rcvdata is properly passed back to the apply callback. For example, this
fixes a bug where an error from rcvmsg/rcvdata would not previously
propagate back to a libnetgraph consumer when the message was queued.
rstone [Fri, 8 Mar 2013 19:01:26 +0000 (19:01 +0000)]
MFC r240923
Some aac(4) adapters will always report that a direct access device is
offline in response to a INQUIRY command that does not retreive vital
product data(I personally have observed the behaviour on an Adaptec 2405
and a 5805). Force the peripheral qualifier to "connected" so that upper
layers correctly recognize that a disk is present.
This bug was uncovered by r216236. Prior to that fix, aac(4) was
accidentally clearing the peripheral qualifier for all inquiry commands.
This fixes an issue where passthrough devices were not created for
disks behind aac(4) controllers suffering from the bug. I have
verified that if a disk is not present that we still properly detect
that and not create the passthrough device.
marius [Fri, 8 Mar 2013 12:57:58 +0000 (12:57 +0000)]
MFC: r247632
- Complete r231621 (MFC'ed to stable/9 in r232092) by also blacklisting the
bridge used by VMware for PCIe devices. While at it, update the comment
now that we know that MSI-X doesn't work with ESXi 5.1 for Intel 82576
either and the underlying issue is a bug in the MSI-X allocation code of
the hypervisor.
Reported by: Harald Schmalzbauer
- Make the nomatch table const.
marius [Fri, 8 Mar 2013 12:50:29 +0000 (12:50 +0000)]
MFC: r247600, r247620
- While Netra X1 generally show no ill effects when registering a power
fail interrupt handler, there seems to be either a broken batch of them
or a tendency to develop a defect which causes this interrupt to fire
inadvertedly. Given that apart from this problem these machines work
just fine, add a tunable allowing the setup of the power fail interrupt
to be disabled.
While at it, remove the DEBUGGER_ON_POWERFAIL compile time option and
make that behavior also selectable via the newly added tunable.
- Correct an incorrect argument to shutdown_nice(9).
- Mark unused parameters as such.
- Use NULL instead of 0 for pointers.
marius [Fri, 8 Mar 2013 12:44:08 +0000 (12:44 +0000)]
MFC: r247579
- Move reporting of failures to disable RX/TX MAC under bootverbose as at
least the Saturn chips of 501-6738 cards may fail to do so the first
time, which isn't fatal though.
Reported by: Paul Keusemann
- Explain why we don't enable infinite bursts on sparc64.
- Given that these chips support memory write invalidate, make sure that
it's enabled in the command register. Also make sure that PERR# and
SERR# assertion is enabled.
marius [Fri, 8 Mar 2013 12:11:37 +0000 (12:11 +0000)]
MFC: r247573
- Remove an unused header.
- Use NULL instead of 0 for pointers.
- Let ofw_pcib_probe() return BUS_PROBE_DEFAULT instead of 0 so specialized
PCI-PCI-bridge drivers may attach instead.
- Add WARs for PLX Technology PEX 8114 bridges and PEX 8532 switches.
Ideally, these should live in MI code but at least for the latter we're
missing the necessary infrastructure there.
marius [Fri, 8 Mar 2013 11:41:48 +0000 (11:41 +0000)]
MFC: r247571
- Apparently, r186520 was just wrong and the clock of Oxford OX16PCI958 is
neither DEFAULT_RCLK * 2 nor DEFAULT_RCLK * 10 but plain DEFAULT_RCLK
and there's no (open) source indicating otherwise. This was tested with
an EXSYS EX-41098-2, whose clock is not configurable and identifies as:
puc0@pci0:5:1:0: class=0x070200 card=0x06711415 chip=0x95381415 rev=0x01 hdr=0x00
vendor = 'Oxford Semiconductor Ltd'
class = simple comms
subclass = multiport serial
Note that this exactly matches the card mentioned in PR 129665 so no
sub-device/sub-vendor based quirking of the latter is possible. So maybe
we should grow some sort of tunable, in case non-default cards such as
the latter aren't configurable either (this also wouldn't be the first
time an allegedly tested commit turns out to be wrong though).
- Make the TiMedia tables const.
marius [Fri, 8 Mar 2013 11:32:58 +0000 (11:32 +0000)]
MFC: r247565, r247590
- Make tables, device ID strings etc const.
- Use NULL instead of 0 for pointers.
- Remove redundant bzero(9)'ing of the softc.
- Remove redundant/unused softc members.
- Don't allocate MSI/MSI-X as RF_SHAREABLE.
- Re-use bus accessor macros instead of duplicating them.
- In bce_miibus_{read,write}_reg(), remove superfluous limiting of the PHY
address (missed in r213893).
yongari [Wed, 6 Mar 2013 08:03:26 +0000 (08:03 +0000)]
MFC r247382:
RX checksum offloading on old Yukon controllers seem to cause more
problems. Disable RX checksum offloading on controllers that don't
use new descriptor format but give chance to enable it with
ifconfig(8).
ian [Mon, 4 Mar 2013 16:50:01 +0000 (16:50 +0000)]
MFC r246134 ...
Improve devd startup time, by tweaking some string handling routines that are
heavily used when parsing config files. Mostly these changes avoid making
temporary copies of the strings, and avoid doing byte at a time append
operations, on the most-used code path.
On a 1.2 GHz ARM processor this reduces the time to parse the config files
from 13 to 6 seconds.
gallatin [Mon, 4 Mar 2013 15:30:38 +0000 (15:30 +0000)]
MFC r247159
Improvements for newer mxge nics:
- Some mxge nics may store the serial number in the SN2 field of the
EEPROM. These will also have an SN=0 field, so parse the SN2 field,
and give it precedence.
- Skip MXGEFW_CMD_UNALIGNED_TEST on mxge nics which do not require it.
This saves roughly 10ms per port at device attach time.
markj [Sun, 3 Mar 2013 19:02:23 +0000 (19:02 +0000)]
MFC r246978:
- Make sure to set an error code when trying to obtain a data descriptor for
a section of type SHT_NULL.
- Update the man page to reflect the fact that elf_getdata() and
elf_rawdata() may return with an error of ELF_E_SECTION.
dteske [Sun, 3 Mar 2013 17:39:29 +0000 (17:39 +0000)]
MFC r230005:
Use a reasonable-sized buffer when formatting error messages about
installing zoneinfo. While we're in the vicinity, add some missing
error checking to eliminate an unhelpful error message when unlink()
fails.
/me is embarrassed by the quality of his 16-year-old code.
The whole thing is awful and could stand a complete rewrite.