marius [Wed, 30 May 2012 20:04:07 +0000 (20:04 +0000)]
Try to finally get the point in time at which bge_add_sysctls() is called
right; it needs to be called before bge_can_use_msi() but in turn requires
bge_flags to be properly set.
raj [Wed, 30 May 2012 18:05:48 +0000 (18:05 +0000)]
Panic openly if we cannot retrieve memory information from the device tree.
This is a critical condition and can lead to all sorts of misterious hangs if
not handled.
kib [Wed, 30 May 2012 16:42:08 +0000 (16:42 +0000)]
vn_io_fault() is a facility to prevent page faults while filesystems
perform copyin/copyout of the file data into the usermode
buffer. Typical filesystem hold vnode lock and some buffer locks over
the VOP_READ() and VOP_WRITE() operations, and since page fault
handler may need to recurse into VFS to get the page content, a
deadlock is possible.
The facility works by disabling page faults handling for the current
thread and attempting to execute i/o while allowing uiomove() to
access the usermode mapping of the i/o buffer. If all buffer pages are
resident, uiomove() is successfull and request is finished. If EFAULT
is returned from uiomove(), the pages backing i/o buffer are faulted
in and held, and the copyin/out is performed using uiomove_fromphys()
over the held pages for the second attempt of VOP call.
Since pages are hold in chunks to prevent large i/o requests from
starving free pages pool, and since vnode lock is only taken for
i/o over the current chunk, the vnode lock no longer protect atomicity
of the whole i/o request. Use newly added rangelocks to provide the
required atomicity of i/o regardind other i/o and truncations.
Filesystems need to explicitely opt-in into the scheme, by setting the
MNTK_NO_IOPF struct mount flag, and optionally by using
vn_io_fault_uiomove(9) helper which takes care of calling uiomove() or
converting uio into request for uiomove_fromphys().
kib [Wed, 30 May 2012 16:06:38 +0000 (16:06 +0000)]
Add a rangelock implementation, intended to be used to range-locking
the i/o regions of the vnode data space. The implementation is quite
simple-minded, it uses the list of the lock requests, ordered by
arrival time. Each request may be for read or for write. The
implementation is fair FIFO.
alc [Tue, 29 May 2012 15:41:20 +0000 (15:41 +0000)]
Rename pmap_collect() to pmap_pv_reclaim() and rewrite it such that it no
longer uses the active and inactive paging queues. Instead, the pmap now
maintains an LRU-ordered list of pv entry pages, and pmap_pv_reclaim() uses
this list to select pv entries for reclamation.
Note: The old pmap_collect() tried to avoid reclaiming mappings for pages
that have either a hold_count or a busy field that is non-zero. However,
this isn't necessary for correctness, and the locking in pmap_collect() was
insufficient to guarantee that such mappings weren't reclaimed. The new
pmap_pv_reclaim() doesn't even try.
mav [Tue, 29 May 2012 14:28:46 +0000 (14:28 +0000)]
Allow to change number of openings (used tags) for ATA/SATA devices
via `camcontrol tags ... -N ...`. There is no need to tune it in
usual cases, but some users want to have it for debugging purposes.
imp [Tue, 29 May 2012 03:23:18 +0000 (03:23 +0000)]
Compute the master clock frequency, so we no longer need to have it
compiled into the kernel. This allows us to boot the same kernel on
machines with different master clock frequencies, so long as we can
determine the main clock frequency accurately. Cleanup the pmc clock
init function so it can be called in early boot so we can use the
serial port just after we call cninit.
# We have two calls to at91_pmc_clock_init for reasons unknown, that will
# be fixed later -- it is harmless for now.
alc [Tue, 29 May 2012 01:52:38 +0000 (01:52 +0000)]
Replace all uses of the vm page queues lock by a r/w lock that is private
to this pmap.c. This new r/w lock is used primarily to synchronize access
to the TTE lists. However, it will be used in a somewhat unconventional
way. As finer-grained TTE list locking is added to each of the pmap
functions that acquire this r/w lock, its acquisition will be changed from
write to read, enabling concurrent execution of the pmap functions with
finer-grained locking.
Reviewed by: attilio
Tested by: flo
MFC after: 10 days
jilles [Mon, 28 May 2012 19:22:23 +0000 (19:22 +0000)]
libfetch: Avoid SIGPIPE on network connections.
To avoid unexpected process termination from SIGPIPE when writing to a
closed network connection, enable SO_NOSIGPIPE on all network connections.
The POSIX standard MSG_NOSIGNAL is not used since it requires modifying all
send calls to add this flag. This is particularly nasty for SSL connections.
Reviewed by: des
Tested by: bapt
MFC after: 5 days
marius [Mon, 28 May 2012 17:58:10 +0000 (17:58 +0000)]
- Correct the comments regarding the sizes of some partitions in the
DataFlash.
- Add a mapping for the Nut/OS configuration DataFlash partition according
to the board manual (but not known to either Linux or U-Boot (patches).
kib [Mon, 28 May 2012 13:58:08 +0000 (13:58 +0000)]
Disable end of buffer fixup by default. New DDX does not need this, and
since batch_len is unused by Linux driver, it seems that it is sometimes
gets passed wrong. This causes command buffer corruption and GPU hung.
Old GEMified DDX drivers that needs this workaround are not supported.
rea [Mon, 28 May 2012 12:13:04 +0000 (12:13 +0000)]
if_lagg: allow to invoke SIOCSLAGGPORT multiple times in a row
Currently, 'ifconfig laggX down' does not remove members from this
lagg(4) interface. So, 'service netif stop laggX' followed by
'service netif start laggX' will choke, because "stop" will leave
interfaces attached to the laggX and ifconfig from the "start" will
refuse to add already-existing interfaces.
The real-world case is when I am bundling together my Ethernet and
WiFi interfaces and using multiple profiles for accessing network in
different places: system being booted up with one profile, but later
this profile being exchanged to another one, followed by 'service
netif restart' will not add WiFi interface back to the lagg: the
"stop" action from 'service netif restart' will shut down my main WiFi
interface, so wlan0 that exists in the lagg0 will be destroyed and
purged from lagg0; the "start" action will try to re-add both
interfaces, but since Ethernet one is already in lagg0, ifconfig will
refuse to add the wlan0 from WiFi interface.
Since adding the interface to the lagg(4) when it is already here
should be an idempotent action: we're really not changing anything,
so this fix doesn't change the semantics of interface addition.
theraven [Mon, 28 May 2012 12:11:00 +0000 (12:11 +0000)]
Correctly export operator new / delete for things linking against libsupc++ but
not libstdc++.
Unfortunately, it appears that libsupc++ / libstdc++ have a different idea of
the type of size_t to the rest of the world, which may cause problems later
on...
bz [Mon, 28 May 2012 09:30:13 +0000 (09:30 +0000)]
It turns out that too many drivers are not only parsing the L2/3/4
headers for TSO but also for generic checksum offloading. Ideally we
would only have one common function shared amongst all drivers, and
perhaps when updating them for IPv6 we should introduce that.
Eventually we should provide the meta information along with mbufs to
avoid (re-)parsing entirely.
To not break IPv6 (checksums and offload) and to be able to MFC the
changes without risking to hurt 3rd party drivers, duplicate the v4
framework, as other OSes have done as well.
Introduce interface capability flags for TX/RX checksum offload with
IPv6, to allow independent toggling (where possible). Add CSUM_*_IPV6
flags for UDP/TCP over IPv6, and reserve further for SCTP, and IPv6
fragmentation. Define CSUM_DELAY_DATA_IPV6 as we do for legacy IP and
add an alias for CSUM_DATA_VALID_IPV6.
This pretty much brings IPv6 handling in line with IPv4.
TSO is still handled in a different way and not via if_hwassist.
Update ifconfig to allow (un)setting of the new capability flags.
Update loopback to announce the new capabilities and if_hwassist flags.
Individual driver updates will have to follow, as will SCTP.
alc [Sun, 27 May 2012 16:24:00 +0000 (16:24 +0000)]
Replace all uses of the vm page queues lock by a r/w lock that is private
to this pmap.c. This new r/w lock is used primarily to synchronize access
to the PV lists. However, it will be used in a somewhat unconventional
way. As finer-grained PV list locking is added to each of the pmap
functions that acquire this r/w lock, its acquisition will be changed from
write to read, enabling concurrent execution of the pmap functions with
finer-grained locking.
emaste [Sun, 27 May 2012 16:16:28 +0000 (16:16 +0000)]
Add IPPROTO_MPLS (rfc4023) IP protocol definition
There are currently no in-tree consumers; I'm adding it now for use by
vendor code. This matches the change OpenBSD made while implementing
MPLS in gif(4).
marius [Sun, 27 May 2012 16:08:58 +0000 (16:08 +0000)]
- Fix some typos in mmc_acquire_bus() and mmc_send_csd().
- Fix some math errors in mmc_decode_csd_sd().
- Fix incorrect arguments to mmc_send_app_op_cond() in mmc_go_discovery().
- Add reporting of CSD for debug purposes.
- Add detection (and skipping) of password-locked cards.
- Add setting of block length on card if necessary.
mm [Sun, 27 May 2012 16:00:00 +0000 (16:00 +0000)]
Import illumos changeset 13570:3411fd5f1589
1948 zpool list should show more detailed pool information
Display per-vdev information with "zpool list -v".
The added expandsize property has currently no value on FreeBSD.
This changeset allows adding expansion support to individual vdevs
in the future.
rea [Sun, 27 May 2012 06:53:35 +0000 (06:53 +0000)]
OpenSSH: allow VersionAddendum to be used again
Prior to this, setting VersionAddendum will be a no-op: one will
always have BASE_VERSION + " " + VERSION_HPN for VersionAddendum
set in the config and a bare BASE_VERSION + VERSION_HPN when there
is no VersionAddendum is set.
HPN patch requires both parties to have the "hpn" inside their
advertized versions, so we add VERSION_HPN to the VERSION_BASE
if HPN is enabled and omitting it if HPN is disabled.
VersionAddendum now uses the following logics:
* unset (default value): append " " and VERSION_ADDENDUM;
* VersionAddendum is set and isn't empty: append " "
and VersionAddendum;
* VersionAddendum is set and empty: don't append anything.
Approved by: des
Reviewed by: bz
MFC after: 3 days
ken [Sun, 27 May 2012 06:11:09 +0000 (06:11 +0000)]
Work around a race condition in devfs by changing the way closes
are handled in most CAM peripheral drivers that are not handled by
GEOM's disk class.
The usual character driver open and close semantics are that the
driver gets N open calls, but only one close, when the last caller
closes the device.
CAM peripheral drivers expect that behavior to be honored to the
letter, and the CAM peripheral driver code (specifically
cam_periph_release_locked_busses()) panics if it is done incorrectly.
Since devfs has to drop its locks while it calls a driver's close
routine, and it does not have a way to delay or prevent open calls
while it is calling the close routine, there is a race.
The sequence of events, simplified a bit, is:
- devfs acquires a lock
- devfs checks the reference count, and if it is 1, continues to close.
- devfs releases the lock
- 2nd process open call on the device happens here
- devfs calls the driver's close routine
- devfs acquires a lock
- devfs decrements the reference count
- devfs releases the lock
- 2nd process close call on the device happens here
At the second close, we get a panic in
cam_periph_release_locked_busses(), complaining that peripheral
has been released when the reference count is already 0. This is
because we have gotten two closes in a row, which should not
happen.
The fix is to add the D_TRACKCLOSE flag to the driver's cdevsw, so
that we get a close() call for each open(). That does happen
reliably, so we can make sure that our reference counts are
correct.
Note that the sa(4) and pt(4) drivers only allow one context
through the open routine. So these drivers aren't exposed to the
same race condition.
scsi_ch.c,
scsi_enc.c,
scsi_enc_internal.h,
scsi_pass.c,
scsi_sg.c:
For these drivers, change the open() routine to
increment the reference count for every open, and
just decrement the reference count in the close.
Call cam_periph_release_locked() in some scenarios
to avoid additional lock and unlock calls.
scsi_pt.c: Call cam_periph_release_locked() in some scenarios
to avoid additional lock and unlock calls.
bz [Sat, 26 May 2012 23:58:51 +0000 (23:58 +0000)]
Correctly get the payload length in host byte order. While we
already plan to support >64k payload here, the IPv6 header payload
length obviously is only 16 bit and the calculations need to be right.
Reported by: dim
Tested by: dim
MFC after: 1 day
X-MFC: with r235958