Tijl Coosemans [Tue, 14 Feb 2012 12:50:20 +0000 (12:50 +0000)]
Change some headers such that lang/gcc* ports no longer patch them.
The lang/gcc* ports patch headers where they think something is
non-standard. These patched headers override the system headers which means
you have to rebuild these ports whenever you do installworld to make sure
they contain the latest changes.
David Chisnall [Tue, 14 Feb 2012 12:03:23 +0000 (12:03 +0000)]
Cleanup of xlocale:
- Address performance regressions encountered by das@ by caching per-thread
data in TLS where available.
- Add a __NO_TLS flag to cdefs.h to indicate where not available.
- Reorganise the xlocale.h definitions into xlocale/*.h so that they can be
included from multiple places.
- Export the POSIX2008 subset of xlocale when POSIX2008 says it should be
exported, independently of whether xlocale.h is included.
- Fix the bug where programs using ctype functions always assumed ASCII unless
recompiled.
- Fix some style(9) violations.
Reviewed by: brooks (mentor)
Approved by: dim (mentor)
Michael Tuexen [Tue, 14 Feb 2012 12:00:34 +0000 (12:00 +0000)]
Fix a bug where the wrong protocol overhead was used. This can lead
to a deadlock of an association when an IPv6 socket was used to
communcate with IPv4 and an ICMPv4 fragmentation needed message
was received.
While there, simplify the code a bit.
Doug Barton [Tue, 14 Feb 2012 10:51:24 +0000 (10:51 +0000)]
Fix various issues with the NFS and RPC related scripts:
1. Add new functionality to the force_depend method to incorporate the
tests for whether the service is enabled and/or already running.
2. Add a new option to bypass checking only that the service is enabled
at boot time, and always check if it is running.
3. Use this new functionality to greatly simplify the rc.d scripts that
use force_depend.
4. Add a force_depend for statd in lockd
5. Remove the check that either nfs_server or nfs_client is _enable'd
from statd and lockd. This was always overkill, and prevented using
the {one|force}start options, as well as stop'ing on the command line.
6. The yp* scripts had some of their arguments in various weird orders.
Bring them into line with the model.
7. If mountd fails to create /var/db/mountdtab, err out.
Ideas, suggestions, and/or review from delphij and jilles.
Pointy hats are completely my responsibility however.
Alexander Motin [Tue, 14 Feb 2012 09:19:30 +0000 (09:19 +0000)]
Do not handle MOD_SHUTDOWN equally to MOD_UNLOAD in sound kernel module.
MOD_SHUTDOWN is not an end of existence, and there is a life after it.
In particular, code previously called on MOD_SHUTDOWN grabbed lock and
deallocated unit numbering. That caused infinite wait loop if snd_uaudio
tried to destroy its PCM device after that point.
Add a rudimentary test to run through all the available counters on a
system and then execute a program with pmcstat in counting mode.
The program will verify that all counters fire and that the code neither
panics the system nor locks it up. This should be considered a first pass
conformance test for new sets of counters being added to hwpmc(4).
Marius Strobl [Tue, 14 Feb 2012 00:18:35 +0000 (00:18 +0000)]
- As it turns out, MSI-X is broken for at least LSI SAS1068E when passed
through by VMware so blacklist their PCI-PCI bridge for MSI/MSI-X here.
Note that besides currently there not being a quirk type that disables
MSI-X only and there's no evidence that MSI doesn't work with the VMware
pass-through, it's really questionable whether MSI generally works in
that setup as VMware only mention three know working devices [1, p. 4].
Also not that this quirk entry currently doesn't affect the devices
emulated by VMware in any way as these don't claim support MSI/MSI-X to
begin with. [2]
While at it, make the PCI quirk table const and static.
- Remove some duplicated empty lines.
- Use DEVMETHOD_END.
Ed Maste [Mon, 13 Feb 2012 16:48:49 +0000 (16:48 +0000)]
Add a sysctl to report the firmware build number.
Some older firmware versions have issues that can be worked around by
avoiding certain operations. Add a sysctl dev.aac.#.firmware_build to
make it easy for scripts or userland tools to detect the firmware
version.
Ed Schouten [Mon, 13 Feb 2012 11:59:59 +0000 (11:59 +0000)]
Polish diff against upstream.
- Revert unneeded whitespace changes.
- Revert modifications to loginrec.c, as the upstream version already
does the right thing.
- Fix indentation and whitespace of local changes.
Adrian Chadd [Mon, 13 Feb 2012 07:47:36 +0000 (07:47 +0000)]
Correct the 802.11s mesh configuration structure and related tidbits.
* Change the mesh IE size to be IEEE80211_MESH_CONF_SZ rather than the
size of the structure;
* conf_cap is now a uint8_t rather than a uint16_t (uint16_t in D3.0,
uint8_t in the amendment spec);
* Update mesh config capability bits - earlier bits were from draft X,
current is amendment spec;
* Update the following to be an enum rather than #define and added
a VENDOR entry too from the amendment spec;
IEEE80211_MESHCONF_PATH_*
IEEE80211_MESHCONF_METRIC_*
IEEE80211_MESHCONF_CC_*
IEEE80211_MESHCONF_SYNC_*
IEEE80211_MESHCONF_AUTH_*
* Kept IEEE80211_MESHCONF_FORM_* and IEEE80211_MESHCONF_CAP_* as
defines because they are defined in a way that we need to mask in/out
information;
* In IEEE80211_MESHCONF_CAP_* IEEE80211_MESHCONF_CAP_TBTTA is removed
and 0x80 is made reserved as defined in the amendment spec.
Ed Maste [Mon, 13 Feb 2012 01:44:12 +0000 (01:44 +0000)]
Fix panic after "WARNING - ATA_IDENTIFY taskqueue timeout"
When performing a firmware upgrade via atacontrol[1] the subsequent
command may time out producing the error message above. When this
happens the callout could still be active, and the system would then
panic due to a destroyed semaphore.
Instead, ensure that the callout is done first, via callout_drain.
Note that this fix applies to the "old" ata(4) and so isn't applicable
to the default configuration in HEAD. It is still applicable to
stable/8.
Adrian Chadd [Mon, 13 Feb 2012 00:28:41 +0000 (00:28 +0000)]
Attempt to address some potential vap->iv_bss race conditions.
There are unfortunately a number of situations where vap->iv_bss is changed
or freed by some code in net80211. Because multiple threads can concurrently
be doing work (and the vap->iv_bss access isn't at all done behind any kind
of lock), it's quite possible that:
* a change will occur in one thread - eg, by a call through
ieee80211_sta_join1();
* a state change occurs in another thread - eg an RX is scheduled
in the ath tasklet and it calls ieee80211_input_mimo_all(), which
does dereference vap->iv_bss;
* these two executing concurrently, causing things to explode.
Another instance is ath_beacon_alloc() which takes an ieee80211_node *.
It's called with the vap->iv_bss node from ath_newstate(). If the node has
changed in the meantime (say it's been freed elsewhere) the reference
that it grabbed _before_ refcounting it may be stale.
I would _prefer_ that these sorts of things were serialised somewhere but
that may be a bit much to ask. Instead, the best we can (currently) hope
is that the underlying bss node is still (somewhat) valid.
There is a related PR (kern/164382) described by the first case above.
That should be fixed by properly serialising the RX path and reset path
so an RX can't occur at the same time as the vap free/shutdown path.
This is inspired by some related fixes in r212127.
Brooks Davis [Sun, 12 Feb 2012 23:18:05 +0000 (23:18 +0000)]
Prevent periodic scripts that run longer than the expected period from
starting up before the previous script finishes. This prevents an
infinite number of them from piling up and slowing a system down.
Since all the refactoring to make this happen required churning the
indenting of most of this file, make the indentation more consistent.
Ed Schouten [Sun, 12 Feb 2012 18:29:56 +0000 (18:29 +0000)]
Globally replace u_int*_t from (non-contributed) man pages.
The reasoning behind this, is that if we are consistent in our
documentation about the uint*_t stuff, people will be less tempted to
write new code that uses the non-standard types.
I am not going to bump the man page dates, as these changes can be
considered style nits. The meaning of the man pages is unaffected.
Andriy Gapon [Sun, 12 Feb 2012 14:58:50 +0000 (14:58 +0000)]
start watchdogd before most of other daemons/servers
The main benefit is that watchdogd would shutdown after most of other
daemons/servers and thus, for example, would remedy a system hang caused
by unlucky X server shutdown.
Ed Schouten [Sat, 11 Feb 2012 21:49:23 +0000 (21:49 +0000)]
Attempt to implement who -a.
According to POSIX, -a is equal to -bdlprtTu. It seems this is not true
in practice, as -b normally restricts the output to BOOT_TIME entries
and all implementations that I know of don't.
Ed Schouten [Sat, 11 Feb 2012 20:47:16 +0000 (20:47 +0000)]
Move utmpx handling out of init(8).
This has the following advantages:
- During boot, the BOOT_TIME record is now written right after the file
systems become writable, but before users are allowed to log in. This
means that they can't cause `hidden logins' by logging in right before
init(8) kicks in.
- The pututxline(3) function may potentially block on file locking,
though this is very rare to occur. By placing it in an rc script, the
user can still kill it with ^C if needed.
- Most importantly: jails don't use init(8). This means that a force
reboot of a system running jails will leave stale entries in the
accounting database of the jails individually.
Bjoern A. Zeeb [Sat, 11 Feb 2012 20:43:01 +0000 (20:43 +0000)]
MFp4 204292:
Ignore the NAT_T extension types so we can at least dump the SADB from
the in-base libipsec/setkey without error when NAT_T support is present
in the kernel, though not printing the additional information yet.
However in case there is no NAT_T support in kernel still consider them
to be an error.
Ed Schouten [Sat, 11 Feb 2012 20:28:42 +0000 (20:28 +0000)]
Replace utxrm(8) by utx(8).
At first, I added a utility called utxrm(8) to remove stale entries from
the user accounting database. It seems there are cases in which we need
to perform different operations on the database as well. Simply rename
utxrm(8) to utx(8) and place the old code under the "rm" command.
In addition to "rm", this tool supports "boot" and "shutdown", which are
going to be used by an rc-script which I am going to commit separately.
Close a race due to dropping of the map lock between creating map entry
for a shared mapping and marking the entry for inheritance.
Other thread might execute vmspace_fork() in between (e.g. by fork(2)),
resulting in the mapping becoming private.
Marius Strobl [Sat, 11 Feb 2012 12:03:44 +0000 (12:03 +0000)]
Flesh out support for SAS1078 and SAS1078DE (which are said to actually
be the same chip):
- The I/O port resource may not be available with these. However, given
that we actually only need this resource for some controllers that
require their firmware to be up- and downloaded (which excludes the
SAS1078{,DE}) just handle failure to allocate this resource gracefully
when possible. While at it, generally put non-fatal resource allocation
failures under bootverbose.
- SAS1078{,DE} use a different hard reset protocol.
- Add workarounds for the 36GB physical address limitation of scatter/
gather elements of these controllers.
Ed Schouten [Sat, 11 Feb 2012 11:11:13 +0000 (11:11 +0000)]
Set read buffer size to multiple of sizeof(struct futx).
If the utmpx database gets updated while an application is reading it,
there is a chance the reading application processes partially
overwritten entries. To solve this, make sure we always read a multiple
of sizeof(struct futx) at a time.
Bjoern A. Zeeb [Sat, 11 Feb 2012 08:33:52 +0000 (08:33 +0000)]
Start to try to hide LRO (and some TSO) bits behind #ifdefs as especially
the symbols are not there when compiling a kernel without IP support and
we do have users doing so.
Doug Barton [Sat, 11 Feb 2012 06:21:16 +0000 (06:21 +0000)]
In the days before r208307 addswap was running early in the second stage
of rcorder. Somehow in the intervening period addswap got moved to the
very end, which is almost certainly not what we want.
This change moves it to right after kld so that for users who need it,
they'll get it ASAP.
Bjoern A. Zeeb [Sat, 11 Feb 2012 06:05:40 +0000 (06:05 +0000)]
Switch getifaddrs(3) to the new API introduced in r231505. Also remove
conditional code parts not used by or applicable to FreeBSD.
The new implementation is supposed to be able to cope with changes to
the 'l' versions of the msghdr structs now used as well as to if_data
allowing future changes without breaking things.
This restores carp(4) config support in HEAD after r231504.
Bjoern A. Zeeb [Sat, 11 Feb 2012 06:02:16 +0000 (06:02 +0000)]
Introduce a new NET_RT_IFLISTL API to query the address list. It works
on extended and extensible structs if_msghdrl and ifa_msghdrl. This
will allow us to extend both the msghdrl structs and eventually if_data
in the future without breaking the ABI.
Bump __FreeBSD_version to allow ports to more easily detect the new API.
Bjoern A. Zeeb [Sat, 11 Feb 2012 05:59:54 +0000 (05:59 +0000)]
Backout changes from r228571. Remove if_data from struct ifa_msghdr again.
While this breaks carp on HEAD temporary, it restores the upgrade path from
stable, and head before 20111215.
In cpu_set_user_tls(), consistently set PCB_FULL_IRET pcb flag for
both 64bit and 32bit binaries, not for 64bit only.
The set of the flag is not neccessary there, because the only current
user of the cpu_set_user_tls() is create_thread(), which calls
cpu_set_upcall() before and cpu_set_upcall() itself sets PCB_FULL_IRET.
Change the function for consistency and preserve existing KPI for now.
Ed Schouten [Fri, 10 Feb 2012 12:35:57 +0000 (12:35 +0000)]
Remove direct access to si_name.
Code should just use the devtoname() function to obtain the name of a
character device. Also add const keywords to pieces of code that need it
to build properly.
Adrian Chadd [Fri, 10 Feb 2012 10:01:09 +0000 (10:01 +0000)]
Add in a new driver feature to allow the TX and RX chainmask to be
overridden at attach time.
Some 802.11n NICs may only have one physical antenna connected.
The radios will be very upset if you try enabling radios which aren't
connected to antennas.
This allows hints to override the TX and RX chainmask.
These hints are:
hint.ath.X.rx_chainmask
hint.ath.X.tx_chainmask
They can be set at either boot time or in kenv before the module is loaded.
This and the previous HAL commit were sponsored in late 2011 by Hobnob, Inc.
Tim Kientzle [Fri, 10 Feb 2012 05:05:42 +0000 (05:05 +0000)]
Implement -print-file-name=include (which is undocumented
but used by some Linux boot loaders). This option prints
out the directory holding the include files needed by
a freestanding program. The default implementation of
this doesn't work on FreeBSD because of the different
include file layout. But it's easy to implement:
just return /usr/include (or the cross-compiling equivalent).