Don't use memcpy() to copy several bytes.
Store IDs is host order. It is not so important to bloat code for it.
Combine m_adj() and M_PREPEND() into single M_PREPEND().
des [Mon, 28 Jul 2008 21:33:34 +0000 (21:33 +0000)]
Add collation sequence for no_NO. The same sequence applies to nb_NO
and nn_NO, which are symlinked to no_NO.
The patch in the PR contained a number of errors apparently based on
(sometimes incorrect) pronunciation; for instance, v and w are
distinct letters and should be collated in that order, even if they
are pronounced the same, while <u:> should be collated with u, even
though it is often mispronounced as y. For lack of a solid reference,
I have taken sv_SE and simply changed the last three letters of the
alphabet.
Fix a few bugs with the _gettemp() routine which implements mkstemp(),
mkstemps(), and mkdtemp().
- Add proper range checking for the 'slen' parameter passed to mkstemps().
- Try all possible permutations of a template if a collision is encountered.
Previously, once a single template character reached 'z', it would not wrap
around to '0' and keep going until it encountered the original starting
letter. In the edge case that the randomly generated starting name used
all 'z' characters, only that single name would be tried before giving up.
PR: standards/66531
Submitted by: Jim Luther
Obtained from: Apple
MFC after: 1 week
When choosing a CPU for a thread in a cpuset, prefer the last CPU that the
thread ran on if there are no other CPUs in the set with a shorter per-CPU
runqueue.
Don't attempt authentication at all if it has been disabled via '-a off'.
This works around a bug in HP-UX's telnet client and also gives a much
saner user experience when using FreeBSD's telnet client.
PR: bin/19405
Submitted by: Joel Ray Holveck joelh of gnu.org
MFC after: 1 month
Implement support for cpusets in the 4BSD scheduler.
- When a cpuset is applied to a thread, walk the cpuset to see if it is a
"full" cpuset (includes all available CPUs). If not, set a new
TDS_AFFINITY flag to indicate that this thread can't run on all CPUs.
When inheriting a cpuset from another thread during thread creation, the
new thread also inherits this flag. It is in a new ts_flags field in
td_sched rather than using one of the TDF_SCHEDx flags because fork()
clears td_flags after invoking sched_fork().
- When placing a thread on a runqueue via sched_add(), if the thread is not
pinned or bound but has the TDS_AFFINITY flag set, then invoke a new
routine (sched_pickcpu()) to pick a CPU for the thread to run on next.
sched_pickcpu() walks the cpuset and picks the CPU with the shortest
per-CPU runqueue length. Note that the reason for the TDS_AFFINITY flag
is to avoid having to walk the cpuset and examine runq lengths in the
common case.
- To avoid walking the per-CPU runqueues in sched_pickcpu(), add an array
of counters to hold the length of the per-CPU runqueues and update them
when adding and removing threads to per-CPU runqueues.
Export the hardware type (as number and name), the secondary firmware
revision and (on Prism cards) the primary firmware revision via
sysctl. Move the printing of this information under bootverbose,
since it is relatively easy to get to it now.
Add a check for 1GB page mappings to pmap_change_attr() so that it fails
gracefully. (On K10 family processors the direct map is implemented using
1GB page mappings.)
Enhance pmap_change_attr(). Use pmap_demote_pde() to demote a 2MB page
mapping to 4KB page mappings when the specified attribute change only
applies to a portion of the 2MB page. Previously, in such cases,
pmap_change_attr() gave up and returned an error.
- Move vp test for beeing NULL under IGNORE_LOCK().
- Check if panicstr isn't set, if it is ignore the lock. This helps to avoid
confusion, because lockmgr is a no-op when panicstr isn't NULL, so
asserting anything at this point doesn't make sense and can just race with
other panic.
Add the ability to add new addresses for interfacesto just one FIB
(Other more specific related options will follow)
This allows one to set multiple p2p links to the same place
and select which to use by having each in different FIBS.
sam [Sat, 26 Jul 2008 23:50:27 +0000 (23:50 +0000)]
don't deauth a station because it sends a ps-poll w/ a bogus aid in it;
turns out some devices do this and since we otherwise validate the station
is associated and don't use the aid for anything being lenient here allows
them to function
Increase UDBHASHSIZE from 16 to 128 items.
Previous value was chosen 10 years ago and not very effective now.
This change gives several percents speedup on 1000 L2TP mpd links.
According to in_pcb.h protocol binding information has double locking.
It allows access it while list travercing holding only global pcbinfo lock.
This relaxed locking noticably increses receive socket lookup performance.
2020447 IPFilter's NAT can undo name server random port selection
(fix output port range, was a random number in [0,max-min]
(byteswapped on litle endian), instead of [min,max])
Add a few clarifications to this man page. Also, add a table that
describes the minimum versions of each feature and each chipset type
supported by this driver. Basically, unless you have a very modern
version of firmware on a Prism card, you won't be able to use these
cards for much on modern networks that have any kind of protection
enabled, except for the few WEP-only stranglers that appear at some
conferences...
Add hash table lookup for a fully connected raw sockets.
This gives significant performance improvements when many raw sockets used.
Benchmarks of mpd handeling 1000 simultaneous PPTP connections show up to 50%
performance boost. With higher number of connections benefit becomes even
bigger. PopTop snd others should also get some benefits.
Register the interrupt handler at the end of wi_attach rather than at
the beginning. There's a race in the shared interrutp case. If
another interrupt happens after the interrupt is setup, then we'd try
to lock an uninitialized mutex. In addition, if we bailed out due to
a too old version of firmware, we'd leave the interrupt enabled with
all the fun that ensues....
Return arc4_i = arc4_j = 0; line from previous backing out since
just found it in OpenBSD, see their sys/crypto/arc4.c, function
rc4_keysetup, line ctx->x = ctx->y = 0;
Per rwatson's request:
"If you don't get a review within a day or two, I would firmly recommend
backing out the changes"
back out all my changes, i.e. not comes from merging from OpenBSD as
unreviewed by secteam@ yet.
(OpenBSD changes stays in assumption they are reviewd by OpenBSD)
Yes, it means some old bugs returned, like not setted rs_stired = 1 in
arc4random_stir(3) causing double stirring.
ed [Fri, 25 Jul 2008 14:31:00 +0000 (14:31 +0000)]
Move ttyinfo() into its own C file.
The ttyinfo() routine generates the fancy output when pressing ^T. Right
now it is stored in tty.c. In the MPSAFE TTY code it is already stored
in tty_info.c. To make integration of the MPSAFE TTY code a little
easier, take the same approach.
This makes the TTY code a little bit more readable, because having the
proc_*/thread_* routines in tty.c is very distractful.
ed [Fri, 25 Jul 2008 12:37:46 +0000 (12:37 +0000)]
Remove <sgtty.h> now it has no practical usage.
When I turned sgtty into a binary-only interface (last month), I added
this explicit #error to the header file, to make sure nobody forgot to
remove the header file after updating world.
I think it is now a good moment to remove this header file.
Try again not to use a userspace pointer in the kernel when trying to record
the hostname which we need for NLM requests. The previous patch was incomplete.
ed [Thu, 24 Jul 2008 09:54:10 +0000 (09:54 +0000)]
Don't include <sys/tty.h> in non-TTY drivers.
The kbd, kbdmux, ugen and uhid drivers included <sys/tty.h>, because
they needed clists, which have been moved to <sys/clist.h> some time
ago. In the MPSAFE TTY branch, <sys/tty.h> does not include
<sys/clist.h>, which means we have to teach these drivers to include
this header file directly.
ed [Wed, 23 Jul 2008 17:47:44 +0000 (17:47 +0000)]
Add TIOCPKT and TIOCSPTLCK to the Linuxolator.
We're very lucky, because the flags used by our TIOCPKT implementation
are the same as flags used by Linux. We can safely enable TIOCPKT,
assuming EXTPROC is not used.
TIOCSPTLCK is used by unlockpt(). Because we don't need unlockpt() in
our implementation, make this ioctl a no-op.
Approved by: philip (mentor, implicit), rdivacky
Obtained from: P4 (//depot/projects/mpsafetty/...)
des [Wed, 23 Jul 2008 16:40:07 +0000 (16:40 +0000)]
pjd@'s r180759 was intended to revert r180755 due to ipfilter breakage,
but removed too much, breaking the build in other places instead. Now
that the ipfilter issue has been fixed (or hacked around), address the
second issue by restoring r180755, with one small change. I don't feel
comfortable using assert(3) in a header that will be included in userland
code that may or may not already have an assertion mechanism in place,
so KASSERT() evaluates to a no-op in the !_KERNEL case.
des [Wed, 23 Jul 2008 16:34:53 +0000 (16:34 +0000)]
For unfathomable reasons, ipfilter abuses kernel data structures for its
own purposes. To pull this off, it defines _KERNEL before including the
headers where these structures are defined. This leads to no end of
trouble when some of these headers, or other headers that they include,
change, as demonstrated by r180755.
The quick fix in this particular case is to define _WANT_FILE instead of
_KERNEL, conditional on __FreeBSD__. A better long-term fix is left as
an exercise to the reader.
Further refine the probe order of devices to more closely match the previous
behavior. Specifically, probe Host-PCI bridges in the order they are
encountered in the tree. For CPUs, just use an order of 100000 and assume
that no Host-PCI bridges will be more than 10000 levels deep in the
namespace. This fixes an issue on some boxes where the HPET timer stopped
attaching.
The ffs_balloc_ufs{1,2} functions call bdwrite() while having several
vnode buffers locked at once. In particular, there are indirect buffers
among locked ones. The bdwrite() may start the flushing to keep dirty
buffer list at the bounds. If any buffer on the dirty list requires
translation from logical to physical block number, code may ends up
trying to lock an indirect buffer already locked in ffs_balloc_ufsX.
Prevent the bdflush() activity when several buffers are locked at once
by setting the TDP_INBDFUSH for the problematic code blocks.
Reported and tested by: pho, Josef Buchsteiner at Juniper
In collaboration with: kan
MFC after: 1 month
Unbreak cc -pg support on i386. In gcc 4.2, %ecx is used as the arg pointer
when stack realignment is turned on (it is ALWAYS on for main), however
in a profiling build %ecx would be clobbered by mcount(), this would lead
to a segmentation fault when the code tries to reference any argument.
This fix changes mcount() to preserve %ecx.
when NDIS framework try to query/set informations NDIS drivers can
return NDIS_STATUS_PENDING. In this case, it's waiting for 5 secs to
get the response from drivers now. However, some NDIS drivers can send
the response before NDIS framework gets ready to receive it so we might
always be blocked for 5 secs in current implementation. NDIS framework
should reset the event before calling NDIS driver's callback not after.
SATA device on some nForce based boards could get confused if MSI is not
used but MSI to HyperTransport IRQ mapping is enabled, and would act as
if MSI is turned on, resulting in interrupt loss.
This commit will,
1. enable MSI mapping on a device only when MSI is enabled for that
device and the MSI address matches the HT mapping window.
2. enable MSI mapping on a bridge only when a downstream device is
allocated an MSI address in the mapping window
Do the pargs_hold() on the copy of the pointer to the p_args of the
child process immediately after bulk bcopy() without dropping the
process lock.
Since process is not single-threaded when forking, dropping and
reacquiring the lock allows an other thread to change the process title
of the parent in between, and results in hold being done on the invalid
pointer. The problem manifested itself as the double free of the old
p_args.