Add drm2 module makefile.

(riastradh)

sync with HEAD and actually get the drm2 import

(riastradh)

sync with HEAD

(riastradh)

When reading from /dev/random, block at most once in cprng_strong.

We are not obligated to return exactly as many bytes as requested,
and many applications -- notably those that use stdio or otherwise
buffered I/O to read from /dev/random -- try to read many more than
32 bytes at a time from /dev/random even if all they are about to use
is 32 bytes.

In this case, blocking until we have enough entropy to fill a large
buffer causes needless application delays, e.g. causing cgdconfig
(which reads from /dev/random with stdio) to hang at boot when trying
to configure a random-keyed device for swap.

Patch tested by Aran Clauson.  Fixes PR kern/48028.

(riastradh)

Rework cprng(9) man page to reflect the current state of affairs.

- Remove defunct cprng_strong_getflags/setflags.
- Remove defunct cprng_strong_ready.
- Document CPRNG_HARD.
- Omit cprng_strong structure, which is now opaque.
- Specify what can sleep and under what conditions.
- Be a little more consistent about some markup.

This is not the whole story (select/kqueue stuff for /dev/random is
still omitted), and I plan to change it some more (to split
cprng_strong into one routine that unconditionally guarantees as many
bytes as you asked, and another routine that may block or return
partial reads), but this will do until I find the time for those.

(riastradh)

Remove now-needless rump cprng_strong_deplete stub.

These rump stubs need to go away at some point...

(riastradh)

Fix races in /dev/u?random initialization and accounting.

- Push /dev/random `information-theoretic' accounting into cprng(9).
- Use percpu(9) for the per-CPU CPRNGs.
- Use atomics with correct memory barriers for lazy CPRNG creation.
- Remove /dev/random file kmem grovelling from fstat(1).

(riastradh)

Replace consttime_bcmp/explicit_bzero by consttime_memequal/explicit_memset.

consttime_memequal is the same as the old consttime_bcmp.
explicit_memset is to memset as explicit_bzero was to bcmp.

Passes amd64 release and i386/ALL, but I'm sure I missed some spots,
so please let me know.

(riastradh)

Use static inline.  Fixes i386/ALL and something else I've forgetten.

(riastradh)

Add consttime_bcmp(3) and explicit_bzero(3) to makefile and set lists.

(riastradh)

Include <sys/lwp.h> for curlwp.

(riastradh)

Add man pages and xrefs for consttime_bcmp and explicit_bzero.

ok wiz

(riastradh)

Fix dangling sentence vestige in cprng(9).

(riastradh)

Account bytes drawn from initial call to libkern arc4random.

XXX This unlocked initialization looks rather bogus...

(riastradh)

Rework rndsink(9) abstraction and adapt arc4random(9) and cprng(9).

rndsink(9):
- Simplify API.
- Simplify locking scheme.
- Add a man page.
- Avoid races in destruction.
- Avoid races in requesting entropy now and scheduling entropy later.

Periodic distribution of entropy to sinks reduces the need for the
last one, but this way we don't need to rely on periodic distribution
(e.g., in a future tickless NetBSD).

rndsinks_lock should probably eventually merge with the rndpool lock,
but we'll put that off for now.

cprng(9):
- Make struct cprng_strong opaque.
- Move rndpseudo.c parts that futz with cprng guts to subr_cprng.c.
- Fix kevent locking.  (Is kevent locking documented anywhere?)
- Stub out rump cprng further until we can rumpify rndsink instead.
- Strip code to grovel through struct cprng_strong in fstat.

(riastradh)

Fix amd64 feraiseexcept so that it actually traps.

The call to fwait got lost somewhere along the line; the i387 code
has it.

XXX pullup to 6

(riastradh)

Fix sense of fegetexcept on x86.

Somehow I overlooked this when I fixed feenableexcept and
fedisableexcept last summer.

XXX pullup to 6

(riastradh)

Set bp->b_resid to bp->b_bcount on error in swstrategy as required.

(riastradh)

Make partial read/write error messages more consistent in vndcompress.

(riastradh)

Add __printflike to vsnprintf_ss.

(riastradh)

Add -l option to synopsis for vndcompress(1) man page.

(riastradh)

Fix sign-compare in compress_blocks.

Not sure why my builds didn't reveal this one -- they revealed
several others during development.

(riastradh)

Rewrite vndcompress to support SIGINFO and restart after interrupt.

Make it generally more robust in the process.

No objection (or comment) on tech-userlevel.

ok christos

(riastradh)

Fix (harmless) typo in struct genfs_rename_ops::gro_lookup prototype.

(riastradh)

First draft of a long-awaited man page for genfs_rename(9).

(riastradh)

Move addition of -g to CSHLIBFLAGS later in bsd.lib.mk.

It is now past any conditionals where its emptiness or nonemptiness can
influence the decision of what objects get built.

Fixes MKDEBUG=yes build on vax (PR 47700).

(riastradh)

Fix documentation of rbtree(3) iteration.

. Fix sense of rb_tree_iterate(rbt, NULL, ...).
. Document RB_TREE_MIN/RB_TREE_MAX to avoid relying on that sense.
. Document RB_TREE_FOREACH and RB_TREE_FOREACH_REVERSE to simplify
iteration.

Addresses PR lib/46034 and PR lib/47144.  It would have been nice to
make `x = NULL; while ((x = rb_tree_iterate(t, NULL, ...)) != NULL)'
DTRT to traverse t, but it's too much late for that now.

We probably ought to have an RB_TREE_FOREACH{,_REVERSE}_SAFE too.

ok christos

(riastradh)

Fix argument order for config_attach_loc in autoconf(9).

(riastradh)

Touch e_aes.c to force a rebuild with new compiler flags for AES-NI.

(riastradh)

Fix build goo for OpenSSL AES-NI support.

OpenSSL now supports AES-NI in evp, not in an engine.  We can now get
rid of the no longer maintained aesni engine, which was broken last
summer.  Not only can OpenSSL now use AES-NI for everything it did
before we broke it last summer, but it can also use AES-NI for more
encryption modes than before, such as CTR.

Tested on amd64, both vanilla and in an i386 chroot.

ok christos

(riastradh)

Fix some screw cases in cmsg file descriptor passing.

- Don't leave garbage in the control buffer if allocating file
descriptors fails in unp_externalize.

- Scrub the space between CMSG_LEN and CMSG_SPACE to avoid kernel
memory disclosure in unp_externalize.

- Don't read past cmsg_len when closing file descriptors that
couldn't get delivered, in free_rights.

ok christos

(riastradh)

Add pthread_condattr_setclock(3) man page link.

(riastradh)

Set NOMAN before including <bsd.own.mk> to really fix PR 47500.

(Hi, christos.)

(riastradh)

Assert equality, not assignment, in selrecord.

Code inspection suggests that this fix is not likely to reveal any
latent problems.

(riastradh)

Assert equality, not assignment, in rnd_hwrng_test.

Not tested, but by inspection, the only caller, rnd_process_events,
clearly guarantees the condition.

(riastradh)

Assert equality, not assignment, in fsplit, and say who guarantees it.

(riastradh)

Move assignment out of assertion in cap_mkdb.

(riastradh)

Assert equality, not assignment, in puffs__fsframe_cmp.

(riastradh)

Fix typo, unless `ro raising' is something I don't know about.

(riastradh)

Match the C600's ichlpcib.

(riastradh)

Match the C600's other smbus controller.

(riastradh)

Nix `-b 1024' from ssh_keygen_flags in /etc/defaults/rc.conf.

No objections on tech-security:

http://mail-index.NetBSD.org/tech-security/2012/09/03/msg000548.html

The security of generating keys at boot remains questionable, but
this change makes that no worse.

(riastradh)

After mount_lfs_parseargs, reset getopt for lfs_cleaner_main.

(riastradh)

Initialize b_resid before biodone in wdioctlstrategy error branch.

Otherwise we may panic when physio's biodone callback kasserts
something about b_resid, since nothing will have initialized it.

(riastradh)

Add some ATA SECURITY commands to atactl(8).

(riastradh)

Name the ATA SECURITY commands.

(riastradh)

Show pass1 SIGINFO output on stderr like other passes, not on stdout.

(riastradh)

Implement TI AM335x's SDHC reset quirk.  Beaglebone SDHC works now!

On the AM335x, we first must wait for the controller to acknowledge
the reset; then we can wait for the reset to complete.

I believe this quirk also applies to the OMAP4 ES, but I don't have
one of those to test and we don't seem to have an obvious conditional
for it anyway.

This quirk may work for controllers that don't require it too, but I
am nervous about doing it by default because if we miss the reset
acknowledgement, then we'll just time out even though everything is
really hunky-dory.

Also, for all sdhc, don't bother writing 0 in sdhc_soft_reset while
waiting for the reset to complete; there is no need.

ok matt

(riastradh)

Avoid the 16-bit read in sdhc_host_found for 32-bit-only devices.

This change makes sdhc_host_found allocate hp up front before showing
the identification, in order to avoid having to open-code a copy of
HREAD2 before hp is available.

(riastradh)

Hack up the BEAGLEBONE config and add a BEAGLEBONE_INSTALL.

This is a total mess right now, and BEAGLEBONE_INSTALL isn't useful
except to boot ramdisks for testing, but this hash of a config works
to get my beaglebone booted into sysinst, at least.

(riastradh)

Initial kludgerific attempt at making sdhc work on the AM335x.

There's a timing issue somewhere in the whole sdmmc thing so that
this doesn't actually work without SDMMC_DEBUG or some specially
crafted debugging messages, and the SoC-specific code needs to be
separated out here another way, but it's a little progress.

(riastradh)

Detect the AM335x CPU frequency on the beaglebone.

(riastradh)

Hook up all the new AM335x files.

(riastradh)

Make the interrupt controller and PRCM critical for the AM335x.

This adds some more crufty #ifdefs to a mess that needs to be
rototilled, but it'll do to make progress for now.

(riastradh)

AM335x driver attachment goop for omapdmtimer.

(riastradh)

New driver for OMAP-style dual-mode timers, for both v1 and v2 cores.

(riastradh)

Implement prcm_module_enable and prcm_module_disable for the AM335x.

(riastradh)

Export a little more from the omap2 prcm driver.

This is in anticipation of drivers that need to enable (and, for
future power management, perhaps disable) modules on an SoC.  Each
SoC has a different notion of enabling and disabling modules and will
need to implement prcm_module_enable and prcm_module_disable
separately.

(riastradh)

Document that ENVSYS_STEMP is in microkelvins.

(riastradh)

Fix some error branches in ugen.

There remains some cruft that should perhaps be better organized, but
at least this should reduce some memory leaks in screw cases, and at
least this does fix panics when plugging in and unplugging a USB
device with a botched configuration (a beaglebone with a hosed sd
card).

(riastradh)

Match the Intel 7 Series SMBus Controller in ichsmb(4).

Tested on a Thinkpad W530.

(riastradh)

Rework firmware reference counting and error messages in wpi(4).

. Clarify the shared firmware abstraction in wpi_cached_firmware
  and its new sibling wpi_release_firmware.
. Fix typo in wpa_cache_firmware error branch leading to free NULL.
. Fix leak in wpi_load_firmware error branch.
. Sprinkle some kasserts to executably document invariants.
. A little KNF here and there.

Based on a patch from dh in PR kern/44144.

(riastradh)

Match LEN0068 in thinkpad(4) too, as newer Thinkpads report.

Tested on a W530.  Lots of stuff is missing, though.

(riastradh)

Use 0 for MS_NOMNTTAB since we have no mnttab anyway.

(riastradh)

Let's try that again without the static...

(riastradh)

Add omitted secpolicy_vnode_utime_modify, for zfs.

(riastradh)

No, we can't elide the fs-wide rename lock for same-directory rename.

rename("a/b", "a/c") and rename("a/c/x", "a/b/y") will deadlock.

Darn.

(riastradh)

Back out accidental commit of errno kludge for rmdir(".") &c.

Solaris returns EEXIST, whereas we want to return ENOTEMPTY (POSIX
allows both), but this got included in an unrelated commit and should
be separated into a common commit for other related error code fixes.

(riastradh)

Take a first whack at making zfs permissions work.

zfs_access uses secpolicy_vnode_access, so it makes no sense for the
latter to call VOP_ACCESS!

Everything seems to return EACCES instead of EPERM, probably because
that's what kauth returns.  This should be fixed, but that may
require some nontrivial surgery to zfs's calls to secpolicy_*, which
is where kauth gets involved.

This commit imports some code from illumos to implement the routine
secpolicy_vnode_setattr.  This shouldn't be outside dist/, but for
now it is expedient to do so.  We ought to fix that, along with all
the other CDDL code outside dist/, when we next import a newer
version of zfs.

(riastradh)

Restore _PC_NO_TRUNC in zfs_netbsd_pathconf.

Accidentally clobbered it in a previous commit while moving patches
around between source trees -- oops.

(riastradh)

Fail lookup of .. in a deleted directory with ENOENT.

Don't try to zget the parent, whose znode id may have been recycled
by now.

It's not clear to me how Solaris avoids this, but maybe I'm just
missing something obvious.

(riastradh)

Fix various issues in zfs life cycle, locking, and vop protocol.

- Restore some zfs locking and unlocking that got lost randomly.

- Enable use of the BSD vnode lock.  Lock order: all BSD vnode locks
are taken before all zfs internal locks.  There remains an issue with
O_EXCL, to be solved later (famous last words).  KASSERT the locking
scheme up the wazoo.

- Take our cruft out of zfs_lookup and move it to zfs_netbsd_lookup.
Restore much of the way zfs_lookup looked to make merging future
versions easier.  Disable use of the namecache for now because its
locking dance is too scary to contemplate.

- Implement BSD semantics for rename, to appease our tests.  This is
a provisional kludge; eventually we need VOP_RENAME to take a flag
specifying whether to use BSD semantics or POSIX semantics.

- Simplify zfs_netbsd_reclaim and make it work.  Now that getnewvnode
never tries to vclean anything itself, we need not worry about
recursion of ZFS_OBJ_MUTEX locks.

- Clarify and fix genfs node initialization and destruction.

zfs passes most of our atf vfs tests now, including the rename races.

Still to do:

- fix the impedance mismatch between our permissions model and zfs's;
- fix O_EXCL (nontrivial);
- throw dirconc at it and see how badly it explodes;
- find why zpool sometimes wedges itself during mkfs; and
- find why pool caches sometimes seem to get corrupted.

(riastradh)

secpolicy_vnode_access must be called with the vnode lock held.

Don't unlock the vnode and then tell the caller about what the world
was like while the vnode was locked.  The world changes fast.

(riastradh)

Simplify zfs dirlock reference counting.

No need to pass dzp around; dl has a pointer to it.

(riastradh)

Do reference counting for zfs range lock waiters.

Avoid cv_broadcast(&cv); cv_destroy(&cv); which works in Solaris only
by abuse of the condvar abstraction.

There are parts of this code that should be factored into smaller
subroutines, mainly range lock allocation and initialization, but
that would make it harder to merge newer versions of zfs, so for now
I've just expanded those parts further in-line.

(riastradh)

Do reference counting for zfs range lock waiters.

Avoid cv_broadcast(&cv); cv_destroy(&cv); which works in Solaris only
by abuse of the condvar abstraction.

There are parts of this code that should be factored into smaller
subroutines, mainly range lock allocation and initialization, but
that would make it harder to merge newer versions of zfs, so for now
I've just expanded those parts further in-line.

(riastradh)

Disentangle do_sys_rename.

Elide the fs-wide rename lock for single-directory renames.  This
required changing the order of lookups, so that we know what the
directories are before we lock the nodes.

Clean up error branches, explain why various nonsense happens and
what it does and doesn't do, and note some of what needs to change.

(riastradh)

Add a remarkably percipient fortune cookie.

(riastradh)

Fix formatting of Churchill quote on declarations of war.

(riastradh)

Use db_read_value to read the trapframe fields in db_nextframe.

Fixes SIGSEGV on bt in crash(8) when the stack trace ends in syscall,
and probably other problems as well.

ok christos

(riastradh)

On failure in _rtld_dlcheck, call _rtld_error, not xwarnx.

Callers need not call _rtld_error themselves now.

Fixes PR lib/41760.  Should add the test case from the PR too.

(riastradh)

Rewrite section on DNS so it makes sense, is true, and is useful.

This may need to be adapted for openresolv too...

(riastradh)

Uncomment call to genfs_rename_cache_purge.

Fixes a side issue mentioned in PR kern/46990.

I left this commented to preserve the old behaviour of tmpfs_rename,
but it is obviously broken to omit the cache purge, and I'm surprised
nobody had encountered any problems with it until now.

(riastradh)

Check tmpfs_rmdired_p after tmpfs_vnode_get when walking up the tree.

tmpfs_vnode_get drops all locks except possibly the reclaiming bit
lock to keep the tmpfs node from being reclaimed while we're still
interested in it.  Consequently, it does not keep the directory's
existence invariant, so we must check that after tmpfs_vnode_get.

Fixes PR kern/46990.  Tested by Wolfgang Stukenbrock.

(riastradh)

Fix typo: tftpd, not tfptd.

(riastradh)

Use as large a packet size as the ftdi endpoint specifies.

Now I can read from my beaglebone!

(riastradh)

Use separate names for the multitudinous uses of `q' in exit1.

Now I can follow which process is which in this routine.

If I jiggle the whitespace so line numbers don't change, there is no
change in the output of `objdump -d kern_exit.o' for amd64.

ok abp

(riastradh)

Quote arguments to .Cd in uatp(4) man page.

(riastradh)

Add man page for uatp(4).

(riastradh)

Force sys_close not to restart by returning ERESTART.

Print a diagnostic message if we ever get ERESTART out of fd_close
and convert it to EINTR instead.

Even if fd_close fails, it has already closed the file descriptor, so
restarting the system call is a mistake, with dangerous consequences
for multithreaded programs.

Should probably turn the message into a kassert eventually, and maybe
add one deeper in fd_close in order to more easily debug it before
all the data structures are destroyed.

(riastradh)

Add hdaudio modules to the md.evbppc set list too.

(riastradh)

Add uatp to the md.evbppc modules set list too.

(riastradh)

Build hdaudio and hdafg modules.

(riastradh)

...and be sure to config_pending_decr if kthread_create fails.

This error branch looks suspect, though.  Shouldn't we bail at this
point rather than blithely try to proceed?

(riastradh)

Restore config pending stuff, adjusted to avoid the race.

Leave a note about what this code probably should look like for
anyone who comes rummaging around with the intent to really fix it.
I would do that myself if I had time and resources to test this at
the moment, and more comfort with our firewire stack, but I don't.

(riastradh)

Fix ffs_fstest_delfs's error branch for rump_pub_etfs_remove.

rump_pub_etfs_remove returns an error code, rather than setting errno
and returning -1.

(riastradh)

Clarify description of ->ucom_read method.

From when I was trying to figure out what's wrong with uftdi reading
from a beaglebone last month.  (Still haven't figured that out!)

(riastradh)

Use my @NetBSD.org email address in the bwi(4) man page.

(riastradh)

Add uatp(4) to i386 ALL kernel configuration.

(riastradh)

Add a module for uatp(4).

(riastradh)

Add uatp(4), a driver for USB Apple trackpads.

This is a work-in-progress driver for USB trackpads found in Apple
laptops since 2005, theoretically covering more models than pbms(4)
and supporting more features.  However, the motion smoothing and
acceleration formulae are still pretty sketchy, and I have tested
this only on one model of MacBook from 2006.  Feedback welcome.

Should the smooting, acceleration, tapping, &c., be done in userland?
Probably, but we don't have the necessary interface for that to work
well yet -- wsmouse isn't enough as is.

(riastradh)

Fix error branches in ath pci attachment.

Should fix double pci_intr_disestablish if ath_attach fails.

(Still not sure why ath_attach sometimes fails here.)

(riastradh)

Implement link(2) in rumphijack.  Add a couple trivial test cases.

(riastradh)

Fix error branches and config pending races in firewire init.

This way, if anything fails, it just fails; you don't panic.  This can
happen if suspending and resuming of firewire is broken (e.g., as I
encountered in PR kern/44581).

(riastradh)

Mask off the result of feenableexcept and fedisableexcept on x86.

These shouldn't return bits outside FE_ALL_EXCEPT.

(riastradh)

Fix error branch in wpi(4) firmware loading.

Fixes panic if firmware is not available.

(riastradh)

Fix i386 `boot' command to try the usual set of kernel names.

Without this, the `boot' command will try only `netbsd', not
`netbsd.gz', `netbsd.old', &c.

Patch from PR port-i386/44562.

(riastradh)

Fix null pointer dereference in msk_rxeof.

This triggers when processing an RX interrupt that was queued while
stopping the interface, which caused my machine to panic last night.
In this case, just drop the packet.

From OpenBSD's if_msk.c rev. 1.71.

(riastradh)

Clean up #includes in progress(1).

Sort them, omit needless ones, and add omitted needful ones.

Omissions revealed by pilfering the code and trying to build it on
Mac OS X.  We ought to have a better way to reveal these omissions...

(riastradh)

Elaborate and clarify rules about #include order.

ok christos

(riastradh)

Omit outdated cryptographic claims as noted in PR 44160.

Interested readers can follow the references or read Wikipedia; this
is the wrong place to explain cryptographic hash functions and give
security advice.

(riastradh)

Take the kernel lock in atapibusdetach just like atapibuschilddet.

Fixes kassert in scsipi_lookup_periph when I press the power button
on one of my laptops (and maybe another one) to power it off.

ok mrg

(riastradh)

Attribute the `Consequences, Schmonsequences...' quote to Daffy Duck.

(riastradh)

Use two separate comments for stub where IN_RENAME was.

(riastradh)

Kill the IN_RENAME in-core inode flag in ufs and ext2fs.

Now that rename works we need not to wave this sort of voodoo at it.

ok dholland

(riastradh)

Fix typo in comment: bp->b_bcount, not bp->b_count.

(riastradh)

Fix ext2fs's scary cross-block directory message too.

(See rev. 1.3 of sys/ufs/ufs/ufs_rename.c for the analysis.)

(riastradh)

Kill scary message about cross-block directories and fix its cause.

Add a bunch of kasserts to check more stringently that ufs_direnter
did not compact across directory blocks.  Don't bother fetching
subsequent I/O blocks from the directory: ufs_lookup guarantees that
it's not necessary, and the kasserts check this to be sure.

The message fired when we were looking at the start of an I/O block,
not when we crossed from the end of one to the start of another.  I
believe it fired only when tulr->ulr_offset was a multiple of the I/O
block size (fs_bsize), which can happen if ufs_lookup either finds an
entry or finds free space at the start of an I/O block.

If ufs_lookup found an entry, none of this ulr recalculation logic
should kick in -- if tvp != NULL, then tulr->ulr_count is garbage, so
it's not merely unnecessary but wrong (although I suspect harmless in
the end) to read it in ufs_rename_overlap_p in consideration of
whether to recalculate fulr.

Discussed with chuq and dholland.

ok dholland

(riastradh)

Tidy up some typos and vestiges in comments after the ulr changes.

(riastradh)

Swap byte order of ext2fs_direct fields in ext2fs_rename_recalculate_fulr.

Symptom found and fix tested by martin.

ok martin

(riastradh)

Disable scary but probably harmless printf.

Still need to find why this harmless-but-shouldn't-happen case is
happening, but in the mean time, we can stop scaring people with it.

(riastradh)

Update ffs, ext2fs, and tmpfs module makefiles to add *_rename.c.

ok martin

(riastradh)

Welcome to 6.99.7, courtesy of genfs_rename.

ok martin

(riastradh)

Fix omitted VN_KNOTE(fvp, NOTE_RENAME) in tmpfs_rename.

Missed that line while refactoring tmpfs_rename to use genfs_rename.

(riastradh)

Adjust t_renamerace now that ext2fs and ffs have good rename.

(riastradh)

Adapt ffs, lfs, and ext2fs to use genfs_rename.

ok dholland, rmind

(riastradh)

Adapt tmpfs_rename to use genfs_rename.

(riastradh)

Implement a genfs_rename abstraction.

First major step in incrementally adapting all the file systems to a
saner rename VOP protocol.

(riastradh)

Fix error reporting in puffs_framev_enqueue_waitevent and callers.

ok christos

(riastradh)

Match adm1030 too in dbcool_ki2c_match.

PR port-macppc/41916: dbcool at ki2c should support adm1030

ok macallan

(riastradh)

KASSERT(!mutex_owned(...)) is not kosher.

(riastradh)

Add test for `ln -s / foo && cd foo'.

(riastradh)

No more xfail for PR kern/36681.

(riastradh)

Fix tmpfs_rename locking.

Fixes PR kern/36681.  tmpfs now survives dirconc, all our vfs/tmpfs
tests and rename races in atf, and a bunch of hand-written tests
that I'd commit if atf didn't find them highly indigestible.

ok dholland

(riastradh)

Handle absolute symlinks to the root.

Fixes panic on `ln -s / foo && cd foo' found by ober by trying to run
wine.

ok dholland

(riastradh)

Cache vattr in psshfs's setattr.

This means within the cache window, a setattr that wouldn't change the
remote file's attributes from our current view of them will not be
relayed to the server and wait for the server to answer.  Thus, e.g., a
process with a periodic timer interrupt that calls open(2) in a loop
can make progress with much higher probability than without caching.

XXX The test case doesn't work, so it's currently disabled.  It needs
to stop the child of sshd that is handling an sftp session, not sshd
itself, and it's not obvious how to do that.

ok pooka

(riastradh)

Need root for /dev/power.

(riastradh)

i386 and amd64 are on GCC 4.5 too now.

(riastradh)

In ufs_rename, declare oldparent and newparent ino_t, not int.

XXX There should be an automatic test for this somewhere.

ok dholland

(riastradh)

Fix renaming over mismatched non-directory types in tmpfs.

Renaming a file of any non-directory type over another file of any
other non-directory type is OK -- they need not match as long as
neither is a directory, so loosen the kassert to reflect this.

XXX Need to write test cases for this.

ok dholland, rmind

(riastradh)

Release mutex before unmapping DMA memory.

Fixes a LOCKDEBUG panic triggered by `hdaudioctl set'.

ok jmcneill

(riastradh)

Fix typo: `symbolically' takes three l's, not two.

(riastradh)

Make `mount_ptyfs -c' actually work.

The ALTF_* are bit masks, not indices, so use 4, not 3, for the one
after 2.  Use 0x on all the values to emphasize this.

ok rmind

(riastradh)

Add a cmsg.3 man page symlink, since I can never remember its name.

ok wiz

(riastradh)

Fix some typos and incorrect idioms with the cmsg botch^WAPI.

Add a couple cross-references too.

ok wiz

(riastradh)

Expand tests for unaligned writes to cgd.  No more xfail.

PR kern/44515
PR kern/44965

(riastradh)

Reject unaligned writes to cgd.

Fixes the following PRs:

PR kern/44515 (cgd dies on non-aligned writes to the raw device)
PR kern/44964 (cgd seems to panic on unaligned writes instead of giving EINVAL)

ok christos

(riastradh)

Initialize psshfs root node's readdir waiters queue.

Every other node's readder waiters queue is initialized in makenode,
but the root node is created specially.

ok pooka

(riastradh)

Add my name to the list.

(riastradh)