Commit Graph

123231 Commits

Author SHA1 Message Date
Robert Watson
f7f45ac8e2 Annotate uses of fgetsock() with indications that they should rely
on their existing file descriptor references to sockets, rather than
use fgetsock() to retrieve a direct socket reference.

MFC after:	3 months
2006-04-01 15:25:01 +00:00
Robert Watson
ac45e92ff2 Change protocol switch pru_abort() API so that it returns void rather
than an int, as an error here is not meaningful.  Modify soabort() to
unconditionally free the socket on the return of pru_abort(), and
modify most protocols to no longer conditionally free the socket,
since the caller will do this.

This commit likely leaves parts of netinet and netinet6 in a situation
where they may panic or leak memory, as they have not are not fully
updated by this commit.  This will be corrected shortly in followup
commits to these components.

MFC after:      3 months
2006-04-01 15:15:05 +00:00
Christian Brueffer
c66235a6ab Convert the SYNOPSIS section to look like the ones used in other driver
manpages, mention module support.

MFC after:	3 days
2006-04-01 11:21:00 +00:00
Robert Watson
fa4c5373ce Add comment to accept1() that it should use getsock() instead of fgetsock()
to avoid additional mutex operations, and also to avoid use of soref/sorele
which are now not preferred.

MFC after:	3 months
2006-04-01 11:14:56 +00:00
Robert Watson
197b35d717 Mark fgetsock() and fputsock() as depcrecated: callers should rely on
the file descriptor reference, rather than paying additional lock
operations to acquire a socket reference from the file descriptor.
This will also help to ensure that file descriptor based socket
requests are not delivered to a socket after close.  Most consumers
have already been converted to this model.

MFC after:	3 months
2006-04-01 11:09:54 +00:00
Christian Brueffer
449e2f5c77 Convert the SYNOPSIS section to look like the ones used in other driver
manpages, mention module support.
Also add the crypto and cryptodev devices as the drivers are kind of useless
without them.

MFC after:	3 days
2006-04-01 10:56:36 +00:00
Christian Brueffer
1f12b08ab4 Convert the SYNOPSIS section to look like the ones used in other driver
manpages.

MFC after:	3 days
2006-04-01 10:55:00 +00:00
Robert Watson
c43a4e8a83 Add a comment describing SS_PROTOREF in detail. This will eventually be
in socket(9).

MFC after:	3 months
2006-04-01 10:54:51 +00:00
Robert Watson
7f689de232 Assert so->so_pcb is NULL in sodealloc() -- the protocol state should not
be present at this point.  We will eventually remove this assert because
the socket layer should never look at so_pcb, but for now it's a useful
debugging tool.

MFC after:	3 months
2006-04-01 10:45:52 +00:00
Robert Watson
220c1357ed Add a somewhat sizable comment documenting the semantics of various kernel
socket calls relating to the creation and destruction of sockets.  This
will eventually form the foundation of socket(9), but is currently in too
much flux to do so.

MFC after:	3 months
2006-04-01 10:43:02 +00:00
Christian Brueffer
7573f75456 Convert the SYNOPSIS section to look like the ones used in other driver
manpages.  Don't mention the include file, it's not important for the
operation of this driver.

MFC after:	3 days
2006-04-01 10:37:35 +00:00
Christian Brueffer
c541fe558c Add a standard boilerplate to the SYNOPSIS section that mentions how
to load a kernel module.

MFC after:	3 days
2006-04-01 10:10:36 +00:00
Christian Brueffer
7bffbceefe Mention the module in the synopsis. 2006-04-01 09:55:37 +00:00
Matt Jacob
5e073106d5 Fix some of the previus changes 'better'.
There's something strange going on with async events. They seem
to be be treated differently for different Fusion implementations.
Some will really tell you when it's okay to free the request that
started them.  Some won't. Very disconcerting.

This is particularily bad when the chip (FC in this case) tells you
in the reply that it's not a continuation reply, which means you
can free the request that its associated with. However, if you do
that, I've found that additional async event replies come back for
that message context after you freed it. Very Bad Things Happen.

Put in a reply register debounce. Warn about out of range context
indices. Use more MPILIB defines where possible. Replace bzero with
memset. Add tons more KASSERTS. Do a *lot* more request free list
auditting and serial number usages. Get rid of the warning about
the short IOC Facts Reply.  Go back to 16 bits of context index.

Do a lot more target state auditting as well. Make a tag out
of not only the ioindex but the request index as well and worry
less about keeping a full serial number.
2006-04-01 07:12:18 +00:00
Marcel Moolenaar
a75b305ac6 Build the scc(4) module with EBus and SBus attachments for sparc64
only and build the scc(4) module with MacIO attachment for powerpc.
2006-04-01 04:54:47 +00:00
Marcel Moolenaar
2825701d6c Add the MacIO attachment for scc(4). 2006-04-01 04:53:08 +00:00
Marcel Moolenaar
ffc08b2147 Add a MacIO bus attachment. The Z8530 as present in the Mac needs
a different register shift and is fed by a different clock than
we use for UltraSPARC hardware. To deal with this, the regshft and
rclk fields in the class structure are removed and bus frontends
now pass the right regshft and rclk to the probe function where
they're put in the BAS and passed in to subordinate drivers.
2006-04-01 04:51:56 +00:00
Daniel Eischen
9fe1984e64 Add __gdtoa to the list of FreeBSD private symbols. Unfortunately,
this is used by some 3rd party applications when {e,f,g}cvt() are
not found.  POSIX defines the xcvt() funtions but says they are
deprecated in favor or sprintf().  We'll import these functions
from OpenBSD and remove __gdtoa() from the exported interfaces
when libc version is bumped.
2006-04-01 02:56:09 +00:00
Jeff Roberson
b2282f9a3f - Busy the filesystem in nfs_statfs to prevent us from creating a new
vnode after vflush() has succeeded.  This would cause a dangling vnode
   panic at unmount time otherwise.  Other filesystems may have this problem
   via their VFS_VGET() routines.

Found by:	kris
Sponsored by:	Isilon Systems, Inc.
2006-04-01 01:15:23 +00:00
Jeff Roberson
0af2472199 - Add an assert to vgone. It is illegal to call vgone without a reference
to the vnode.  Without a reference the vnode will never be vdestroy'd
   and the memory will never be reclaimed.

Sponsored by:	Isilon Systems, Inc.
2006-03-31 23:39:26 +00:00
Jeff Roberson
ba5eb429e3 - When there are dangling vnodes at unmount print them before we panic.
Sponsored by:	Isilon Systems, Inc.
2006-03-31 23:38:15 +00:00
Jeff Roberson
23b77994f2 - Add a bogus vhold/vdrop around vgone() in devfs_revoke. Without this
the vnode is never recycled.  It is bogus because the reference really
   should be associated with the devfs dirent.
2006-03-31 23:37:29 +00:00
Maksim Yevmenkin
b643101293 Add kbdmux(4) to GENERIC on amd64
Requested by:	scottl
Tested by:	scottl
2006-03-31 23:04:48 +00:00
Marcel Moolenaar
19ef0b1184 Fix cut-n-paste braino in previous commit: s/puc/scc/g
Pointy hat: marcel@
2006-03-31 21:55:53 +00:00
Scott Long
c2e1d5d04c Remove the USB keyboard hack now that KBDMUX is enabled by default. Allow
it to be disabled if Safe Mode is selected.
2006-03-31 21:36:17 +00:00
Maksim Yevmenkin
9216fccdd9 Add kbdmux(4) to GENERIC
Requested by:	scottl
2006-03-31 19:03:37 +00:00
Marcel Moolenaar
1b0c333020 Add a DRIVER_MODULE declaration for fhc(4) as this attachement is
also used for the FHC bus.

Pointed out by: marius@
2006-03-31 17:39:49 +00:00
Ruslan Ermilov
d6acfaf5e6 Sort bootstrap-tools entries. 2006-03-31 14:10:35 +00:00
Julian Elischer
905c41b537 Amazing.. two screwups in one commit.
I'm piling on thise pointy hats on top of each other.
At least they nest..
2006-03-31 12:54:17 +00:00
Julian Elischer
dbadd6b028 I can't believe that no-one noticed that I broke ipfw table del
for over a month!
put {}  around if clause with multiple statements
2006-03-31 12:40:09 +00:00
Ariff Abdullah
a1d444e140 MEGA Fixes / Cleanup
--------------------

- Seal the fate of long standing memory leak (4 years, 7 months) during
  pcm_unregister(). While destroying cdevs, scan / detect possible
  children and free its SLIST placeholder properly.
- Optimize channel allocation / numbering even further. Do brute cyclic
  checking only if the channel numbering screwed.
- Mega vchan create/destroy cleanup:
  o Implement pcm_setvchans() so everybody can use it freely instead
    of implementing their own, be it through sysctl or channel auto
    allocation.
  o Increase vchan creation/destruction resiliency:
    + it's possible to increase/decrease total vchans even during
      busy playback/recording. Busy channel will be left alone, untouched.
      Abusive test sample:
      # play whatever...
      #
         while : ; do
           sysctl hw.snd.pcm0.vchans=1
           sysctl hw.snd.pcm0.vchans=10
           sysctl hw.snd.pcm0.vchans=100
           sysctl hw.snd.pcm0.vchans=200
         done
      # Play something else, leave above loop running frantically.
    + Seal another 4 years old bug where it is possible to destroy (virtual)
      channel even when its cdevs being referenced by other process.
      The "First Come First Served" nature of dsp_clone() is the main
      culprit of this issue, and usually manifest itself as dangling
      channel <-> process association. Ensure that all of its cdevs
      are free from being referenced before destroying it (through
      ORPHAN_CDEVT() macross).

All these fixes (including previous fixes) will be MFCed, later.
2006-03-31 10:36:36 +00:00
Ariff Abdullah
9c271eba78 - Increase snddev refcount earlier while accessing through sysctl interface
to avoid possible device unregister race (impossible to reproduce, yet
  possible).
- Extra sanity check to ensure proper parent channel is being selected.
- Reset parent channel once all of its children gone.
2006-03-31 10:27:47 +00:00
Søren Schmidt
cd8a592bb3 Make the ATAPI sense data accessible when using the ioctl interface
MFC candidate.
2006-03-31 08:09:05 +00:00
Jeff Roberson
3bbd6d8ae6 - Release the references acquired by VOP_GETWRITEMOUNT and vfs_getvfs().
Discussed with:	tegge
Tested by:	kris
Sponsored by:	Isilon Systems, Inc.
2006-03-31 03:54:20 +00:00
Jeff Roberson
94bc95db3c - Hold a reference from the time vfs_busy starts until vfs_unbusy is
called.
 - vfs_getvfs has to return a reference to prevent the returned mountpoint
   from changing identities.
 - Release references acquired via vfs_getvfs.

Discussed with:	tegge
Tested by:	kris
Sponsored by:	Isilon Systems, Inc.
2006-03-31 03:53:25 +00:00
Jeff Roberson
c5fcce21c5 - GETWRITEMOUNT now returns a referenced mountpoint to prevent its
identity from changing.  This is possible now that mounts are not freed.

Discussed with:	tegge
Tested by:	kris
Sponsored by:	Isilon Systems, Inc.
2006-03-31 03:52:24 +00:00
Jeff Roberson
a218edceb2 - Allocate mounts from a uma zone that uses UMA_ZONE_NOFREE to prevent
mount memory from being reclaimed.  This resolves a number of race
   conditions described in vfs_default.c and introduced with the
   VFS_LOCK_GIANT macros.
 - Let the mtx and lock remain valid after the mount structure has been
   freed by using init and fini calls.  Technically fini will never be
   called but is included for completeness.
 - Consistently use lockmgr directly rather than lockmgr to lock and
   vfs_unbusy to unlock.

Discussed with:	tegge
Tested by:	kris
Sponsored by:	Isilon Systems, Inc.
2006-03-31 03:49:51 +00:00
Jeff Roberson
9eee260594 - Define mnt_startzero and mnt_endzero as a range that excludes mnt_mtx
and mnt_lock so that the mountpoint can be explicitly zeroed on
   creation.

Discussed with: tegge
Tested by:      kris
Sponsored by:   Isilon Systems, Inc.
2006-03-31 03:49:16 +00:00
Jeff Roberson
fdf86b2dcd - LK_RETRY means nothing when passed to VOP_LOCK. Call vn_lock instead.
- Move the vn_lock of the dvp until after we've unbusied the filesystem
   to avoid a LOR with the mount point lock.
 - In the v_mountedhere while loop we acquire a new instance of giant each
   time through without releasing the first.  This would cause us to leak
   Giant.

Sponsored by:	Isilon Systems, Inc.
2006-03-31 02:59:23 +00:00
Jeff Roberson
084d64ac21 - Add the B_NEEDSGIANT flag which is only set if the vnode that owns a buf
requires Giant.  It is set in bgetvp and cleared in brelvp.
 - Create QUEUE_DIRTY_GIANT for dirty buffers that require giant.
 - In the buf daemon, only grab giant when processing QUEUE_DIRTY_GIANT and
   only if we think there are buffers in that queue.

Sponsored by:	Isilon Systems, Inc.
2006-03-31 02:56:30 +00:00
Marcel Moolenaar
c9e0c5b9c4 Build uart(4) on PowerPC. 2006-03-31 01:50:23 +00:00
Marcel Moolenaar
ce8f00136f Allow uart(4) to be built on PowerPC. 2006-03-31 01:42:55 +00:00
Marcel Moolenaar
bc1fefaa7d Add a dummy implementation of bus_space_map(). 2006-03-31 01:39:50 +00:00
John-Mark Gurney
b5fbf2189d remove these vestiges of gdbserver.. It hasn't built in a long time, and
now really belongs in gdb, not binutils..  Plus, these don't resemble what
the new gdbserver looks like.
2006-03-31 00:37:42 +00:00
Sam Leffler
00537061dd fixup error handling in taskqueue_start_threads: check for kthread_create
failing, print a message when we fail for some reason as most callers do
not check the return value (e.g. 'cuz they're called from SYSINIT)

Reviewed by:	scottl
MFC after:	1 week
2006-03-30 23:06:59 +00:00
Marcel Moolenaar
1411591a14 Include the sbus attachment of scc(1) when either fhc(4) or sbus(4)
is configured.
2006-03-30 21:39:36 +00:00
Christian S.J. Peron
ae4f0fc3a9 Currently, if writing out a log entry fails, we unlink that log entry from our
internal list of logfiles. So if writev(2) fails for potentially transient
errors like ENOSPC, syslogd requires a restart, even if the filesystem has
purged.

This change allows syslogd to ignore ENOSPC space errors, so that when the
filesystem is cleaned up, syslogd will automatically start logging again
without requiring the reset. This makes syslogd(8) a bit more reliable.

MFC after:	1 week
2006-03-30 21:04:52 +00:00
Jason Evans
cd70100e5d Optimize runtime performance, primary using the following techniques:
* Avoid choosing an arena until it's certain that an arena is needed
    for allocation.

  * Convert division/multiplication to bitshifting where possible.

  * Avoid accessing TLS variables in single-threaded code.

  * Reduce the amount of pointer dereferencing.

  * Move lock acquisition in critical paths to only protect the the code
    that requires synchronization, and completely remove locking where
    possible.
2006-03-30 20:25:52 +00:00
Hajimu UMEMOTO
eb27a220f8 Expose res_update and friends again. At least, ports/mail/spamilter
uses them.
Now, we have res_nupdate and res_nmkupdate as well, but they are
still based on our old resolver for binary backward compatibility.
So, they don't provide new features such as TSIG support.

Reported by:	pointyhat via kris
2006-03-30 20:15:53 +00:00
Scott Long
cadfc102f1 Revert to using acpi_max_threads instead of the hardcoded value of '3'. 2006-03-30 19:22:45 +00:00