dragonfly.git
11 hours agohammer2 - fix I/O race, flesh out bulkfree master
Matthew Dillon [Fri, 30 Jan 2015 02:26:34 +0000 (18:26 -0800)]
hammer2 - fix I/O race, flesh out bulkfree

* Fix an I/O race which could cause hammer2 to block on an iocb forever.
  A TAILQ was being tested outside of a spin-lock and could race the
  related atomic op on the flag.

* Add signal checking for the bulkfree scan, allowing it to be interrupted.

* Loop sbase/sstop to cover the whole volume.  The in-memory bitmap must
  be bzero()d on each loop.

19 hours agohammer2 - work on bmap->avail tracking, fix DIO issue w/block reuse
Matthew Dillon [Thu, 29 Jan 2015 18:21:46 +0000 (10:21 -0800)]
hammer2 - work on bmap->avail tracking, fix DIO issue w/block reuse

* Start working on tracking bmap->avail in a way that the bulkfree
  code can undo.

* DIO was allowing mixed 16/64K buffers, but bulkfree doesn't invalidate
  cached DIOs and the re-use case might want a larger buffer.  For now
  deal with it by just having DIO use 64K buffers.

40 hours agoUPDATING: Mention recently removed kernel options.
Sascha Wildner [Wed, 28 Jan 2015 21:02:46 +0000 (22:02 +0100)]
UPDATING: Mention recently removed kernel options.

41 hours agoRemove <sys/_posix.h>. It serves no purpose anymore.
Sascha Wildner [Wed, 28 Jan 2015 20:58:53 +0000 (21:58 +0100)]
Remove <sys/_posix.h>. It serves no purpose anymore.

41 hours agokernel: Retire the P1003_1B and _KPOSIX_VERSION kernel options.
Sascha Wildner [Wed, 28 Jan 2015 20:56:43 +0000 (21:56 +0100)]
kernel: Retire the P1003_1B and _KPOSIX_VERSION kernel options.

41 hours ago<sys/semaphore.h>: Remove dead p1003_1b code.
Sascha Wildner [Wed, 28 Jan 2015 19:25:08 +0000 (20:25 +0100)]
<sys/semaphore.h>: Remove dead p1003_1b code.

41 hours ago<sys/sysctl.h>: Remove <sys/_posix.h> include. It is not needed here.
Sascha Wildner [Wed, 28 Jan 2015 19:27:02 +0000 (20:27 +0100)]
<sys/sysctl.h>: Remove <sys/_posix.h> include. It is not needed here.

41 hours agokernel/mib: Remove <sys/_posix.h> dependency.
Sascha Wildner [Wed, 28 Jan 2015 20:20:34 +0000 (21:20 +0100)]
kernel/mib: Remove <sys/_posix.h> dependency.

This commit bumps the value of the kern.posix1version sysctl from 199309 to
200112. I know no port which checks this sysctl.

42 hours agokernel/snd_hda: Enable uncacheable DMA attribute as default again.
Imre Vadasz [Wed, 28 Jan 2015 19:32:18 +0000 (20:32 +0100)]
kernel/snd_hda: Enable uncacheable DMA attribute as default again.

* Set the HDAC_F_DMA_NOCACHE flag as default (Reinserts a line that was
  removed in commit 403fad76b1ecab3ce8ef68b48445be3a9b0a4e74).

* Allocate at least PAGE_SIZE aligned memory in hdac_dma_alloc() when
  HDAC_F_DMA_NOCACHE is set.

42 hours ago<sys/mman.h>: Some cleanup.
Sascha Wildner [Wed, 28 Jan 2015 19:11:20 +0000 (20:11 +0100)]
<sys/mman.h>: Some cleanup.

* Handle _P1003_1B_VISIBLE and _POSIX_SOURCE checks with __POSIX_VISIBLE.

* Separate BSD specific defines/prototypes better.

47 hours ago<sys/mman.h>: Whitespace cleanup (no functional change).
Sascha Wildner [Mon, 26 Jan 2015 19:19:57 +0000 (20:19 +0100)]
<sys/mman.h>: Whitespace cleanup (no functional change).

2 days agobsd.hostlib.mk: Fix case where SRCS is defined but empty
John Marino [Wed, 28 Jan 2015 11:47:55 +0000 (12:47 +0100)]
bsd.hostlib.mk: Fix case where SRCS is defined but empty

This is similar to the recent changes to bsd.lib.mk and bsd.prog.mk.
The check covers two augmentations to OBJS, plus let's add some indents
for readability.  Note the second augmentation didn't have a bug in it,
so including it inside the wrap only saves having to evaluate it.

2 days agoinet6: Dispatch nd6_timer to netisr0 to run
Sepherosa Ziehau [Sat, 24 Jan 2015 05:42:14 +0000 (13:42 +0800)]
inet6: Dispatch nd6_timer to netisr0 to run

This function now accesses ifindex2ifnet global variable in netisr0.

2 days agolib/libhammer: cleanups
Tomohiro Kusumi [Sun, 25 Jan 2015 14:13:04 +0000 (23:13 +0900)]
lib/libhammer: cleanups

Using pointer dereference instead of several mntbuf[curmount].f_xxx
makes code clear and also makes lines fit within 80 characters.

2 days agolib/libhammer: cleanups
Tomohiro Kusumi [Sun, 25 Jan 2015 13:41:59 +0000 (22:41 +0900)]
lib/libhammer: cleanups

This patch has some minor cleanups of libhammer.

hunk1: Remove parentheses
hunk2: Remove parentheses
hunk3: Remove forward declaration of obsolete function.
       find_pfs_mount() was re-implemented as a part of libhammer
       and commit 4f09feab replaces it with libhammer version.

2 days agolib/libhammer: fix minor memory leaks
Tomohiro Kusumi [Sun, 25 Jan 2015 13:38:50 +0000 (22:38 +0900)]
lib/libhammer: fix minor memory leaks

This patch fixes minor memory leaks of libhammer.

hunk1: This isn't necessary since it never gets here.
       _libhammer_malloc() exit(1) if it failed to calloc().
hunk2: Add missing close(fd).
hunk3: Add missing free(mntbuf).

2 days agohammer: some trivial fixes
Tomohiro Kusumi [Sat, 24 Jan 2015 15:55:21 +0000 (00:55 +0900)]
hammer: some trivial fixes

This patch fixes some trivial things.

hunk1: Cleanup incomplete sentence
hunk2: Sync comment with changes made in 4a2796f3
hunk3: Use right function name
hunk4: Add missing comment
hunk5: Add missing \n

Closes: #2777

2 days agobsd.*.mk: Fix bug when SRCS is explicitly defined empty
John Marino [Tue, 27 Jan 2015 22:20:44 +0000 (23:20 +0100)]
bsd.*.mk: Fix bug when SRCS is explicitly defined empty

There exists a valid possibility of setting a makefile to build a program
or library exclusively with OBJS defined.  In the case of bsd.prog.mk,
not defining SRCS will cause the definition of OBJS to be overwritten
with ${PROG}.c or ${PROG}.cc.  Defining SRCS as an empty list will stop
OBJS from being overwritten, but there's a bug.  It adds ".o" to the end
of the list.

For example:
======================
OBJS= first.o second.o
SRCS=
.include <bsd.prog.mk>
======================
make -V OBJS
first.o second.o .o

This is caused by a transformation assuming SRCS has at least one item
in the list.  It's fixed by only performing the transformation if SRCS
is not empty.  I'm adding this check to both bsd.prog.mk and bsd.lib.mk

2 days agolibc/sigaltstack: Return EPERM upon disabling the active stack, not EINVAL.
Sascha Wildner [Tue, 27 Jan 2015 21:54:57 +0000 (22:54 +0100)]
libc/sigaltstack: Return EPERM upon disabling the active stack, not EINVAL.

Per POSIX, FreeBSD and OpenBSD agree too. NetBSD doesn't, hmmm.

2 days agokernel/sound: Always create the devfs aliases when loading a sound driver.
Imre Vadasz [Tue, 27 Jan 2015 18:36:16 +0000 (19:36 +0100)]
kernel/sound: Always create the devfs aliases when loading a sound driver.

3 days agokernel/snd_hda: Fix module unloading.
Sascha Wildner [Tue, 27 Jan 2015 09:28:49 +0000 (10:28 +0100)]
kernel/snd_hda: Fix module unloading.

Submitted-by: ivadasz
3 days agokernel/snd_hda: Re-apply 6d233c5d040778c2655923f2a20676e7e519ceca.
Sascha Wildner [Tue, 27 Jan 2015 09:27:59 +0000 (10:27 +0100)]
kernel/snd_hda: Re-apply 6d233c5d040778c2655923f2a20676e7e519ceca.

It still occurs for us.

Pointed-out-by: ivadasz
3 days agoClarify message about 'make rescue' after installworld.
Sascha Wildner [Tue, 27 Jan 2015 08:12:20 +0000 (09:12 +0100)]
Clarify message about 'make rescue' after installworld.

3 days agoopenssh: Adjust DragonFly version to actual openssh p1 release date.
Robin Hahling [Tue, 27 Jan 2015 05:56:42 +0000 (06:56 +0100)]
openssh: Adjust DragonFly version to actual openssh p1 release date.

3 days agolibusb: Sync with FreeBSD.
Sascha Wildner [Tue, 27 Jan 2015 05:31:19 +0000 (06:31 +0100)]
libusb: Sync with FreeBSD.

3 days agohammer2 - Adjust TODO
Matthew Dillon [Tue, 27 Jan 2015 02:07:05 +0000 (18:07 -0800)]
hammer2 - Adjust TODO

* Add bulkfree TODO items.

3 days agohammer2 - Add bulkfree directive.
Matthew Dillon [Tue, 27 Jan 2015 02:04:30 +0000 (18:04 -0800)]
hammer2 - Add bulkfree directive.

* Add a bulkfree directive for testing the bulkfree code.

3 days agohammer2 - bulkfree work 2/many
Matthew Dillon [Tue, 27 Jan 2015 02:00:53 +0000 (18:00 -0800)]
hammer2 - bulkfree work 2/many

* Compare the in-memory bitmap against the on-media bitmap and issue
  transitions.

  in-memory 00 - 11 (allocated) -> 10 (staged)
  in-memory 00 - 10 (staged)    -> 00 (free)
  in-memory 11 - 10 (staged)    -> 11 (allocated) (handles live races)

3 days agohammer2 - Stabilize new I/O infrastructure and work on bulkfree
Matthew Dillon [Mon, 26 Jan 2015 23:48:03 +0000 (15:48 -0800)]
hammer2 - Stabilize new I/O infrastructure and work on bulkfree

* Stabilize the new asynchronous hammer_io infrastructure.

* Document the new simpler bulkfree scan algorithm.  The new algorithm
  scans meta-data in two passes, rebuilding the freemap in kernel memory.
  It then compares the map against the live media.

  The first pass transitions the freemap from (11 = allocated) to
  (10 = staged for deletion), the second pass transitions (10) either back
  to (11) if the block is found to not actually be freeable or to (00) if
  the block winds up being actually freeable.

  The two pass mechanism deals with any races and allows the scan to occur
  essentially unlocked,

* Start work on the bulkfree scan algorithm.

3 days agokernel - Add swap-backed kernel memory allocator API
Matthew Dillon [Mon, 26 Jan 2015 23:46:25 +0000 (15:46 -0800)]
kernel - Add swap-backed kernel memory allocator API

* Add kmem_alloc_swapbacked() and kmem_free_swapbacked().  Allocates kernel
  memory which is pageable and backed by swap.

3 days agokernel - limit backtrace-on-panic to 6 call levels
Matthew Dillon [Mon, 26 Jan 2015 23:45:11 +0000 (15:45 -0800)]
kernel - limit backtrace-on-panic to 6 call levels

* Limit the backtrace-on-panic to 6 call levels to try to keep more
  information leading up to the crash (such as the panic message) on
  row-limited consoles.

4 days agopxeboot.8: Small adjustments.
Sascha Wildner [Mon, 26 Jan 2015 12:41:23 +0000 (13:41 +0100)]
pxeboot.8: Small adjustments.

4 days agoopenssh: Fix version stuff.
Peter Avalos [Sun, 25 Jan 2015 22:00:28 +0000 (14:00 -0800)]
openssh: Fix version stuff.

The previous code was trying to get too fancy by trying to use
configuration options in places where those options weren't sent yet.

Fix a case in sshconnect.c where version_addendum wasn't being used.

Update README.DRAGONFLY to reflect the latest upgrade instructions.

4 days agoMerge branch 'openssh'
Sascha Wildner [Sun, 25 Jan 2015 20:05:45 +0000 (21:05 +0100)]
Merge branch 'openssh'

4 days agolibssh: remove some unused files vendor/OPENSSH
Robin Hahling [Sun, 25 Jan 2015 19:49:14 +0000 (20:49 +0100)]
libssh: remove some unused files

4 days agolibssh: Fix a remaining warning and raise WARNS to 2.
Sascha Wildner [Sun, 25 Jan 2015 19:03:55 +0000 (20:03 +0100)]
libssh: Fix a remaining warning and raise WARNS to 2.

4 days agolibssh: Silence some redefinition warnings.
Sascha Wildner [Sun, 25 Jan 2015 18:33:15 +0000 (19:33 +0100)]
libssh: Silence some redefinition warnings.

Taken-from:   FreeBSD
Submitted-by: Robin Hahling <robin.hahling@gw-computing.net>
5 days agofile(1): Unbreak buildworld.
Sascha Wildner [Sun, 25 Jan 2015 12:54:27 +0000 (13:54 +0100)]
file(1): Unbreak buildworld.

The issue is the missing definition of wchar_t in <inttypes.h> (required
by POSIX too) which file(1) winds up needing because of new code in this
version.

The commit 3e581a72263dd3b726387b5dc2399a066511099f fixes it correctly
but mkmagic.nx is a build-tool and so uses host includes (which don't
have that <inttypes.h> fix yet at this point).

I see no other way to fix this easily right now than just including
<wchar.h>. This has a wchar_t definition, too.

Reported-by: Robin Hahling <robin.hahling@gw-computing.net>
5 days agodrm: Implement and use the Linux version of kfree()
François Tigeot [Sun, 24 May 2015 07:53:21 +0000 (08:53 +0100)]
drm: Implement and use the Linux version of kfree()

5 days agossh: Raise WARNS back to 2.
Sascha Wildner [Sun, 25 Jan 2015 03:09:47 +0000 (04:09 +0100)]
ssh: Raise WARNS back to 2.

5 days agossh: Add a missing include for bindresvport_sa().
Sascha Wildner [Sun, 25 Jan 2015 02:50:33 +0000 (03:50 +0100)]
ssh: Add a missing include for bindresvport_sa().

5 days agorc.d/sshd: Simplify key generation.
Peter Avalos [Sat, 24 Jan 2015 23:10:51 +0000 (15:10 -0800)]
rc.d/sshd: Simplify key generation.

ssh-keygen(1) already provides the functionality we were previously
trying to do here with its -A flag, except now it will reduce the
maintenance burden of having to update this file when there are new key
types.

5 days agoUpdate files for file-5.22 import.
Peter Avalos [Sat, 24 Jan 2015 22:56:40 +0000 (14:56 -0800)]
Update files for file-5.22 import.

5 days agoImport file-5.22.
Peter Avalos [Sat, 24 Jan 2015 20:48:01 +0000 (12:48 -0800)]
Import file-5.22.

5 days agointtypes.h: Add typedef for wchar_t.
Peter Avalos [Sat, 24 Jan 2015 22:36:29 +0000 (14:36 -0800)]
inttypes.h: Add typedef for wchar_t.

5 days agoMerge new origin commits.
Peter Avalos [Sat, 24 Jan 2015 20:15:30 +0000 (12:15 -0800)]
Merge new origin commits.

6 days agoliblvm: Request _XOPEN_SOURCE=700 for alphasort() prototype.
Sascha Wildner [Sat, 24 Jan 2015 13:09:48 +0000 (14:09 +0100)]
liblvm: Request _XOPEN_SOURCE=700 for alphasort() prototype.

6 days ago<dirent.h>: Don't check _POSIX_SOURCE.
Sascha Wildner [Sat, 24 Jan 2015 13:19:59 +0000 (14:19 +0100)]
<dirent.h>: Don't check _POSIX_SOURCE.

Use the __POSIX_VISIBLE, __XSI_VISIBLE and __BSD_VISIBLE macros, per
<sys/cdefs.h>.

6 days agolibc: Change dirfd() into a function (needed per POSIX).
Sascha Wildner [Sat, 24 Jan 2015 13:15:08 +0000 (14:15 +0100)]
libc: Change dirfd() into a function (needed per POSIX).

At the same time, make struct _dirdesc private to libc. If access
to dd_fd is needed, dirfd() has to be used. Inside libc, the former
dirfd() macro is preserved as _dirfd().

Also, in struct _dirdesc, change the lock from "void *" to
"struct pthread_mutex *". This allows us to eliminate some casts
in readdir() and friends.

Our documentation already describes dirfd() as a function.

Based-on: FreeBSD's similar work

6 days ago<dirent.h>: Adjust whitespace better (no functional change).
Sascha Wildner [Sat, 24 Jan 2015 13:11:54 +0000 (14:11 +0100)]
<dirent.h>: Adjust whitespace better (no functional change).

6 days agodrm: Improve linux/jiffies.h
François Tigeot [Sat, 24 Jan 2015 11:11:27 +0000 (12:11 +0100)]
drm: Improve linux/jiffies.h

6 days agodrm: Improve linux/pci.h
François Tigeot [Sat, 24 Jan 2015 10:56:16 +0000 (11:56 +0100)]
drm: Improve linux/pci.h

* Add struct pci_device_id

* Add PCI_DEVFN, obtained from FreeBSD's OFED subsystem

6 days agodrm: Add linux/backlight.h
François Tigeot [Sat, 24 Jan 2015 10:52:03 +0000 (11:52 +0100)]
drm: Add linux/backlight.h

6 days agoRemove blacklisted keys support from OpenSSH.
Peter Avalos [Sat, 24 Jan 2015 07:25:04 +0000 (23:25 -0800)]
Remove blacklisted keys support from OpenSSH.

As time progresses, this code becomes less useful. I left the code that
still scans the configuration option, and in the future it can be
removed as well.

The original author made a decent case for removing this in 2013:
https://lists.debian.org/debian-devel/2013/09/msg00240.html

6 days agoUpdate files for OpenSSH-6.7p1 import.
Peter Avalos [Sat, 24 Jan 2015 05:38:03 +0000 (21:38 -0800)]
Update files for OpenSSH-6.7p1 import.

This also updates the HPN patch to the most recent, which brings in the
multi-threaded cipher for improved performance.

6 days agosys/vfs/hammer: make description on low level storage layout up-to-date with code
Tomohiro Kusumi [Tue, 13 Jan 2015 11:13:28 +0000 (20:13 +0900)]
sys/vfs/hammer: make description on low level storage layout up-to-date with code

- This patch fixes description regarding hammer's low level storage
  layout based on two levels of blockmap layer. It aims to make the
  description explicit and up-to-date with what the actual code
  does.

Closes: #2767

6 days agopxeboot - Add option to improve NFS performance
Matthew Dillon [Fri, 23 Jan 2015 19:28:15 +0000 (11:28 -0800)]
pxeboot - Add option to improve NFS performance

* Allow nfs.read_size=N to be specified in /boot/loader.conf
  (1024-4096 allowed).  This allows the pxe server to tell
  the pxeboot code to use larger NFS data packets, significantly
  reducing kernel binary and module load times.

7 days agopxeboot - Workaround BIOS breakage
Matthew Dillon [Fri, 23 Jan 2015 08:30:38 +0000 (00:30 -0800)]
pxeboot - Workaround BIOS breakage

* Some BIOSes (the one in my Gigabyte H97N-WIFI) continue to use the original
  negotiated by the TFTP preloader even after we have negotiated a new IP
  and used it in the udpopen.  This will cause our dest_ip filter in
  udpread() to implode.  Detect and handle this case.

* Why does dhcpd hand out different IPs?  I tracked one down to the PXE
  BIOS sending a Client-ID option.  However, our PXEBOOT DHCP and our
  dhclient don't seem to send that option and also get different IPs.

* This won't fix post-boot breakage if dhclient rips the IP out from under
  the NFS root mount in a pxebooted situation.

7 days agolibstand - Don't allow the port# to underflow.
Matthew Dillon [Fri, 23 Jan 2015 08:29:26 +0000 (00:29 -0800)]
libstand - Don't allow the port# to underflow.

* Recycle port numbers if our decrement would underflow a reasonable
  low port.

7 days agoMerge branch 'vendor/OPENSSH'
Peter Avalos [Fri, 23 Jan 2015 05:54:26 +0000 (21:54 -0800)]
Merge branch 'vendor/OPENSSH'

7 days agoImport OpenSSH-6.7p1.
Peter Avalos [Wed, 26 Nov 2014 20:54:34 +0000 (12:54 -0800)]
Import OpenSSH-6.7p1.

7 days agokernel/usb4bsd: Remove unused malloc type declaration.
Sascha Wildner [Thu, 22 Jan 2015 22:39:27 +0000 (23:39 +0100)]
kernel/usb4bsd: Remove unused malloc type declaration.

7 days agokinfo.3: Fix prototype.
Sascha Wildner [Thu, 22 Jan 2015 22:23:07 +0000 (23:23 +0100)]
kinfo.3: Fix prototype.

7 days agoalc.4: Update the manual page for today's upgrade.
Sascha Wildner [Thu, 22 Jan 2015 14:59:00 +0000 (15:59 +0100)]
alc.4: Update the manual page for today's upgrade.

8 days agoinet6: Remove detecting of aggregatable unicast address (RFC2374).
Sepherosa Ziehau [Thu, 22 Jan 2015 07:54:24 +0000 (15:54 +0800)]
inet6: Remove detecting of aggregatable unicast address (RFC2374).

It has been deprecated for a long time.

Submitted-by: gpr
Dragonfly-bug: 2774

8 days agoecc: No need to go through parent device explicitly
Sepherosa Ziehau [Wed, 21 Jan 2015 14:17:05 +0000 (22:17 +0800)]
ecc: No need to go through parent device explicitly

8 days agoalc: Fix up MSI configuration
Sepherosa Ziehau [Thu, 22 Jan 2015 07:51:54 +0000 (15:51 +0800)]
alc: Fix up MSI configuration

Tested-by: dillon@
8 days agokernel - Update if_alc
Matthew Dillon [Thu, 22 Jan 2015 04:47:07 +0000 (20:47 -0800)]
kernel - Update if_alc

* Bring in everything FreeBSD has for if_alc.  This adds a few more
  generations including the one used (for the hardport) by my new
  Gigabyte H97N-WIFI mobo.

* Update mii as well.

* Disable MSI for now, doesn't seem to work with my test box.

Taken-From: FreeBSD

8 days agodrm - Add another Haswell graphics chipid
Matthew Dillon [Thu, 22 Jan 2015 04:45:09 +0000 (20:45 -0800)]
drm - Add another Haswell graphics chipid

* Add the gpu identified by the i3-4130 (haswell)

8 days agokernel - audio Makefile needs opt_snd.h
Matthew Dillon [Thu, 22 Jan 2015 04:44:30 +0000 (20:44 -0800)]
kernel - audio Makefile needs opt_snd.h

* Audio Makefile depends on opt_snd.h.

8 days agokernel - Add more definitions to pcireg.h
Matthew Dillon [Thu, 22 Jan 2015 04:43:54 +0000 (20:43 -0800)]
kernel - Add more definitions to pcireg.h

* Add more definitions to pcireg.h

Taken-From: FreeBSD

8 days agoacpi: Don't set non-exist EXT IRQ
Sepherosa Ziehau [Thu, 22 Jan 2015 01:33:04 +0000 (09:33 +0800)]
acpi: Don't set non-exist EXT IRQ

Certain mobo, e.g. gigabyte H97N-WIFI, reports EXT IRQ 37 and 38,
which do not exist at all (max legacy IRQ is 23).  Don't set bus
resource for this kind of IRQ resources.

Tested-by: dillon@
8 days agoecc: Use identify to add ecc device for E3-1200 memory controllers
Sepherosa Ziehau [Wed, 21 Jan 2015 12:50:25 +0000 (20:50 +0800)]
ecc: Use identify to add ecc device for E3-1200 memory controllers

Misc
- Disable ecc for X3400 temporarily; it seems to use E5 style ecc
  registers accessing method, which needs further investigation.
- Ecc should not be forcefully added to hostb

Tested-by: dillon@ on E3-1200 and E3-1200v3
8 days agoAdd a manual page for libkinfo (one for libkcore is still missing).
Sascha Wildner [Wed, 21 Jan 2015 23:48:00 +0000 (00:48 +0100)]
Add a manual page for libkinfo (one for libkcore is still missing).

8 days agorconfig - A couple additions to hammer.sh
Antonio Huete Jimenez [Wed, 21 Jan 2015 22:42:41 +0000 (23:42 +0100)]
rconfig - A couple additions to hammer.sh

Submitted-by: Rumko
Closes: #1173

8 days agolibkvm - Remove unused file
Antonio Huete Jimenez [Wed, 21 Jan 2015 22:33:02 +0000 (23:33 +0100)]
libkvm - Remove unused file

9 days agosysctl.3: Use .Vt instead of .Va for types.
Imre Vadasz [Tue, 20 Jan 2015 23:23:45 +0000 (00:23 +0100)]
sysctl.3: Use .Vt instead of .Va for types.

9 days agosysctl.3: Fix documentation of KERN_FILE.
Imre Vadasz [Tue, 20 Jan 2015 22:27:48 +0000 (23:27 +0100)]
sysctl.3: Fix documentation of KERN_FILE.

9 days agosys/vfs/hammer: make btree_search() climb up btree slightly faster
Tomohiro Kusumi [Thu, 8 Jan 2015 15:13:59 +0000 (00:13 +0900)]
sys/vfs/hammer: make btree_search() climb up btree slightly faster

- This cleanup patch lets cursor climb up deep/fat btree slightly faster
  (lightweight), if not slower.
- Remove a duplicate check that traces back to a moment where there was
  btree_node_is_almost_full() and the differentiation was needed.

Closes: #2765

9 days agosbin/hammer: fix potential memory leak
Tomohiro Kusumi [Mon, 12 Jan 2015 16:37:09 +0000 (01:37 +0900)]
sbin/hammer: fix potential memory leak

- This patch properly frees calloc/malloced memory.

Closes: #2766

9 days agosys/vfs/hammer: fix comments regarding obsolete "spike" code
Tomohiro Kusumi [Thu, 8 Jan 2015 12:29:12 +0000 (21:29 +0900)]
sys/vfs/hammer: fix comments regarding obsolete "spike" code

- This patch fixes comments regarding "spike" code that only existed in
  early days of hammer development.
- The spike comments have been left unchanged even after the code was
  removed in 47197d71536907482d2d215e222600eed3aedc0e.

Closes: #2764

9 days agosbin/hammer: remove obsolete "spike" code from hammer show command
Tomohiro Kusumi [Wed, 7 Jan 2015 12:45:43 +0000 (21:45 +0900)]
sbin/hammer: remove obsolete "spike" code from hammer show command

- Remove spike related code from print_btree_node() as it is no longer
  relevant since commit 47197d.
- Not calling print_btree_node() twice makes 'show' directive more
  efficient.

Closes: #2761

10 days agosys/vfs/hammer: fix off-by-one error in hammer volume-add ioctl
Tomohiro Kusumi [Sun, 18 Jan 2015 18:26:47 +0000 (03:26 +0900)]
sys/vfs/hammer: fix off-by-one error in hammer volume-add ioctl

- This patch should fix off-by-one error regarding maximum volume number.
- Correct how hammer_ioc_volume_add() tests current # of volumes before
  it adds a new volume.

Closes: #2773

10 days agoFix the man page: make sure all list items are bolded and a grammar tweak.
Justin C. Sherrill [Tue, 20 Jan 2015 02:24:54 +0000 (21:24 -0500)]
Fix the man page: make sure all list items are bolded and a grammar tweak.

Based on OpenBSD and FreeBSD changes that are exactly the same.

http://cvsweb.openbsd.org/cgi-bin/cvsweb/ports/editors/ee/patches/patch-ee_1?rev=1.4&content-type=text/x-cvsweb-markup&f=h

https://svnweb.freebsd.org/base/head/contrib/ee/ee.1?r1=276230&r2=277328&pathrev=277328

10 days agoecc: Implement detach and shutdown methods
Sepherosa Ziehau [Mon, 19 Jan 2015 13:30:04 +0000 (21:30 +0800)]
ecc: Implement detach and shutdown methods

While I'm here, clean up indentation.

10 days agosysctl.3: Remove KERN_UPDATEINTERVAL identifier. Fix a copy-paste error.
Imre Vadasz [Mon, 19 Jan 2015 22:35:00 +0000 (23:35 +0100)]
sysctl.3: Remove KERN_UPDATEINTERVAL identifier. Fix a copy-paste error.

10 days agokernel - limit DMA for older ALI chips to 256KB
Matthew Dillon [Mon, 19 Jan 2015 17:09:05 +0000 (09:09 -0800)]
kernel - limit DMA for older ALI chips to 256KB

* For ancient machines using old ALI chipsets which do
  not support 48-bit DMA, also limit the maximum DMA
  size to 256KB.

10 days agoipfw2 - Fix buffer overflow with large tables
Matthew Dillon [Mon, 19 Jan 2015 17:04:07 +0000 (09:04 -0800)]
ipfw2 - Fix buffer overflow with large tables

* Fix according to PR 1753.

Taken-from: FreeBSD

11 days agohammer - Attempt to fix improper call to vsetisdirty() in hammer flush
Matthew Dillon [Mon, 19 Jan 2015 04:17:42 +0000 (20:17 -0800)]
hammer - Attempt to fix improper call to vsetisdirty() in hammer flush

* The hammer flush occurs in the background.  Because of this it is possible
  for a vnode to be reclaimed just prior to the flush code getting to it.

* Attempt to fix an improper call to vsetisdirty() which can panic the machine
  on vp->v_mount becoming NULL.  The vnode reclamation should theoretically be
  interlocked against the inode lock so it should suffice to simply check
  to see if it has been reclaimed yet or not.

* Also don't bother marking the vnode dirty if it is already marked dirty.

11 days agoinet: Link ifa after ifa_{addr,dstaddr,netmask} are setup
Sepherosa Ziehau [Sun, 18 Jan 2015 13:53:36 +0000 (21:53 +0800)]
inet: Link ifa after ifa_{addr,dstaddr,netmask} are setup

While I'm here, remove unused in_interfaces.

11 days agobnx/bge: Add miibus module dependency
Sepherosa Ziehau [Sun, 18 Jan 2015 13:19:27 +0000 (21:19 +0800)]
bnx/bge: Add miibus module dependency

11 days agommap.2: A little markup..
Sascha Wildner [Sun, 18 Jan 2015 20:53:54 +0000 (21:53 +0100)]
mmap.2: A little markup..

11 days agoptrace.2: Document the PT_IO request.
Imre Vadasz [Sun, 18 Jan 2015 13:06:43 +0000 (14:06 +0100)]
ptrace.2: Document the PT_IO request.

Taken-from: FreeBSD

11 days agosysctl.3: Fix a typo. Document the HW_MACHINE_PLATFORM identifier.
Imre Vadasz [Sun, 18 Jan 2015 13:07:34 +0000 (14:07 +0100)]
sysctl.3: Fix a typo. Document the HW_MACHINE_PLATFORM identifier.

11 days agosysctl.3: Use .Dv where appropriate, use nicer syntax in tables.
Imre Vadasz [Sun, 18 Jan 2015 09:15:35 +0000 (10:15 +0100)]
sysctl.3: Use .Dv where appropriate, use nicer syntax in tables.

Mostly-taken-from: OpenBSD

11 days agommap.2: Update manpage to reflect the introduction of MAP_ANONYMOUS
Markus Pfeiffer [Sun, 18 Jan 2015 17:23:31 +0000 (17:23 +0000)]
mmap.2: Update manpage to reflect the introduction of MAP_ANONYMOUS

11 days agokernel: Exchange MAP_ANON and MAP_ANONYMOUS
Markus Pfeiffer [Sun, 18 Jan 2015 17:06:58 +0000 (17:06 +0000)]
kernel: Exchange MAP_ANON and MAP_ANONYMOUS

* In commit a2ed5072 I defined MAP_ANONYMOUS to take the role of
  MAP_ANON, and defined MAP_ANON as an alias. This lead to compatibility
  problems with programs that #define MAP_ANONYMOUS MAP_ANON, so
  #define MAP_ANON as before and let MAP_ANONYMOUS be the alias.

Reported-By: John Marino <draco@marino.st>
12 days agodrm: Sync drm_edid.c and drm_modes.c with Linux 3.9.11
François Tigeot [Sun, 18 Jan 2015 09:50:27 +0000 (10:50 +0100)]
drm: Sync drm_edid.c and drm_modes.c with Linux 3.9.11

12 days agodrm: Move vbl_lock initialization to the right place
François Tigeot [Sun, 18 Jan 2015 08:00:07 +0000 (09:00 +0100)]
drm: Move vbl_lock initialization to the right place

Reducing differences with Linux 3.9.11

12 days agosysctl.3: Document the KERN_PROC_CWD sysctl name. Explain KERN_PROC_ARGS.
Imre Vadasz [Sat, 17 Jan 2015 22:02:31 +0000 (23:02 +0100)]
sysctl.3: Document the KERN_PROC_CWD sysctl name. Explain KERN_PROC_ARGS.

Partly-taken-from: FreeBSD