dragonfly.git
10 years agodrm: Merge the drm and drm2 implementations
François Tigeot [Sun, 1 Sep 2013 19:38:11 +0000 (21:38 +0200)]
drm: Merge the drm and drm2 implementations

* Move the drm2 i915 driver to sys/dev/drm/i915kms

* The legacy i915 driver remains untouched

* Only sys/dev/drm/ shall remain

10 years agodrm/mach_64: Fix white spaces.
jkim [Tue, 23 Apr 2013 18:30:33 +0000 (18:30 +0000)]
drm/mach_64: Fix white spaces.

This brings us on part with FreeBSD commit r249816 (2013-04-23 20:30:33 +0200)

10 years agokernel: Rewrite kmem_alloc_attr()
François Tigeot [Sun, 1 Sep 2013 09:20:14 +0000 (11:20 +0200)]
kernel: Rewrite kmem_alloc_attr()

This prevents random kernel panics after Xorg exit with the new
drm scather-gather memory allocation code

10 years agodrm: Partial sync with FreeBSD
François Tigeot [Fri, 30 Aug 2013 06:46:37 +0000 (08:46 +0200)]
drm: Partial sync with FreeBSD

10 years agopppd.8: conversion to mdoc(7)
Franco Fichtner [Sat, 31 Aug 2013 10:34:57 +0000 (12:34 +0200)]
pppd.8: conversion to mdoc(7)

Major overhaul of a well-written manual page.  The content has
been shuffled around and subtly cleaned where appropriate.

Suggested-by: swildner
10 years agokernel: Fix some void functions returning the results of void functions.
Sascha Wildner [Sat, 31 Aug 2013 07:56:00 +0000 (09:56 +0200)]
kernel: Fix some void functions returning the results of void functions.

It was found with -pedantic, since -Wreturn-type will not warn about this
as long as the calling function has a return type of void, too.

10 years agosystat - cleanup netbw/pftop (2)
Matthew Dillon [Fri, 30 Aug 2013 23:34:40 +0000 (16:34 -0700)]
systat - cleanup netbw/pftop (2)

* Cleanup pass, fix a few bugs.

10 years agosystat - cleanup netbw/pftop
Matthew Dillon [Fri, 30 Aug 2013 23:22:57 +0000 (16:22 -0700)]
systat - cleanup netbw/pftop

* Cleanup pass, fix a few bugs.

10 years agosystat - Add -pftop feature to display active PF state table traffic
Matthew Dillon [Fri, 30 Aug 2013 22:30:19 +0000 (15:30 -0700)]
systat - Add -pftop feature to display active PF state table traffic

* Add the -pftop (:pftop) feature.  This feature displays all active
  PF states.  PF (packet filter) must be active for this feature to be
  useful.

  Only those states which are actively passing traffic are displayed,
  which makes this a lot more useful than a raw 'pfctl -s state' dump.

* Display is sorted by highest aggregate receive+transmit bandwidth.

10 years agoclog.8: break line after a sentence
Franco Fichtner [Fri, 30 Aug 2013 08:22:32 +0000 (10:22 +0200)]
clog.8: break line after a sentence

10 years agosyslog - add rexmit to -netbw
Matthew Dillon [Fri, 30 Aug 2013 06:41:08 +0000 (23:41 -0700)]
syslog - add rexmit to -netbw

* Add aggregate byte retransmission rate

10 years agosystat - further enhance -netbw
Matthew Dillon [Fri, 30 Aug 2013 06:36:21 +0000 (23:36 -0700)]
systat - further enhance -netbw

* Print IPs in a more readable fashion
* A better IP:port sort order
* Better alignment
* Add interesting state and TCP flags

10 years agosystat - Add -netbw option
Matthew Dillon [Fri, 30 Aug 2013 04:26:09 +0000 (21:26 -0700)]
systat - Add -netbw option

* Add a new option -netbw (or :netbw when running).  This option displays
  the aggregate and per-connection tcp receive and send data rate.

  Only active connections are shown, making it easy to pick out which
  connections are hogging the line.

10 years agokernel/ed: Fix typo in a variable name.
Sascha Wildner [Fri, 30 Aug 2013 02:48:45 +0000 (04:48 +0200)]
kernel/ed: Fix typo in a variable name.

10 years agotest/vkernel - Additions to leaf target
Antonio Huete Jimenez [Fri, 30 Aug 2013 00:06:38 +0000 (17:06 -0700)]
test/vkernel - Additions to leaf target

* Add nameservers to resolv.conf

10 years agofix mandoc(1) warnings in usr.sbin/
Franco Fichtner [Tue, 27 Aug 2013 20:38:19 +0000 (22:38 +0200)]
fix mandoc(1) warnings in usr.sbin/

10 years agotools/accept_connect: Remove unused .gitignore
Sepherosa Ziehau [Thu, 29 Aug 2013 06:15:36 +0000 (14:15 +0800)]
tools/accept_connect: Remove unused .gitignore

10 years agotcp: Implement random initial msgport
Sepherosa Ziehau [Wed, 28 Aug 2013 07:11:43 +0000 (15:11 +0800)]
tcp: Implement random initial msgport

After e368a6e95e2cd9556a3e0fc43167d2dcf3a8253f, it is found that the
left bottle neck of nonblocking TCP connect(2) performance is that all
socket(2) and initial TCP connect operation (bind laddr and lport) are
all carried out in netisr0; CPU0 is 100% busy during test.

The idea of random initial msgport for TCP is that instead of using
netisr0's msgport as initial msgport, we could use any of the available
netisr msgport to carry out socket(2) and initial TCP connect operation.

Most of parts of TCP are already ready for random initial msgport, only
TCP pru_listen requires trivial modification to fix the socket msgport
to netisr0's msgport (which is required to perform global wild hashtable
updating).

As of this commit, the current CPU's netisr msgport will be selected as
TCP socket's initial msgport, if random initial msgport is enabled.

Sysctl node kern.ipc.rand_initport is added to disable this optimization.
It is enabled by default.

This commit improves both nonblocking TCP connect(2) and blocking TCP
connect(2) performance.

Nonblocking connect(2) performance measurement (i7-2600 w/ bnx(4)), using
tools/tools/netrate/accept_connect/kq_connect_client:

    kq_connect_client -4 SERVADDR -p SERVPORT -i 8 -c 32 -l 30
    (8 processes, each creates 32 connections simultaniously)

16 run average:
    random initial msgport          netisr0 msgport
      263915.17 conns/s            220979.89 conns/s

This commit gives ~19% performance improvement for nonblocking connect(2)

Blocking connect(2) performance measurement (i7-2600 w/ bnx(4)), using
tools/tools/netrate/accept_connect/connect_client:

    connect_client -4 SERVADDR -p SERVPORT -i 256 -l 30
    (256 processes)

16 run average:
    random initial msgport          netisr0 msgport
      240235.23 conns/s            198312.87 conns/s

This commit gives ~21% performance improvement for blocking connect(2)

10 years agoinpcb: Make sure that inp_cpcbinfo always points to the correct pcbinfo
Sepherosa Ziehau [Wed, 28 Aug 2013 06:45:05 +0000 (14:45 +0800)]
inpcb: Make sure that inp_cpcbinfo always points to the correct pcbinfo

10 years agokernel: Fix LINT too.
Sascha Wildner [Wed, 28 Aug 2013 06:40:07 +0000 (08:40 +0200)]
kernel: Fix LINT too.

10 years agoproto: Fix comment
Sepherosa Ziehau [Wed, 28 Aug 2013 05:47:18 +0000 (13:47 +0800)]
proto: Fix comment

10 years agotcp: Fix comment
Sepherosa Ziehau [Wed, 28 Aug 2013 05:37:00 +0000 (13:37 +0800)]
tcp: Fix comment

10 years agobuild: Unbreak LINT64 building
Sepherosa Ziehau [Wed, 28 Aug 2013 05:17:44 +0000 (13:17 +0800)]
build: Unbreak LINT64 building

10 years agotcp: Implement asynchronous pru_connect
Sepherosa Ziehau [Tue, 20 Aug 2013 12:52:35 +0000 (20:52 +0800)]
tcp: Implement asynchronous pru_connect

This is mainly used to improve TCP nonblocking connect(2) performance.

Before this commit the user space thread uses nonblocking connect(2)
will have to wait for the netisr completes the SYN output.  This could
be performance hit for nonblocking connect(2).  First, the user space
thread is put into sleep, even if the connect(2) is nonblocking.
Second, it does not make too much sense for nonblocking connect(2) to
wait for the SYN output.

TCP's asynchronous pru_connect implementation will set ISCONNECTING
before dispatching netmsg to netisr0.  The errors like EADDRNOTAVAIL,
i.e. out of local port space, will be notified through kevent(2) or
getsockopt(2) SOL_SOCKET/SO_ERROR.

NFS and other kernel code still use old synchronized pru_connect.  This
commit only affects connect(2) syscall.

Sysctl node kern.ipc.soconnect_async is added to enable and disable
asynchronous pru_connect.  It is enabled by default.

The performance measurement (i7-2600 w/ bnx(4)), using
tools/tools/netrate/accept_connect/kq_connect_client:

    kq_connect_client -4 SERVADDR -p SERVPORT -i 8 -c 32 -l 30
    (8 processes, each creates 32 connections simultaniously, run 30 secs)

16 runs average:

    asynchronous pru_connect        synchronized pru_connect
        220979.89 conns/s               189106.88 conns/s

This commit gives ~16% performance improvement for nonblocking connect(2)

10 years agotools: Add test for nonblocking connect(2) performance measurement
Sepherosa Ziehau [Wed, 28 Aug 2013 03:16:49 +0000 (11:16 +0800)]
tools: Add test for nonblocking connect(2) performance measurement

10 years agovga_pci: Add API to map the Video BIOS
François Tigeot [Tue, 27 Aug 2013 10:17:12 +0000 (12:17 +0200)]
vga_pci: Add API to map the Video BIOS

Obtained-from: FreeBSD
SVN commit r254882 and a few missing bits

10 years agokernel/i386: Implement pmap_mapbios()
François Tigeot [Sun, 25 Aug 2013 19:57:49 +0000 (21:57 +0200)]
kernel/i386: Implement pmap_mapbios()

10 years agokernel/i386: Add PAT support, mapdev_attr, kmem_alloc_attr
François Tigeot [Tue, 27 Aug 2013 06:49:11 +0000 (08:49 +0200)]
kernel/i386: Add PAT support, mapdev_attr, kmem_alloc_attr

10 years agokernel/i386: add pmap_invalidate* functions
François Tigeot [Sun, 25 Aug 2013 11:54:12 +0000 (13:54 +0200)]
kernel/i386: add pmap_invalidate* functions

They are roughly similar to the amd64 implementations but due to
i386 pmap shortcomings we have to completely flush cpu caches in
some cases, which is not ideal performance-wise.

10 years agotop.1: Remove unnecessary .Nm argument.
Sascha Wildner [Tue, 27 Aug 2013 10:05:29 +0000 (12:05 +0200)]
top.1: Remove unnecessary .Nm argument.

10 years agokernel: Add three missing ')', two of them in (yet) unused code.
Sascha Wildner [Mon, 26 Aug 2013 18:51:02 +0000 (20:51 +0200)]
kernel: Add three missing ')', two of them in (yet) unused code.

10 years agokernel/cam: Fix I/O performance regression seen at least in VMware ESXi.
Sascha Wildner [Sun, 25 Aug 2013 18:15:37 +0000 (20:15 +0200)]
kernel/cam: Fix I/O performance regression seen at least in VMware ESXi.

This brings in FreeBSD's revisions 188570, 188670, 188671 and 188688.
188670 (most changed lines in the patch) is for debugging purposes only,
while 188570 and 188671 fix the actual issue and 188688 fixes gcc whining.

Reported-and-tested-by: Dongsheng Song <dongsheng.song@gmail.com>
Extra credit to vsrinivas who had actually pointed out 188570 to us a
while ago before it hit us now but at the time I hadn't noticed.

Quoting FreeBSD's commit messages:

r188570
-------
In the case that the probe has determined that it can't query the device
for a serial number, fall through to the next case so that initial
negotiation still happens.  Without this, devices were showing up with
only 1 available tag opening, leading to observations of very poor I/O
performance.

This should fix problems reported with VMWare Fusion and ESX.  Early
generation MPT-SAS controllers with SATA disks might also be affected.
HP CISS controllers are also likely affected, as are many other
pseudo-scsi disk subsystems.

r188671
-------
Fix parallel SCSI negotiation in the CAM_NEW_TRAN_CODE world order.
Overzealous sanity checks were locking the sync_rate and offset values
to zero, thanks to a twisty maze of recursive code.

10 years agofix mandoc(1) warnings in sbin/
Franco Fichtner [Sat, 24 Aug 2013 11:58:44 +0000 (13:58 +0200)]
fix mandoc(1) warnings in sbin/

dump.8 bits taken from FreeBSD.

10 years agogroff(1): disable hyphenation and change bullet representation
Franco Fichtner [Sat, 24 Aug 2013 10:49:11 +0000 (12:49 +0200)]
groff(1): disable hyphenation and change bullet representation

In order to switch to mandoc(1) eventually we want groff(1) to
output manual pages as similar as possible.  Hyphenation is the
biggest offender.

Approved-by: swildner
Taken-from:  OpenBSD

10 years agogroff: Commit our local patches.
Sascha Wildner [Sat, 24 Aug 2013 18:52:02 +0000 (20:52 +0200)]
groff: Commit our local patches.

10 years agoAvoid calling chn_wrfeed with an empty soft buffer
Francis GUDIN [Fri, 3 May 2013 12:28:27 +0000 (14:28 +0200)]
Avoid calling chn_wrfeed with an empty soft buffer

10 years agokernel: Add some missing sound devices to the LINT configs.
Sascha Wildner [Sun, 25 Aug 2013 06:00:13 +0000 (08:00 +0200)]
kernel: Add some missing sound devices to the LINT configs.

10 years agokernel: Remove unused lines in pc64's 'files' file.
Sascha Wildner [Sat, 24 Aug 2013 01:16:44 +0000 (03:16 +0200)]
kernel: Remove unused lines in pc64's 'files' file.

10 years agokernel: Add 'targ' and 'targbh' devices to the LINT configs.
Sascha Wildner [Sat, 24 Aug 2013 00:27:44 +0000 (02:27 +0200)]
kernel: Add 'targ' and 'targbh' devices to the LINT configs.

10 years agokernel: Remove some unused lines from 'files' files.
Sascha Wildner [Sat, 24 Aug 2013 00:17:07 +0000 (02:17 +0200)]
kernel: Remove some unused lines from 'files' files.

10 years agokernel: Remove another syslink related file which was forgotten earlier.
Sascha Wildner [Fri, 23 Aug 2013 23:54:33 +0000 (01:54 +0200)]
kernel: Remove another syslink related file which was forgotten earlier.

Approved-by: dillon
10 years agokernel/usb4bsd: Remove the Makefile for usbdevs{,_data}.h generation.
Sascha Wildner [Fri, 23 Aug 2013 23:37:58 +0000 (01:37 +0200)]
kernel/usb4bsd: Remove the Makefile for usbdevs{,_data}.h generation.

10 years agokernel/usb4bsd: Switch to generating usbdevs{,_data}.h during the build.
Sascha Wildner [Fri, 23 Aug 2013 23:33:22 +0000 (01:33 +0200)]
kernel/usb4bsd: Switch to generating usbdevs{,_data}.h during the build.

Taken-from: FreeBSD

10 years agokernel/drm2: Fix for kernels that don't have DDB compiled in.
Sascha Wildner [Thu, 22 Aug 2013 11:55:41 +0000 (13:55 +0200)]
kernel/drm2: Fix for kernels that don't have DDB compiled in.

10 years agotop.1: mdoc conversion
Sascha Wildner [Wed, 21 Aug 2013 17:36:00 +0000 (19:36 +0200)]
top.1: mdoc conversion

unixtop has gone to sleep since 2008 and most changes we did to its
manual page were documenting stuff we added ourselves.

Convert it to mdoc to make this process easier in the future.

Submitted-by: Franco Fichtner <franco@lastsummer.de>
10 years agousb4bsd: Add a usb_quirk(4) manual page (from FreeBSD).
Sascha Wildner [Tue, 20 Aug 2013 14:36:42 +0000 (16:36 +0200)]
usb4bsd: Add a usb_quirk(4) manual page (from FreeBSD).

10 years agouipc: Reference the originator thread if asynchronized pru_send is used
Sepherosa Ziehau [Tue, 20 Aug 2013 10:59:08 +0000 (18:59 +0800)]
uipc: Reference the originator thread if asynchronized pru_send is used

It is required for UDP asynchronized pru_send, which could access the
originator's thread.

10 years agokernel/usb4bsd: Bring in urtwn(4) and firmware.
Sascha Wildner [Mon, 19 Aug 2013 19:53:36 +0000 (21:53 +0200)]
kernel/usb4bsd: Bring in urtwn(4) and firmware.

It passes initial testing, but of course there might still be uncaught
issues.

Taken-from: FreeBSD

10 years agoiwi.4: Fix wording.
Sascha Wildner [Mon, 19 Aug 2013 19:29:07 +0000 (21:29 +0200)]
iwi.4: Fix wording.

10 years agohier.7: Move .Pp to the right place.
Sascha Wildner [Mon, 19 Aug 2013 03:01:37 +0000 (05:01 +0200)]
hier.7: Move .Pp to the right place.

10 years agokernel/usb4bsd: Remove forgotten wlan_serialize_exit() in run(4).
Sascha Wildner [Mon, 19 Aug 2013 02:46:43 +0000 (04:46 +0200)]
kernel/usb4bsd: Remove forgotten wlan_serialize_exit() in run(4).

Spotted-by: sephe
10 years agohier.7: Add missing .Pp
Sascha Wildner [Mon, 19 Aug 2013 00:45:39 +0000 (02:45 +0200)]
hier.7: Add missing .Pp

10 years agoAnother typo fixing round (in messages and manpages) across the tree.
Sascha Wildner [Mon, 19 Aug 2013 02:41:27 +0000 (04:41 +0200)]
Another typo fixing round (in messages and manpages) across the tree.

10 years agokernel/usb4bsd: Warning fix and minor adjustments from FreeBSD for xhci.
Sascha Wildner [Sun, 18 Aug 2013 18:46:25 +0000 (20:46 +0200)]
kernel/usb4bsd: Warning fix and minor adjustments from FreeBSD for xhci.

10 years agokernel/usb4bsd: Remove unnecessary NULL checks in rum(4) and run(4).
Sascha Wildner [Sun, 18 Aug 2013 16:12:56 +0000 (18:12 +0200)]
kernel/usb4bsd: Remove unnecessary NULL checks in rum(4) and run(4).

10 years agokernel/usb4bsd: Port run(4), add firmware too.
Sascha Wildner [Sun, 18 Aug 2013 13:53:40 +0000 (15:53 +0200)]
kernel/usb4bsd: Port run(4), add firmware too.

Some minor issues might remain, but it works good here with various
compatible adapters.

While here, sync with FreeBSD's latest.

Taken-from: FreeBSD

10 years agokernel: Align backslashes in 'files' a bit better for firmwares.
Sascha Wildner [Sun, 18 Aug 2013 00:41:42 +0000 (02:41 +0200)]
kernel: Align backslashes in 'files' a bit better for firmwares.

10 years agokernel/firmware: Add NORMAL_FW and NORMAL_FWO for firmware.
Sascha Wildner [Sun, 18 Aug 2013 00:16:55 +0000 (02:16 +0200)]
kernel/firmware: Add NORMAL_FW and NORMAL_FWO for firmware.

Taken-from: FreeBSD

10 years agokernel/usb4bsd: Remove unused variable.
Sascha Wildner [Sat, 17 Aug 2013 21:40:19 +0000 (23:40 +0200)]
kernel/usb4bsd: Remove unused variable.

10 years agokernel/usb4bsd: Omit 'if_' in building directories of network drivers.
Sascha Wildner [Sat, 17 Aug 2013 18:52:53 +0000 (20:52 +0200)]
kernel/usb4bsd: Omit 'if_' in building directories of network drivers.

10 years agokernel/usb4bsd: Port rum(4).
Sascha Wildner [Sat, 17 Aug 2013 18:01:15 +0000 (20:01 +0200)]
kernel/usb4bsd: Port rum(4).

The driver behaves mostly stable but a few issues remain, such as
kldunloading cleanly.

10 years agokernel/files: Comment out two lines for old USB wlan drivers.
Sascha Wildner [Sat, 17 Aug 2013 17:44:03 +0000 (19:44 +0200)]
kernel/files: Comment out two lines for old USB wlan drivers.

In preparation for some usb4bsd wlan work.

10 years agosbin: Use socklen_t in a few places.
Sascha Wildner [Sat, 17 Aug 2013 17:18:17 +0000 (19:18 +0200)]
sbin: Use socklen_t in a few places.

10 years agoRemove last two references to legacy make
John Marino [Sat, 17 Aug 2013 14:45:10 +0000 (16:45 +0200)]
Remove last two references to legacy make

This last two LEGACY MAKE conditions slipped by.  The others were removed
in a previous commit.

10 years agomake.conf: Use MACHINE_ARCH instead of ARCH
François Tigeot [Sat, 17 Aug 2013 12:46:11 +0000 (14:46 +0200)]
make.conf: Use MACHINE_ARCH instead of ARCH

It is already defined by bmake itself

10 years agomake.conf: Switch dports to building new Xorg/KMS packages by default
François Tigeot [Sat, 17 Aug 2013 11:41:00 +0000 (13:41 +0200)]
make.conf: Switch dports to building new Xorg/KMS packages by default

10 years agoBump __DragonFly_version for i915 KMS support
François Tigeot [Sat, 17 Aug 2013 10:51:07 +0000 (12:51 +0200)]
Bump __DragonFly_version for i915 KMS support

10 years agokernel/i386: Remove BDE_DEBUGGER remains.
Sascha Wildner [Sat, 17 Aug 2013 02:51:37 +0000 (04:51 +0200)]
kernel/i386: Remove BDE_DEBUGGER remains.

10 years agoMerge branch 'master' of ssh://crater.dragonflybsd.org/repository/git/dragonfly
Matthew Dillon [Sat, 17 Aug 2013 01:22:47 +0000 (18:22 -0700)]
Merge branch 'master' of ssh://crater.dragonflybsd.org/repository/git/dragonfly

10 years agoRemove the uhidev.4 manual page. We never had this driver.
Sascha Wildner [Sat, 17 Aug 2013 01:21:17 +0000 (03:21 +0200)]
Remove the uhidev.4 manual page. We never had this driver.

10 years agokernel - Fix SMP race in mtx*() code.
Matthew Dillon [Sat, 17 Aug 2013 01:21:34 +0000 (18:21 -0700)]
kernel - Fix SMP race in mtx*() code.

* Fix a SMP race due to forgetting to set the PINTERLOCKED flag.

10 years agosmp.4: Mention that APIC I/O is disabled by default in virtual machines.
Sascha Wildner [Sat, 17 Aug 2013 00:08:35 +0000 (02:08 +0200)]
smp.4: Mention that APIC I/O is disabled by default in virtual machines.

10 years agoath_hal.4: Comment out some config lines which we do not offer.
Sascha Wildner [Fri, 16 Aug 2013 23:57:20 +0000 (01:57 +0200)]
ath_hal.4: Comment out some config lines which we do not offer.

10 years agoAdd ath(4) and associated options/devices to our default configs.
Sascha Wildner [Fri, 16 Aug 2013 23:53:17 +0000 (01:53 +0200)]
Add ath(4) and associated options/devices to our default configs.

10 years agostyle.9: Remove wcswcs(3) reference.
Sascha Wildner [Fri, 16 Aug 2013 14:54:49 +0000 (16:54 +0200)]
style.9: Remove wcswcs(3) reference.

10 years agokernel - Fix bug in SO_NOSIGPIPE when socket fp called from kernel thread
Matthew Dillon [Fri, 16 Aug 2013 04:16:12 +0000 (21:16 -0700)]
kernel - Fix bug in SO_NOSIGPIPE when socket fp called from kernel thread

* SO_NOSIGPIPE assumed that soo_write() was being called from a user
  process.  It can also be called from a kernel thread.

* Fixes hammer2 kdmsg thread crashes.

10 years agovkernel - Module loading for vkernel64
Matthew Dillon [Fri, 16 Aug 2013 02:35:52 +0000 (19:35 -0700)]
vkernel - Module loading for vkernel64

* Fix a missing vm_object_drop() in the kldload error path.

* Load modules using sbrk()'d memory (for now).

* Hack rtld to accept a new environment variable, LD_SHAREDLIB_BASE,
  which forces shared libraries to be mmap()'d at the specified address.

* Hack the vkernel64 binary to re-exec itself in order to set
  LD_SHAREDLIB_BASE to low memory.  This forces shared libraries into
  low memory which the KLD module loader can relocate against, instead
  of high memory which it can't.

* test/vkernel/Makefile now builds and installs modules by default

With-fixes-from: tuxillo

10 years agokernel - Add missing reltoken/LWPRELE in signal error path
Matthew Dillon [Thu, 15 Aug 2013 21:08:51 +0000 (14:08 -0700)]
kernel - Add missing reltoken/LWPRELE in signal error path

* Add missing releases in the signal error path which can be triggered
  when a dying process is signalled.

10 years agoUpdate the pciconf(8) database.
Sascha Wildner [Thu, 15 Aug 2013 13:08:17 +0000 (15:08 +0200)]
Update the pciconf(8) database.

August 9, 2013 snapshot from http://pciids.sourceforge.net/

10 years agohammer - Fix filesystem selection path bug in cleanup directive
Matthew Dillon [Thu, 15 Aug 2013 07:57:53 +0000 (00:57 -0700)]
hammer - Fix filesystem selection path bug in cleanup directive

* When being asked to cleanup a hammer filesystem, hammer checks the
  /var/hammer/<filesystem>/ directory for manual snapshots and is able
  to extract the target filesystem from the softlinks it finds.

  If no softlinks are found hammer incorrectly calls 'hammer prune'
  with the /var/hammer/* path instead of the target filesystem path
  which causes the prune code to prune the wrong filesystem.

* Issue only happened when snapshots were disabled but snapshot aging
  was being enforced.  e.g. in the hammer config for the filesystem,
  something like 'snapshots 0d 60d'.  This feature tends to be used
  only on slaves that are the target of a mirror-stream backup.
  Other combinations will either have created snapshot softlinks
  in /var/hammer/* or would have disabled snapshots entirely (which
  causes a prune-everything to happen and doesn't have the bug).

10 years agokernel/linux: Use #ifdef (fixes -Wundef).
Sascha Wildner [Wed, 14 Aug 2013 21:41:02 +0000 (23:41 +0200)]
kernel/linux: Use #ifdef (fixes -Wundef).

10 years agomps.4: Sync with FreeBSD to document the current state of affairs.
Sascha Wildner [Wed, 14 Aug 2013 15:38:22 +0000 (17:38 +0200)]
mps.4: Sync with FreeBSD to document the current state of affairs.

10 years agocputimer: Allow MP synchronized TSC to become cputimer on i386
Sepherosa Ziehau [Wed, 14 Aug 2013 14:41:59 +0000 (22:41 +0800)]
cputimer: Allow MP synchronized TSC to become cputimer on i386

See also 8d23b56caf513f5c15f56fb053479cc00fe2eaf4

10 years agodrm/i915: Use a spinlock for i915_file_priv locking
François Tigeot [Wed, 14 Aug 2013 08:13:03 +0000 (10:13 +0200)]
drm/i915: Use a spinlock for i915_file_priv locking

* This is also what Linux uses

* Sync struct drm_i915_file_private and i915_gem_ring_throttle() a bit
  with Linux while I'm at it

10 years agonrelease - Add delay before umounting
Matthew Dillon [Tue, 13 Aug 2013 21:22:29 +0000 (14:22 -0700)]
nrelease - Add delay before umounting

* Add a delay before unmounting the usb image filesystem.

10 years agosysctl_add_oid.9: Reduce diffs to Free a bit.
Sascha Wildner [Mon, 12 Aug 2013 20:04:19 +0000 (22:04 +0200)]
sysctl_add_oid.9: Reduce diffs to Free a bit.

10 years agoacpica: Fix buildworld issue introduced by previous commit
François Tigeot [Mon, 12 Aug 2013 17:39:40 +0000 (19:39 +0200)]
acpica: Fix buildworld issue introduced by previous commit

10 years agoacpica: Replace some include paths by full ones
François Tigeot [Sun, 11 Aug 2013 16:29:00 +0000 (18:29 +0200)]
acpica: Replace some include paths by full ones

To allow the ACPI header files to be used in the whole kernel

10 years agokernel: Import pci_save_state and pci_restore_state from FreeBSD
François Tigeot [Sun, 11 Aug 2013 08:45:10 +0000 (10:45 +0200)]
kernel: Import pci_save_state and pci_restore_state from FreeBSD

10 years agokernel: Add bitcount32 to sys/systm.h
François Tigeot [Sat, 10 Aug 2013 17:22:20 +0000 (19:22 +0200)]
kernel: Add bitcount32 to sys/systm.h

10 years agodrm2: Remove dubious AGP locks
François Tigeot [Mon, 12 Aug 2013 06:28:00 +0000 (08:28 +0200)]
drm2: Remove dubious AGP locks

They are not present in the original Linux code.

10 years agodrm2: remove DRM_SPINXXX macros
François Tigeot [Sat, 10 Aug 2013 21:51:04 +0000 (23:51 +0200)]
drm2: remove DRM_SPINXXX macros

* They did not apply 1:1 to DragonFly locking directives; keeping them
  in the drivers would have meant lockmgr locks would have had to be be
  used in situations where spinlocks could have done the job.

* Besides the obvious performance pessimization issues, they had the
  unfortunate effect of obscuring the source code

* The original DRM code in the Linux kernel doesn't even use them, they
  were purely a FreeBSD addition

10 years agouipc: Set up netmsg's socket properly before perform notify abort
Sepherosa Ziehau [Mon, 12 Aug 2013 02:42:04 +0000 (10:42 +0800)]
uipc: Set up netmsg's socket properly before perform notify abort

Mainly to make the netisr so_port change chasing work

10 years agokernel/x86_64: Display Standard Extended Features during CPU setup.
Sascha Wildner [Sun, 11 Aug 2013 21:15:46 +0000 (23:15 +0200)]
kernel/x86_64: Display Standard Extended Features during CPU setup.

Taken-from: FreeBSD
Tested-by: Mihai Carabas <mihai.carabas@gmail.com>, ftigeot
10 years agoiicbb: Update from FreeBSD
François Tigeot [Sat, 10 Aug 2013 15:15:14 +0000 (17:15 +0200)]
iicbb: Update from FreeBSD

10 years agotest: Test push
Sepherosa Ziehau [Sun, 11 Aug 2013 12:23:11 +0000 (20:23 +0800)]
test: Test push

10 years agoFix two prototypes in manual pages.
Sascha Wildner [Sun, 11 Aug 2013 01:42:51 +0000 (03:42 +0200)]
Fix two prototypes in manual pages.

10 years agokernel: Start removing syslink.
Sascha Wildner [Sat, 10 Aug 2013 20:19:05 +0000 (22:19 +0200)]
kernel: Start removing syslink.

It's tied in with a few things that we don't want to break, so care is
needed.

Approved-by: dillon
10 years agoGC a manpage backup that was created by passing -b when installing it.
Sascha Wildner [Sat, 10 Aug 2013 11:18:31 +0000 (13:18 +0200)]
GC a manpage backup that was created by passing -b when installing it.

10 years agoshare/mk: Fix the creation of MLINKS for manpages of fschg'd stuff.
Sascha Wildner [Sat, 10 Aug 2013 11:07:15 +0000 (13:07 +0200)]
share/mk: Fix the creation of MLINKS for manpages of fschg'd stuff.

We cannot pass INSTALLFLAGS directly for manual page installation because
when INSTALLFLAGS contains -fschg (such as for init(8)), it will cause the
manual page to be fschg'd too which breaks subsequent MLINKS creation.

Therefore, use MANINSTALLFLAGS for manual pages. While here, document it
too.