dragonfly.git
20 months agoigb: Utilize mbuf's header length to setup TX context
Sepherosa Ziehau [Thu, 2 Aug 2012 11:53:07 +0000 (19:53 +0800)]
igb: Utilize mbuf's header length to setup TX context

While im here, also fix the case that VLAN hardware tagging is enabled
but IP checksum offloading is not enable.

20 months agostatclock: Fix intr% counting on heavily interrupted systems
Sepherosa Ziehau [Thu, 2 Aug 2012 09:26:30 +0000 (17:26 +0800)]
statclock: Fix intr% counting on heavily interrupted systems

statclock() has high probability to run from crit_exit() on heavily
interrupted systems and 'frame' is NULL in this case, so intr% is
mis-counted as sys%.

To fix this, TDF_INTTHREAD is used to determine whether intr% should
be counted or not, in addition to testing gd_intr_nest_level and
'frame'

20 months agoem/emx: Reduce the default interrupt rate to 6000Hz
Sepherosa Ziehau [Thu, 2 Aug 2012 01:43:38 +0000 (09:43 +0800)]
em/emx: Reduce the default interrupt rate to 6000Hz

20 months agotools/pktgen: Setup link header length properly
Sepherosa Ziehau [Thu, 2 Aug 2012 01:27:09 +0000 (09:27 +0800)]
tools/pktgen: Setup link header length properly

20 months agoemx: Add EMX_TSO_DEBUG
Sepherosa Ziehau [Wed, 1 Aug 2012 09:09:58 +0000 (17:09 +0800)]
emx: Add EMX_TSO_DEBUG

It is used to measure TSO context reuse rate.  Setting up TX context
interferes TX data fetching pipelining, thus reduces performance, so
TX context setting up should be avoided as much as possible.

According to limited measurement:
- 1 one direction TX stream, reuse rate is ~40%
- 16 one direction TX streams, reuse rate is ~20%
- 16 bi-direction streams (total 32), reuse rate is ~10%

20 months agoemx: Remove dependency on opt_rss.h; RSS is always enabled
Sepherosa Ziehau [Wed, 1 Aug 2012 08:52:44 +0000 (16:52 +0800)]
emx: Remove dependency on opt_rss.h; RSS is always enabled

20 months agoemx: Add TSO support
Sepherosa Ziehau [Wed, 1 Aug 2012 08:51:02 +0000 (16:51 +0800)]
emx: Add TSO support

20 months agoemx: Remove unused macros
Sepherosa Ziehau [Wed, 1 Aug 2012 04:21:53 +0000 (12:21 +0800)]
emx: Remove unused macros

20 months agotools/pktgen: Setup header length properly
Sepherosa Ziehau [Wed, 1 Aug 2012 03:24:53 +0000 (11:24 +0800)]
tools/pktgen: Setup header length properly

20 months agoemx: Utilize mbuf's header lengthes
Sepherosa Ziehau [Wed, 1 Aug 2012 03:20:49 +0000 (11:20 +0800)]
emx: Utilize mbuf's header lengthes

20 months agombuf: segsz -> tso_segsz, which is more expressive
Sepherosa Ziehau [Wed, 1 Aug 2012 02:54:12 +0000 (10:54 +0800)]
mbuf: segsz -> tso_segsz, which is more expressive

20 months agobnx: Utilize header lengthes in mbuf to setup TSO
Sepherosa Ziehau [Wed, 1 Aug 2012 02:50:00 +0000 (10:50 +0800)]
bnx: Utilize header lengthes in mbuf to setup TSO

20 months agombuf: Regroup fields, mainly to improve cache utilization
Sepherosa Ziehau [Wed, 1 Aug 2012 02:47:53 +0000 (10:47 +0800)]
mbuf: Regroup fields, mainly to improve cache utilization

20 months agombuf: Save linker layer, IP and TCP/UDP header length
Sepherosa Ziehau [Wed, 1 Aug 2012 02:35:06 +0000 (10:35 +0800)]
mbuf: Save linker layer, IP and TCP/UDP header length

This could ease most drivers's TSO operation and avoid extra data
area accessing during TSO setting up.

This could also help Intel's 1000M/10G drivers' hardware checksum
offloading, which requires protocol header length.

20 months agoixgbe: remove some FreeBSD version checks
François Tigeot [Tue, 31 Jul 2012 13:19:57 +0000 (15:19 +0200)]
ixgbe: remove some FreeBSD version checks

20 months agoLINT/LINT64: Add BNX_TSO_DEBUG.
Sascha Wildner [Tue, 31 Jul 2012 11:39:05 +0000 (13:39 +0200)]
LINT/LINT64: Add BNX_TSO_DEBUG.

20 months agoopendir(3): use O_CLOEXEC
François Tigeot [Tue, 31 Jul 2012 08:32:24 +0000 (10:32 +0200)]
opendir(3): use O_CLOEXEC

20 months agokernel: Implement O_CLOEXEC
François Tigeot [Mon, 30 Jul 2012 20:39:46 +0000 (22:39 +0200)]
kernel: Implement O_CLOEXEC

* Using fcntl(2) just after open(2) is not enough to avoid race conditions
  in programs doing fork+exec sequences. Child processes may be created
  before fcntl() is run and inherit the parent's file descriptors.

* In some circonstances this behavior may even create security issues.

* O_CLOEXEC can be used to atomically set the close-on-exec flag for new
  file descriptors, avoiding the whole mess in the first place.

* Fixes issue #2356

Inspired-from: NetBSD

20 months agokdump(1)/truss(1): Fix occasional quickworld breakage.
Sascha Wildner [Mon, 30 Jul 2012 17:51:07 +0000 (19:51 +0200)]
kdump(1)/truss(1): Fix occasional quickworld breakage.

Whenever we messed with our ioctls or certain definitions
(like fcntl(2)'s O_* flags), kdump and truss could break
quickworld because their generated ioctl.c and kdump_subr.c
files in /usr/obj would still have removed stuff.

Fix this by regenerating them every time, no matter what.

20 months ago<sys/fcntl.h>: Mark O_DIRECTORY as being new in POSIX.1-2008.
Sascha Wildner [Mon, 30 Jul 2012 16:28:00 +0000 (18:28 +0200)]
<sys/fcntl.h>: Mark O_DIRECTORY as being new in POSIX.1-2008.

It is also visible in our default environment, just not if the code
explicitly requests an older POSIX version.

20 months agoopendir(3) - use O_DIRECTORY
Alex Hornung [Mon, 30 Jul 2012 15:58:05 +0000 (15:58 +0000)]
opendir(3) - use O_DIRECTORY

20 months agolinkat(2) man page
Alex Hornung [Mon, 30 Jul 2012 11:37:09 +0000 (11:37 +0000)]
linkat(2) man page

20 months agoimplement linkat
Alex Hornung [Mon, 30 Jul 2012 11:21:53 +0000 (11:21 +0000)]
implement linkat

20 months agoopen.2 - document O_DIRECTORY
Alex Hornung [Mon, 30 Jul 2012 14:02:26 +0000 (14:02 +0000)]
open.2 - document O_DIRECTORY

20 months agoopen - implement O_DIRECTORY
Alex Hornung [Mon, 30 Jul 2012 11:20:00 +0000 (11:20 +0000)]
open - implement O_DIRECTORY

20 months agokernel: remove O_FRNONBLOCKING
François Tigeot [Sun, 29 Jul 2012 17:26:02 +0000 (19:26 +0200)]
kernel: remove O_FRNONBLOCKING

* Like O_ROOTCRED it was effectively unused and removing it frees up
  a precious bit among the various O_xxx flags

* Partially fixes issue #2357

20 months agokernel: remove O_ROOTCRED
François Tigeot [Sun, 29 Jul 2012 16:23:31 +0000 (18:23 +0200)]
kernel: remove O_ROOTCRED

* This flag was DragonFly-specific and effectively unused

* Removing it frees up flag space for interacting with the rest of the world

* Partially fixes issue #2357

20 months agotcp: Increase tcp_sosend_agglim from 2 to 3
Sepherosa Ziehau [Mon, 30 Jul 2012 06:36:52 +0000 (14:36 +0800)]
tcp: Increase tcp_sosend_agglim from 2 to 3

This change makes better use of TSO.

When 2 is used, most of the large TCP segments' size are 2*MSS or 3*MSS.
When 3 is used, most of the large TCP segments' size are 4*MSS or 5*MSS.

The math is quite simple, given 1448B segment size:
2 == 4096B == 2 segments
3 == 6144B == 4 segments

Increasing it to a higher value improves single full speed stream's CPU
utilization, but does not have much effect on multiple streams'.

20 months agotools: Add toeplitz, which calculate toeplitz on given addresses/ports
Sepherosa Ziehau [Mon, 30 Jul 2012 04:29:31 +0000 (12:29 +0800)]
tools: Add toeplitz, which calculate toeplitz on given addresses/ports

20 months agobnx: Add TSO support
Sepherosa Ziehau [Sun, 29 Jul 2012 05:29:49 +0000 (13:29 +0800)]
bnx: Add TSO support

Obtained-from: FreeBSD if_bge.c

For TSO header information, ether_tso_pullup() is used instead of
making our own wheel.

BNX_TSO_DEBUG is added to help testing various TCP parameters tuning

20 months agoixgbe: Unbreak compile
Sepherosa Ziehau [Mon, 30 Jul 2012 05:58:57 +0000 (13:58 +0800)]
ixgbe: Unbreak compile

20 months agotcp: Add TSO support for IPv4
Sepherosa Ziehau [Fri, 27 Jul 2012 06:07:27 +0000 (14:07 +0800)]
tcp: Add TSO support for IPv4

It is implemented mainly according to NetBSD's TSO implementation.

Following stuffs are only in DragonFly
- Add comment about devices' expected behaviour upon PUSH and FIN flags
  Obtained-from: Microsoft's LSO online document
- Don't use TSO, if there are SACK or DSACK blocks to report
- Don't use TSO, if congestion window needs validation
- Don't use TSO, if URG flag is to be set
- Take IP and TCP header sizes into consideration when calculate the
  large TCP segment size
- Pseudo checksum for the large TCP segment is calculated using only
  source address, destination address and IPPROTO_TCP according to
  Microsoft's LSO online document.  This fashion of pseudo checksum
  calculation seems to be adopted by several NIC chips.

Several driver helper functions are added:
- tcp_tso_pullup(), which extracts IPv4 and TCP header's location and
  length.  And make sure that IPv4 and TCP headers are in contiguous
  memory.
- ether_tso_pullup(), in addition to what tcp_tso_pullup() does, it
  also extracts ethernet header's length and make sure that ethernet,
  IPv4 and TCP headers are in contiguous memory.

Sysctl node net.inet.tcp.tso could be used to globally disable TSO.
TSO is by default on.

tso/-tso are added to ifconfig(8), which could be used to enable or
disable TSO on the specific interface.

20 months agociss.4: Document some tunables.
Sascha Wildner [Sun, 29 Jul 2012 19:06:30 +0000 (21:06 +0200)]
ciss.4: Document some tunables.

20 months agociss(4): Sync with FreeBSD.
Sascha Wildner [Sun, 29 Jul 2012 18:49:26 +0000 (20:49 +0200)]
ciss(4): Sync with FreeBSD.

Tons of good stuff. MPSAFEness, MSI interrupts, better performance,
and a lot more (I'm too lazy to go through FreeBSD's logs :P).

Support for a number of adapters has been added, too:

* HP Smart Array P220i
* HP Smart Array P222
* HP Smart Array P420
* HP Smart Array P420i
* HP Smart Array P421
* HP Smart Array P700m
* HP Smart Array P712m
* HP Smart Array P721m
* HP Smart Array P822

I've tested it with a HP Smart Array P212.

Many thanks to FreeBSD.

20 months agomptable - use the table's cpu_flags
Alex Hornung [Sun, 29 Jul 2012 15:28:00 +0000 (15:28 +0000)]
mptable - use the table's cpu_flags

 * Sometimes the CPUs are marked as disabled in the APIC table, and for
   good reason (e.g. a machine having only 24 cores, but 32 appearing in
   the APIC table, 8 of them disabled).

 * Just in case, provide an override that falls back to the old
   behaviour. This is the tunable hw.lapic_force_enable. It will force
   all lapic entries to be marked as enabled. It is however not the
   default.

 * When parsing, don't error out when a lapic id is invalid (255) but it
   is disabled.

Reported-by: Mihai Carabas, Francois Tigeot
20 months agoacpidump - be a bit more lenient and don't assert
Alex Hornung [Sun, 29 Jul 2012 15:23:47 +0000 (15:23 +0000)]
acpidump - be a bit more lenient and don't assert

 * Asserting in this case causes the dump not to continue, although it
   is safe to do so.

Reported-by: Mihai Carabas
20 months agox86_64 mptable - maximum lapic id is 64
Alex Hornung [Sun, 29 Jul 2012 12:00:53 +0000 (12:00 +0000)]
x86_64 mptable - maximum lapic id is 64

 * In x86_64, lapic ids with HT can be up to 64, not 32 as in i386.

Reported-by: Mihai Carabas, Francois Tigeot
20 months agoRemove some more unused headers.
Sascha Wildner [Sun, 29 Jul 2012 00:27:27 +0000 (02:27 +0200)]
Remove some more unused headers.

20 months agoRemove unused header.
Sascha Wildner [Sat, 28 Jul 2012 23:40:18 +0000 (01:40 +0200)]
Remove unused header.

20 months agogcore(1): Make it work again.
Sascha Wildner [Sat, 28 Jul 2012 20:29:43 +0000 (22:29 +0200)]
gcore(1): Make it work again.

Two commits broke things:

64fce751acfda11e21846e900fc774e88f32f52e which changed us to no longer
claiming that we're FreeBSD in core files.

0b684e3f9af3abc5ded676c43a1a5a478b45325c which added another field to
/proc/<pid>/map (the vnode path for vn-backed vm objects).

Both changes need adjustments in gcore(1).

20 months agobsd-family-tree: Sync with FreeBSD.
Sascha Wildner [Fri, 27 Jul 2012 16:11:45 +0000 (18:11 +0200)]
bsd-family-tree: Sync with FreeBSD.

20 months agobnx: Utilize BNX_IS_5717_PLUS
Sepherosa Ziehau [Thu, 26 Jul 2012 09:56:05 +0000 (17:56 +0800)]
bnx: Utilize BNX_IS_5717_PLUS

20 months agobnx: Enable software workaround for possiblely hosed interrupt
Sepherosa Ziehau [Thu, 26 Jul 2012 09:19:09 +0000 (17:19 +0800)]
bnx: Enable software workaround for possiblely hosed interrupt

It is only needed for BCM57765 family chips and BCM5717/BCM5718 chips

According-to: tg3

20 months agobnx: Dispatch state timer to the same CPU as MSI/legacy interrupt CPU
Sepherosa Ziehau [Tue, 24 Jul 2012 12:47:22 +0000 (20:47 +0800)]
bnx: Dispatch state timer to the same CPU as MSI/legacy interrupt CPU

Mainly avoid possible serializer contention

20 months agocallout: Add callout_reset_bycpu, which dispatches callout to the specified CPU
Sepherosa Ziehau [Tue, 24 Jul 2012 10:59:31 +0000 (18:59 +0800)]
callout: Add callout_reset_bycpu, which dispatches callout to the specified CPU

20 months agobnx: Use MPSAFE callout
Sepherosa Ziehau [Tue, 24 Jul 2012 09:21:26 +0000 (17:21 +0800)]
bnx: Use MPSAFE callout

20 months agobnx: Add support for BCM57766 chips
Sepherosa Ziehau [Tue, 24 Jul 2012 08:57:41 +0000 (16:57 +0800)]
bnx: Add support for BCM57766 chips

According-to: tg3

20 months agopcidevs: regen
Sepherosa Ziehau [Tue, 24 Jul 2012 08:52:41 +0000 (16:52 +0800)]
pcidevs: regen

20 months agopcidevs: Add more PCI ids for BCM57785 family chips
Sepherosa Ziehau [Tue, 24 Jul 2012 08:52:11 +0000 (16:52 +0800)]
pcidevs: Add more PCI ids for BCM57785 family chips

20 months agobnx: BCM57791 and BCM57795 are 10/100 only
Sepherosa Ziehau [Tue, 24 Jul 2012 06:09:28 +0000 (14:09 +0800)]
bnx: BCM57791 and BCM57795 are 10/100 only

While im here, remove some unused code

20 months agobnx: Reconfigure DMA read/write control register
Sepherosa Ziehau [Tue, 24 Jul 2012 04:53:52 +0000 (12:53 +0800)]
bnx: Reconfigure DMA read/write control register

- Use default DMA write watermark (256bytes by default, was 128bytes)
- Don't touch reserved bits
- Disable 32bytes cache align; 64bytes cache align is left enabled

According-to: tg3

20 months agobnx: Enable MBUF low attention on buffer manager
Sepherosa Ziehau [Mon, 23 Jul 2012 09:17:58 +0000 (17:17 +0800)]
bnx: Enable MBUF low attention on buffer manager

According-to: tg3

20 months agobnx: Using 57765_FAMILY to conf Low Watermark Maximum Receive Frame Register
Sepherosa Ziehau [Sun, 22 Jul 2012 13:23:39 +0000 (21:23 +0800)]
bnx: Using 57765_FAMILY to conf Low Watermark Maximum Receive Frame Register

20 months agobnx: Properly configure PCI-E PL/DL registers and MAC clock speed
Sepherosa Ziehau [Sun, 22 Jul 2012 13:20:11 +0000 (21:20 +0800)]
bnx: Properly configure PCI-E PL/DL registers and MAC clock speed

According-to: tg3

20 months agobnx: Adjust RX/TX ring limit for BCM57785 family
Sepherosa Ziehau [Sun, 22 Jul 2012 10:52:55 +0000 (18:52 +0800)]
bnx: Adjust RX/TX ring limit for BCM57785 family

20 months agobnx: Remove unused code
Sepherosa Ziehau [Fri, 20 Jul 2012 09:43:24 +0000 (17:43 +0800)]
bnx: Remove unused code

20 months agobge/bnx: Avoid using magic register name
Sepherosa Ziehau [Fri, 20 Jul 2012 09:07:30 +0000 (17:07 +0800)]
bge/bnx: Avoid using magic register name

20 months agobnx: UDP hardware checksum offloading works
Sepherosa Ziehau [Fri, 20 Jul 2012 08:49:06 +0000 (16:49 +0800)]
bnx: UDP hardware checksum offloading works

20 months agobnx: Utilize BNX_IS_5717_PLUS
Sepherosa Ziehau [Fri, 20 Jul 2012 08:14:43 +0000 (16:14 +0800)]
bnx: Utilize BNX_IS_5717_PLUS

- Correctly disable 4 sending rings for BCM5718 family
- Correct NVRAM read offset.  NVRAM reading as no effect currently.

20 months agobnx: Add macros to further differentiate BCM57785 and BCM5718 families
Sepherosa Ziehau [Fri, 20 Jul 2012 07:31:12 +0000 (15:31 +0800)]
bnx: Add macros to further differentiate BCM57785 and BCM5718 families

21 months agowpa_cli(8): Add libedit support (using libedit's readline API).
Sascha Wildner [Tue, 24 Jul 2012 21:22:41 +0000 (23:22 +0200)]
wpa_cli(8): Add libedit support (using libedit's readline API).

Unfortunately, this introduces a warning which required lowering
WARNS to 0 for wpa_cli(8). I didn't feel like modifying contrib/
source for it.

Taken-from: FreeBSD
Tested-by: Johannes Hofmann <Johannes.Hofmann@gmx.de>
21 months agorc.d/netif: Require cleanvar.
Sascha Wildner [Tue, 24 Jul 2012 21:19:23 +0000 (23:19 +0200)]
rc.d/netif: Require cleanvar.

Before the ipfilter removal, netif required ipfilter, which required
ipmon, which required cleanvar. In other words, cleanvar was run before
netif.

With the ipfilter requirement removed from netif, cleanvar was now
ordered at some place after netif, which caused the PID file and
control socket of wpa_supplicant(8) (which is started by netif if an
ifconfig_xxx="WPA" line is found in /etc/rc.conf) to be removed from
/var/run by cleanvar after they were created.

This caused trouble for wpa_cli(8), which needs the socket.

Found-by: Johannes Hofmann <Johannes.Hofmann@gmx.de>
21 months agotftp(1): Add libedit support.
Sascha Wildner [Tue, 24 Jul 2012 12:31:40 +0000 (14:31 +0200)]
tftp(1): Add libedit support.

Taken-from:  FreeBSD

21 months agopam_exec.8: Fix references to nonexistent manual pages.
Sascha Wildner [Mon, 23 Jul 2012 20:50:51 +0000 (22:50 +0200)]
pam_exec.8: Fix references to nonexistent manual pages.

They exist in section 3, not 8, so reference those.

21 months agobc(1): Raise WARNS to 3 and fix a warning.
Sascha Wildner [Mon, 23 Jul 2012 17:38:29 +0000 (19:38 +0200)]
bc(1): Raise WARNS to 3 and fix a warning.

21 months agobc(1): Add libedit support.
Sascha Wildner [Mon, 23 Jul 2012 17:23:55 +0000 (19:23 +0200)]
bc(1): Add libedit support.

Reported-by: Pierre Abbat <phma@phma.optus.nu>
Taken-from:  FreeBSD

21 months agoaic7xxx: fix a pointer sign warning
François Tigeot [Mon, 23 Jul 2012 07:29:25 +0000 (09:29 +0200)]
aic7xxx: fix a pointer sign warning

21 months agoixgbe(4): Use MSI by default
François Tigeot [Sun, 22 Jul 2012 07:04:26 +0000 (09:04 +0200)]
ixgbe(4): Use MSI by default

21 months agompt.4: Mention that mpt(4) is available in VMware (player in my case).
Sascha Wildner [Sat, 21 Jul 2012 19:52:19 +0000 (21:52 +0200)]
mpt.4: Mention that mpt(4) is available in VMware (player in my case).

21 months agokenv.9: Improve wording after my last commit made it sound weird.
Sascha Wildner [Sat, 21 Jul 2012 18:56:46 +0000 (20:56 +0200)]
kenv.9: Improve wording after my last commit made it sound weird.

21 months agompt.4: Add missing .El
Sascha Wildner [Sat, 21 Jul 2012 18:48:23 +0000 (20:48 +0200)]
mpt.4: Add missing .El

21 months agompt(4): Use MSI on SAS adapters that support it.
Sascha Wildner [Sat, 21 Jul 2012 18:43:31 +0000 (20:43 +0200)]
mpt(4): Use MSI on SAS adapters that support it.

The hw.mpt.msi.enable tunable can be used to override this behavior.

Tested-by: ftigeot
21 months agokenv.9: kgetenv_*() functions actually return 1 if successful.
Sascha Wildner [Sat, 21 Jul 2012 18:33:03 +0000 (20:33 +0200)]
kenv.9: kgetenv_*() functions actually return 1 if successful.

21 months agoSync various files with tzcode2012c from ftp://ftp.iana.org/tz/releases
Sascha Wildner [Sat, 21 Jul 2012 09:35:27 +0000 (11:35 +0200)]
Sync various files with tzcode2012c from ftp://ftp.iana.org/tz/releases

Not much to see here. Mainly removal of RCS IDs and some minor
adjustments in localtime(3).

21 months agoaac(4): Sync with FreeBSD.
Sascha Wildner [Fri, 20 Jul 2012 20:42:07 +0000 (22:42 +0200)]
aac(4): Sync with FreeBSD.

Nothing spectacular. Race fixes, some cleanup and a sysctl for displaying
the firmware version.

21 months agotests: make the code in poll_1.c a little bit clearer
Markus Pfeiffer [Fri, 20 Jul 2012 19:02:26 +0000 (19:02 +0000)]
tests: make the code in poll_1.c a little bit clearer

* Add a smalÃl comment at the top of poll_1.c and use constants from
  poll.h
* While I am here also fix a typo in the Makefile

21 months agokernel - another attempt at fixing checkloop panic
Markus Pfeiffer [Thu, 19 Jul 2012 15:49:02 +0000 (15:49 +0000)]
kernel - another attempt at fixing checkloop panic

21 months agokernel/acpi_thermal: Clarify sysctl description and remove duplicate header.
Sascha Wildner [Fri, 20 Jul 2012 16:01:54 +0000 (18:01 +0200)]
kernel/acpi_thermal: Clarify sysctl description and remove duplicate header.

Taken-from: FreeBSD

21 months agoSync zoneinfo database with tzdata2012d from ftp://ftp.iana.org/tz/releases
Sascha Wildner [Fri, 20 Jul 2012 08:22:22 +0000 (10:22 +0200)]
Sync zoneinfo database with tzdata2012d from ftp://ftp.iana.org/tz/releases

* Removes the SCCS IDs

* africa: Morocco does not observe DST from Jul 20 03:00 to Aug 20 02:00.

21 months agolapic: Allow lapic enumeration to fail
Sepherosa Ziehau [Fri, 20 Jul 2012 02:59:08 +0000 (10:59 +0800)]
lapic: Allow lapic enumeration to fail

It is mainly used to workaround for some old brain dead BIOS, which
will set BSP's LAPIC apic id to 255, but all MADT LAPIC entries are
valid

Reported-by: klanger on #dragonflybsd
21 months agoapic: APIC ID should not be 255, which is used for broadcast
Sepherosa Ziehau [Fri, 20 Jul 2012 02:31:04 +0000 (10:31 +0800)]
apic: APIC ID should not be 255, which is used for broadcast

21 months agoLINT/LINT64: Fix some comments.
Sascha Wildner [Thu, 19 Jul 2012 22:28:44 +0000 (00:28 +0200)]
LINT/LINT64: Fix some comments.

21 months agokernel: Remove two duplicate words in messages.
Sascha Wildner [Thu, 19 Jul 2012 21:54:20 +0000 (23:54 +0200)]
kernel: Remove two duplicate words in messages.

21 months agobnx.4: Add a if_bnx(4) MLINK.
Sascha Wildner [Thu, 19 Jul 2012 07:40:11 +0000 (09:40 +0200)]
bnx.4: Add a if_bnx(4) MLINK.

21 months agobnx: Add manpage
Sepherosa Ziehau [Thu, 19 Jul 2012 07:29:30 +0000 (15:29 +0800)]
bnx: Add manpage

Reminded-by: swildner@
21 months agobnx: Add BCM5718 and BCM57785 chip families support
Sepherosa Ziehau [Sun, 15 Jul 2012 10:39:29 +0000 (18:39 +0800)]
bnx: Add BCM5718 and BCM57785 chip families support

Broadcom's new generation of chips will support RSS (multiple RX rings),
multiple TX rings (BCM5718 family only) and multi-vector MSI-X.  The old
chips supported by bge(4) is becoming burden, so new driver, bnx(4), is
created to support new chip families features.

Related changes obtained from FreeBSD if_bge.c:
214428 221445 221818 226871 228479

Also, PCI-E 1.0a mode is not forced for BCM5718 and BCM57785 chip families

21 months agoRemove more useless -offset in manual pages.
Sascha Wildner [Thu, 19 Jul 2012 02:51:02 +0000 (04:51 +0200)]
Remove more useless -offset in manual pages.

21 months agomfiutil.4: -width -indent -> -width indent
Sascha Wildner [Thu, 19 Jul 2012 02:33:53 +0000 (04:33 +0200)]
mfiutil.4: -width -indent -> -width indent

21 months agobluetooth.4: Remove useless -offset
Sascha Wildner [Thu, 19 Jul 2012 02:32:11 +0000 (04:32 +0200)]
bluetooth.4: Remove useless -offset

21 months agomii: regen
Sepherosa Ziehau [Thu, 19 Jul 2012 01:25:54 +0000 (09:25 +0800)]
mii: regen

21 months agomii: Update description string of BCM5720C
Sepherosa Ziehau [Thu, 19 Jul 2012 01:25:33 +0000 (09:25 +0800)]
mii: Update description string of BCM5720C

21 months agobrgphy: Add BCM5718C, BCM54680, BCM54880, BCM54640, BCM54685 and BCM54881
Sepherosa Ziehau [Wed, 18 Jul 2012 23:42:59 +0000 (07:42 +0800)]
brgphy: Add BCM5718C, BCM54680, BCM54880, BCM54640, BCM54685 and BCM54881

21 months agomii: regen
Sepherosa Ziehau [Wed, 18 Jul 2012 23:42:29 +0000 (07:42 +0800)]
mii: regen

21 months agomii: Add more model ids for PHYs from Broadcom
Sepherosa Ziehau [Wed, 18 Jul 2012 23:40:22 +0000 (07:40 +0800)]
mii: Add more model ids for PHYs from Broadcom

- BCM5717C should be BCM5720C according to the datasheet
- Add BCM5718C
- Add BCM54680, BCM54880, BCM54640, BCM54685 and BCM54881

21 months agoconfig_intrhook.9: Use .Dv for SI_* identifier.
Sascha Wildner [Wed, 18 Jul 2012 13:29:43 +0000 (15:29 +0200)]
config_intrhook.9: Use .Dv for SI_* identifier.

21 months agokernel/lwkt_msgport: Improve comments a bit.
Nuno Antunes [Mon, 16 Jul 2012 13:11:04 +0000 (14:11 +0100)]
kernel/lwkt_msgport: Improve comments a bit.

Reviewed-by: Sacha Wildner <swildner@>
21 months agobge: More register value definition
Sepherosa Ziehau [Wed, 18 Jul 2012 08:52:08 +0000 (16:52 +0800)]
bge: More register value definition

21 months agopci: regen pcidevs
Sepherosa Ziehau [Tue, 17 Jul 2012 23:56:34 +0000 (07:56 +0800)]
pci: regen pcidevs

21 months agopci: Add device ids for Broadcom BCM57785 and BCM5718 families
Sepherosa Ziehau [Tue, 17 Jul 2012 23:55:53 +0000 (07:55 +0800)]
pci: Add device ids for Broadcom BCM57785 and BCM5718 families

21 months agobge: Add register and chip id values for BCM57785 and BCM5718 families
Sepherosa Ziehau [Tue, 17 Jul 2012 13:29:23 +0000 (21:29 +0800)]
bge: Add register and chip id values for BCM57785 and BCM5718 families

Obtained-from: FreeBSD