1
0
mirror of https://git.FreeBSD.org/src.git synced 2024-12-23 11:18:54 +00:00
Commit Graph

101790 Commits

Author SHA1 Message Date
Robert Watson
2a8c860fe3 In order to reduce use of M_EXT outside of the mbuf allocator and
socket-buffer implementations, introduce a return value for MCLGET()
(and m_cljget() that underlies it) to allow the caller to avoid testing
M_EXT itself.  Update all callers to use the return value.

With this change, very few network device drivers remain aware of
M_EXT; the primary exceptions lie in mbuf-chain pretty printers for
debugging, and in a few cases, custom mbuf and cluster allocation
implementations.

NB: This is a difficult-to-test change as it touches many drivers for
which I don't have physical devices.  Instead we've gone for intensive
review, but further post-commit review would definitely be appreciated
to spot errors where changes could not easily be made mechanically,
but were largely mechanical in nature.

Differential Revision:	https://reviews.freebsd.org/D1440
Reviewed by:	adrian, bz, gnn
Sponsored by:	EMC / Isilon Storage Division
2015-01-06 12:59:37 +00:00
Hans Petter Selasky
fe657d68f5 Fixes and updates for the Linux compatibility layer:
- Remove unsupported "bus" field from "struct pci_dev".
- Fix logic inside "pci_enable_msix()" when the number of allocated
  interrupts are less than the number of available interrupts.
- Update header files included from "list.h".
- Ensure that "idr_destroy()" removes all entries before destroying
  the IDR root node(s).
- Set the "device->release" function so that we don't leak memory at
  device destruction.
- Use FreeBSD's "log()" function for certain debug printouts.
- Put parenthesis around arguments inside the min, max, min_t and max_t macros.
- Make sure we don't leak file descriptors by dropping the extra file
  reference counts done by the FreeBSD kernel when calling falloc()
  and fget_unlocked().

MFC after:	1 week
Sponsored by:	Mellanox Technologies
2015-01-06 10:02:14 +00:00
Hans Petter Selasky
b2d77b2e23 Fix kernel build after r276730. 2015-01-06 09:39:42 +00:00
Craig Rodrigues
4de985af0b Instead of creating a purge thread for every vnet, create
a single purge thread and clean up all vnets from this thread.

PR:                     194515
Differential Revision:  D1315
Submitted by:           Nikos Vassiliadis <nvass@gmx.com>
2015-01-06 09:03:03 +00:00
Craig Rodrigues
c75820c756 Merge: r258322 from projects/pf branch
Split functions that initialize various pf parts into their
    vimage parts and global parts.
    Since global parts appeared to be only mutex initializations, just
    abandon them and use MTX_SYSINIT() instead.
    Kill my incorrect VNET_FOREACH() iterator and instead use correct
    approach with VNET_SYSINIT().

PR:			194515
Differential Revision:	D1309
Submitted by: 		glebius, Nikos Vassiliadis <nvass@gmx.com>
Reviewed by: 		trociny, zec, gnn
2015-01-06 08:39:06 +00:00
Adrian Chadd
bc6ee3cc93 Glue to suck in the QCA9550 HAL support if AH_SUPPORT_QCA9550 is defined. 2015-01-06 07:52:19 +00:00
Adrian Chadd
7b95ec7e1f Add option to support the QCA955x SoC wifi. 2015-01-06 07:51:46 +00:00
Adrian Chadd
87a2f105ee Make the apb.c code optional behind ar71xx_apb rather than standard.
The QCA955x has more mux interrupts going on - and the AR934x actually does,
but I cheated and assigned wlan and pcie to the same interrupt line.
They are, there's just a status register mux that I should've been using.

Luckily this isn't too bad a change in itself - almost all of the
Atheros MIPS configurations use a _BASE file to inherit from.
Except PB92, which I should really fix up at some point.

The AR934x will use the legacy apb for now until I write its replacement.

The QCA955x SoC I'm doing bring-up on will have a separate qca955x_apb.c
implementation that includes hooking into IP2/IP3 and doing further
interrupt demuxing as appropriate.
2015-01-06 07:43:07 +00:00
Adrian Chadd
e090d1f591 Add an APB base/size for the QCA955X for an upcoming QCA955x specific
APB mux.

It's larger than the AR71xx because it needs to replace the nexus
for some devices (notably wifi) and the wifi driver (if_ath_ahb.c)
reads the SPI data directly at early boot whilst it's memory mapped
in.

I'm eventually going to rip it out and replace it with a firmware
interface similar to what exists for the if_ath_pci.c path -
something early on (likely something new that I'll write) will
suck in the calibration data into a firmware API blob and that'll
be accessed from if_ath_ahb.c.

But, one thing at a time.

Tested:

* QCA955x SoC, AP135 development board
2015-01-06 07:37:33 +00:00
Adrian Chadd
0723a49181 The QCA955x USB init path doesn't require any of this, so delete it.
Obtained from:	Linux/OpenWRT
2015-01-06 07:35:05 +00:00
Warner Losh
a91275f72f Remove old ioctl use and support, once and for all. 2015-01-06 05:28:37 +00:00
Craig Rodrigues
a3f9f9a5a8 Add some GEOM options. 2015-01-06 05:28:22 +00:00
Craig Rodrigues
0db7f806be Refactor common options from MALTA and MALTA64 kernel
configs into MALTA_COMMON file.
2015-01-06 05:26:13 +00:00
Adrian Chadd
cc6dd788c6 Refactor and split out the net80211 software scan engine from the rest
of the scan API.

The eventual aim is to have 'ieee80211_scan.c' have the net80211 and
driver facing scan API to start, finish and continue doing scanning
while 'ieee80211_swscan.c' implements the software scanner that
runs the scan task, handles probe request/reply bits, configures
the VAP off-channel, changes channel and does the scanning bits.

For NICs that do no scanning at all, the existing code is needed.
ath(4) and most of the other NICs (dumb USB ones in particular)
do little to no scan offload - it's all done in software.

Some NICs may do single channel at a time scanning; I haven't really
checked them out in detail.

iwn(4), the upcoming 7260 driver stuff, the new Qualcomm Atheros
11ac chipsets and the Atheros mobile/USB full-offload chips all
have complete scan engines in firmware.  We don't have to drive
any of it at all - the firmware just needs to be told what to scan,
when to scan, how long to scan.  It'll take care of going off
channel, pausing TX/RX appropriately, sending sleep notification
to the AP, sending probe requests and handling probe responses.
It'll do passive/active scan itself.  It's almost completely
transparent to the network stack - all we see are scan notifications
when it finishes scanning each channel and beacons/probe responses
when it does its thing.  Once it's done we get a final notification
that the scan is complete, with some scan results in the message.
The iwn(4) NICs handle doing active scanning too as an option
and will handle waiting appropriately on 5GHz passive channels
before active scanning.

There's some more refactoring, tidying up and lock assertions to
sprinkle around to tidy this whole thing up before I turn swscan.c
into another set of ic methods to override by the driver or
alternate scan module.  So in theory this is all one big no-op
commit.  In theory.

Tested:

* iwn(4) 5200, STA mode
* ath(4) 6205, STA mode
* ath(4) - various NICs, AP mode
2015-01-06 02:08:45 +00:00
Navdeep Parhar
f8c479085f cxgbe/tom: use vmem(9) as the DDP page pod allocator.
MFC after:	1 month
2015-01-06 01:30:32 +00:00
Navdeep Parhar
008015d2f4 cxgbe(4): fix the description of a strange bunch of counters.
MFC after:	1 week
2015-01-05 23:43:24 +00:00
Mark Johnston
bbd685e3a5 Use crcopysafe(9) to make a copy of a process' credential struct. crcopy(9)
may perform a blocking memory allocation, which is unsafe when holding a
mutex.

Differential Revision:	https://reviews.freebsd.org/D1443
Reviewed by:	rwatson
MFC after:	1 week
Sponsored by:	EMC / Isilon Storage Division
2015-01-05 23:07:22 +00:00
Nathan Whitehorn
08b96b9ff5 Restore use of ofw_bus_intr_to_rl() in the pseries vdevice driver after fixing
ofw_bus_intr_to_rl() to match the spec for unspecified interrupt-parent
properties.
2015-01-05 21:39:35 +00:00
John Baldwin
531d65e139 Trim trailing whitespace. 2015-01-05 20:50:44 +00:00
John Baldwin
92597e064b On some Intel CPUs with a P-state but not C-state invariant TSC the TSC
may also halt in C2 and not just C3 (it seems that in some cases the BIOS
advertises its C3 state as a C2 state in _CST).  Just play it safe and
disable both C2 and C3 states if a user forces the use of the TSC as the
timecounter on such CPUs.

PR:		192316
Differential Revision:	https://reviews.freebsd.org/D1441
No objection from:	jkim
MFC after:	1 week
2015-01-05 20:44:44 +00:00
Hans Petter Selasky
b217d18412 Add 64-bit DMA support in the XHCI controller driver.
- Fix some comments and whitespace while at it.

MFC after:	1 month
Submitted by:	marius@
2015-01-05 20:22:18 +00:00
John Baldwin
f045564dbb Use struct thread * directly instead of d_thread_t. This driver is not
likely to be merged back to stable/4.

Reviewed by:	delphij
2015-01-05 19:49:20 +00:00
Marius Strobl
799c49196c Exclude drivers which build but don't actually work with PAE enabled,
missed in r276377
2015-01-05 19:15:35 +00:00
Nathan Whitehorn
efabbd4c02 Revert r272109 locally, which is not quite equivalent in how it deals with
missing interrupt-parent properties. A better solution will come later,
but this restores pseries in QEMU for the time being.
2015-01-05 18:15:16 +00:00
Nathan Whitehorn
1d67d16647 Missed change in r276688. Apologies for build breakage. 2015-01-05 16:52:25 +00:00
Ruslan Bukin
1aac28e753 o Switch to use non-mergeable RX buffers to avoid mbuf adjustment needs
o Operate with copy of iov as we expect later it was not modified
2015-01-05 16:43:22 +00:00
Hans Petter Selasky
ece4b0bd43 Make a bunch of USB debug SYSCTLs tunable, so that their value(s) can
be set before the USB device(s) are probed.
2015-01-05 15:04:17 +00:00
Warner Losh
9aca3e0f2b Clang's 3.5 integrated assembler now handles these files correctly (it
has support for the .codeXX directives). However, it is desirable, for
a time, to allow kernels to be built with clang 3.4. Historically, it
has been advantageous to allow stable X-1 to build kernels the old
way (so long as the impact of doing so is small), and this restores
that ability.

Also, centralize the addition of ${ASM_CFLAGS.${.IMPSRC}}, place it in
kern.mk rather than kern.pre.mk so that all modules can benefit, and
give the same treatment to CFLAGS in kern.mk as well.
2015-01-05 12:28:22 +00:00
Bjoern A. Zeeb
5334f8d74f With r276692 m_align() moved and became an inline function.
Remove the (now) redundant function declaration.
2015-01-05 11:22:58 +00:00
Robert Watson
ed6a66ca6c To ease changes to underlying mbuf structure and the mbuf allocator, reduce
the knowledge of mbuf layout, and in particular constants such as M_EXT,
MLEN, MHLEN, and so on, in mbuf consumers by unifying various alignment
utility functions (M_ALIGN(), MH_ALIGN(), MEXT_ALIGN() in a single
M_ALIGN() macro, implemented by a now-inlined m_align() function:

- Move m_align() from uipc_mbuf.c to mbuf.h; mark as __inline.
- Reimplement M_ALIGN(), MH_ALIGN(), and MEXT_ALIGN() using m_align().
- Update consumers around the tree to simply use M_ALIGN().

This change eliminates a number of cases where mbuf consumers must be aware
of whether or not mbufs returned by the allocator use external storage, but
also assumptions about the size of the returned mbuf. This will make it
easier to introduce changes in how we use external storage, as well as
features such as variable-size mbufs.

Differential Revision:	https://reviews.freebsd.org/D1436
Reviewed by:	glebius, trasz, gnn, bz
Sponsored by:	EMC / Isilon Storage Division
2015-01-05 09:58:32 +00:00
Adrian Chadd
cd470fead8 Remove the remnants of the OpenWRT/Linux bits that this was based off
of.

Obtained from:	Linux/OpenWRT
2015-01-05 05:30:07 +00:00
Adrian Chadd
3eda0cb18c Oops - missed refclk.
Tested:

* AP135, QCA955x SoC
2015-01-05 05:26:57 +00:00
Nathan Whitehorn
a4f7863238 Add kexec-based Linux shim loader for use with the firmware petitboot
loader on PS3 and POWER8 systems. This is reasonably portable to other
architectures, especially FDT-based ones, if similar features are useful
elsewhere.

Netboot support is missing for now and will be added in a future commit,
at which time loader.ps3 will be garbage collected.

Done at:	Hackathon
2015-01-05 04:33:39 +00:00
Adrian Chadd
855c46100d Add initial Qualcomm Atheros QCA955x SoC support.
This adds the initial frequency poking and configures up enough
for it to boot and spit out data over the console.

There's still a whole bunch of work to do in the reset path
and devices to support this thing, but hey, it's alive!

ath> go 0x80050100
## Starting application at 0x80050100 ...
CPU platform: Atheros AR9558 rev 0
CPU Frequency=720 MHz
CPU DDR Frequency=600 MHz
CPU AHB Frequency=200 MHz
platform frequency: 720 MHz
CPU reference clock: 0 MHz
CPU MDIO clock: 40 MHz

Done at:	hackathon
Obtained from:	Linux OpenWRT, Qualcomm Atheros
2015-01-05 02:06:26 +00:00
Adrian Chadd
917d2c3c3b ACK interrupts on the new SoCs. 2015-01-05 02:00:41 +00:00
Adrian Chadd
dc6fff9835 add QCA955x SoC types. 2015-01-05 01:59:44 +00:00
Adrian Chadd
55b32e75df Add QCA955x series register definitions.
There's likely a bunch of register offsets that I have to add the
register window base to before I use them.

Done at:	Hackathon
Obtained from:	Linux OpenWRT
2015-01-05 01:44:23 +00:00
Nathan Whitehorn
fd520b6e21 Use FDT properties to identify a PS3 if present. 2015-01-05 01:05:35 +00:00
Nathan Whitehorn
f1d2752f31 Make PS3 work with the userspace kboot loader. loader.ps3 will disappear
from the tree in the near future.

Done at:	Hackathon
2015-01-05 00:50:16 +00:00
Ruslan Bukin
2499610b9a Do not configure Altera PIO device on ARM startup.
PIO is a device implemented in soft-core and becomes
available after flashing FPGA only.
2015-01-04 23:14:04 +00:00
Ian Lepore
bbae23a725 Eliminate uninitialized variable warnings in kernel and module builds when
building with gcc 4.2

This has been requested several times over the past few months by several
people (including me), because gcc 4.2 just gets it wrong too often.  It's
causing us to litter the code with lots of bogus initializers just to
squelch the warnings.  We still have clang and coverity telling us about
uninitialized variables, and they do so more accurately.
2015-01-04 20:08:24 +00:00
Justin T. Gibbs
5b326a32ce Prevent live-lock and access of destroyed data in taskqueue_drain_all().
Phabric:	https://reviews.freebsd.org/D1247
Reviewed by:	jhb, avg
Sponsored by:	Spectra Logic Corporation

sys/kern_subr_taskqueue.c:
	Modify taskqueue_drain_all() processing to use a temporary
	"barrier task", rather than rely on a user task that may
	be destroyed during taskqueue_drain_all()'s execution.  The
	barrier task is queued behind all previously queued tasks
	and then has its priority elevated so that future tasks
	cannot pass it in the queue.

	Use a similar barrier scheme to drain threads processing
	current tasks.  This requires taskqueue_run_locked() to
	insert and remove the taskqueue_busy object for the running
	thread for every task processed.

share/man/man9/taskqueue.9:
	Remove warning about live-lock issues with taskqueue_drain_all()
	and indicate that it does not wait for tasks queued after
	it begins processing.
2015-01-04 19:55:44 +00:00
Kevin Lo
aad37c8145 Change the type of blksize_t from uint32_t to a POSIX compliant int32_t
http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/sys_types.h.html
2015-01-04 15:26:02 +00:00
Dmitry Chagin
1beb1a8e13 Regen for r276654 (__getcwd()). 2015-01-04 10:40:23 +00:00
Dmitry Chagin
9f7a06f27e Indeed, instead of hiding the kern___getcwd() bug by bogus cast
in r276564, change path type to char * (pathnames are always char *).
And remove bogus casts of malloc().
kern___getcwd() internally doesn't actually use or support u_char *
paths, except to copy them to a normal char * path.

These changes are not visible to libc as libc/gen/getcwd.c misdeclares
__getcwd() as taking a plain char * path.

While here remove _SYS_SYSPROTO_H_ for __getcwd() syscall as
we always have sysproto.h.

Pointed out by:	bde

MFC after:	1 week
2015-01-04 10:34:02 +00:00
Justin Hibbits
41ddc6ea3d Truncate DB_SMALL_VALUE_MAX to a much lower value.
Unlike the other architectures, the PowerPC kernel is loaded under the 2GB
boundary.

MFC after:	2 weeks
2015-01-04 01:45:26 +00:00
Ian Lepore
c6f7b90092 Don't allow clang to use fpu instructions or registers in kernel modules.
Submitted by:	Keith White <kwhite@site.uottawa.ca>
2015-01-04 00:12:30 +00:00
Ian Lepore
fbf1ef4849 Switch all arm kernels with option SMP to use SCHED_ULE instead of 4BSD. 2015-01-03 23:57:03 +00:00
Ian Lepore
dad0c32719 Remove commented-out options SMP and APIC_IO from old arm boards that will
never be able to support smp.
2015-01-03 23:30:29 +00:00
Justin Hibbits
031a27e354 The radeon DRI module doesn't really work correctly on powerpc, so don't build
it.  Hopefully we'll get radeonkms eventually.

Discussed with:	nwhitehorn
MFC after:	2 weeks
2015-01-03 22:36:18 +00:00