Commit Graph

21 Commits

Author SHA1 Message Date
adrian
67670e1f19 [bhnd] handle unknown bhnd port type. 2016-05-05 19:54:17 +00:00
adrian
b53c7f24e6 [bhnd] quieten gcc warning. 2016-05-05 19:38:53 +00:00
adrian
251aa10f1c [bhnd] handle unknown port type.
Reported by: emaste
2016-05-05 19:30:14 +00:00
emaste
b01f61207a bhnd: another build fix for GCC-using architectures
Further to r299119. GCC architectures failed with
bcma_subr.c:138: warning: control reaches end of non-void function

Sponsored by:	The FreeBSD Foundation
2016-05-05 17:47:03 +00:00
emaste
a846752fb2 bhnd: fix build on gcc architectures
"make tinderbox" fails on sparc64 GENERIC-NODEBUG with:
bhnd_subr.c:188: warning: control reaches end of non-void function

Sponsored by:	The FreeBSD Foundation
2016-05-05 15:07:40 +00:00
adrian
1927244465 [bwn] [bhnd] initial support for using bhnd for if_bwn devices.
This is an initial work in progress to use the replacement bhnd
bus code for devices which support it.

* Add manpage updates for bhnd, bhndb, siba
* Add kernel options for bhnd, bhndbus, etc
* Add initial support in if_bwn_pci / if_bwn_mac for using bhnd
  as the bus transport for suppoted NICs
* if_bwn_pci will eventually be the PCI bus glue to interface to bwn,
  which will use the right backend bus to attach to, versus direct
  nexus/bhnd attachments (as found in embedded broadcom devices.)

The PCI glue defaults to probing at a lower level than the bwn glue,
so bwn should still attach as per normal without a boot time tunable set.

It's also not fully fleshed out - the bwn probe/attach code needs to be
broken out into platform and bus specific things (just like ath, ath_pci,
ath_ahb) before we can shift the driver over to using this.

Tested:

* BCM4311, STA mode
* BCM4312, STA mode

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D6191
2016-05-04 23:38:27 +00:00
pfg
eed4bd22ad sys/dev: minor spelling fixes.
Most affect comments, very few have user-visible effects.
2016-05-03 03:41:25 +00:00
adrian
e760d28035 [bhnd] fix more module dependencies.
Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D6180
2016-05-02 22:30:46 +00:00
adrian
9a744dd2b2 [bhnd] add module dependencies.
Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D6179
2016-05-02 21:01:08 +00:00
adrian
fbc1338474 [bhnd] Fix ChipCommon probing.
ChipCommon probing uses mapping table "chipc_devices". It calls bhnd_device_lookup,
which iterate over mapping table with end condition:

entry->desc != NULL

So if mapping table contains row with description equals to NULL, it will
stop processing of mapping. I.e. description is mandatory field and should
be not NULL.

This patch corrects mapping table for ChipCommon.

Submitted by:	Michael Zhilin <mizhka@gmail.com>
Differential Revision:	https://reviews.freebsd.org/D6088
2016-04-25 16:53:13 +00:00
adrian
2b13469532 [bhnd] use correct bus deallocation method
Submitted by:	Mizhka Zhilin <mizhka@gmail.com>
Differential Revision:	https://reviews.freebsd.org/D5647
2016-04-22 16:28:41 +00:00
adrian
2d9c76e920 [bhnd] Update bhndb format strings to match long -> rman_res_t size change.
Does what it says on the tin; this unbreaks 32-bit builds.

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D5764
2016-04-22 16:27:35 +00:00
adrian
d8ac63dee1 [bhnd] Add a common bhnd_pci driver shared by both bhnd_pcib and bhnd_pci_hostb
This extracts common code from bhndb_pci, bhnd_pcib, and bhnd_pci_hostb into a
simpler shared bhnd_pci base driver, and should enable SoC-side implementation
of bhnd_pcib root complex support.

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D5763
2016-04-22 16:26:53 +00:00
adrian
7b1f91ce8b [bhnd] Standardize bhnd device tables and quirk matching.
This add a bhnd device table mechanism that standardizes matching of
devices on the bhnd(4) bus, discovery of device quirk flags, and should
be pluggable into the new PNPINFO machinery.

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D5759
2016-04-19 15:56:39 +00:00
adrian
240c5e5159 [bhnd] Clean up bhnd resource handling and inherited bus methods
To facilitate use by SoC implementors working with bhnd-inheriting fdt/nexus
drivers:

* Splits bhnd_bus method implementations into generic bus implementations
  (bhnd_bus_generic) and generic bhnd(4) driver implementations (bhnd_generic)
* Simplifies bhnd resource handling, allowing bhnd bus implementations to
  support bhnd resource activation by implementing the standard BUS_*
  resource APIs and BHND_BUS_ACTIVATE_RESOURCE().

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D5758
2016-04-19 15:53:57 +00:00
adrian
d094c9b381 [bhnd] Add support for specifying the address space used by bhndb children
This adds support for specifying the address space used by a bridge child;
this will either be the bridged SoC address space, or the host address space
required by children that map non SoC-address ranges from the PCI BAR.

This is necessary to support SROM/OTP child devices that live directly
beneath the bhndb device and require access to host resources, instead
of the standard behavior of delegating access to the bridged SoC address
space.

Submitted by:	Landon Fuller <landonf@landonf.org>
Differential Revision:	https://reviews.freebsd.org/D5757
2016-04-19 15:52:55 +00:00
pfg
b63211eed5 Cleanup unnecessary semicolons from the kernel.
Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
jhibbits
c55aa7292d Fix the resource_list_print_type() calls to use uintmax_t.
Missed a bunch from r297000.
2016-03-22 22:25:08 +00:00
jhibbits
70aaabfeac Replace all resource occurrences of '0UL/~0UL' with '0/~0'.
Summary:
The idea behind this is '~0ul' is well-defined, and casting to uintmax_t, on a
32-bit platform, will leave the upper 32 bits as 0.  The maximum range of a
resource is 0xFFF.... (all bits of the full type set).  By dropping the 'ul'
suffix, C type promotion rules apply, and the sign extension of ~0 on 32 bit
platforms gets it to a type-independent 'unsigned max'.

Reviewed By: cem
Sponsored by:	Alex Perez/Inertial Computing
Differential Revision: https://reviews.freebsd.org/D5255
2016-03-03 05:07:35 +00:00
jhibbits
48d897bc00 Replace some more default range checks with RMAN_IS_DEFAULT_RANGE().
This is a follow-on to r295832.
2016-03-02 03:26:56 +00:00
adrian
8e7c909559 Bring over the initial rewrite of the broadcom bus support found in their
SoCs and various chips (including, famously, their wifi chips.)

This is "just" (all 20,000 lines of it) code to enumerate the various
versions of busses inside these devices, including the PCI bridge and
the direct SIBA bridge found in MIPS chips.

It also includes shared code for some bus operations (suspend, resume, etc);
EEPROM/SROM/etc parsing and other things that are shared between chips.

Eventually this'll replace the code that bwi/bwn uses for the internal
bus, as well as some apparently upcoming mips74k broadcom SoC support
which uses bwn!

Thanks to Landon Fuller <landonf@landonf.org> for all this work!

Obtained from:	https://github.com/landonf/freebsd/compare/user/landonf/bcm4331-CURRENT
2016-02-26 03:34:08 +00:00