Commit graph

15409 commits

Author SHA1 Message Date
Johan Hedberg
470349c25a Bluetooth: settings: Add support for per-submodule handlers
Add a new linker section for a list of submodule settings handlers,
and iterate the list from the various settings callbacks.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
2018-05-04 17:26:05 -04:00
Johan Hedberg
d22b7c9f2d Bluetooth: Remove bt_storage API
The same functionality is now supported by the settings-based
solution, so remove bt_storage out of the way. There were stubs in
bt_storage to handle per-peer information (e.g. pairing keys) but this
was never actually implemented in full. The next step is to add this
support to the settings-based solution.

Leave the code for generating temporary IRK and identity address in
case BT_SETTINGS is not enabled. Also leave the code for using vendor
HCI to read the identity address, in which case the settings
implementation will not touch it.

Introduce a new bt_unpair() API to replace the removed
bt_storage_clear(), since the latter was actually doing more than just
storage management: it was also handling runtime storage of pairing
information. Later, the bt_unpair() implementation will be extended to
clear settings-based pairing storage.

There is one feature that the bt shell module looses: the ability to
give a specific identity address to the "init" command as a parameter.
We might look later in the future if this is really needed, and add a
separate API for this.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
2018-05-04 17:26:05 -04:00
Manivannan Sadhasivam
26babfc622 drivers: led: Add system call handler support
Add system call handler support to LED subsystem. No buffers are
involved in any of the API's and hence the syscall support is
straightforward.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
2018-05-04 10:07:39 -07:00
Jukka Rissanen
128557896c net: if: Add functions to get correct IPv4 address
Add functions that will return correct source IPv4 address
according to given destination address. This is done similar
way as for IPv6.

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
2018-05-03 12:47:52 +03:00
Johan Hedberg
6a71a69f8d Bluetooth: GATT: Fix CCC handling
The 'valid' member of struct bt_gatt_ccc_cfg was redundant, since
setting 'peer' to BT_ADDR_LE_ANY does the same job. What's worse, the
handling of 'valid' was also buggy in that some places looking for
valid CCC structs only matched the address, meaning it might yield a
positive match for invalid entries.

Fix these issues by removing the 'valid' struct member, and solely
using the 'peer' member to identify valid entries. Also simplify the
code by acknowledging that no CCC entry is essentially the same as the
value '0' written to CCC.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
2018-05-03 09:53:16 +01:00
Andy Ross
15c400774e kernel: Rework SMP irq_lock() compatibility layer
This was wrong in two ways, one subtle and one awful.

The subtle problem was that the IRQ lock isn't actually globally
recursive, it gets reset when you context switch (i.e. a _Swap()
implicitly releases and reacquires it).  So the recursive count I was
keeping needs to be per-thread or else we risk deadlock any time we
swap away from a thread holding the lock.

And because part of my brain apparently knew this, there was an
"optimization" in the code that tested the current count vs. zero
outside the lock, on the argument that if it was non-zero we must
already hold the lock.  Which would be true of a per-thread counter,
but NOT a global one: the other CPU may be holding that lock, and this
test will tell you *you* do.  The upshot is that a recursive
irq_lock() would almost always SUCCEED INCORRECTLY when there was lock
contention.  That this didn't break more things is amazing to me.

The rework is actually simpler than the original, thankfully.  Though
there are some further subtleties:

* The lock state implied by irq_lock() allows the lock to be
  implicitly released on context switch (i.e. you can _Swap() with the
  lock held at a recursion level higher than 1, which needs to allow
  other processes to run).  So return paths into threads from _Swap()
  and interrupt/exception exit need to check and restore the global
  lock state, spinning as needed.

* The idle loop design specifies a k_cpu_idle() function that is on
  common architectures expected to enable interrupts (for obvious
  reasons), but there is no place to put non-arch code to wire it into
  the global lock accounting.  So on SMP, even CPU0 needs to use the
  "dumb" spinning idle loop.

Finally this patch contains a simple bugfix too, found by inspection:
the interrupt return code used when CONFIG_SWITCH is enabled wasn't
correctly setting the active flag on the threads, opening up the
potential for a race that might result in a thread being scheduled on
two CPUs simultaneously.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-05-02 10:00:17 -07:00
Johan Hedberg
92749e7a92 Bluetooth: Mesh: Remove unnecessary #ifdefs from header file
The #ifdef protections are not needed, and in fact prevent helpers
such as IS_ENABLED() from being used when calling APIs which are
optional.

Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
2018-05-02 15:45:16 +01:00
Alberto Escolar Piedras
e3efbd54c6 POSIX arch: Fix linker -T warning
Modify the linker script with the command INSERT.
It instructs the linker to augment the default linker script
SECTIONS with the ones provided with the one provided in
this script.
It also modified the meaning of the -T switch, so it no longer
replaces the default linker script

Signed-off-by: Alberto Escolar Piedras <alpi@oticon.com>
2018-05-02 10:19:09 -04:00
Tomasz Gorochowik
3da4d37003 net: Implement VLAN priority to packet priority conversion
According to IEEE 802.1Q the VLAN priority (PCP) is not directly mapped
to the network packet priority. The Best Effort priority has a PCP value
of 0. The lowest priority (Background) has a PCP value of 1.

All the values are mapped according to the following table:

  +-----+-----+---------+
  | PCP | PRI | Acronym |
  +-----+-----+---------+
  |  1  |  0  |    BK   |
  |  0  |  1  |    BE   |
  |  2  |  2  |    EE   |
  |  3  |  3  |    CA   |
  |  4  |  4  |    VI   |
  |  5  |  5  |    VO   |
  |  6  |  6  |    IC   |
  |  7  |  7  |    NC   |
  +-----+-----+---------+

Signed-off-by: Tomasz Gorochowik <tgorochowik@antmicro.com>
2018-05-02 13:40:32 +03:00
Savinay Dharmappa
97f721d928 dts: xtensa: Add device tree support for xtensa
Change-Id: I309bc50c6b575caa84fbc7ab98cc9890771b4274
Signed-off-by: Savinay Dharmappa <savinay.dharmappa@intel.com>
2018-05-01 16:46:41 -04:00
Rajavardhan Gundi
5b6f02442b drivers: i2s: introduce CAVS I2S
Implements the driver for Intel CAVS I2S. Only Playback
is currently supported.

Change-Id: I7b816f9736dc35e79a81d3664d6405dc0aac15b4
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-05-01 16:46:41 -04:00
Rajavardhan Gundi
08172cdf83 xtensa: provide XCC compiler support for Xtensa
This patchset provides Xtensa's xcc compiler support for Xtensa
projects in Cmake. This requires the below environment variables
to be defined aptly. The appropriate xcc license information also
need to be supplied.

ZEPHYR_GCC_VARIANT=xcc
TOOLCHAIN_VER=RF-2015.3-linux
XTENSA_CORE=cavs21_LX6HiFi3_RF3_WB16
XTENSA_SYSTEM=/opt/xtensa/XtDevTools/install/tools/
		RF-2015.3-linux/XtensaTools/config/
XTENSA_BUILD_PATHS=/opt/xtensa/XtDevTools/install/builds/

Change-Id: Ib3c10e8095439b0e32276ff37c00eca8420773ec
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-05-01 16:46:41 -04:00
Rajavardhan Gundi
dadf9e7a81 xtensa: intel_s1000: implement interrupt mechanism
intel_s1000 has multiple levels of interrupts consisting of core, CAVS
Logic and designware interrupt controller. This patchset modifies
the regular gen_isr mechanism to support these multiple levels.

Change-Id: I0450666d4e601dfbc8cadc9c9d8100afb61a214c
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-05-01 16:46:41 -04:00
Rajavardhan Gundi
74016bb64c drivers: interrupts: introduce CAVS interrupt logic
CAVS interrupt logic is an intel IP that combines several sources of
interrupt into one line that is then routed to the parent controller.
CAVS stands for "connected Audio, Voice and Speech". This IP supports
4 lines which can have a max of 32 interrupts each.

Change-Id: Ia6be51428bedf1011d148ae1fc5d4c34252c05da
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-05-01 16:46:41 -04:00
Rajavardhan Gundi
47ff96593f intel_s1000: uart: configure UART for intel_s1000
intel_s1000 uses DesignWare IP for UART. National Semiconductor
16550 (UART) component specification is followed in this IP.

Change-Id: Ied7df1dc178d55b6dbe71d729d6383ba07274ea4
Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-05-01 16:46:41 -04:00
Leandro Pereira
0ec79d6853 lib: json: Efficiently pack field name, offset, alignment, type
This trades a little bit over 40 bytes (on x86) of text for a lot of
savings in rodata.  This is accomplished by using bitfields to pack the
field name length, offset, alignment, and the type tag into a single
32-bit unsigned integer instead of scattering this information into
four different integers.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-05-01 15:44:09 -04:00
Johann Fischer
46700eaa65 include: usb: add USB DFU class header
Add USB DFU class header.

Signed-off-by: Johann Fischer <j.fischer@phytec.de>
2018-04-30 17:42:40 -04:00
Rajavardhan Gundi
68040c8d78 kernel: sem: Modify the way BUILD_ASSERT is used
BUILD_ASSERT() macro makes use of __COUNTER__ which may not be
supported in some compilers (like xcc). So, multiple uses of
BUILD_ASSERT() in same scope is not possible for such compilers.
Instead, the expression to BUILD_ASSERT can be "&&"ed to achieve
the same purpose.

Signed-off-by: Rajavardhan Gundi <rajavardhan.gundi@intel.com>
2018-04-30 16:46:14 -04:00
Andrew Boie
9faa42f522 slist: abstract node and list implementation
Singly-linked lists work in a specific way, but this will allow
slist semantics to be defined for any data structure that can store
a "next" pointer in any way at all, possibly something not even
a pointer value, but a table index.

The immediate need for this patch is to allow the easy definition
of an slist variant which stores flags in the low-order bits of the
node pointer.

The compiler does not need them, but the function prototypes have
been left in slist.h to make the header easier to understand by
the end user and not confuse Doxygen.

Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
2018-04-30 13:15:51 -04:00
Manivannan Sadhasivam
94bba071d6 drivers: led: Add public API for LED drivers
Add support for LED APIs for controlling the LED devices. This
API can be used by the LED devices present on the chip and connected
externally via buses like I2C, SPI etc...

Following APIs are currently supported:

1. led_blink
2. led_set_brightness
3. led_on
4. led_off

Driver support using these APIs will be added in subsequent patches.

Signed-off-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
2018-04-27 10:46:53 -05:00
Jukka Rissanen
8c561994f2 net: http: Fix client compilation if HTTPS is enabled
The http.h used HTTP server defines when compiling HTTP client.
This causes compilation error.

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
2018-04-26 12:35:05 +03:00
Adithya Baglody
960e97834f x86: linker: Maintain 4K alignment for application memory.
When MMU is enabled and the SOC we are running doesn't have
execute in-place(XIP) the final image will be a monolith which
sits in RAM. In such situations we need to maintain the alignment
for application memory. If not maintained the MMU boot tables
will not be configured properly.

Signed-off-by: Adithya Baglody <adithya.nagaraj.baglody@intel.com>
2018-04-25 14:46:00 -07:00
Leandro Pereira
c200367b68 drivers: Perform a runtime check if a driver is capable of an operation
Driver APIs might not implement all operations, making it possible for
a user thread to get the kernel to execute a function at 0x00000000.

Perform runtime checks in all the driver handlers, checking if they're
capable of performing the requested operation.

Fixes #6907.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-26 02:57:12 +05:30
Andrzej Puzdrowski
9fe30535d7 susbsys: settings: fix coverity issues
API settings_subsys_init call was changed so that it returns
error (so returns int instead of void).
Prototype of storage helper function export_func for
settings_handler::h_export was changed so that it returns error
(so returns int instead of void).
Fixed few other error handling issues by ignoring return
values.

Tests were aligned to above patches.

Signed-off-by: Andrzej Puzdrowski <andrzej.puzdrowski@nordicsemi.no>
2018-04-25 22:25:32 +05:30
Andrew Boie
31bdfc014e userspace: add support for dynamic kernel objects
A red-black tree is maintained containing the metadata for all
dynamically created kernel objects, which are allocated out of the
system heap.

Currently, k_object_alloc() and k_object_free() are supervisor-only.

Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
2018-04-24 12:27:54 -07:00
Leandro Pereira
f5f95ee3a9 kernel: sem: Ensure that initial count is lesser or equal than limit
Ensure this value during static initialization (with build assertions),
and dynamic initializations through system calls.

If initial count is larger than the limit, it's possible for the count
to wraparound, causing locking issues.

Expanding the BUILD_ASSERT() macros after declaring a k_sem struct in
K_SEM_DEFINE() is necessary to support cases where a semaphore is
defined statically.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-24 04:04:36 +05:30
Leandro Pereira
05169a02ce toolchain: common: Allow multiple uses of BUILD_ASSERT() in same scope
BUILD_ASSERT() was always defining a type with the name
__build_assert_failure, causing issues if more than one assertion were
used in the same scope.

Also, use an enum instead of a typedef to avoid (possibly spurious)
warnings such as these:

  variably modified ‘__build_assert_failure1’ at file scope [-Werror]

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-24 04:04:36 +05:30
Leandro Pereira
666274fa60 toolchain: gcc: Only use _Static_assert if building with C11
Zephyr currently builds as C99, whereas _Static_assert is a C11
feature.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-24 04:04:36 +05:30
Leandro Pereira
ae7911cce0 arch: x86: segmentation: Remove unused _far_call() and _far_jump()
These functions were not used throughout the Zephyr code base, and
as such has been removed.  They can be reinstated if there's a need,
but will need to be adapted to use retpolines when CONFIG_RETPOLINE
is set.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-24 04:00:01 +05:30
Leandro Pereira
16472cafcf arch: x86: Use retpolines in core assembly routines
In order to mitigate Spectre variant 2 (branch target injection), use
retpolines for indirect jumps and calls.

The newly-added hidden CONFIG_X86_NO_SPECTRE flag, which is disabled
by default, must be set by a x86 SoC if its CPU performs speculative
execution.  Most targets supported by Zephyr do not, so this is
set to "y" by default.

A new setting, CONFIG_RETPOLINE, has been added to the "Security
Options" sections, and that will be enabled by default if
CONFIG_X86_NO_SPECTRE is disabled.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-24 04:00:01 +05:30
Andy Ross
3f55dafebc kernel: Deprecate k_thread_cancel() API
The only difference between this call and k_thread_abort() (beyond
some minor performance deltas) is that "cancel" will act as a noop in
cases where the thread has begun execution and will return an error.
"Abort" always succeeds, of course.  That is inherently racy when used
as a "stop the thread" API: there's no way in general (or at all in
SMP situations) to know that you're calling this function "early
enough" to catch the thread before it starts.

Effectively, all k_thread_cancel() gives you that k_thread_abort()
doesn't is an indication about whether or not a thread has started.
There are many other ways to get that information that don't require
dangerous kernel APIs.

Deprecate this function.  Zephyr's own code never used it except for
its own unit test.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-04-24 03:57:20 +05:30
Andy Ross
8606fabf74 kernel: Scheduler refactoring: use _reschedule_*() always
There was a somewhat promiscuous pattern in the kernel where IPC
mechanisms would do something that might effect the current thread
choice, then check _must_switch_threads() (or occasionally
__must_switch_threads -- don't ask, the distinction is being replaced
by real English words), sometimes _is_in_isr() (but not always, even
in contexts where that looks like it would be a mistake), and then
call _Swap() if everything is OK, otherwise releasing the irq_lock().
Sometimes this was done directly, sometimes via the inverted test,
sometimes (poll, heh) by doing the test when the thread state was
modified and then needlessly passing the result up the call stack to
the point of the _Swap().

And some places were just calling _reschedule_threads(), which did all
this already.

Unify all this madness.  The old _reschedule_threads() function has
split into two variants: _reschedule_yield() and
_reschedule_noyield().  The latter is the "normal" one that respects
the cooperative priority of the current thread (i.e. it won't switch
out even if there is a higher priority thread ready -- the current
thread has to pend itself first), the former is used in the handful of
places where code was doing a swap unconditionally, just to preserve
precise behavior across the refactor.  I'm not at all convinced it
should exist...

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-04-24 03:57:20 +05:30
Leandro Pereira
81f4b11126 include: toolchain: common: Remove deprecated __stack macro
This macro has been deprecated in favor of K_DECLARE_STACK; should have
been removed by 1.11.

Signed-off-by: Leandro Pereira <leandro.pereira@intel.com>
2018-04-23 10:11:08 -07:00
Andrei Emeltchenko
4f84cf782d usb: Add BOS Descriptors
Add Binary Device Object Store descriptors

Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
2018-04-20 21:04:42 -07:00
Wayne Ren
3d9ba10b5c arch: arc: bug fixes and optimization in exception handling
The original exception handling has space to optimize and
and some bugs need to be fixed.

* define NANO_ESF
   * add the definition of NANO_ESF which is an irq_stack_frame
   * add the corresponding codes in exception entry and handler
   * remove _default_esf
* implement the _ARCH_EXCEPT
   * use trap exception to raise exception by kernel
   * add corresponding trap exception entry
   * add _do_kernel_oops to handle the exception raised by
     _ARCH_EXCEPT.
* add the thread context switch in exception return
   * case: kernel oops may raise thread context switch
   * case: some tests will re-implement SysFatalHandler to raise
     thread context switch.
   * as the exception and isr are handled in kernel isr stack, so
     the thread context switch must be in the return of exception/isr
     , and the exception handler must return, should not be decorated
     with FUNC_NORETURN
* for arc, _is_in_isr should consider the case of exception

Signed-off-by: Wayne Ren <wei.ren@synopsys.com>
2018-04-17 10:50:12 -07:00
Andy Gross
09a8810b33 arm: userspace: Rework system call arguments
This patch changes the ARM system calls to use registers for passing
or arguments.  This removes the possibility of stack issues when
callers do not adhere to the AAPCS.

Fixes #6802

Signed-off-by: Andy Gross <andy.gross@linaro.org>
2018-04-16 10:13:20 -07:00
Erwan Gouriou
4fe3a9776f dts: gpio: create gpio dt-bingings and inlude in stm32 dtsi files
Create a dt-bindings/gpio.h file.
Bindings definitions are extracted from existing gpio.h.
gpio dt-bindings file is required because existing gpio.h file could
not be parsed by dts parser.

Signed-off-by: Erwan Gouriou <erwan.gouriou@linaro.org>
2018-04-13 13:36:45 -05:00
Jukka Rissanen
da8af39302 net: if: Add helper to select src interface for a IPv4 dst addr
If we have multiple network interfaces and we want to send
a IPv4 network packet to certain destination, then this new
helper can be used to figure out what network interface to use.

Note that this commit only adds support to select the correct network
interface according to destination IPv4 address. This does not enable
any automatic routing to happen.

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
2018-04-13 08:29:44 -04:00
Iván Sánchez Ortega
05f02bd038 docs: group the GPIO_* flags into logical groups.
Aesthetical changes to the documentation about GPIO flags - group
them together, add some headers, some minor markup and wording
modifications.

Signed-off-by: Iván Sánchez Ortega <ivan@sanchezortega.es>
2018-04-12 17:54:20 -05:00
Anas Nashif
c7f5cc9bcb license: fix spdx identifier in a few files
Use correct SPDX identifier for Apache 2.0.

Signed-off-by: Anas Nashif <anas.nashif@intel.com>
2018-04-12 15:19:51 -04:00
Tomasz Bursztyka
eab3f168fd net/mgmt/wifi: Add dedicated net mgmt hooks for WiFi offload devices
Exposing connect, disconnect and scan for now.

In case the iface is an instance of a WiFi offload device, the way it
manages scanning, connecting and disconnecting will be specific to that
device (not the mgmt interface obviously). In such case the device will
have to export relevantly a dedicated bunch of function to serve the
mgmt interface in a generic way.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-12 09:56:07 -04:00
Gil Pitney
17b923ae3c net: wifi: Add a wifi.h to hold WiFi definitions
Placeholder to hold generic WiFi definitions.

Signed-off-by: Gil Pitney <gil.pitney@linaro.org>
2018-04-12 09:56:07 -04:00
Patrik Flykt
7bde51d86a net/mgmt: Add initial WiFi management API definitions
Add empty WiFi network management functions that only return -ENETDOWN.
Define management handlers for scan, connect and disconnect requests,
again without any implementation nor parameters defined.

Signed-off-by: Patrik Flykt <patrik.flykt@intel.com>
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-12 09:56:07 -04:00
Tomasz Bursztyka
93ac7ce655 net/mgmt: Move NET_EVENT_INFO_MAX_SIZE into net core's private header
First because nobody needs to know that besides net_mgmt core and
secondary to avoid possible circular dependancy on
net_mgmt.h/net_event.h.

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-12 09:56:07 -04:00
Dario Pennisi
6d387ec98f net: Remove the need for an l2 on offloaded drivers
Drivers will be directly contacted via net_if's offload attribute. No
need for a an extra layer as an L2.

Signed-off-by: Dario Pennisi <dario@iptronix.com>
Signed-off-by: Massimiliano Agneni <massimiliano.agneni@iptronix.com>
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-12 09:56:07 -04:00
Jukka Rissanen
c1e7fd76ef net: stats: Add infrastructure for collecting ethernet stats
Make sure we are able to collect ethernet statistics and query
it via net management API.

Fixes #6899

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
2018-04-11 16:49:48 +03:00
Andy Ross
2ef57f0a1b lib/rbtree: Add a rb_contains() predicate
Returns true if the specified node is in the tree.  Allows the tree to
be used for "set" style semantics along with a lessthan_fn that simply
compares the nodes by their address.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-04-10 12:31:51 -04:00
Andy Ross
193f4feb84 lib: Red/Black balanced tree data structure
A balanced tree implementation for Zephyr as we grow into bigger
regimes where simpler data structures aren't appropriate.

This implements an intrusive balanced tree that guarantees O(log2(N))
runtime for all operations and amortized O(1) behavior for creation
and destruction of whole trees.  The algorithms and naming are
conventional per existing academic and didactic implementations, c.f.:

https://en.wikipedia.org/wiki/Red%E2%80%93black_tree

The implementation is size-optimized to prioritize runtime memory
usage.  The data structure is intrusive, which is to say the struct
rbnode handle is intended to be placed in a separate struct the same
way other such structures (e.g. Zephyr's dlist list) and requires no
data pointer to be stored in the node.  The color bit is unioned with
a pointer (fairly common for such libraries).  Most notably, there is
no "parent" pointer stored in the node, the upper structure of the
tree being generated dynamically via a stack as the tree is recursed.
So the overall memory overhead of a node is just two pointers,
identical with a doubly-linked list.

Code size above dlist is about 2-2.5k on most architectures, which is
significant by Zephyr standards but probably still worthwhile in many
situations.

Signed-off-by: Andy Ross <andrew.j.ross@intel.com>
2018-04-10 12:31:51 -04:00
Tomasz Bursztyka
8d558fb5ea net/ethernet: Add a management interface
This will permit to tweak ethernet L2 and devices settings at runtime.
Currently, only devices settings are tweaked through this interface.

Fixes #6640

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-10 14:45:37 +03:00
Tomasz Bursztyka
f3d8012655 net/ethernet: Add function driver API to change some hw configuration
Depending on what's supported by the device, it could be possible to
configure these at runtime such as auto-negociation, link speed, etc...

Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
2018-04-10 14:45:37 +03:00