Updated the riscv-privilege SOC family to account for SOCs supporting
a Platform-level Interrupt Controller (PLIC) as specified by the
riscv privilege architecture.
riscv-privilege SOCs supporting a PLIC have to implement the following
list of APIs:
void riscv_plic_irq_enable(uint32_t irq);
void riscv_plic_irq_disable(uint32_t irq);
int riscv_plic_irq_is_enabled(uint32_t irq);
void riscv_plic_set_priority(uint32_t irq, uint32_t priority);
int riscv_plic_get_irq(void);
Change-Id: I0228574967348d572afc98a79257c697efc4309e
Signed-off-by: Jean-Paul Etienne <fractalclone@gmail.com>
added the riscv-privilege SOC_FAMILY, under which all
riscv SOCs supporting the riscv privilege architecture
specifcation shall reside. These SOCs shall notably have
a common base for handling IRQs.
Moved riscv32-qemu under the riscv-privilege SOC_FAMILY
Change-Id: I5372cb38e3eaed78886f22b212ab4f881ef30b3f
Signed-off-by: Jean-Paul Etienne <fractalclone@gmail.com>
The params command is not printing the arguments, which makes this not
so useful to verify how things work, so added it to the sample.
Change-Id: I2d71ee064f12c7353ba83144824ae46803e2ad39
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
Since MSYS2 is far more stable, reliable, maintained and easy
to set up than MinGW as a Windows development environment,
default to it for the official Windows Getting Started guide,
including instructions to compile for both x86 and ARM with it.
Change-Id: Iaba61498b3b16a8ee1edbff953de16c275282c7b
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Tested on nrf52_pca10040 and nrf51_pca10028 using hello_world sample.
Change-Id: I7cdf1d21e7f8232da737a06e5afbfb1eaec05cde
Signed-off-by: Michał Kruszewski <mkru1992@gmail.com>
At present 'make flash' target for boards that support flashing with
OpenOCD works only with Zephyr SDK. This patch enables flashing
without Zephyr SDK as long as openocd itself is installed.
Change-Id: I4594d795225e54b7638839d590e4c0620f9ca88c
Jira: ZEP-1893
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
The only copy of the release notes for releases prior to 1.6 is on
the zephyrproject.org/downloads site. Preserving the release notes
history in the git repo would be a good thing. Here are the notes
for 1.5 (more to come).
Also added an auto table-of-contents header to make it easier to
navigate.
Change-Id: Ibd8b5a8d3816f7f5c376216adc2d52bb9467214d
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
A recent mailing list question asked for clarify about
the duration and period parameters for starting a timer.
Change-Id: I9bf8dd93dbbb9bbb94c95c2d7072f446ea1d6b01
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
MSYS2 is a modern fork of Cygwin that is widely supported
and provides frequent releases and good support. It is
also the framework used by the official Windows Git port.
This patch adds MSYS2 support so that builds on Windows using
MSYS2 work properly.
Change-Id: Ia5743a410d1cff983a7aab37f8e3d8228cb8ae8e
Signed-off-by: Carles Cufi <carles.cufi@nordicsemi.no>
k_delayed_work_init shall only be called once per work and not every
time before as that can trash the workqueue list if the work is pending.
Change-Id: Ib04b6689c59d03328687b0b579e80012dbd2f853
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
k_delayed_work_init shall only be called once per work and not every
time before as that can trash the workqueue list if the work is pending.
Change-Id: I454f40b63133bb7d6239dbe902858932bacfc454
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Allow user to set the four last bytes of a MAC address.
By default the bytes are selected randomly but for testing
purposes we need a way to specify the address in advance.
Change-Id: Ic9820a8026b327d1ee5025ab77fdc77f5778de67
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The RX callback and timeout parameters were reversed in call to
mbedtls_ssl_set_bio().
Change-Id: I285694e57a024ac3ce735e496daad690f94737ac
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Print TCP header information in one line as there is really no
need to use multiple lines. Also use debug level when printing
the header info so that it is only seen if debugging is activated.
Change-Id: I27f314ca060239545769dec07148897da3426436
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
k_delayed_work_init shall only be called once per work and not every
time before as that can trash the workqueue list if the work is pending.
Change-Id: Ib6398a6a843dd4d820529487ad76a9dc9d9e152e
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Due to commit 8308b9bd2d ("net/http: Add the HTTP/1.1 API")
every user of CONFIG_HTTP_PARSER would need to add CONFIG_HTTP to
their .conf files. Which is fine for intree samples/tests as they
have been adjusted, but the rest of world working on Zephyr apps
will need to make this changes as well.
Instead, we should have each of the following select HTTP instead of
depend on it, which will make future use of these configs and their
dependencies more intuitive:
HTTP_SERVER
HTTP_CLIENT
HTTP_PARSER
NOTE: As cleanup, this commit also removes the CONFIG_HTTP added to
samples and test .conf files.
Change-Id: I81cfaa19e37333b1bf98778f8147814780e7f77c
Signed-off-by: Michael Scott <michael.scott@linaro.org>
This fixes warnings when using %u to print the result of sizeof.
Change-Id: I5391456c855ec2785af22467d9d0f355bbb1e577
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Networking stack has split one global DATA pool to RX and TX DATA pools
and also added net_buf pool support to each context. Update the driver
to support this new design. Since the GMAC TX descriptor list has a fixed
size but the number of TX DATA buffers is no longer limited updating the
TX descriptor list has to be guarded by a semaphore.
Tested on Atmel SMART SAM E70 Xplained board
Change-Id: I181e1cdd183e173b85d5d1711b6e78cd5165666d
Signed-off-by: Piotr Mienkowski <piotr.mienkowski@gmail.com>
The return code from net_icmpv{4|6}_send_error() was not correct
if the error message could be sent. Now 0 is returned if sending
succeed, and <0 otherwise.
Change-Id: Iff67f097a9d9519c9f11d4cbc9cf428a7c74ec1b
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Update statistics for every sent, received and dropped
ICMPv{4|6} packet.
Change-Id: Ibe6f02e8222adb3db1f1dbd0cde1ed251710eb43
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we are replying ICMPv{4|6} message, then do not wait forever
for a free buffer. In a busy system, this might lead to non-progress
in RX path if we receive lot of packet from the network and never get
a free buffer.
Change-Id: Iaef92541b8745f872a07bc6e2052d0393d4d1e8b
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
commit 2e3e93dccb ("net: 6lo: Fix compress bit calculation") adds
a bit shift to the compress bit calculation, however it's typo'd as
a greater than comparison causing get_6co_compress() to return a
boolean value rather than the shifted version of the calculation.
Change-Id: Idacac2c032a03f1f75a78c94bed4c70428b8c77a
Signed-off-by: Michael Scott <michael.scott@linaro.org>
The IPv6 prefix address length was incorrectly set in net_if
after RA was received from the network. The incorrect "len"
variable was used instead of correct "prefix_len" in the prefix
info struct.
Change-Id: Ifeaf150b9960414d2ad200053a3bd5290dc8d365
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If the earlier prefix timeout is not cancelled before setting
it again, the system will hang.
Change-Id: I6e271294c8e5d43e3ceae4780b5d1c26bab2b296
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It is possible that the slip->last pointer is not set during
the startup of the slip driver if we receive some garbage bytes.
This in turn caused memory to be overwritten in random places.
Jira: ZEP-1737
Change-Id: Ia2ce193b81a8f707b7ca63497d497ea70699bb32
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The caller semaphore was released too early, this can cause the caller
to re-use the data and possibly corrupt the memory, if caller yields
and is run before this management thread. Solution is to first remove
the node from the list and then unlock the semaphore.
Change-Id: I02cef53559d776f32a5959380e6b7122cd5198c5
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Add CONFIG_NET_IPV6_NBR_CACHE option that is the base for both
neighbor discovery (ND) and duplicate address detection (DAD).
Both ND and DAD can be disabled if needed. If NBR cache is
disabled, then ND and DAD are disabled too.
Note that it makes not much sense to disable DAD or NBR cache
as IPv6 will not work properly without them. It is possible
to disable ND but then the neighbor information needs to come
via other sources like RPL.
Change-Id: I57c8668ad828b3a153dfc58eea78bf5f7ac3938a
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Although very unlikely, I saw crashes at startup that could
be because of the buf or iface were null.
Change-Id: I8649eeb4f24fcd7d9f2acaeefaba4e6593388e91
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
No need to always look through all the network interfaces
when receiving packet as we support only one interface and
it is always the same one.
Change-Id: Ic0bfabc8409160352081cac1a6c3dbbfa35fe390
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The fragment pointer is good to print here.
Change-Id: If09e684bbe0f9f5d3f961c4af8f86fb3ff364671
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If for some reason the neighbor solicitation cannot be sent,
then we need to unref the pending packet as that would never
be freed otherwise.
Change-Id: Ied12a9b3a027a2aad3fafd0b5b36c56b5f042c28
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we are trying to send a IPv6 network packet and there
is no data, then print error.
Change-Id: Ia0e7bebae513f1bcf984e189566ae7e10a90bbfc
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
When going through fragments in update_ll_reserve(), use
net_nbuf_unref() in order to get better memory allocation
debug prints.
Change-Id: Ia7ee15b32800acdebeb28125465515132c55e179
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
No need to receive anything if network interface is down.
Change-Id: I22d62aeaf4fbef54608818bf6c8073ec2e9a7c09
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Make sure that both the TX and RX threads are running before
we turn the network interface up.
Change-Id: Ie7e5938403e90e1ae76047c221cf9fd2f5a1167b
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The current state of affairs will not allow to set a MAC address and
thus will initialize a random MAC address.
However, when we are in a situation where we are trying to interface
Zephyr OS QEMU VMs with physical targets for testing, this is not
practical. We need to create a macvtap interface, on which we need to
set the MAC address of the Zephyr OS SLIP device so packets can be
routed. There is no way to guess which random address is going to be
used, and there is no practical way to query it.
So this patch adds a CONFIG_SLIP_MAC_ADDR config
setting. slip_iface_init() is changed to take the code from
slip_get_mac() for initializing the MAC address (which should be
happening there anyway) and if CONFIG_SLIP_MAC_ADDR is set, it takes
that one -- otherwise goes to generate a random one.
Change-Id: I55a68f2743604f310d41c934783eca82084d7e94
Signed-off-by: Inaky Perez-Gonzalez <inaky.perez-gonzalez@intel.com>
Use the proper MAC address space (00-00-5E-00-53-xx) dedicated
for documentation and specified in RFC 7042 ch 2.1.2.
Change-Id: If8ef9e4ee4e041ad005060664ebafe60df0a6bf9
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Test net_mgmt_event_wait() and net_mgmt_event_wait_on_iface()
Change-Id: I82ba70a697f0d13aa2e1eef2110e139b3621c119
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
In case of callback based event listener, it is easy for the callback to
filter on the given interface. But in case of the synchronous call it's
not: it would need, after a failed comparison on the interface pointer
to loop by itself on the net_mgmt_event_wait() which is a little bit
heavy (reinstalling the event listener, with the semaphore and all) and
a bit of a burden for the caller itself.
Instead, net_mgmt provides a dedicated call
net_mgmt_event_wait_on_iface() which does it the right way, so the
callback and the related semaphore are destroyed if only the iface
matches the one given as parameter (besides the timeout obviously).
Change-Id: Iab05c3249586f4f4d0447eea42fdac72b8428f2e
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
Instead of creating a handler and a related callback structure on an
event_mask: net_mgmt_event_wait() can be used to wait synchronously on
such event_mask. The core mgmt part will seamlessly reuse the struct
net_mgmt_event_callback so the whole internal notification mechanism is
using the same code.
Change-Id: I426d782c770e75e5222aa3c5b703172b1f1f2e5e
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
It's not enough to know just network address and port number to
disambiguate an endpoint, protocol (TCP vs UDP, etc.) should be
known too. Without this patch, there is a confusing output from
e.g. echo_server if both TCP and UDP is enabled.
Jira: ZEP-1086
Change-Id: I247a2dfc87df634ceac3b8733d7349b1d6aca80c
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
This function is useful for logging in other parts of network stack,
so allow to reuse it. (Will be used for net_context logging in a
follow-up patch).
Change-Id: I8238874584f8f039c94e5d827265944d97d13f6b
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
NET_IF_UP may change during the lifetime of k_pool which means we would
have to reconfigure everytime the flag changes but NET_IF_UP is already
checked during net_if_send_data thus it should never reach the queue in
the first place making this check unnecessary.
Jira: ZEP-1888
Change-Id: Iaa8471bee886a6f7e701a1dd243fb199def26589
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Unref the buffer if net_contect_send_to() fails to send.
Change-Id: I182125e64aabd171cd86f42fd9e1de036dd68544
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Current finalize api's takes buf as input parameter and returns
the finalized buf. But if there are any issue while finalizing,
it failed to throw an error.
Change-Id: I6db54b7453eec41a8051fab50d5c0dc937debd54
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
Anas Nashif