Add net_nbuf_split() function that can be used to split
an existing net_buf fragment into two arbitrary length pieces,
and return the two new fragments to the caller. The data from
the original fragment is copied into these two new fragments.
Existing fragment is not modified.
Change-Id: I463e675232c6e19c2a42929f480893a6d1265873
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Following changes are made:
1. Output using NET_INFO logging level. This function is not part of
automatic logging, and must be called explicitly by a user application.
But if it outputs using NET_DEBUG, then DEBUG level needs to be
enabled, and the output of this function will be drowned in logging
spam. So, let user to enable just INFO level.
2. Show entire structure of the fragment chain, *including* the head
net_buf which holds net_nbuf structure. It is numbered as -1 in the
output to preserve existing order (and not change existing size
calculations).
3. Show owning pool (and its properties, like buffer size/user_data
size) for each buffer.
4. Check for NULL pointer, e.g. for convenience of calling directly
from net_context receive callback (which will be called with NULL
on TCP peer closed connection event).
With these changes, a newcomer from one look at the output of this
function will be able to have a clear basic picture of network
buffer management in Zephyr, and recurring user will be able to
recall any details at once.
Change-Id: I8f9562748329d749f765cc6af7989a448256d7e0
Signed-off-by: Paul Sokolovsky <paul.sokolovsky@linaro.org>
It is possible that CONFIG_NET_CONTEXT_NBUF_POOL is set but
application has not defined any pools. In this case the tx and
data pool pointers will be NULL in net_context struct and we
must use the default pools instead.
Change-Id: I286f34c87d9182aace71e0a61f038945810e4916
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Print information who freed the net_buf last time if this
can be figured out. This info is not fully accurate if there
is lot of allocations going on but could give some indicatation
who was using the net_buf previously.
Change-Id: I3d4463c54a9b944847a532d951fb6d2303fc06cf
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If net_nbuf_read() or net_nbuf_write() offset adjustment fails,
print the offset value in order to make debugging easier.
Change-Id: I899998bdcbc91a25ed7d71a599a5052a6fa4ee36
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Show only what would be freed when debugging the allocations.
Let the net_buf_unref() to delete the individual fragments.
Change-Id: I1a39a553f12d73b78c4ba08be0e8e8252b666c8e
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Create net_nbuf_frag_del() so that we can track net_buf frees.
If CONFIG_NET_DEBUG_NET_BUF is not defined, then this new function
will call net_buf_frag_del() directly, and if not, then it will
track memory usage.
Change-Id: I5f382436cebc71fdaf12baf7bf964fb63bee7aca
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
User can track memory allocations and deallocations.
This is enabled by CONFIG_NET_DEBUG_NET_BUF option.
Change-Id: I9d83b9b63fb2b141c9a283887f1770bb4232f61c
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
No need to set the buf to NULL as it will get its value
from net_buf_alloc() call anyway.
Change-Id: Ib3a32b6e4cb4b446761b6a97a88c20fc52d3d683
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Instead of always calling net_nbuf_get_reserve_tx() with zero
ll_reserve (first parameter) and then setting the link layer
reserve separately, pass the reserve to that function which can
then set the ll_reserve in buf itself.
Change-Id: I21c14cb7e2e6c36d170c09998bca0207ecf65c75
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The net_buf pool address is not very useful so print the
user friendly name of the pool instead.
Change-Id: I1c64621c816832d2459819490ec5609f5c39f9f1
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
If we receive lot of packets, it might happen that we exhaust
all the DATA buffers in the system. This would prevent from
us sending anything to the network.
Change this by splitting the DATA buffer pool into RX and TX
parts. This way RX flooding cannot consume all DATA buffers
that needs to be sent.
Change-Id: I8e8934c6d5fdd47b579ffa6268721b5eb3d64b6d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
It makes no sense to modify the original net_buf so change the
copy function not to touch the original buffer.
Change-Id: I5d22445ce50cee62994c36567f0e995a500cb89d
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The first parameter of net_nbuf_copy() must be the head
of the buffer chain i.e., it must contain the user_data
part. If a fragment is given, then we do not know enough
information to allocate the data fragments.
Change-Id: I052b183d8c63d7326b320254f36f00b2fc48b0a0
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Print information about various network buffer pools in the
system. This is useful in debugging the buffer allocations.
Change-Id: I31123c6f1f6647f77503f32e268c174330762128
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
User can configure its own pools for data that needs to
be transmitted out (TX). This helps to avoid deadlocking
the system if user space application uses all the buffers
in the system, and the core IP stack tries to get buffer
that needs to be sent out.
By default the net_buf pool support in net_context is turned
off as application developer needs to create the pools and
tie them to desired contexts.
Change-Id: Ida4a1771d34d6c250974e56fba4f0e0b2592cb29
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
As the pool contains now more accounting information, there is
no need to keep track of this data in nbuf.c.
Change-Id: I7fd44f9feda8fd9816356556d1e788ad56e6eedd
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
As the function does not change the original data, make
the corresponding parameter const.
Change-Id: I1125a2f9205dc73de2f0aac0c30110591baace1e
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Current API description of net_nbuf_compact() is not very clear.
The first parameter needs to be the first net_buf in the chain.
The changes to this API are needed in order to clarify following
use cases:
1) User provides fragment that is not first of the chain and compact is
successfully done. In this case there is no free space in fragment list
after the input fragment. But there might be empty space in previous
fragments. So fragment chain is not completely compacted.
2) What if input fragment has been deleted and api returns the same
buf?
So this commit simplifies the API behavior. Now net_nbuf_compact()
expects the first parameter to be either TX or RX net_buf and then it
compacts it. It fails only if the input fragment is a data fragment.
Change-Id: I9e02dfcb6f3f2e2998826522a25ec207850a8056
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
The net_nbuf_push() API is not used by anyone. Semantics are not
clear and following patch requires changes to push api, so removing
this API for now. If needed this can be re-introduced later.
Change-Id: I1d669c861590aa9bc80cc1ccb08144bd6020dac5
Signed-off-by: Ravi kumar Veeramally <ravikumar.veeramally@linux.intel.com>
If we are in the end of the fragment chain, then we can just
bail out as there is nothing more to do. There will be a
double free if we continue as the last entry is already
removed at this point.
Change-Id: I0f9782b408244d283dc7e3e087359dd00bada7a9
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
This commit changes the net_buf getter functions in nbuf.h
by adding a timeout parameter. These function prototypes
are changed to accept a timeout parameter.
net_nbuf_get_rx()
net_nbuf_get_tx()
net_nbuf_get_data()
net_nbuf_get_reserve_rx()
net_nbuf_get_reserve_tx()
net_nbuf_get_reserve_data()
net_nbuf_copy()
net_nbuf_copy_all()
net_nbuf_push()
net_nbuf_append()
net_nbuf_write()
net_nbuf_insert()
Following convinience functions have not been changed
net_nbuf_append_u8
net_nbuf_append_be16
net_nbuf_append_be32
net_nbuf_insert_u8
net_nbuf_insert_be16
net_nbuf_insert_be32
net_nbuf_write_u8
net_nbuf_write_be16
net_nbuf_write_be32
so they call the base function using K_FOREVER. Use the
base function if you want to have a timeout when net_buf
is allocated.
Change-Id: I20bb602ffb73069e5a02668fce60575141586c0f
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The net_nbuf_get*() functions were calling net_buf_alloc()
with K_FOREVER. This can cause issue if called from isr
context. The fix is to check if we are in isr and then try
to alloc net_buf with K_NO_WAIT.
Change-Id: I809170f2cd059480d436763e19a35386d9bdf048
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
There is dead code that is never executed so removing it.
Coverity-CID: 157585
Change-Id: I6926289b5735b78fcb99ad493d1b05198b9e36cc
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Wrong pool pointer was used when printing pool information during
the unref in nbuf.c. During nbuf init, wrong pool pointer was used
when printing memory pool sizes.
Change-Id: I9ed08cf7afa3c841c97ea981b97fff37aa40a984
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
In net_nbuf_get(), check that context pointer value is not
null before accessing data via it.
Coverity-CID: 157600
Change-Id: I7e7ea19a85f6fbef129e9ce699ea740d3be84cb8
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
The context parameter might be NULL so we need to check
its value before accessing its content.
Coverity-CID: 157571
Change-Id: I7f75323d9d261a77421688f37a40bb44ff3ca2bd
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Replace the existing Apache 2.0 boilerplate header with an SPDX tag
throughout the zephyr code tree. This patch was generated via a
script run over the master branch.
Also updated doc/porting/application.rst that had a dependency on
line numbers in a literal include.
Manually updated subsys/logging/sys_log.c that had a malformed
header in the original file. Also cleanup several cases that already
had a SPDX tag and we either got a duplicate or missed updating.
Jira: ZEP-1457
Change-Id: I6131a1d4ee0e58f5b938300c2d2fc77d2e69572c
Signed-off-by: David B. Kinder <david.b.kinder@intel.com>
Signed-off-by: Kumar Gala <kumar.gala@linaro.org>
This commit adds one routine previously found at the DNS resolver
library. The net_nbuf_linear_copy routine allows to "linearize"
an IP stack network buffer. This routine is required by functions
that must jump between big chunks of data that in this case may
lie between many fragments. Tracking fragments may be a tedious
task, so getting a linear copy of the buffer will reduce code
complexity altough it increases memory consumption.
The DNS client library is updated to reflect these changes.
Change-Id: Iae321f99fa9b05fae7e722b6d41baac427d82d7e
Signed-off-by: Flavio Santes <flavio.santes@intel.com>
Let's make net stack having its own level of debugging through sys_log.
It replaces NET_DEBUG by NET_LOG_ENABLED, which is then semantically
better: someone wanting to log the errors might want that not only for
debugging.
Along with it, CONFIG_NET_LOG_GLOBAL option is added, in order to enable
all available logging in network stack. It is disabled by default but
might be found useful when warning/errors need to be logged, so it is
then unnecessary to selectively enable by hand all CONFIG_NET_DEBUG_*
options.
It is possible, locally, to override CONFIG_SYS_LOG_NET_LEVEL by setting
the level one want to NET_SYS_LOG_LEVEL. This can be useful on samples
or tests.
Change-Id: I56a8f052340bc3a932229963cc69b39912093b88
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
It's possible to know from which pool the nbuf comes from so
let's reduce nbuf size by removing its internal type attribute which
becomes then useless.
When a data buffer is not coming from nbuf data buffer pool, let's call
it "EXTERNAL", just to make debugging handling a little simple in the
code.
Change-Id: I6931394c8c4f594137f6380be0a0ba5cea371040
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
After commit 71c7c01819 about buf pool, we
started to get spurious behaviors on various places of the code on
different boards (a101, frdm...) BUT on qemu. Basically, outgoing ip/udp
packets were full of garbage. Or sometimes it was the abilitty to parse
incoming packet that was happening.
The difference between qemu and actualy boards is - afaik, at least, let
me know if I am wrong - that qemu provide initialized memory (full of
0s). Following this asssumption, this patch just reset the nbuf right
after it got allocated. And all started to work again as thought.
It's in fact a good thing to reset nbuf memory. Even before the above
commit: after being used more than once, a buffer would have ended up
with old content, and this could have been generating a bug. So let's be
on the safe side and always intialize nbuf content.
Change-Id: I50647d9e9b82a4ed340a5ceb0d69409b0194dddd
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
In order to keep the initialization process light-weight, remove
net_buf_pool_init() and instead perform the initialization of the pool
and buffers in a "lazy" manner. This means storing more information
in the pool, and removing any 'const' members from net_buf. Since
there are no more const members in net_buf the buffer array can be
declared with __noinit, which further reduces initialization overhead.
Change-Id: Ia126af101c2727c130651b697dcba99d159a1c76
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
This is only for use with custom destroy callbacks, so that the
application gets isolated away from the details of how exactly the
buffers are managed. This opens up the possibility of switching away
from k_fifo to potentially better solutions, such as k_lifo.
Change-Id: I0d8322fdec3500d8ae060ae471b9448aeaa4572a
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Until now it has been necessary to separately define a k_fifo and
an array of buffers when creating net_buf pools. This has been a bit
of an inconvenience as well as blurred the line of what exactly
constitutes the "pool".
This patch removes the NET_BUF_POOL() macro and replaces it with a
NET_BUF_POOL_DEFINE() macro that internally expands into the buffer
array and new net_buf_pool struct with a given name:
NET_BUF_POOL_DEFINE(pool_name, ...);
Having a dedicated context struct for the pool has the added benefit
that we can start moving there net_buf members that have the same
value for all buffers from the same pool. The first such member that
gets moved is the destroy callback, thus shrinking net_buf by four
bytes. Another potential candidate is the user_data_size, however
right not that's left out since it would just leave 2 bytes of padding
in net_buf (i.e. not influence its size). Another common value is
buf->size, however that one is also used by net_buf_simple and can
therefore not be moved.
This patch also splits getting buffers from a FIFO and allocating a
new buffer from a pool into two separate APIs: net_buf_get and
net_buf_alloc, thus simplifying the APIs and their usage. There is no
separate 'reserve_head' parameter anymore when allocating, rather the
user is expected to call net_buf_reserve() afterwards if something
else than 0 headroom is desired.
Change-Id: Id91b1e5c2be2deb1274dde47f5edebfe29af383a
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
Fix compilation issues that show up if SYS_LOG is mapped to printk
instead of printf. Unlike printf, printk is annotated so that the
compiler catches incorrect format specifiers passed to it.
Change-Id: Iab7cc6da110e9c98720211a6f773dcf055a3a411
Signed-off-by: Johan Hedberg <johan.hedberg@intel.com>
* Moved networking code into subsys/net.
* Renamed net/yaip to net/ip at the same time.
* Fixed the tests/net to compile
* Fixed the Makefiles and Kconfig files in subsys/net
to use the new location of the IP stack
Change-Id: Ie45d9e8cb45a93fefdf969b20a81e3b1d3c16355
Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
