Before, we used tho have a BTs object split into 2 parts, a C
gprs_rlcmac_bts struct and a C++ BTS struct, and "bts_data" naming was
used to distinguish them in variable names. Nowadays the struct is
finally combined into one, so there's no point in using this "bts_data"
terminology, we use always "bts".
Change-Id: I9852bf439292d1abc70711bea65698b21bde0ee8
Previous work on BTS class started to get stuff out of the C++ struct
into a C struct (BTS -> struct gprs_glcmac_bts) so that some parts of
it were accessible from C code. Doing so, however, ended up being messy
too, since all code needs to be switching from one object to another,
which actually refer to the same logical component.
Let's instead rejoin the structures and make sure the struct is
accessible and usable from both C and C++ code by rewriting all methods
to be C compatible and converting 3 allocated suboject as pointers.
This way BTS can internally still use those C++ objects while providing
a clean APi to both C and C++ code.
Change-Id: I7d12c896c5ded659ca9d3bff4cf3a3fc857db9dd
Currently the BTS object (and gprs_rlcmac_bts struct) are used to hold
both PCU global fields and BTS specific fields, all mangled together.
The BTS is even accessed in lots of places by means of a singleton.
This patch introduces a new struct gprs_pcu object aimed at holding all
global state, and several fields are already moved from BTS to it. The
new object can be accessed as global variable "the_pcu", reusing and
including an already exisitng "the_pcu" global variable only used for
bssgp related purposes so far.
This is only a first step towards having a complete split global pcu and
BTS, some fields are still kept in BTS and will be moved over follow-up
smaller patches in the future (since this patch is already quite big).
So far, the code still only supports one BTS, which can be accessed
using the_pcu->bts. In the future that field will be replaced with a
list, and the BTS singletons will be removed.
The cur_fn output changes in TbfTest are actually a side effect fix,
since the singleton main_bts() now points internally to the_pcu->bts,
hence the same we allocate and assign in the test. Beforehand, "the_bts"
was allocated in the stack while main_bts() still returned an unrelated
singleton BTS object instance.
Related: OS#4935
Change-Id: I88e3c6471b80245ce3798223f1a61190f14aa840
When allocating multiple slots for a UE the following example
is not allowed 'UU----UU' for a UE class 12.
The time slot number can not roll over 7 and move to 0.
44.060 or 45.002 only specifies contigous however it was unclear
it this is an allowed pattern.
Only the example 45.002 B.3 in release 12 cleared this up.
It gives an example for a multi slot class 5 UE which has 7 possible
configuration this means the rolled over is not allowed.
Multislot class type 2 UE doesn't have this limitation.
Further if a UE supports 8 time slots this is not a limitation because
the window size (45.002 B.1) can include all time slots.
Releated: SYS#5073
Change-Id: I16019bdbe741b37b83b62749b840a3b7f4ddc6c7
As we integrate osmo-pcu more and more with libosmocore features, it
becomes really hard to use them since libosmocore relies heavily on C
specific compilation features, which are not available in old C++
compilers (such as designated initializers for complex types in FSMs).
GprsMs is right now a quite simple object since initial design of
osmo-pcu made it optional and most of the logic was placed and stored
duplicated in TBF objects. However, that's changing as we introduce more
features, with the GprsMS class getting more weight. Hence, let's move
it now to be a C struct in order to be able to easily use libosmocore
features there, such as FSMs.
Some helper classes which GprsMs uses are also mostly move to C since
they are mostly structs with methods, so there's no point in having
duplicated APIs for C++ and C for such simple cases.
For some more complex classes, like (ul_,dl_)tbf, C API bindings are
added where needed so that GprsMs can use functionalitites from that
class. Most of those APIs can be kept afterwards and drop the C++ ones
since they provide no benefit in general.
Change-Id: I0b50e3367aaad9dcada76da97b438e452c8b230c
In those cases since a string pointer was passed, it always printed
"single" instead of whatever really was being used, since the string
pointer was not NULL.
Change-Id: Idab7d18e8f519e10fc3df4007634661c46f9256d
In general we want to see explicitly the kind of requested allocation at
the start. The MS class is not needed since it's printed in the previous
log line in any case.
Change-Id: I9eb0a592c15be96da9d140ff373c1afead76b18c
Fixes consistent crash under some specific scenarios explained in
OS#4756.
The crash was caused due to a bug in channel allocator algorithm
incorrectly populating tbf->pdch[] array as a result of mismatching
first_ts and resulting pdch selected slot bitmask.
The issue happens because when allocating a UL TBF in allocator B, the
subset is always further forced into allocating one single TS. As a
result, on that branch several variables are updated, but first_ts was
not.
The field used to be updated in older versions, but a bug was introduced
during code refactoring in commit listed below (31 Jan 2018).
Fixes: 0cc7212cfd
Related: OS#4756
Change-Id: I79596803f7dab6f21b58bfe39c2af65d9c5b39d5
It's super annoying seeing lots of functions being called everywhere
only to find out they are only incrementing a counter. Let's drop all
those functions and increment the counter so people looking at code
doesn't see dozens of code paths evyerwhere.
Most of the commit was generated by following sh snippet:
"""
#!/bin/bash
grep -r -l ^CREATE_COUNT_INLINE . | xargs cat | grep "^CREATE_COUNT_INLINE("| tr -d ",;" | tr "()" " " | awk '{ print $2 " " $3 }' >/tmp/hello
while read -r func_name ctr_name
do
#echo "$func_name -> $ctr_name"
files="$(grep -r -l "${func_name}()" .)"
for f in $files; do
echo "$f: $func_name -> $ctr_name";
sed -i "s#${func_name}()#do_rate_ctr_inc(${ctr_name})#g" $f
done;
done < /tmp/hello
grep -r -l "void do_rate_ctr_inc" | xargs sed -i "/void do_rate_ctr_inc(CTR/d"
grep -r -l "CREATE_COUNT_INLINE" | xargs sed -i "/^CREATE_COUNT_INLINE/d"
"""
Change-Id: I360e322a30edf639aefb3c0f0e4354d98c9035a3
Move timeslot applicability check outside of nested for loop into
separate functions and document them. Add corresponding tests.
This allows us to clarify types used in TS-related computations.
Change-Id: Ic39e848da47dc11357782362fdf6206d2c1457c2
Related: OS#2282
Don't use PDCH from free TFI lookup routine. This allows for simpler
function which can be moved to mslot_class.c alongside with other
similar helpers.
Change-Id: Ie154866900453d232a890f7b9a30911b451525a1
Related: OS#2282
The PDCH class and corresponding functions are rather self-contained and
independent from BTS implementation. Let's move them into separate file
to make bts.cpp more manageable. As additional benefit it allow us to
somewhat untangle all the different cross-dependent includes.
Change-Id: Ie05e25361e6741a81b024679f9675c98d4923683
Related: OS#1539
* move USF allocation into separate function
* document USF allocation
This allows to clearly see where selected UL TS is forced into single TS
in algorithm B allocator.
Change-Id: I563dc10827ce68295553f88f3bf2e1fc0ba595c1
Related: OS#2282
* generalize TS allocation and move it into separate function
* move single-slot allocation into separate function
* use common functions for TS allocation on both UL and DL
Change-Id: Ied45ae380c345bc76fe9d6fd9a6184d1109f83f2
Related: OS#2282
Move computation of RX mask into separate function and document it. This
allows to significantly shrink find_multi_slot() function and overall
improve code readability.
Since the test output requires cosmetic adjustment anyway due to change
in the sequence of log messages, use this opportunity to better group
and format log message.
Change-Id: I731726a096bba7ee97499e5cbe3e7401869d7392
Related: OS#2282
Many files include unnecessary headers and don't include headers which
are actually used. Because of that combined with the fact that OsmoPCU
is a mixture of C and C++, it makes it hard to modularize code. Fix
this (using iwyu [1] tool):
* add missing headers
* remove unused headers
[1] https://include-what-you-use.org/
Related: OS#1539
Change-Id: I8c9f488a43b099c72b2d30d3245e7ba50872fc00
Move TRX check inside local tfi_find_free() wrapper to make main
algorithm easier to follow.
Change-Id: I02da2b8ba8c9c8815dae0e39e1fed277ca0df171
Related: OS#2282
* document used parameters and return values
* use consistent formatting
* constify function parameters where appropriate (adjusting parameter
types if necessary)
Change-Id: I211b10b4da59c73d509b719346774515c761886a
Related: OS#2282
* define and use constant for occupied TFI instead copying the same
magic number all over the place
* use libosmocore's define for bit pretty-printer
Change-Id: I2699ceebf0cbec01652a02fa68ccc9e9419d0293
Related: OS#2282
Assign reserved_*_slots only when multislot masks are found to avoid
reassignment and make code easier to follow.
Change-Id: I9b0482f4ea75ead9855cd78e33c8e70d0ccf4484
Related: OS#2282
* drop unused parameters (from both functions and structs)
* document used parameters and return values
* tighten types used for parameters
* use consistent formatting
Tests are adjusted accordingly but test results are left untouched to
avoid regressions.
Change-Id: I39d81ab64ff790b9c4c2d0312a574485cd83e755
Related: OS#2282
The table B.1 is copy-pasted from 3GPP TS 45.002 and reformatted via
Emacs macros into C struct to avoid typos. The test output expanded
accordingly.
The allocation test expectations and output are adjusted accordingly.
Note: classes 35-45 which need TA offset are not properly supported
yet. This can be extended once we have such devices available for tests.
Change-Id: I1ef2eb99c517f25e7d1e71b985a3e0eb3879eb2c
Related: OS#2282
To facilitate testing and addition of support for new multislot classes,
hide multislot class struct internals:
* introduce mslot_class_get_*() functions
* use those functions instead of direct access to array of structs
* use ms_class as a parameter to find_multi_slot() instead of entire
object
Change-Id: Id796bcff1322b1e273a0e3236c66c23b9da8fac6
Add header similar to the one used in the standard, reformat to
facilitate further extention.
Change-Id: I786df6b154c0668d2cefa0ea84d7dea336b0da1d
Related: OS#2282
These functions are currently defined in src/gprs_rlcmac_ts_alloc.cpp
but will be needed elsewhere.
Turn them into template functions to support different types and move
them to pcu_utils.h.
Sponsored-by: On-Waves ehf
Currently casts from gprs_rlcmac_tbf to gprs_rlcmac_ul_tbf and
gprs_rlcmac_dl_tbf are done by using static_cast. This doesn't provide
protection against converting a gprs_rlcmac_ul_tbf pointer to a
gprs_rlcmac_dl_tbf pointer and vice versa.
This commit provides two functions as_ul_tbf and as_dl_tbf, that
behave similar like dynamic_cast but use the direction field instead
of RTTI.
Sponsored-by: On-Waves ehf
Currently the values Bmax/R default MS are computed under the
assumption than min(4, N_PDCH) DL slots are allocated for an MS, even
if multislot assignment is not enabled.
This commit changes the computation to assume 1 DL slot if algorithm
A is selected or the dynamic algorithm is used and has disabled
multislot assigment due to high load.
Sponsored-by: On-Waves ehf
Currently there is no persistent state being used in
alloc_algorithm_dynamic. So algorithm B is even used in persistent
high usage scenarios. If there are many active TBFs, multislot
assigments are not fair, because MS of a "higher" multislot class get
higher troughputs. On the other hand, as long as all PDCH are busy no
bandwidth will be wasted even if all MS use algorithm A.
This commit modifies alloc_algorithm_dynamic to disable algorithm B
when that call fails. It then keeps it disabled until there is a
single PDCH which is idle (it is considered idle, if there is at most
one active DL TBF assigned to it).
Sponsored-by: On-Waves ehf
Currently algorithm A can select an TBF even when there is no free
TBF in the reverse direction. While this does not necessarily lead to
an allocation failure, the probabily is higher. In addition, the
current slot reservations are not taken into account.
This commit changes the selection algorithm to prefer slots where TFI
are available in both directions and which are less reserved.
Sponsored-by: On-Waves ehf
Currently each tx_window combination is checked only once by using a
set containing the sets of TX slots that have been checked already.
This approach does not ensure, that num_tx and ul_ts really match the
tx_window being tested. This does not make a difference with the
current test cases probably because num_tx starts with 1 and is
increased each iteration. Since the bitmap optimization is equivalent
to a cache optimization strategy that only uses tx_window as key. On
the other hand, ul_ts, num_tx, and rx_mask cannot be derived from
tx_window, but these values are also refered to after the call to
test_and_set_bit(). This makes it difficult to prove that correctness
of the caching. While this will not lead to a defect, the results
might be less optimal.
This commit changes the optimization strategy to skip all tx_window
where ul_ts and ul_ts+num_tx-1 are not both contained. This provides
a similar degree of optimization like the set approach (only the
iteration with num_ts == 8 is not optimized, which only applies to to
ms class 18 and 29 MS) but ensures that the values of the related
variables have a clear relationship.
Note that the bitset based optimization for rx_window does not suffer
from a possible cache inconsistency, since only tx_window and
rx_window (tx_slot_count and rx_slot_count can be derived from the
windows and thus are covered by the cache key) are used after the call
to test_and_set_bit(). tx_window is constant over the whole lifetime
of the cache.
Sponsored-by: On-Waves ehf
The idea behind this meta algorithm is to automatically select one of
the other algorithms based on the system state. Basically algorithm B
will be selected if the PDCH usage is low to improve throughput and
latency. Algorithm A will be selected to support more concurrent MS.
This commit adds a first simple state-less version of this algorithm
that always tries B first and only if that fails A is tried
afterwards.
The following VTY command is added to the 'pcu' node:
- alloc-algorithm dynamic
Ticket: #1934
Sponsored-by: On-Waves ehf
This part of algorithm_b has already been disabled. Further work may
depend on this, but it is going out of sync. So this commit removes
it completely.
Sponsored-by: On-Waves ehf
Currently these algorithms modify other objects (MS, TBF, PDCH) even
if the allocation will fail later on. To implement an algorithm that
dynamically tries another algorithm on failure (e.g. A after B), the
first (failing) algorithm should not change or damage anything.
This commit refactors algorithm A and B to delay the actual allocation
until it is known that the allocation will not fail.
Ticket: #1934
Sponsored-by: On-Waves ehf
Currently this code path is only used, if an allocation has been
taken place in a former call to an allocation algorithm function.
If this was for an DL TBF, the first common TS was selected,
otherwise the least used common TS was selected for an UL TBF.
The shrinking of the UL set (to 1<<first_common_ts) is done in the
latter case.
This commit removes an additional code path that aligns the UL set to
first_common_ts, because it has no more influence on the set of
common TS after both UL and DL TBF have been allocated.
Sponsored-by: On-Waves ehf
Currently the TBFs are registered in a TFI indexed array within the TRX
objects. TBFs can be searched globally by TFI and TRX number. This
conflicts with the use of the same TFI for different TBF on different
PDCH. This use case requires the specification of the PDCH as
additional search dimension.
This commit moves the TFI index TBF arrays into the PDCH objects. The
related methods are updated accordingly.
Ticket: #1793
Sponsored-by: On-Waves ehf
Currently the TFI are managed per TRX, thus only a maximum of 32 TBF
per direction and per TRX are possible simultaneously.
This commit modifies algorithm_a() to allow the sharing of TFI
between different PDCH. Since algorithm A only assigns a single slot
to each TBF, the TFI of each PDCH can be assigned independently.
This increases the maximum to 32 TBF per direction and per PDCH
concerning the TFI allocation.
Ticket: #1793
Sponsored-by: On-Waves ehf
Currently the TFI and the TRX have to be determined before the actual TBF
allocation function is called, passing TFI and TRX number as
parameters. This does fit to TFI reuse for different slots, since
this were tightly coupled with the slot selection.
This commit just moves the TFI selection into the alloc_algorithm
functions. The tfi parameter is removed from the the TFI alloc
functions. The trx parameter is changed into use_trx to optionally
limit the trx selection (same semantics like in tfi_find_free).
Sponsored-by: On-Waves ehf
Currently the handling of MS_B and MS_C is not compliant with TS
45.002, annex B.1. These values may only interpreted as 0, if
frequency hopping is not enabled and if there is no change from Rx to
Tx or vice-versa.
This commit sets Ttb/Trb to 1 if the table entry is MS_B/MS_C, since
only combined down/up access modes are supported.
Sponsored-by: On-Waves ehf
Pau Espin
Alexander Couzens
Vadim Yanitskiy