1. 13 Feb, 2024 40 commits
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/ors: Fix ssb_pos_bitmap default in schema · 9979fa63
      Kirill Smelkov authored
      Noticed accidentaly: ssb_pos_bitmap is string, not integer - the default should be of the same type.
      9979fa63
    • Kirill Smelkov's avatar
      software/ors-amarisoft: ue: Generalize UEsim · 30179be5
      Kirill Smelkov authored
      Rework UEsim to be able to work with multiple cells, multiple radio units(*),
      multiple UE all at the same time. RU, CELLs and UEs are now configured,
      simiularly to eNB, via shared instances. Add tests.
      
      Contrary to ORS don't care about backward compatibility here because currently
      we have just a few UEsim deployments and migrating them should be easy.
      
      Please see added schemas, tests and updates slapos-render-config on how to use
      the new system.
      
      (*) contrary to eNB UEsim does not allow to use one RU for two cells,
      30179be5
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Try to support NR/FDD a bit · 4abb4101
      Kirill Smelkov authored
      It should be already generally supported, but many parameters needs in NR case
      are currently hardcoded to their TDD values. I've adjusted only a few of them
      and stopped, becuase there is currently no practical case at hand for me to
      test it. Still I think it makes sense to save this first step.
      
      For ORS, who uses TDD, rendered config stays practically the same:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff -C -C config/{old,out}/ors/
      ```
      
      ```diff
      diff --git a/config/old/ors/gnb/enb.cfg b/config/out/ors/gnb/enb.cfg
      index ead7f0160..9a260d73f 100644
      --- a/config/old/ors/gnb/enb.cfg
      +++ b/config/out/ors/gnb/enb.cfg
      @@ -97,6 +97,7 @@
      
               prach: {
      +          ra_response_window: 20,
               },
      
               pdcch: {
      @@ -180,7 +181,6 @@
             preamble_received_target_power: -110,
             preamble_trans_max: 7,
             power_ramping_step: 4,
      -      ra_response_window: 20,
             restricted_set_config: "unrestricted_set",
             ra_contention_resolution_timer: 64,
             ssb_per_prach_occasion: 1,
      ```
      4abb4101
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Intra-ENB handover · 5c300a7d
      Kirill Smelkov authored
      If a base station has multiple cells, it should be possible for UE to be handed
      over from one cell into another, for example when UE crosses border of sectors.
      
      So far we had only Inter-ENB HO and now we are also adding configuration for
      HO in between all own cells.
      
      Add tests for everything.
      
      NOTE: we use allowed_meas_bandwidth, antenna_port_1 in dst=LTE case because
      otherwise, e.g. for NR->LTE HO if they are not present, lteenb complains:
      
          enb.cfg:260: expecting 'allowed_meas_bandwidth' field
          enb.cfg:260: expecting 'antenna_port_1' field
      
      for both Intra-ENB and Inter-ENB handovers.
      
      For Intra-ENB case we can compute those numbers from cell definition. For
      Inter-ENB case, when we don't know target context, we use conservative values.
      5c300a7d
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Full Carrier Aggregation · 881b5104
      Kirill Smelkov authored
      We already had CA for LTE+LTE case. Let's also setup it for NR+NR and LTE+NR
      cases as well. Add tests for everything.
      881b5104
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Take frequencies control into our hands · 37b1b28d
      Kirill Smelkov authored
      Currently when setting up cells we allow users to input
      
      - dl_earfcn for LTE, and
      - dl_nr_arfcn and nr_band for NR
      
      and from that lteenb automatically computes
      
      - ul_earfcn for LTE, and
      - ul_nr_arfcn and ssb_nr_arfcn for NR
      
      everything kind of works out of the box for eNB case when there is simple SDR Radio Unit.
      
      Then there are also the following cases:
      
      1. we also need to set UL frequency when configuring Lopcomm RU, or any other ORAN-based Radio Unit.
      2. we also need to specify DL/UL frequencies in MHz when configuring Lopcomm RU.
      3. when configuring NR peercell it is required to set ssb_nr_arfcn. It is also
         required to set both dl_nr_arfcn and ul_nr_arfcn, and so far we were setting
         only DL one and using it for both assuming TDD band.
      4. when configuring UEsim we need to specify both dl_nr_arfcn and ssb_nr_arfcn for NR cells.
      
      So the problem is that even though lteenb automatically computes UL and SSB
      frequencies, there is no way for us to reuse results of that computation for
      scenarios outside of "simple enb" case.
      
      As the result we kind of workaround that currently with exposing additional
      parameters and asking users to look into enb, probably its `cell phy` output,
      with the following:
      
           ---- 8< ---- (from UEsim)
           "ssb_nr_arfcn": {
             "title": "SSB NR ARFCN",
             "description": "SSB NR ARFCN, you can retrieve from ENB/GNB side",
      
      or use additional parameters just for ul_earfcn and frequency-in-MHz:
      
           ---- 8< ---- (from ru/lopcomm)
           "txa0cc00_center_frequency": {
             "title": "Lopcomm ORAN DL Center Frequency in MHz (TXA0CC00)",
             "description": "Lopcomm ORAN Center Frequency in MHz (TXA0CC00)",
             "type": "number",
             "default": 2140
           },
           "rxa0cc00_center_frequency_earfcn": {
             "title": "Lopcomm ORAN UL Center Frequency EARFCN (RXA0CC00)",
             "description": "Lopcomm ORAN Center Frequency EARFCN (RXA0CC00)",
             "type": "number",
             "default": 18300
           },
           "rxa0cc00_center_frequency": {
             "title": "Lopcomm ORAN UL Center Frequency in MHz (RXA0CC00)",
             "description": "Lopcomm ORAN Center Frequency in MHz (RXA0CC00)",
             "type": "number",
             "default": 1950
           },
      
      which prevents automation, opens the door for inconsistencies and puts the load
      to resolve all that on users.
      
      The root cause of the problem is that there is no way to access at
      instantiation time what lteenb would compute internally. But DL<->UL conversion
      and DL->SSB conversion is not a difficult task and we can do that on our own.
      
      -> Do that here and solve all the problems listed above in one go.
      
      For frequency computations we use my xlte.earfcn and xlte.nrarfcn modules and
      upstream nrarfcn egg. For this xlte is updated(*) to primarily pick up
      
      kirr/xlte@6cb9d37f
      kirr/xlte@b8065120
      
      and the rest of the conversion is in slaplte to use corresponding dl2ul and
      dl2ssb routines.
      
      For consistency ul_earfcn and ul_nr_arfcn now become input parameters for LTE
      and NR cells, but optional. If they are absent - they are computed with
      defaults, but a user can now control them explicitly. The same applies for
      ssb_nr_arfcn.
      
      This patch does not convert UEsim yet, because UEsim does not use rulib yet and
      will be handled all in one go in a follow-up step.
      
      The computation routines are thoroughly tested. First they have unit tests
      inside XLTE itself, then we also update our tests in generic test/test.py here
      with explicitly checking that correct numbers are emitted for UL and SSB
      frequencies, and third I've also verified SSB computation results with respect
      to https://tech-academy.amarisoft.com/OutOfBox_UEsim_SA.html#Tips_SSB_Frequency .
      
      All this creates a base to be sure that the computations are correct and we are
      indeed safe to switch our frequencies computation modules.
      
      (*) full upgrade brings kirr/xlte@e716ab51...8e606c64 ;
          nrarfcn egg: https://pypi.org/project/nrarfcn/
      37b1b28d
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/ors: test: Use valid frequencies · 7a22aaa0
      Kirill Smelkov authored
      So far in the unit tests we were only generating enb.cfg without semantically
      processing as running lteenb is not possible on testnodes due to licensing
      restrictions. That allowed to use arbitrary numbers, even invalid, for
      frequencies without hitting an error.
      
      However in the next patch we are going to start handling frequencies ourselves,
      and invalid frequencies will lead to instantiation failures like
      
          ... in dl2ul(dl_earfcn)
          KeyError: 'no band that corresponds to DL EARFCN=325320'
      
      Avoid that by switching to valid LTE and NR frequency numbers in the test.
      7a22aaa0
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Introduce txrx_active generic RU parameter · 45378fec
      Kirill Smelkov authored
      Radio Units like Lopcomm ORAN require explicit activation for the radio.
      Generalize this and introduce a generic parameter for explicit activation of
      the radio on any RU. Default it to _INACTIVE_ so that radio happens only when
      it is explicitly asked to be on for real. But preserve backward compatibility
      for ORS which used to do "radio on" by default. disable_sdr becomes
      ORS-specific parameter.
      
      Lopcomm-specific txa0cc00_active and rxa0cc00_active go away and are replaced
      by txrx_active usage.
      45378fec
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Make tx_gain/rx_gain to be generic RU parameters · a6772923
      Kirill Smelkov authored
      I.e. apply not only to SDR but to all Radio Units and expecting that they apply
      to them with the same uniform semantic of controlling Tx/Rx amplifiers on on
      the RU.
      
      - For SDR nothing is changed.
      - For Lopcomm txa0cc00_gain goes away and is replaced by tx_gain. For rx_gain
        there is a question as I could not find offhand how to control aplifier on RX
        circuit there.
      - For Sunwave tx_gain/rx_gain, as with lots of more(*) is TODO.
      a6772923
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Interlink RU/CELLs by publishing info who uses who · 4eb6a873
      Kirill Smelkov authored
      For example by publishing on RU the list of CELLs that are using it.
      
      While at it and as we are introducing per-RU publication, fix a long-standing
      TODO in ru/lopcomm to publish RU-specific information in per-RU way instead of
      global.
      4eb6a873
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Publish information about a cell on the cell shared instance · e35d1d40
      Kirill Smelkov authored
      Instead of publishing it on enb with prefixing by cell reference. It is more
      logical to have cell input parameters on the cell instance and also have output
      parameters about that cell published on that same instance.
      
      For now it is all only what was input. But in the next patches we'll extend set
      of published information with also computed things, so it won't remain trivial.
      e35d1d40
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Allow SDR RU to consist of multiple SDR boards · 97ecce86
      Kirill Smelkov authored
      For example if we need to do 4T4R but have SDR boards with only 2T2R ports, we
      need to combine two such boards into single RF port. Allow to do that via
      changing SDR schema from taking sdr_dev into accepting sdr_dev_list and
      adjusting the code everywhere correspondingly.
      
      For ORS full backward compatibility is preserved via adjusting "ORS mode" proxy
      correspondingly.
      
      One example in slapos-render-config.py is changed from 2T2R to be 4T2R on such
      RF port combined from two SDR boards. For that example the diff for rendered
      enb.cfg is as follows:
      
          --- a/config/old/enb/iRU1_SDR_tLTE2_tNR/enb.cfg
          +++ b/config/out/enb/iRU1_SDR_tLTE2_tNR/enb.cfg
          @@ -8,7 +8,7 @@
             rf_driver: {
                 // __CELL__a__ru 4T2R  (sdr)
                 name: "sdr",
          -      args: "dev0=/dev/sdr0",
          +      args: "dev0=/dev/sdr0,dev1=/dev/sdr1",
             },
             tx_gain: [51, 51, 51, 51],
             rx_gain: [52, 52],
      
      All other cases, including ORS, remain without any change.
      97ecce86
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Protect from buildout code injection · c83fa3cb
      Kirill Smelkov authored
      We currently have at least the following problem:
      
      1. a malfored shared instance reference leads to instantiation failure and
         potential buildout code injection. For example if reference of shared
         instance contains space then buildout fails with something like
      
             zc.buildout.configparser.ParsingError: File contains parsing errors: .../instance-enb.cfg
                 [line 45]: '[promise-testing partition 0.RU-sdr-busy]\n'
      
         since, when requesting, it is possible to use arbitrary characters for
         references, including newline, it also opens the door for code injection attacks.
      
      2. we currently use `json` directive of slapos.recipe.template e.g. as
      
           extra-context =
               json iru_dict   {{ rulib.iru_dict   | tojson }}
      
         this also potentially leads to instantiation failure if JSON(rulib.iru_dict)
         contains buildout control characters.
      
      Solve this problems.
      
      For 1 we develop "buildout encoding" that encodes arbitrary string into form
      that is safe to be used as a name of buildout section. This encoding never fails
      and does not loose information, which means it is safe to be applied
      unconditionanly and there is no chance for two string inputs to result in the
      same encoded form. The encoding also has practical property that it leaves most
      of the strings, that we use in our buildot profiles, intact. Which means that
      the result of the encoding is readable and the encoding can be applied almost
      unconditionally to all strings without hurting instantiation. Then this
      encoding is applied wherever reference of a shared instance is used to form
      e.g. name of buildout section or inside other codes. We apply this encoding
      universally - in config files too to avoid breakage there - e.g. in enb.cfg to
      protect from e.g. \n being inserted in the middle of comment resulting in YAML
      breakage. See documentation of added xbuildout.py for details.
      
      For 2 we switch from `json` to using `dumps` on an object and referring that
      dumps result via `key` directive.
      
      See the following links for preliminary discussions with Jérome and Vincent on
      this topic:
      
      kirr/slapos@d5d4f7c8
      nexedi/slapos@bf6b2ef3
      c83fa3cb
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Add basic tests · b168cb33
      Kirill Smelkov authored
      As promised add tests for generic enb that exercise enb instantiation with
      multiple Radio Units and cells in all kind of combinations:
      
         FDD|TDD x LTE|NR x SDR|Lopcomm|Sunwave
      
      and not only one particular case, but verifying instantiation with many
      combinations at the same time - verifying how cells of different kinds are
      handled and how radio units of different kinds are handled too. Using multiple
      cells on top of several radio units of the same type is verified too.
      
      For Lopcomm we also exercise generated cu_config.xml a bit.
      
      See added tests for details.
      
      Enb/generic tests are bootstrapped now.
      
      From now one every fix to enb/generic and every semantic change will go on only
      if it is accompanied test change or enhancement.
      b168cb33
    • Kirill Smelkov's avatar
      software/ors-amarisoft: ru: amarisoft-stats.jinja2.py: Don't hardcode... · 34126027
      Kirill Smelkov authored
      software/ors-amarisoft: ru: amarisoft-stats.jinja2.py: Don't hardcode n_antenna_ul=2 on testing=True
      
      This usually matches ORS expectations, but for testing generic enb it is
      inappropriate and work work good, because in those tests there are more uplink
      channels.
      
      -> Fix the code to detect how many uplink channels are there automatically.
      34126027
    • Kirill Smelkov's avatar
      software/ors-amarisoft: slapos-render-config: enb: Also render CELL-drb.cfg and CELL-sib23.asn · f2927790
      Kirill Smelkov authored
      As promised in indicated TODO.
      
      It is useful for debugging and for being able to take all the rendered files,
      copy them to a callbox, and to be able to run lteenb directly on them.
      
      Besides enb.cfg we need to also have those drb and asn23, because enb.cfg
      references them.
      f2927790
    • Kirill Smelkov's avatar
      software/ors-amarisoft: slapos-render-config: enb: Add various examples of configuring RU and CELLs · 525b1036
      Kirill Smelkov authored
      Extend slapos-render-config to simulate not only what ORS does, but also
      simulate multiple radio units and/or multiple cells with RU being SDR or
      CPRI-based Lopcomm. It is useful as an example on how to configure things in
      new "shared instances" MultiRU, and to verify handling of all those combinations.
      525b1036
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Switch configuration of RU, cell, peer... · 80017f8c
      Kirill Smelkov authored
      software/ors-amarisoft: enb/generic: Switch configuration of RU, cell, peer and peercell to shared instances
      
      Previously we had cell_list, ncell_list and peers parameters and we now remove
      them and rework the software release to accept configuration for said objects
      via shared instances. In other words now it is possible to add Radio Units,
      cells, peers and peer cells on on top of main eNB service.
      
      Schema for parameters of those shared instances is based on what we just recently
      generalized and established in ru/ cell/ peer/ and peer/cell/ input schemas. We
      only add cell_kind=enb|enb_peer addition field to be able to distinguish a cell
      from peercell object, and we add a way for cell to reference RU one way or another.
      
      RU-CELL relation is no longer 1-1: one RU can be generally serving multiple
      cells. For example transmission bandwidth of SDR100 board is ~ 100MHz while
      bandwidth of one LTE cell is max 20MHz. This way it is possible to put several
      cells whose frequencies are nearby each other to be run on the same SDR
      board(*). And in general one Radio Unit can be serving transmission/reception in
      multiple frequency ranges, thus providing ability for multiple cells to be
      served by RU. This way RU is split off from CELL object and each CELL needs to
      say which RU object it wants to use.
      
      There are 3 ways to configure CELL->RU links:
      
      1. CELL references a RU object. This is the most general.
      2. CELL embeds definition of RU object. This variant is provided for simplicity
         of management for users that do not want to split CELL/RU relations.
      3. CELL2 references another cell object and says to use the same RU as CELL1.
         Again, this variant is provided for simplicity of management when users want to
         add a cell on top of same RU when they already started with "2".
      
      Full backward compatibility is provided for ORS. For this instance-ors-enb.jinja2.cfg
      is adjusted to inject synthetic RU/CELL/PEER/PEERCELL shared instances instead
      of translating to slapparameter_dict. As the result for ORS users it works
      without visible change as parameters with original ORS schema are handled as expected.
      
      Rendered files for ORS also stay practically the same as before this patch.
      
      The main change is in slaplte.jinja2 - in the loading routines. The other
      changes are mostly straightforward adaptation because details of how ru_dict,
      cell_dict, peer_dict and peercell_dict are loaded were already localized to
      slaplte and the rest of the code independent from that.
      
      Now, once we switched to shared instances, we will be finally able to add tests
      for updated enb. Please see a soon follow-up patch for that.
      
      And for existing test_ors.py we also temporary workaround breakage of tests,
      because currently generic code does not handle well special characters in
      partition references. For example it currently breaks on spaces with buildout
      complaining that sections with spaces in their name are invalid. We will fix
      that in another soon-followup patch as well, but apply a workaround for now.
      
      (*) see https://tech-academy.amarisoft.com/SDR_MultiCell_OneSdr.html for details.
      
      --------
      
      Appendix. Diff for rendered ORS enb.cfg and gnb.cfg before and after this patch:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/ors/enb/enb.cfg b/config/out/ors/enb/enb.cfg
      index 15f3b68e9..6046d366e 100644
      --- a/config/old/ors/enb/enb.cfg
      +++ b/config/out/ors/enb/enb.cfg
      @@ -6,7 +6,7 @@
      
         // Radio Units
         rf_driver: {
      -      // CELL-RU 2T2R  (sdr)
      +      // RU 2T2R  (sdr)
             name: "sdr",
             args: "dev0=/dev/sdr0",
             rx_antenna:"tx_rx",
      @@ -28,7 +28,7 @@
         // LTE cells
         cell_list: [
      
      -    // CELL  (CELL-RU)
      +    // CELL  (RU)
           {
             rf_port:      0,
             n_antenna_dl: 2,
      @@ -46,7 +46,7 @@
               // Inter-ENB HO
               {
                 rat:          "eutra",
      -          cell_id:      0x12345,  // -> 1
      +          cell_id:      0x12345,  // -> PEERCELL1
                 n_id_cell:    35,
                 dl_earfcn:    700,
                 tac:          123,
      diff --git a/config/old/ors/gnb/enb.cfg b/config/out/ors/gnb/enb.cfg
      index ac564db6c..9473f3207 100644
      --- a/config/old/ors/gnb/enb.cfg
      +++ b/config/out/ors/gnb/enb.cfg
      @@ -6,7 +6,7 @@
      
         // Radio Units
         rf_driver: {
      -      // CELL-RU 2T2R  (sdr)
      +      // RU 2T2R  (sdr)
             name: "sdr",
             args: "dev0=/dev/sdr0",
             rx_antenna:"tx_rx",
      @@ -35,7 +35,7 @@
         // NR cells
         nr_cell_list: [
      
      -      // CELL  (CELL-RU)
      +      // CELL  (RU)
             {
               rf_port:      0,
               n_antenna_dl: 2,
      @@ -58,7 +58,7 @@
               // Inter-ENB HO
               {
                 rat:          "nr",
      -          nr_cell_id:   0x77712, // -> 1
      +          nr_cell_id:   0x77712, // -> PEERCELL2
                 gnb_id_bits:  22,
                 n_id_cell:    75,
                 dl_nr_arfcn:  520000,
      ```
      80017f8c
    • Kirill Smelkov's avatar
      software/ors-amarisoft: slapos-render-config: Add peers · 69602afe
      Kirill Smelkov authored
      We already have peer cells in the setup of generated enb/gnb, but there were no
      X2/XN peers there. Add them to be able to see how handling of peers works.
      
      Rendered files change as expected:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/ors/enb/enb.cfg b/config/out/ors/enb/enb.cfg
      index 252c2e379..15f3b68e9 100644
      --- a/config/old/ors/enb/enb.cfg
      +++ b/config/out/ors/enb/enb.cfg
      @@ -22,7 +22,7 @@
           },
         ],
         gtp_addr: "127.0.1.1",
      -  x2_peers: [],
      +  x2_peers: ["44.1.1.1", "44.1.1.2"],
         enb_id: 0x1A2D0,
      
         // LTE cells
      diff --git a/config/old/ors/gnb/enb.cfg b/config/out/ors/gnb/enb.cfg
      index b401f3513..ac564db6c 100644
      --- a/config/old/ors/gnb/enb.cfg
      +++ b/config/out/ors/gnb/enb.cfg
      @@ -22,7 +22,7 @@
           },
         ],
         gtp_addr: "127.0.1.1",
      -  xn_peers: [],
      +  xn_peers: ["55.1.1.1", "55.1.1.2"],
         gnb_id_bits: 28,
         gnb_id: 0x12345,
         en_dc_support: true,
      ```
      69602afe
    • Kirill Smelkov's avatar
      software/ors-amarisoft: slapos-render-config: Diverge enb.cfg and ue.cfg generation · 1e3e198d
      Kirill Smelkov authored
      The procedures to generate those are currently very similar, and that's why we
      employ a shared `do` function to render them. But enb and ue rendering will
      become different from each other in the next step.
      
      -> Prepare to that by pushing necessary bits of do into do_enb and do_ue insides.
      
      No change in rendered files.
      1e3e198d
    • Kirill Smelkov's avatar
      software/ors-amarisoft: slapos-render-config: Retarget currently generated... · 28ab315b
      Kirill Smelkov authored
      software/ors-amarisoft: slapos-render-config: Retarget currently generated enb.cfg and gnb.cfg to out/ors/
      
      Besides ORS we will soon add other cases to enb.cfg generation. It will be
      handy to clearly see what is what and to keep ORS case to be also alive and distinguished.
      
      Prepare to that by emitting generated enb.cfg and gnb.cfg to out/ors/enb/ and
      out/ors/gnb instead of just out. Name the file itself enb.cfg always to match
      what happens for real in the SR.
      
      There is no change to generated files - it is only renaming:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/enb.cfg b/config/out/ors/enb/enb.cfg
      similarity index 100%
      rename from config/old/enb.cfg
      rename to config/out/ors/enb/enb.cfg
      diff --git a/config/old/gnb.cfg b/config/out/ors/gnb/enb.cfg
      similarity index 100%
      rename from config/old/gnb.cfg
      rename to config/out/ors/gnb/enb.cfg
      ```
      28ab315b
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Switch cell type to be a runtime parameter · 08731df7
      Kirill Smelkov authored
      In other words merge enb and gnb into sole enb that can handle both LTE and NR
      cells at the same time and simultaneously.
      
      Previously which type of Radio Access Technology to use - LTE or NR - was
      static parameter of particular software - it was possible to instantiate cells
      only with selected RAT for particular template.
      
      In MultiRU it is possible to generally instantiate all kind of cells -
      LTE/NR and TDD/FDD all at the same time and each served by a different kind of Radio Unit.
      
      -> Switch cell type to be runtime parameter as the final step to be able to do that.
      
      For generic cell definitions are now required to come with cell_type=lte|nr
      parameter. In schemas cell_list now includes oneOf for lte|nr cell instead of
      particular cell type. Similar adaption is done for ncell_list and
      x2_peers/xn_peers are merged into just peers with similar oneOf.
      
      Global gnb-specific parameters of went-away gnb are also added as global ones to enb:
      
        - gnb_id
        - gnb_id_bits
        - amf_list
        - plmn_list (as plmn_list_5g not to conflict with plmn_list)
        - nssai
      
      They must be present if there are NR cells.
      
      Full backward compatibility is preserved for ORS: now it injects cells to
      generic with particular cell_type annotations and translates ORS-specific
      x2_peers/xn_peers and gnb/plmn_list to proper generic parameters.
      
      In ORS-gnb schema gnb_stats_fetch_period and gnb_drb_stats_enabled had to be
      renamed to enb_stats_fetch_period and enb_drb_stats_enabled, but once again,
      full backward compatibility is preserved because ORS mode already had the
      following in its proxy:
      
          {#- backward compatibility: if ORS is running in gnb mode, and gnb_* parameters
              are present, replace their generic enb_* counterparts with gnb_* ones #}
          ...
      
      The diff for rendered enb.cfg and gnb.cfg, that simulate ORS, before and after
      hereby patch is also empty.
      08731df7
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Start to generalize the code to handle multiple peer cells · a9101bd9
      Kirill Smelkov authored
      As with Radio Units and cells organize a registry of peer cells and start to handle
      that registry generally everywhere. The registry is still populated with only
      LTE cells on enb and only NR cells on gnb, but the only place that is aware of
      that is registry loading routine.
      
      In the old code defaults were noticed for tac to be 1 if not provided, and so
      those defaults are pushed to live in ORS now. ORS schemas are also updated
      because that defaults were not declared there.
      
      Generic switches to use peer/cell/lte directly which means that now it accepts
      e_cell_id instead of cell_id as the name for full E-UTRAN cell identity. Full
      backward compatibility is provided for ORS which implements translation of its
      original schema to the new generic one.
      
      Tests will be added later as full tests for generic MultiRU.
      
      Rendered enb.cfg and gnb.cfg change a bit, but the changes are mostly due to
      reordering of fields. The only change with semantic value is
      
          -          ssb_nr_arfcn: ,
          +          ssb_nr_arfcn: 520090,
      
      in gnb.cfg .
      
      The old configuration was invalid anyway because if it is only
      
                    ssb_nr_arfcn: ,
      
      then enb does not start with the following error:
      
          # ../lteenb gnb.cfg
          gnb.cfg:119: unexpected token: ','
      
      Like with many other handover-related fields
      https://tech-academy.amarisoft.com/lteenb.doc#prop.ncell_list.ssb_nr_arfcn
      documents that ssb_nr_arfcn "Must be present" for Inter-gNB HO case. The old
      code was erroneously doing `ncell.get(name, '')` with yielding empty string in
      case a parameter is missing, which is corrected here.
      
      --------
      
      Appendix. Diff for rendered enb.cfg and gnb.cfg before and after this patch:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff -w config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/enb.cfg b/config/out/enb.cfg
      index 1a2c41b6b..252c2e379 100644
      --- a/config/old/enb.cfg
      +++ b/config/out/enb.cfg
      @@ -46,9 +46,9 @@
               // Inter-ENB HO
               {
                 rat:          "eutra",
      +          cell_id:      0x12345,  // -> 1
                 n_id_cell:    35,
                 dl_earfcn:    700,
      -          cell_id:      0x12345,
                 tac:          123,
               },
             ],
      diff --git a/config/old/gnb.cfg b/config/out/gnb.cfg
      index 635c4baf2..b401f3513 100644
      --- a/config/old/gnb.cfg
      +++ b/config/out/gnb.cfg
      @@ -58,14 +58,14 @@
               // Inter-ENB HO
               {
                 rat:          "nr",
      +          nr_cell_id:   0x77712, // -> 1
      +          gnb_id_bits:  22,
      +          n_id_cell:    75,
                 dl_nr_arfcn:  520000,
      -          ssb_nr_arfcn: ,
      +          band:         38,
      +          ssb_nr_arfcn: 520090,
                 ul_nr_arfcn:  520000,
      -          n_id_cell:    75,
      -          gnb_id_bits:  22,
      -          nr_cell_id:   0x77712,
                 tac:          321,
      -          band:         38,
                 ssb_subcarrier_spacing: 30,
                 ssb_period: 20,
                 ssb_offset: 0,
      ```
      a9101bd9
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/*: Deduplicate ncell_list via peer/cell · bf66de55
      Kirill Smelkov authored
      Move schema of peer cells into under peer/cell/lte and peer/cell/nr with
      keeping in mind that peer/cell will become a separate entity in MultiRU.
      
      NR deduplication is straightforward, but for LTE original ncell_list used
      `cell_id` name for full E-UTRAN cell identity. peer/cell/lte uses e_cell_id
      name for that and content of that peer/cell/lte cannot be applied directly to
      ORS's ncell_list and ncell_list in current enb/generic. To keep things working
      we pickup peer/cell/lte field by field with mapping e_cell_id into cell_id. For
      the reference for NR both original ncell_list and new peer/cell/nr use
      nr_cell_id for full 5G cell identity.
      
      Backward compatibility: nothing practically changes for ORS and generic.
      bf66de55
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Start to generalize the code to handle multiple peers · 58ff8e24
      Kirill Smelkov authored
      Add X2 peers support for symmetry with XN peers.
      
      As with Radio Units and cells organize a registry of peers and start to handle
      that registry generally everywhere. The registry is still populated with only
      LTE peers on enb and only NR peers on gnb, but the only place that is aware of
      that is registry loading routine.
      
      Tests will be added later as full tests for generic MultiRU.
      
      Rendered gnb.cfg stays the same, but there is a small change in rendered enb
      but the config remains with the previous semantic as before hereby patch:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff -w config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/enb.cfg b/config/out/enb.cfg
      index 5842dfc22..1a2c41b6b 100644
      --- a/config/old/enb.cfg
      +++ b/config/out/enb.cfg
      @@ -22,6 +22,7 @@
           },
         ],
         gtp_addr: "127.0.1.1",
      +  x2_peers: [],
         enb_id: 0x1A2D0,
      
         // LTE cells
      ```
      58ff8e24
    • Kirill Smelkov's avatar
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/ors: Deduplicate schemas · fa93f10e
      Kirill Smelkov authored
      Use $ref to share definitions taken from either generic enb or cell object.
      fa93f10e
    • Kirill Smelkov's avatar
      software/ors-amarisoft: Add notices to ORS and generic softwares what they do... · 5ba7e5ac
      Kirill Smelkov authored
      software/ors-amarisoft: Add notices to ORS and generic softwares what they do and intended usage area
      5ba7e5ac
    • Kirill Smelkov's avatar
      software/ors-amarisoft: Stop doing software*-base.cfg · 241c7116
      Kirill Smelkov authored
      Both software.cfg and software-ors.cfg became trivial wrappers over their bases
      - so move bases to be directly software.
      241c7116
    • Kirill Smelkov's avatar
      software/ors-amarisoft: Stop rendering software*.cfg · 6b252b1c
      Kirill Smelkov authored
      After previous patch this rendering became unneeded. And after we stop
      render-templates becomes unneeded too, so say goodbye to it as well.
      6b252b1c
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb+ue: Switch to JSON schemas as the primary source... · bd55f250
      Kirill Smelkov authored
      software/ors-amarisoft: enb+ue: Switch to JSON schemas as the primary source of defaults; stop rendering them
      
      JSON schemas for eNB/gNB and UE specify defaults for many parameters and the
      software release use those defaults when parameters are not explicitly
      specified on the instantiation.
      
      Some primary defaults - for bandwidth and n_antenna DL/UL - were setup in
      render-templates and propagated through all places to avoid duplicating them.
      
      Defaults for many other parameters were duplicated in both JSON schemas and in
      the code that handles those parameters.
      
      To avoid this duplication we either need to extend render-templates and put
      defaults for all the other many parameters there, or just switch to JSON
      schemas to be primary source of those defaults, and use the schemas on
      instantiation by automatically loading them and extracting defaults for all
      parameters from there.
      
      I decided to go the second way because it avoids additional layer of rendering
      and the need to keep on remembering not to put raw defaults in the JSON
      schemas. Evidently this is the same approach that Rapid CDN is going to take
      (see !1380 for details).
      
      In this patch we stop rendering JSON schemas and keep their last rendered
      result. In the later patches we will handle formed duplication there and go on
      to further merge enb and gnb.
      
      UE is handled brutally for now, because it will be much reworked after.
      
      Actually loading JSON schemas is left as TODO as I do not have resources to
      implement that now. Still this step is needed as precondition for further
      patches. I hope bit of duplication for enb/gnb parameters is ok, given that
      they are now centrally maintained and corresponding TODO warning is in place.
      bd55f250
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Move LTE and NR cell definitions to... · 253f36e0
      Kirill Smelkov authored
      software/ors-amarisoft: enb/generic: Move LTE and NR cell definitions to schemas under cell/lte/ and cell/nr/
      
      And reuse bits from there for ORS cell-related parameters.
      
      Preparatory step to handle cell as separate object and a bit less code duplication.
      253f36e0
    • Kirill Smelkov's avatar
      b0d1a1ea
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: lte: Switch bandwidth to MHz · 84e417ac
      Kirill Smelkov authored
      For uniformity with NR which already has bandwidth specified as number in MHz.
      
      For ORS mode everything stays as is which inputs bandwidth as string with MHz
      suffix, e.g. "20 MHz" instead of number `20`.
      84e417ac
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Tighten defaults · 95e1a3bd
      Kirill Smelkov authored
      We currently have 2 kinds of software:
      
      - ORS, whose intended usage is small private networks, and
      - generic, whose intended usage is small-to-medium networks.
      
      and currently they both share defaults for their parameters inherited from ORS.
      But what is appropriate for small private networks, might be not so appropriate
      for medium networks.
      
      For example ORS has default enb_id/gnb_id, but in a network with 100 nodes, it
      is better to force every node to be explicitly assigned an id and error
      otherwise. Similarly for cells ORS has defaults for cell_id, but in network
      with many nodes, cell_id needs to be explicitly assigned. With having a default
      there is also a chance to misspell the parameter name, and do not notice it
      because the software will instantiate without an error but work incorrectly.
      
      So in this patch we tighten the defaults for generic:
      
      - enb_id, mme_list become non-optional for LTE
      - gnb_id, afm_list become non-optional for NR
      - plmn_list becomes non-optional for both LTE and NR. The format is different
        in between RATs and so later we will use plmn_list and plmn_list_5g when
        merging enb and gnb.
      
      - the following parameters of LTE cells now needs to be explicitly configured:
        rf_mode, dl_earfcn, bandwidth, cell_id, pci, tac.
      - the following parameters of  NR cells now needs to be explicitly configured:
        rf_mode, dl_nr_arfcn, nr_band, bandwidth, cell_id, pci.
      
      We rework ORS mode to translate its own set of parameters and defaults into
      generic enb/gnb parameters.
      
      We similarly pull some global generic defaults into instance-enb.cfg.jinja2
      with the idea that they could be maintained in one central place. In the future
      it would be good to automatically load them from JSON schemas to avoid
      duplication.
      
      This patch should be backward compatible for ORS, but it introduces the change
      in cell_id and pci in rendered enb.cfg which become 1 instead of 0:
      
          --- a/config/old/enb.cfg
          +++ b/config/out/enb.cfg
          @@ -33,9 +33,9 @@
                 n_antenna_dl: 2,
                 n_antenna_ul: 2,
      
          -      cell_id:    0x00,
          +      cell_id:    0x01,
                 tac:        0x0001,
          -      n_id_cell:  0,
          +      n_id_cell:  1,
                 root_sequence_index: 204,
                 dl_earfcn:  36100,
                 inactivity_timer: 10000,
      
      The defaults for those parameters according to instance-enb-input-schema.json
      are 0x01 and 1. And it looks like enb.jinja2.cfg was changed to emit them starting from
      0 instead of 1 in c4d0958e due to probable thinko in that patch because before
      that patch those parameters were emitted as 0x01 and 1:
      
          nexedi/slapos@c4d0958e
      
      and, once again, the schema says their defaults should be 0x01 and 1 as well.
      
      The rest of the changes in rendered enb.cfg and gnb.cfg should not introduce
      any semantic difference. Please see the appendix for full details.
      
      --------
      
      Appendix. Diff for rendered enb.cfg and gnb.cfg before and after this patch:
      
      ```
      $ ./pythonwitheggs slapos-render-config.py && xdiff -w config/{old,out}
      ```
      
      ```diff
      diff --git a/config/old/enb.cfg b/config/out/enb.cfg
      index 89862f1d9..5842dfc22 100644
      --- a/config/old/enb.cfg
      +++ b/config/out/enb.cfg
      @@ -33,9 +33,9 @@
             n_antenna_dl: 2,
             n_antenna_ul: 2,
      
      -      cell_id:    0x00,
      +      cell_id:    0x01,
             tac:        0x0001,
      -      n_id_cell:  0,
      +      n_id_cell:  1,
             root_sequence_index: 204,
             dl_earfcn:  36100,
             inactivity_timer: 10000,
      @@ -102,8 +102,11 @@
         ],
         cell_default: {
           plmn_list: [
      -      "00101",
      -
      +      {
      +        plmn: "00101",
      +        reserved: false,
      +        attach_without_pdn: false,
      +      },
           ],
           cyclic_prefix: "normal",
      
      diff --git a/config/old/gnb.cfg b/config/out/gnb.cfg
      index 1df699a22..635c4baf2 100644
      --- a/config/old/gnb.cfg
      +++ b/config/out/gnb.cfg
      @@ -22,7 +22,7 @@
           },
         ],
         gtp_addr: "127.0.1.1",
      -
      +  xn_peers: [],
         gnb_id_bits: 28,
         gnb_id: 0x12345,
         en_dc_support: true,
      @@ -142,6 +142,7 @@
      
         nr_cell_default: {
           ssb_period: 20,
      +
           plmn_list: [
             {
               plmn: "00101",
      @@ -151,10 +152,10 @@
                 {
                   sst: 1,
                 },
      -
               ],
             },
           ],
      +
           si_window_length: 40,
           cell_barred: false,
           intra_freq_reselection: true,
      ```
      95e1a3bd
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb: Set rf_driver.args="---" if no RUs are configured · 9e7c6809
      Kirill Smelkov authored
      Else enb treats args="" unset and implicitly uses /dev/sdr0 by default.
      
      We do not want implicit default behaviour -> we want to see an error.
      9e7c6809
    • Kirill Smelkov's avatar
      software/ors-amarisoft: enb/generic: Do not expose handover measurement config publicly yet · 5b2a2eee
      Kirill Smelkov authored
      We currently have support for a3_offset and a3_time_to_trigger only, while
      there is also a3_hysteresis plus at least a1_*, a2_* and probably other parameters.
      
      Simple A3-only based meas_config tuning might be appropriate for ORS, but for
      generic enb it is better to expose full control which I do not currently feel
      good to implement right now. So do not expose it at all on generic to avoid
      establishing a public API yet.
      
      enb.jinja2.cfg already has corresponding comments:
      
        {#  TODO fully expose lte meas_config_desc in generic SR #}
          meas_config_desc: {
          ...
      
        {#  TODO fully expose nr meas_config_desc in generic SR #}
          meas_config_desc: {
      5b2a2eee
    • Kirill Smelkov's avatar
      software/ors-amarisoft: Start to generalize software.cfg.json · 7d4a76ad
      Kirill Smelkov authored
      - Indicate that it is not for ORS, but for generic case.
      - Move ue to the end, to where it logically belongs.
      - Indicate shared instances clearly with clear visually appearing indent.
        For now it is only about SIM card, but later we will be adding more kinds of
        shared instances, so it will be needed to clearly see on the UR if an item is
        a kind of service, or a shared instance of a service.
      7d4a76ad
    • Kirill Smelkov's avatar
      software/ors-amarisoft: Stop autogenerating software*.cfg.json · 681a8d6e
      Kirill Smelkov authored
      There is now only 2 kinds of softwares - ORS and generic, and their cfg.json
      will diverge with ORS staying as it is now for backward compatibility, and
      generic gaining more features and more kind of shared instances. That means
      that it no longer makes sense to render those two from a common source.
      
      -> Stop rendering them and leave previous rendered result in place for further manual edits.
      681a8d6e
    • Kirill Smelkov's avatar
      software/ors-amarisoft: software.cfg.json: Start index from 1, not 0 · ace534b8
      Kirill Smelkov authored
      Previously, e.g. for ORS it was:
      
          enb                 index=0
          gnb                 index=1
          core-network        index=2
          core-network-slave  index=3
      
      which resulted in the following appearance order on SlapOS Master UI:
      
          gnb
          core-network
          core-network-slave
          enb
      
      Note that enb comes last instead of being first as originally intended.
      
      That happens because SlapOS Master treats index=0 as being unset and puts all
      items with unset index in the end.
      
      -> Fix this by starting explicitly-set indices from 1.
      
      Now, for ORS, the order on UI is
      
          enb
          gnb
          core-network
          core-network-slave
      ace534b8