Update patch set 8

Patch Set 8:

(2 comments)

Patch-set: 8

a92e9c966b20200950a8a4581b1c94e72b6f9954

@@ -154,6 +154,24 @@
       "revId": "a92e9c966b20200950a8a4581b1c94e72b6f9954",
       "serverId": "035e6965-6537-41bd-912c-053f3cf69326"
     },
+    {
+      "unresolved": true,
+      "key": {
+        "uuid": "06639526_6fdc3633",
+        "filename": "/PATCHSET_LEVEL",
+        "patchSetId": 3
+      },
+      "lineNbr": 0,
+      "author": {
+        "id": 1000005
+      },
+      "writtenOn": "2024-04-02T22:55:36Z",
+      "side": 1,
+      "message": "it\u0027s not so much moving, but introducing either an async step between parsing the counters and storing them in rate_ctr; or a mutex lock around storing them in rate_ctr. Pau suggested osmo_it_q.\n\nthe hard part is that we evaluate all counters at the same time. if we submit all in one it_q entry, it could become a fairly large allocation+data copy. On the other extreme, neither do we want each counter causing a separate select() trigger.\nwe could submit them in batches, e.g. one it_q entry copies 100 counters.\n\nI\u0027m still thinking about the alternatives I see and how complex/performant they would be:\n- osmo_it_q in batches\n- a global mutex lock around rate_ctr (do we have multi-threaded rate_ctr code yet?)\n- volatile counter storage per struct hnb_context instance with a flag signalling the main thread to copy to rate_ctr.",
+      "parentUuid": "4d943566_3a1d15b6",
+      "revId": "a92e9c966b20200950a8a4581b1c94e72b6f9954",
+      "serverId": "035e6965-6537-41bd-912c-053f3cf69326"
+    },
     {
       "unresolved": true,
       "key": {

b32a0bb81839d14b55e3d6ddedec1075fb58331b

@@ -51,6 +51,24 @@
       "message": "current problem is getting it to pass the gerrit verification, which has problems with the nftables dependencies on debian 10.",
       "revId": "b32a0bb81839d14b55e3d6ddedec1075fb58331b",
       "serverId": "035e6965-6537-41bd-912c-053f3cf69326"
+    },
+    {
+      "unresolved": true,
+      "key": {
+        "uuid": "93e5e8f5_68a4524d",
+        "filename": "/PATCHSET_LEVEL",
+        "patchSetId": 8
+      },
+      "lineNbr": 0,
+      "author": {
+        "id": 1000005
+      },
+      "writtenOn": "2024-04-02T22:55:36Z",
+      "side": 1,
+      "message": "right, it appears we need a compiler switch for the gerrit verification, too. So now is the time, after all.",
+      "parentUuid": "6905dda3_d02e24f7",
+      "revId": "b32a0bb81839d14b55e3d6ddedec1075fb58331b",
+      "serverId": "035e6965-6537-41bd-912c-053f3cf69326"
     }
   ]
 }