75 lines
3.2 KiB
Plaintext
75 lines
3.2 KiB
Plaintext
{
|
|
"comments": [
|
|
{
|
|
"unresolved": true,
|
|
"key": {
|
|
"uuid": "ba4f9ca7_879604ff",
|
|
"filename": "src/osmo-bsc-nat/bssap_conn.c",
|
|
"patchSetId": 4
|
|
},
|
|
"lineNbr": 47,
|
|
"author": {
|
|
"id": 1000005
|
|
},
|
|
"writtenOn": "2022-06-08T10:58:03Z",
|
|
"side": 1,
|
|
"message": "above we traverse the TLV structure once to store tlv_parsed entries.\ntlv_parsed entries are good for \"random access\", find a specific IE.\nThen we traverse the TLV again to store the tags in their order,\ncaching in an array. Then we iterate the array again, which is sort\nof identical to traversing the TLV data itself.\n\nit seems to me that it makes more sense to not use tlv_parse() at all,\nbut only do tlv_parse_one() and tlv_encode_one() in this loop.\nThen we don\u0027t need to use a random-access API for sequential access.\n\nHence not needing new libosmocore API.\n\nI guess this doesn\u0027t really affect performance in a noticeable way,\nmaking it a bikeshed of sorts?",
|
|
"revId": "88cc1a2dbad497c84f09ec3b513cdca57b314160",
|
|
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
|
|
},
|
|
{
|
|
"unresolved": true,
|
|
"key": {
|
|
"uuid": "ba2b9a7d_efa30c14",
|
|
"filename": "src/osmo-bsc-nat/bssap_conn.c",
|
|
"patchSetId": 4
|
|
},
|
|
"lineNbr": 47,
|
|
"author": {
|
|
"id": 1000005
|
|
},
|
|
"writtenOn": "2022-06-08T11:03:30Z",
|
|
"side": 1,
|
|
"message": "About leaving all other known and unknown IEs unchanged: I think there can theoretically also be unknown IEIs in the TLV which tlv_parse() skips? not entirely sure about that though.",
|
|
"parentUuid": "ba4f9ca7_879604ff",
|
|
"revId": "88cc1a2dbad497c84f09ec3b513cdca57b314160",
|
|
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
|
|
},
|
|
{
|
|
"unresolved": true,
|
|
"key": {
|
|
"uuid": "7ada68e1_4223f2f1",
|
|
"filename": "src/osmo-bsc-nat/bssap_conn.c",
|
|
"patchSetId": 4
|
|
},
|
|
"lineNbr": 47,
|
|
"author": {
|
|
"id": 1000074
|
|
},
|
|
"writtenOn": "2022-06-08T11:12:21Z",
|
|
"side": 1,
|
|
"message": "The best approach is probably locating the important IE start of buffer by using tlv_parse, then simply modify that IE buffer with the new content and resend the buffer as it is, no need to encode it again?",
|
|
"parentUuid": "ba2b9a7d_efa30c14",
|
|
"revId": "88cc1a2dbad497c84f09ec3b513cdca57b314160",
|
|
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
|
|
},
|
|
{
|
|
"unresolved": true,
|
|
"key": {
|
|
"uuid": "cbae2127_e8bcd74b",
|
|
"filename": "src/osmo-bsc-nat/bssap_conn.c",
|
|
"patchSetId": 4
|
|
},
|
|
"lineNbr": 47,
|
|
"author": {
|
|
"id": 1000147
|
|
},
|
|
"writtenOn": "2022-06-08T11:35:04Z",
|
|
"side": 1,
|
|
"message": "The length of the IE is different depending on IPv4 or IPv6 address being encoded (3GPP TS 48.008 § 3.2.2.102). So I can\u0027t just modify the existing buffer without re-encoding, I\u0027ll implement Neels approach. Thanks for the feedback!",
|
|
"parentUuid": "7ada68e1_4223f2f1",
|
|
"revId": "88cc1a2dbad497c84f09ec3b513cdca57b314160",
|
|
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
|
|
}
|
|
]
|
|
} |