libosmocore/4e289a9ff84904e1e5144942ab7...

181 lines
6.5 KiB
Plaintext

{
"comments": [
{
"unresolved": false,
"key": {
"uuid": "b6aff731_e86e10c4",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000074
},
"writtenOn": "2023-11-27T16:47:28Z",
"side": 1,
"message": "I have the feeling this patch is wrong, in the sense that we are adding tons of SCTP specific stuff here which actually seem to be more generic (cmsg) in the existing layer.\nSo I\u0027d appreciate the idea of thinking whether we can keep the API generic (cms based) instead of putting lots of SCTP stuff in there. Becuase in the end we are not really calling any libsctp API....",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": false,
"key": {
"uuid": "9c9e892b_96dd3b72",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000004
},
"writtenOn": "2023-11-27T17:34:20Z",
"side": 1,
"message": "we are not *calling* libsctp, no. That is not possible. We are emulating what libsctp does, so we can have the libsctp user semantics of sctp_send + sctp_recvmsg, but built on top of io_uring. \n\nNeither we nor the user can use libsctp, as it has no clue about io_uring\n\nIf we don\u0027t do it here, every user will have to reimplement it themselves. Is that really what you want? Where would the implementation be from your point of view?",
"parentUuid": "b6aff731_e86e10c4",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": true,
"key": {
"uuid": "7eb640c4_aaa36159",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000074
},
"writtenOn": "2023-11-28T13:45:05Z",
"side": 1,
"message": "We can probably have a read/write cb passing a cmsg, and then a separate API/cb_helper constructing/filling the SCTP specific bits in the msg.",
"parentUuid": "b6aff731_e86e10c4",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": false,
"key": {
"uuid": "f49c782a_029c0880",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000074
},
"writtenOn": "2023-11-28T13:45:05Z",
"side": 1,
"message": "Sorry I left a previous comment I wrote at the time as draft and didn\u0027t press the submit button, doing now.",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": true,
"key": {
"uuid": "68965ecd_9b8d7bfa",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000231
},
"writtenOn": "2024-02-22T15:04:16Z",
"side": 1,
"message": "Issue will be fixed later.",
"parentUuid": "7eb640c4_aaa36159",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": false,
"key": {
"uuid": "540843c4_f78524db",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000004
},
"writtenOn": "2024-02-22T21:22:56Z",
"side": 1,
"message": "I\u0027ve discussed this to some extent with jolly earlier today, and decided I\u0027d want to tryo how a pure sendmsg/recvmsg implementation looks like. I\u0027ve now pushed that implementation. I\u0027m not really sure if it\u0027s really the best abstraction given that we already do lots of sctp specific stuff in libosmocore, but now don\u0027t do the stream/ppid part in osmo_io. However, it also allows us to avoid any clashes in sctp_msg_* #defines, and that is a very clear benefit IMHO.",
"parentUuid": "68965ecd_9b8d7bfa",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": false,
"key": {
"uuid": "92615ee1_81208eba",
"filename": "/PATCHSET_LEVEL",
"patchSetId": 3
},
"lineNbr": 0,
"author": {
"id": 1000074
},
"writtenOn": "2024-02-23T13:28:35Z",
"side": 1,
"message": "yay! this looks much better from architectural point of view imho.",
"parentUuid": "540843c4_f78524db",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": true,
"key": {
"uuid": "91b91767_2a98b2e5",
"filename": "include/osmocom/core/osmo_io.h",
"patchSetId": 3
},
"lineNbr": 17,
"author": {
"id": 1000074
},
"writtenOn": "2024-02-20T10:11:33Z",
"side": 1,
"message": "did this clash disappear magically?",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": true,
"key": {
"uuid": "9b4e3afc_1811364c",
"filename": "include/osmocom/core/osmo_io.h",
"patchSetId": 3
},
"lineNbr": 17,
"author": {
"id": 1000004
},
"writtenOn": "2024-02-20T19:10:08Z",
"side": 1,
"message": "I explicitly requested somewhere (review comment or redmine issue) to test build all osmocom software to see if the clash ever appears in practice. It seems it hasnt? Would be good to get explicit confirmation from jolly.",
"parentUuid": "91b91767_2a98b2e5",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
},
{
"unresolved": false,
"key": {
"uuid": "39dac0ae_a82f7b2f",
"filename": "include/osmocom/core/osmo_io.h",
"patchSetId": 3
},
"lineNbr": 17,
"author": {
"id": 1000231
},
"writtenOn": "2024-02-22T15:04:16Z",
"side": 1,
"message": "I tested compilation with libsccp, osmo-bsc and osmo-msc. There was no clash, so I removed the comment.",
"parentUuid": "9b4e3afc_1811364c",
"revId": "4e289a9ff84904e1e5144942ab77d03a93dfe39c",
"serverId": "035e6965-6537-41bd-912c-053f3cf69326"
}
]
}