TCPSocket.bufferedAmount represents the number of bytes which have previously been buffered by calls to TCPSocket.send().
It should be updated and eventually become 0 when the data buffer is sent.
It works as expected on main process but the value wouldn't become 0 if it isn't on main process.

Steps to Reproduce:
Call web API TCPSocket.send on any kind of non-main process.

Here is an example.

1. Open Firefox browser (Firefox 83.0)
2. Type about:debugging#/runtime/this-firefox to the URL bar
3. Inspect a Firefox extension (ex: BugzillaJS)
4. Send data to any TCP server via TCPSocket in console

socket = navigator.mozTCPSocket.open(IP_OF_TCP_SERVER, PORT_OF_TCP_SERVER, { binaryType: "arraybuffer", });
smallUint8Array = new Uint8Array(128);
for (let i = 0; i < smallUint8Array.length; i++) {
  smallUint8Array[i] = i;
}
socket.send(smallUint8Array.buffer, 0, smallUint8Array.length);

5. Wait for 1 or 2 seconds and check the value of socket.bufferedAmount
   socket.bufferedAmount is still 128
   P.S. It would be updated to 0 if it's on main process.

The issue can be observed since gecko lands the changes of Bug 1569183.
It seems like the IPC of TCPSocket doesn't cover the update of TCPSocket::mTrackingNumber.

Though, I'm not sure whether all the use cases on Firefox Browser use TCPSocket on main process.
Thanks

(In reply to Jamin Liu [:jaliu][:Jamin] from comment #1)

The issue can be observed since gecko lands the changes of Bug 1569183.
It seems like the IPC of TCPSocket doesn't cover the update of TCPSocket::mTrackingNumber.

Though, I'm not sure whether all the use cases on Firefox Browser use TCPSocket on main process.
Thanks

Thanks for reporting this.
I think it's not common to use TCPSocket on parent process, so mark this as P3.

(In reply to Kershaw Chang [:kershaw] from comment #2)

(In reply to Jamin Liu [:jaliu][:Jamin] from comment #1)

The issue can be observed since gecko lands the changes of Bug 1569183.
It seems like the IPC of TCPSocket doesn't cover the update of TCPSocket::mTrackingNumber.

Though, I'm not sure whether all the use cases on Firefox Browser use TCPSocket on main process.
Thanks

Thanks for reporting this.
I think it's not common to use TCPSocket on parent process, so mark this as P3.

See Jamin's first comment: " It works as expected on main process but the value wouldn't become 0 if it isn't on main process."

EIther way, Jamin can you submit your patch?

(In reply to [:fabrice] Fabrice Desré from comment #3)

EIther way, Jamin can you submit your patch?
Provide a WIP patch that partially undo the change of [1] for correcting the update of TCPSocket.bufferedAmount.
[1] https://phabricator.services.mozilla.com/D45112

The patch is shared here for starting a discussion.
I'm not sure whether the change is suitable for all the use cases of TCPSocket, therefore, I didn't raise the review flag.

Feel free to take this Bug without asking me.
Thanks.