708926 - RIL does not reconnect on socket connection loss

Status: RESOLVED FIXED

(Core :: IPC, defect)

Description

[Kyle Machulis [:qdot] [:kmachulis] (INACTIVE)]

If the RIL for some reason cannot pick up a connection to the rilb2g socket, it does not current retry for connection. This needs to be fixed in the RIL thread.

Comment 1

[Kyle Machulis [:qdot] [:kmachulis] (INACTIVE)]

Attachment: Adding reconnection capabilities to RIL IPC thread

Comment 2

[Andreas Gal :gal]

Put a space between comments and the comment slashes //[ ]comment, and also put the "return" on a new line. I am not fluent in IPC, so lets wait for r=chris, but looks good to me. Nice fix.

Comment 3

[Kyle Machulis [:qdot] [:kmachulis] (INACTIVE)]

Attachment: RIL Patch with comment/return fixes

Comment 4

[Chris Jones [:cjones] inactive; ni?/f?/r? if you need me]

Comment on attachment 581050 [details] [diff] [review]
RIL Patch with comment/return fixes

>diff --git a/ipc/ril/Ril.cpp b/ipc/ril/Ril.cpp

>+void RilReconnectTask::Run() {

Nit:

void
RilReconnectTask::Run()
{

>+    if(sClient->OpenSocket()) {

Nit: "if ("

>@@ -221,6 +234,13 @@ RilClient::OnFileCanReadWithoutBlocking(int fd)
>             int ret = read(fd, mIncoming->mData, 1024);
>             if (ret <= 0) {
>                 LOG("Cannot read from network, error %d\n", ret);
>+                //At this point, assume that we can't actually access
>+                //the socket anymore, and start a reconnect loop.

Nit: Space after "//".

>+                mIncoming.forget();
>+                mReadWatcher.StopWatchingFileDescriptor();
>+                mWriteWatcher.StopWatchingFileDescriptor();
>+                close(mSocket.mFd);

I would prefer if this were packaged into a "*State" kind of struct,
so that it could all be reset as a group, but this is OK for now.

Looks good, r=me with nits picked.

Comment 5

[Kyle Machulis [:qdot] [:kmachulis] (INACTIVE)]

Attachment: Post-Review RIL Patch with nits picked

Comment 6

[Philipp von Weitershausen [:philikon]]

Pushed (with the wrong commit msg, bah!): https://hg.mozilla.org/integration/mozilla-inbound/rev/e6f4331dba0d

Comment 7

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/e6f4331dba0d