IPC-HLE event handling improvements #432
Conversation
|
As far as I know, this looks good to me. |
|
I had to revise this comment because the disc channel stuff was removed. The non-disc channel stuff works perfectly. |
|
Taking out the commit for the disk channel; getting it right requires more changes, and those fixes will be easier to review separately. |
|
Updated to be properly threadsafe. |
| @@ -65,7 +65,7 @@ void CWII_IPC_HLE_Device_hid::handleUsbUpdates(struct libusb_transfer *transfer) | |||
| // Return value | |||
| Memory::Write_U32(ret, replyAddress + 4); | |||
|
|
|||
| WII_IPC_HLE_Interface::EnqueReplyCallback(replyAddress); | |||
| WII_IPC_HLE_Interface::EnqReply_Threadsafe(replyAddress); | |||
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
|
@neobrain if we were to ship a release next week this branch wouldn't be cherry-picked/merged to our stable branch. I don't see the point of that comment. We're not cutting a release next week as far as I know. |
|
I very extensively tested this and ipc-disc-channel2 + dvdlowopenpartition. It's a great set of features that I really like, and the other two rely on this merge. I'm biased because it lets me do really dumb stuff in Dolphin, like use the system menu proficiently. This alone doesn't do much, I don't remember what it fixes alone and will need to retest it. |
| @@ -76,9 +76,9 @@ IWII_IPC_HLE_Device* es_handles[ES_MAX_COUNT]; | |||
| typedef std::deque<u32> ipc_msg_queue; | |||
| static ipc_msg_queue request_queue; // ppc -> arm | |||
| static ipc_msg_queue reply_queue; // arm -> ppc | |||
| static std::mutex s_reply_queue; | |||
|
|
|||
| static int enque_reply; | |||
This comment was marked as off-topic.
This comment was marked as off-topic.
Sorry, something went wrong.
|
Okay, so after retesting it, this PR basically fixes pretty fundamental problems with loading/exiting games from the system menu. With this installed, I can load any channel that runs (except the disc channel, which is part of another branch,) play around in it, exit the channel back to the system menu. This includes Virtual Console games, Mii Channel, Internet Channel, etc. After playing through with this alone, I really don't think it will cause issues. I'm about 95% sure that there are no regressions; whatever regressions there could be are too tricky for me to catch or so obvious I didn't notice them. Considering I had been testing the more complex builds that fixes a lot more and noticed nothing, I can say that this could be merged without issues. |
|
Given that we won't be having a release for another while and that this change seems to be stable, this can be merged soonish. There are still some open PR comments though. |
|
I think I've addressed all the review comments. |
|
Is the second commit a correction to the first one? If so, please squash them together. Otherwise I'll merge this later today. |
This change attempts to make calls to WII_IPC_HLE_Interface::Update() more consistent. We should be calling it whenever there is something we can pop from a queue, and this patch does that. DeviceUpdate() now works independently of Update(): it's now just a periodic timer for devices which need to regularly update state. With this change, we need a non-zero default delay for IPC replies because replies are popped off the queue much more aggressively. This fixes launching channels from the the System Menu.
IPC-HLE event handling improvements
See commits for detailed description. Makes the System Menu work much better: installed channels can now be launched from the System Menu.
This is making changes to historically sensitive timing code, but testing hasn't found any regressions involving Wiimote handling (or anything else).