**** BEGIN LOGGING AT Sat Aug 24 02:59:57 2019 Aug 24 08:48:55 zeekhuge Abhishek_ The kernel oops issue is solved, it was a timing problem. What was happening was that the pruss daemon was receiving requests to send data to the rpmsg device file before the file was even created in /dev. The code worked properly second time onwards because the rpmsg drivers had already been probed and the rpmsg device file would then be created before the daemon would receive requests to Aug 24 08:48:55 write to it. Aug 24 08:51:39 Good. Aug 24 08:51:47 IIRC, that file is tied to the channel creation Aug 24 08:52:50 ds2: you mean the file inside /dev right? Yes it is the tied to the channel. Aug 24 08:53:27 zeekhuge[m]: yep. so it is tied to when the firmware is fully running Aug 24 08:54:55 zeekhuge: Fix: https://github.com/pratimugale/PRUSS-Bindings/commit/bc4ba9070c68a3c3d686c8d28a3f07b83f656d99. Aug 24 08:56:05 hmmmm that doesn't seem right Aug 24 08:56:20 Ooops do not happen if a userland process can't find a /dev entry Aug 24 08:56:49 Ooops are purely a kernel thing.... if the entry doesn't exists, open() should fail not oops out Aug 24 08:58:01 ds2: The message does stop to show if we wait until the /dev entry is created Aug 24 08:58:38 I suspect what you are doing is adding a delay Aug 24 08:59:22 ds2: It is a delay Aug 24 08:59:57 that isn't exactly a fix Aug 24 09:00:52 somehow you are managing to successfully open the /dev entry before everything is initialized in the kernel... fixing the init process with locks, etc is the right thing to do not add an arbitrary delay on the userland end Aug 24 09:01:12 it isn't easy to chase down but.... Aug 24 09:04:48 ds2: So does it mean the issue might be in the rpmsg_pru driver? Aug 24 20:07:08 drhunter: can you take a quick look on that: https://medium.com/@denisobrezkov/xen-on-beagleboard-x15-a517ea7d9bb Aug 24 20:07:08 Today and tomorrow, I am going to prepare the patch. I will also add some picture of the crossbar+gic. Anything else that needs to be added? **** BEGIN LOGGING AT Sat Aug 24 20:28:34 2019 **** ENDING LOGGING AT Sun Aug 25 02:59:57 2019