**** BEGIN LOGGING AT Tue Feb 03 02:59:58 2015 Feb 03 03:01:10 alexanderhiam : the PRU bridge is generic, has nothing to do with pruspeak as such! Feb 03 03:01:47 look at the link I posted in my idea Feb 03 03:02:52 There is a PRU side API which will enable users to develop their own firmware and communicate with the linux userspace. Feb 03 03:04:23 karki: yeah, on the same page there. I'm thinking the target should be an API that covers all the bridge stuff, including rpc and shared memory Feb 03 03:05:08 something like pruspeak could then be implemented with it Feb 03 03:05:28 alexanderhiam : Thats how the project currently stands, right? Feb 03 03:05:49 or did you understand it differently? Feb 03 03:06:26 right Feb 03 03:07:25 alexanderhiam : (i) A easy way to have multi channel communication Feb 03 03:07:38 (ii) have higher level API's for RPCs etc Feb 03 03:07:56 what I have explained in the pru_serial readme page is only (i) Feb 03 03:08:14 alexanderhiam : I'd love a pull request with your addons :) Feb 03 03:09:09 yeah, I need to gather and write down my thoughts on this Feb 03 03:09:37 Once you do that, I will update the ideas page to Feb 03 03:09:52 as of now both the ideas on a new driver seems vague Feb 03 03:10:18 I don't think I have managed to give a good high level view Feb 03 03:10:25 have to change that Feb 03 03:10:33 I'll work on that tomorrow (EST tomorrow that is) Feb 03 03:11:17 and abhisheks does not explain much on how the implementation should be, or rather along what lines the student should think along Feb 03 03:11:29 We have to merge this idea at some point of time Feb 03 03:11:46 that can be done based on how many students apply Feb 03 03:12:07 I already got mails from 2 students who were interested to work on pruspeak Feb 03 03:12:17 great Feb 03 03:12:17 but redirected them to pru bridge Feb 03 03:12:39 we should get jkridner up to speed on this plan soon Feb 03 03:13:21 so (i) figure out what we want the driver to do. (ii) an effective way to get to LKML ( I guess this is what Abhishek is mainly interested in ;) ) Feb 03 03:13:43 Then we can split it properly into 2 projects if reqd. Feb 03 03:13:59 or minimize the work to be done for one student, etc Feb 03 03:14:08 as of now, let both the projects stand Feb 03 03:14:50 sounds good Feb 03 03:14:56 ok I gotta run Feb 03 03:15:06 talk soon Feb 03 15:22:41 evening Feb 03 15:30:02 karki__: Implementation is up to the student to work out details. I just gave high level design guidelines for them to think on. Feb 03 15:37:26 yeah, the problem is that no new comer will know what we need and how to go about it. Feb 03 15:37:47 My only concern is ending up with something no one can use Feb 03 16:39:21 which is also a valid concern Feb 03 17:58:16 Abhishek_ : any ideas on how to accommodate a plugin system in the kernel for the new pruss driver? Feb 03 17:58:39 I didn't like what I did for pruspeak..... Feb 04 00:16:08 karki: It's better to put it down in code than just speculate Feb 04 00:16:44 how's our idea lists coming along? Feb 04 00:19:16 there are two PRU design ideas, I guess you saw the thread Feb 04 00:23:00 wonder if there are any interest in a project to figure out the EDMA stuff for the PRU Feb 04 00:24:21 I personally think that if a framework for PRU based drivers evolves then any kind of driver can be written out easily. Feb 04 00:24:48 that has kind of been done Feb 04 00:25:09 but there doesn't seem to be any docs for EDMA so any large xfers has to burn the other PRU Feb 04 00:25:37 I see Feb 04 00:25:52 you did your xfers to DDR using the other PRU, right? Feb 04 00:25:58 yeah Feb 04 00:27:24 the EDMA should be able to do that for you. EDMA is more or less a processor that knows how to do memcpy's Feb 04 00:28:00 is EDMA supported by the kernel. I mean, do the other devices or periphs use it? Feb 04 00:29:55 i think so Feb 04 02:09:30 ds2: any idea for android based projects or so ? **** ENDING LOGGING AT Wed Feb 04 02:59:58 2015