| 00:28.27 | *** join/#openjtag davygravy_ (n=davygrav@h75-100-90-62.75-100.unk.tds.net) |
| 01:02.48 | *** join/#openjtag davygravy (n=davygrav@h75-100-90-62.75-100.unk.tds.net) |
| 04:14.55 | *** join/#openjtag jkilb_ (n=jkilb@p5B209107.dip0.t-ipconnect.de) |
| 05:59.29 | *** join/#openjtag drath_ (i=vmaster@p5B07E028.dip.t-dialin.net) |
| 09:10.26 | *** join/#openjtag toi (n=pleemans@d5152D3B4.access.telenet.be) |
| 10:17.16 | *** join/#openjtag ali_as (n=as@ambix.plus.com) |
| 10:50.33 | *** join/#openjtag jkilb_ (n=jkilb@p5B20A5A5.dip0.t-ipconnect.de) |
| 13:17.39 | *** join/#openjtag prpplague (n=dave@mail.americanmicrosystems.com) |
| 14:30.08 | *** join/#openjtag ce_geek (n=ce_geek@roger-blanchette-4.umelst.maine.edu) |
| 16:25.22 | prpplague | drath: know of anyone working on cortex-a8 support for openocd? |
| 16:36.20 | ali_as | The OMAP stuff is still very early isn't it, barely sampling? |
| 16:36.50 | prpplague | ali_as: yea, that was where i was headed with the question \ |
| 16:36.59 | prpplague | ali_as: in reference to the beagle |
| 16:37.16 | ali_as | beagle? |
| 16:37.28 | prpplague | beagleboard.org |
| 16:38.09 | drath | prpplague: no, don't know of anyone |
| 16:38.18 | ali_as | Oh that could be very cool. |
| 16:38.42 | ali_as | RISC-OS went open source last year and they are looking for targets to port it to. |
| 16:41.44 | prpplague | only does linux |
| 16:46.27 | ali_as | Linux has a lot of advantages but I miss the time when you told an OS to do something and it did it, not had a think, went through it's virtual memory and then did it. Windows and Linux GUIs are headed in a very bogged down direction. |
| 16:49.40 | prpplague | indeed |
| 16:50.39 | *** join/#openjtag gandhijee (n=root@host-66-202-34-165.spr.choiceone.net) |
| 17:00.50 | ali_as | I'm a little dissapointed ethernet seems to be playing second fiddle to USB on the beagle. |
| 17:02.36 | prpplague | ali_as: as long as you have usb you have ethernet, just plug a dongle in |
| 17:02.59 | prpplague | ali_as: plus you can add a enc28j60 module easy if you wanted to |
| 17:03.22 | *** join/#openjtag bjdooks (n=ben@trinity.fluff.org) [NETSPLIT VICTIM] |
| 17:05.42 | ali_as | That's more functional ethernet, rather than performance ethernet. |
| 17:23.17 | prpplague | ali_as: what kind of application did you have in mind? |
| 17:24.10 | ali_as | The one obsessing in my head at the moment is my jtag interface design but it's not a good fit to the beagle board. |
| 17:25.02 | prpplague | ali_as: just don't see doing a jtag via ethernet needing that much performance or bandwidth |
| 17:25.49 | prpplague | ali_as: do you want ethernet simply for bandwidth or for remote operations? |
| 17:26.12 | ali_as | I see ethernet as the gateway to the world beyond the computer itself. |
| 17:26.37 | prpplague | ali_as: how is that applicable to a jtag device? |
| 17:26.44 | ali_as | Internet, streaming media, the ability for any computer anywhere on the planet to use the device over the internet. |
| 17:27.33 | prpplague | ali_as: well you have to remember, the beagle and other embedded boards are just that, embedded. you shouldn't expect them to perform as desktops or servers |
| 17:27.38 | ali_as | These are seperate objections to the lack of ethernet, but by the by I have briefly debugged a set top box in spain over the internet. |
| 17:28.00 | ali_as | (I'm in the UK). |
| 17:28.12 | prpplague | ali_as: i too have done complete development remotely before |
| 17:28.16 | ali_as | Performance wise there was a bit of lag but I was impressed it worked. |
| 17:28.44 | prpplague | ali_as: but i have rarely seen an instance where 10baseT via a enc28j60 or usb dongle weren't more than adequate |
| 17:29.38 | ali_as | Well one of things I want to try is streaming mpeg video though a jtag interface. |
| 17:29.56 | prpplague | ali_as: uh? why? |
| 17:30.29 | ali_as | Because I think it can be done, and there may be no better interface available on some systems. |
| 17:30.33 | prpplague | ali_as: why would you want to do something like that? |
| 17:31.00 | prpplague | ali_as: doesn't exactly go with the purpose of JTAG |
| 17:31.24 | ali_as | True, but then neither do most debugging solutions, it was intended for board testing. |
| 17:31.58 | ali_as | JTAG has been used and abused by everyone. Just not for video yet (That I know of). |
| 17:32.46 | prpplague | hmm, well whatever makes your boat float |
| 17:32.50 | ali_as | 1Gbit ethernet has higher performance than USB, in my experience with this equipment the drivers are a lot more stable... |
| 17:33.45 | ali_as | To me USB is a crippled version of ethernet. |
| 17:33.48 | prpplague | yea, 1000baseT definetly has higher performance than USB, just don't know you really need it |
| 17:33.54 | ali_as | Powered ethernet would solve everything that USB solved and so much more. |
| 17:34.14 | prpplague | ali_as: yea PoE does help alot |
| 17:34.29 | prpplague | we have several products that are PoE |
| 17:34.41 | ali_as | Cool, where do you work? |
| 17:35.30 | prpplague | http://www.amltd.com and sister company http://www.tincantools.com |
| 17:35.51 | ali_as | I'm biased due to how much trouble I've has with USB. It doesn't work on my main machine at all and when it does work on my second machine it seems to take up a lot of cpu time. |
| 17:36.37 | prpplague | well its the standard people use, we have to support it |
| 17:36.43 | prpplague | ali_as: linux or windows? |
| 17:37.11 | ali_as | Windows but the problem on my main machine I think is a hardware issue. |
| 17:37.17 | drath | ali_as: problems with USB itself? |
| 17:37.25 | drath | ali_as: or with a specific USB device? |
| 17:38.17 | prpplague | my major complaint about USB is that it doesn't have a low power support option, i.e. +3.3v or +2.5v powered devices |
| 17:38.22 | ali_as | USB every device I've tried including a hauphauge DVB-T tuner and an Opal Kelly spartan development board. |
| 17:38.41 | drath | ouch |
| 17:39.29 | ali_as | And every driver I've tried trying to make all of them work. |
| 17:40.04 | ali_as | USB on this laptop works ok. |
| 17:40.15 | drath | guess that't the common case |
| 17:40.30 | drath | it's increasingly popular for internal devices, too |
| 17:40.47 | prpplague | last time i had problems with usb was back in 2.4.18 days |
| 17:40.51 | drath | yeah |
| 17:42.15 | drath | main problem i've seen is devices that don't follow the standard closely enough |
| 17:44.51 | ali_as | USB would be a lot better if you could bypass the protocol and just use it like an ultra fast serial port. |
| 17:45.39 | prpplague | ali_as: you can do that |
| 17:45.58 | prpplague | ali_as: we use it like that do firmware upgrades on one of our handhelds |
| 17:46.08 | ali_as | Do you have a link? |
| 17:46.30 | prpplague | link to what? |
| 17:46.41 | ali_as | Information on how. |
| 17:47.15 | prpplague | ali_as: libusb howto pages is a good start |
| 17:47.17 | ali_as | I remeber there used to be a method for usb 1.0, but the microsoft page on how to do it was pulled. |
| 17:47.53 | ali_as | I'll look at that at. |
| 17:48.01 | drath | prpplague: I guess he's thinking about one or two layers below |
| 17:48.24 | drath | libusb exposes the various transfer mechanisms |
| 17:48.31 | ali_as | I'm thinking at raw bytes, NRZ encoded or whatever on the line. |
| 17:48.43 | drath | yeah - libusb is way above that |
| 17:49.02 | drath | libusb provides full access to the USB base standard |
| 17:49.09 | prpplague | yea, but that give you an idea of how to start |
| 17:49.22 | drath | but not to the physical layer |
| 17:49.31 | prpplague | ours uses a custom host driver along with a custom gadget driver |
| 17:49.48 | prpplague | we get about 8-9mb/s transfer rates |
| 17:49.48 | ali_as | I'll look, I'm not sure the chips will let you have access to the physical layer on USB2.0, but I will check that. |
| 17:49.57 | drath | they don't |
| 17:50.52 | prpplague | there isn't a real need to get that low, but you can use it to do high speed data transfer at very low levels |
| 17:51.02 | drath | what prpplague means is using the control and probably bulk transfer without one of the device classes |
| 17:51.20 | prpplague | drath: yes, sorry didn't explain that well |
| 17:51.26 | ali_as | I have no idea what that means. |
| 17:51.53 | ali_as | The last time I looked at trying to program for it, I hit concepts like 'end points' that I couldn't wrap my head around. |
| 17:51.54 | drath | well, USB is a bus with a shared medium, that needs some kind of arbitration |
| 17:52.03 | drath | ali_as: that's what prpplague means |
| 17:52.34 | drath | on USB, arbitration is implemented by the master sending token packets, upon which a slave may answer |
| 17:53.03 | drath | USB further defines for transfer classes - control, interrupt, bulk and isochronous |
| 17:53.35 | ali_as | Ok. |
| 17:53.37 | drath | and uses the concept of endpoints to identify communication partners (like a TCP port, i guess) |
| 17:54.05 | drath | in the most simple case, a device implements one control endpoint, that's used to identify and configure the device |
| 17:54.09 | ali_as | Like a window in TCP? |
| 17:54.39 | drath | no, like a port in TCP - you're talking to device 192.168.0.1, port 22 |
| 17:54.45 | drath | or port 23 |
| 17:54.48 | ali_as | Oh right. |
| 17:55.18 | drath | the endpoints have special properties, like some being bidrectional (control endpoints) or unidrectional (bulk endpoints) |
| 17:55.28 | drath | but basically they identify who you're talking with |
| 17:55.31 | ali_as | I had the wrong end of the stick entirely then, nowonder I didn't get it. I thought an endpoint was the end of a transfer block, not it's destination. |
| 17:56.06 | drath | a USB stick usually implements three endpoints - one control endpoint, and two bulk endpoints, one for incoming, and one for outgoing data |
| 17:56.43 | ali_as | Ok. |
| 17:57.06 | drath | when you connect the stick, the host is notified, and sends a "read device descriptor" command to the control endpoint 0 of the default address 0 |
| 17:57.14 | drath | the device comes up with address 0 |
| 17:57.21 | drath | and implements control endpoint 0 |
| 17:57.29 | drath | so it answers with its identification |
| 17:57.54 | drath | the host then decides what to do with the device, usually assigning it an address > 0 |
| 17:58.36 | drath | it then configures further things, like the desired configuration, it more than one is supported |
| 17:58.56 | ali_as | That's helpful. I'll have to deal with USB eventually. |
| 17:58.58 | drath | at some point, the usb stick is ready to accept transfers to/from its bulk endpoint |
| 17:59.36 | drath | the meaning of the various endpoints is either completely vendor defined, or follows some of the device classes, like usb mass storage (usb sticks, usb hard disk, ...) |
| 18:00.02 | ali_as | MTP? |
| 18:00.31 | drath | MTP? |
| 18:01.03 | ali_as | My mp3 player connects as an MTP device on this laptop, wondering if thats a generic protocol. |
| 18:01.26 | prpplague | that doesn't ring a bell |
| 18:01.41 | ali_as | USB Mass storage I've heard of as a generic 'driverless' USB protocol. |
| 18:01.50 | prpplague | yea, |
| 18:01.59 | drath | heh, http://en.wikipedia.org/wiki/Media_Transfer_Protocol |
| 18:01.59 | prpplague | mass storage is a well documented class |
| 18:02.08 | prpplague | looks |
| 18:02.25 | drath | apparently it's from our friends in redmond |
| 18:02.40 | prpplague | ahh no wonder i'd never heard of it |
| 18:03.28 | drath | yeah, MTP is supposed to become a generic protocol just like USB mass storage, but it isn't an official device class yet |
| 18:10.35 | ali_as | Incidentally, is anyone aware of any documentation on debugging x86 chips over JTAG? For example athlon or geode processors? |
| 18:13.17 | drath | i guess that's only for testing, not debugging |
| 18:13.33 | prpplague | ali_as: cosmic over in #edev will be able to point you in the right direction |
| 18:16.21 | ali_as | I'll ask, thanks. |
| 18:17.13 | ali_as | Well, I'll idle until someone of that description apears, and then I'll ask ;) |
| 18:30.14 | prpplague | ali_as: he's on/off today |
| 18:30.16 | prpplague | meetings |
| 18:31.35 | ali_as | I'll be sticking around anyway and and a hardware hacking channel sounds fun. |
| 18:33.16 | ali_as | Though taking my Canon hv20 to bits turned out to be something of a mistake. |
| 19:21.37 | *** join/#openjtag davygravy (n=davygrav@h75-100-90-62.75-100.unk.tds.net) |
| 20:09.44 | gandhijee | is something wrong with the svn server? |
| 20:43.49 | gandhijee | ?? |
| 20:48.23 | *** join/#openjtag ali_as_ (n=as@ambix.plus.com) |
| 21:43.41 | *** join/#openjtag ali_as_ (n=as@ambix.plus.com) |
| 21:43.41 | *** join/#openjtag davygravy (n=davygrav@h75-100-90-62.75-100.unk.tds.net) [NETSPLIT VICTIM] |
| 21:43.41 | *** join/#openjtag gandhijee (n=root@host-66-202-34-165.spr.choiceone.net) |
| 21:43.41 | *** join/#openjtag rwhitby (n=rwhitby@nslu2-linux/rwhitby) [NETSPLIT VICTIM] |
| 22:24.26 | *** join/#openjtag ali_as_ (n=as@ambix.plus.com) |
| 22:24.26 | *** join/#openjtag davygravy (n=davygrav@h75-100-90-62.75-100.unk.tds.net) |
| 22:24.26 | *** join/#openjtag gandhijee (n=root@host-66-202-34-165.spr.choiceone.net) |
| 22:24.26 | *** join/#openjtag rwhitby (n=rwhitby@nslu2-linux/rwhitby) |
| 23:36.57 | *** join/#openjtag dwery (n=dwery@nslu2-linux/dwery) |