| 00:03.00 | [g2-lap] | http://pastebin.ca/38619 |
| 00:04.34 | [g2] | AH... |
| 00:09.49 | beewoolie | I don't see how this can work on the loft. There is no setup for your target. |
| 00:10.15 | [g2] | MACH_LOFT |
| 00:10.26 | beewoolie | drivers/ide/Kconfig |
| 00:10.31 | beewoolie | drivers/ide/arm/Kconfig |
| 00:10.44 | [g2] | I've got a patch |
| 00:11.01 | beewoolie | So it isn't in the kernel tree. |
| 00:11.32 | [g2] | the code that has the issue is in the kernel tree |
| 00:11.40 | [g2] | the patch is against 2.6.15 |
| 00:11.47 | beewoolie | Yeah, but the IDE code works. The problem is in the patch. |
| 00:12.13 | [g2] | very true, but it works BE and doesn't LE |
| 00:12.31 | beewoolie | and the patch probably makes a BE assumption. |
| 00:12.32 | [g2] | I'm trying to understand why/how this is happening |
| 00:12.43 | beewoolie | The IDE interface is really dumb. |
| 00:12.49 | [g2] | I know |
| 00:12.59 | beewoolie | It has all sorts of goofy bits that make drivers difficult to write. |
| 00:13.21 | beewoolie | It swaps bytes, it has 8 bit registers even when someone needs to read/write 8 bits. |
| 00:13.43 | beewoolie | It's got several methods for accessing the same status info as well as the data being read/written. |
| 00:13.47 | beewoolie | It's hokey. |
| 00:14.05 | beewoolie | Moreover, it's very vulnerable to differences in the bus access modes. |
| 00:14.36 | beewoolie | so, just because the driver gives an error in the IDE code, doesn't mean that the problem didn't originate in the access routines. |
| 00:15.04 | [g2] | right I'm just trying to understand where the error is occurring |
| 00:15.28 | [g2] | I know an end failure point as it's probing the drive |
| 00:15.30 | [g2] | and it fails |
| 00:15.49 | [g2] | but it could be bad all the way from the first access (and probably is) |
| 00:15.52 | beewoolie | So, it's probably reading the status incorrectly. |
| 00:16.31 | beewoolie | It's not too hard to verify this from APEX, for example, because you can read the registers and check that the right bits are coming out. |
| 00:16.50 | [g2] | well it's LE |
| 00:16.57 | beewoolie | APEX builds LE. |
| 00:16.58 | [g2] | so Reboot is BE |
| 00:17.22 | [g2] | we don't have a IXP425 Loft port for APEX yet |
| 00:17.31 | beewoolie | :-) |
| 00:17.41 | beewoolie | edboot can read the registers, too. |
| 00:17.47 | beewoolie | How are you booting this thing? |
| 00:17.55 | [g2] | I've got a Redboot |
| 00:18.05 | beewoolie | There are commands to read from memory. |
| 00:18.13 | [g2] | I'd actually love to put APEX on there in LE |
| 00:18.29 | beewoolie | One test to perform is to read out the registers and look for the appropriate status values. |
| 00:19.53 | [g2] | does APEX support IDE ? |
| 00:20.08 | beewoolie | Sure does. |
| 00:20.17 | beewoolie | There is a CF driver. |
| 00:20.29 | [g2] | you run in PIO mode ? |
| 00:20.31 | beewoolie | It may need some tweaking on the ixp target. |
| 00:20.39 | beewoolie | Yeah. I have no interrupt/DMA capabilities. |
| 00:20.53 | [g2] | this has int capability |
| 00:20.59 | [g2] | but no DMA |
| 00:21.19 | beewoolie | Sure. APEX supports interrupts now, but I don't have much reason to use it. I think I implemented a timer interrupt for grins. |
| 00:23.00 | [g2] | Hmmmm this plot really thickens |
| 00:23.16 | [g2] | To APEX or not to APEX |
| 00:23.33 | beewoolie | One thing at a time. |
| 00:23.36 | beewoolie | Let's look at the patch. |
| 00:31.49 | [g2] | beewoolie I was going to send you the patch for the LE verison |
| 00:32.16 | [g2] | since I added the byteswap to the BE version and that works, I was going to build the LE version with the same patch |
| 00:32.31 | [g2] | and see if I get the identical failure |
| 00:46.42 | [g2] | ok same deal |
| 00:48.34 | [g2] | beewoolie e-mail you the patch |
| 00:49.05 | [g2-lap] | same failure |
| 00:49.09 | [g2-lap] | hda: probing with STATUS(0x00) instead of ALTSTATUS(0x0a) |
| 00:49.09 | [g2-lap] | hda: probing with STATUS(0xa1) instead of ALTSTATUS(0x0a) |
| 00:50.14 | [g2-lap] | which is the line in drivers/ide/ide-probe.c |
| 01:01.31 | beewoolie | [g2-lap]: that message isn't an error |
| 01:01.38 | beewoolie | It's normal for CF ide. |
| 01:02.23 | beewoolie | IIRC, that message occurs because the CF interface cannot perform byte reads. |
| 01:03.00 | [g2] | beewoolie I don't get the message on BE and it works |
| 01:03.22 | beewoolie | I get that message on all of the Sharp targets and they work fine. |
| 01:03.28 | beewoolie | That may be a red herring. |
| 01:04.56 | [g2] | well the probe fails and no IRQ is assigned |
| 01:05.27 | ka6sox-office | [g2] I've seen that on something else that had CF only. |
| 01:06.08 | beewoolie | [g2]: Just because the door-bell rings and you drop a glass on the floor doesn't mean that the doorbell breaks glasses. :-) |
| 01:07.11 | [g2-lap] | hda: probing with STATUS(0x00) instead of ALTSTATUS(0x0a) |
| 01:07.11 | [g2-lap] | hda: probing with STATUS(0xa1) instead of ALTSTATUS(0x0a) |
| 01:07.11 | [g2-lap] | hda: no response (status = 0xa1), resetting drive |
| 01:07.12 | beewoolie | [g2]: I need to spend a little more time with this to make sure of what I might proclaim. |
| 01:07.17 | [g2] | that's the next line |
| 01:07.23 | beewoolie | BTW, who wrote the patch? |
| 01:08.08 | [g2] | it's an adapted patch by th GW kernel guy |
| 01:08.58 | beewoolie | OK. I'll keep my mouth shut until I can read it. |
| 01:14.56 | *** join/#openjtag key2 (n=key2@gob75-2-81-56-64-17.fbx.proxad.net) |
| 01:21.10 | [g2] | beewoolie are the latest APEX changes for IXP42x up on your site ? |
| 01:22.37 | *** join/#openjtag vmaster_ (i=vmaster@p549B5C46.dip.t-dialin.net) |
| 01:29.54 | beewoolie | [g2]: Yeah. The last code changes have been posted. |
| 01:34.59 | [g2] | cool |
| 01:35.51 | [g2] | beewoolie I'll take a shot at modifying it to support the Loft/Intel/Avila boards |
| 01:36.10 | beewoolie | Excellent. |
| 01:36.34 | [g2] | I'll leave the CF/IDE out to start |
| 01:37.05 | [g2] | and IAL/CSR/OSAL networking of course |
| 01:38.10 | [g2] | so it should be pretty straight forward |
| 01:42.55 | beewoolie | [g2]: The CF driver isn't really that hard. |
| 01:43.02 | beewoolie | It's only got a couple of parameters. |
| 01:43.33 | [g2] | beewoolie so it'll be really easy to add then :) |
| 01:43.52 | beewoolie | Yeah. There is a header to be defined in the mach-directory. |
| 01:44.01 | beewoolie | It may need some hacking to get it right for the target. |
| 01:44.18 | beewoolie | For example, there is nothing having to do endian-ness right now. |
| 04:21.28 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
| 04:51.20 | *** join/#openjtag g2_lap_ (n=g2@cpe-066-057-008-035.nc.res.rr.com) |
| 06:31.08 | *** join/#openjtag ka6sox (n=ka6sox@nslu2-linux/ka6sox) |
| 07:18.24 | ka6sox | lennert: ping |
| 07:18.37 | ka6sox | email failed to send to you...(big files) |
| 09:08.24 | *** join/#openjtag bullet (n=bullet@64.50.76.83.cust.bluewin.ch) |
| 11:59.05 | lennert | ka6sox: pong |
| 11:59.18 | lennert | ka6sox: what files? |
| 11:59.30 | lennert | ka6sox: i think my limit is 10MB-ish |
| 12:46.56 | *** join/#openjtag prpplague (n=billybob@72.22.143.75) |
| 12:48.13 | *** join/#openjtag ByronT (i=byron-po@nslu2-linux/ByronT) |
| 15:30.01 | ka6sox | lennert, the design+ supporting files |
| 15:30.16 | lennert | oh :-/ |
| 15:30.18 | lennert | hm |
| 15:30.24 | lennert | can you put it online somewhere and msg me the url? |
| 15:30.34 | g2_lap_ | me too :) |
| 15:31.55 | vmaster | g2_lap_: seems like GDB is planning to do something similar to what you suggested yesterday (if i really understood what you want) |
| 15:33.04 | vmaster | a "target available features interface" |
| 15:39.11 | g2_lap_ | vmaster is that in the GDB roadmap or something ? |
| 15:39.34 | vmaster | on the gdb mailing list |
| 15:40.13 | vmaster | http://sources.redhat.com/ml/gdb/2006-01/msg00257.html |
| 15:41.39 | vmaster | and that's the may posting daniel is refering to: http://sourceware.org/ml/gdb/2005-05/msg00171.html |
| 15:45.00 | g2_lap_ | nod. it does |
| 15:45.12 | g2_lap_ | vmaster thx for noticing and the urls! |
| 15:50.57 | vmaster | g2_lap_: is your loft already being sold? the website seems a bit outdated |
| 15:54.43 | g2_lap_ | vmaster yes it is |
| 15:54.59 | g2_lap_ | it is being sold and the website is out of date |
| 16:01.56 | *** join/#openjtag velinp (n=velinp@c88-163.cable.netissat.bg) |
| 16:41.37 | AchiestDragon | a lot of mail boxes only allow for a 1mb file limit per email |
| 16:54.04 | *** join/#openjtag ByronT (i=byron-po@nslu2-linux/ByronT) |
| 18:24.13 | prpplague | http://www.sparkfun.com/commerce/product_info.php?products_id=606 |
| 18:24.20 | prpplague | any of you guys seen that? |
| 18:24.27 | prpplague | its suppose to be from olimex |
| 18:27.42 | vmaster | yes, it's from olimex |
| 18:28.18 | vmaster | and it uses a ftdi ft232 (not 2232) |
| 18:29.10 | prpplague | interesting |
| 18:29.17 | vmaster | not really sure how the msp430 jtag works |
| 18:29.37 | vmaster | but i suspect they're only shifting data in |
| 18:30.20 | prpplague | hmm |
| 18:30.26 | prpplague | i thought about ordering one |
| 18:32.01 | prpplague | vmaster: i assume it would be the BM chip right? |
| 18:32.41 | vmaster | yeah |
| 18:34.42 | prpplague | interesting, i can't imagin how they have that working |
| 18:37.33 | vmaster | just checking that... |
| 18:44.02 | *** part/#openjtag dwery-away (n=dwery@nslu2-linux/dwery) |
| 18:44.27 | vmaster | there driver packages comes with ftdi's ftd2xx, but i can't find the site where someone mentioned the ft232 |
| 18:44.32 | vmaster | *their |
| 18:44.43 | prpplague | ahh |
| 18:45.49 | *** join/#openjtag dwery (n=dwery@nslu2-linux/dwery) |
| 18:47.32 | vmaster | mhh... their avr usb jtag comes with a ft232bm, maybe i was wrong about the msp430 |
| 18:48.25 | prpplague | vmaster: well i ordered a couple of the ftdi ft2232 chips |
| 18:48.40 | prpplague | vmaster: was planning to tinker around with them with the jtag stuff |
| 18:49.09 | prpplague | vmaster: i've got a dual usb->rs232 adapter with the chip which i've been meaning to document for a conversion |
| 18:51.03 | vmaster | mhh... how much was the usb->rs232? the dlp modules aren't that expensive |
| 18:52.10 | *** part/#openjtag velinp (n=velinp@c88-163.cable.netissat.bg) |
| 18:53.42 | prpplague | vmaster: $40.00 http://www.usbgear.com/USBG-232FT-1.html |
| 18:54.32 | prpplague | hmm looks like they have two new models as well |
| 18:57.19 | prpplague | vmaster: i had plans to leave one serial port, but convert the second to jtag |
| 18:57.37 | vmaster | ah, ok |
| 18:58.35 | prpplague | vmaster: the ft2232 chip uses the same commands as the ft232bm according to ftdi |
| 18:58.52 | prpplague | vmaster: so its possible that they are using the ft232bm driver with the ft2232 |
| 19:02.32 | vmaster | only the ft2232 has the mpsse |
| 19:02.43 | vmaster | and there's only one driver |
| 19:03.01 | vmaster | could be that the msp430 usb jtag uses a 2232 |
| 19:03.28 | prpplague | yea |
| 19:03.46 | prpplague | vmaster: ah i didn't know that the ft2232 has mcu bus emulation |
| 19:04.22 | prpplague | you could program nand flash with those |
| 19:07.52 | prpplague | jeeze, i should have been hacking on the ft2232 stuff instead of the ez-usb chips |
| 19:08.17 | vmaster | yeah |
| 19:10.44 | prpplague | ho ho ho - http://www.usbgear.com/USB-8COMi-RM.html |
| 19:10.44 | prpplague | how about 4 ft2232 chips in one box |
| 19:11.20 | vmaster | hahaha |
| 19:11.41 | vmaster | 320$ |
| 19:11.42 | vmaster | nice |
| 19:12.19 | prpplague | hehe |
| 19:12.34 | prpplague | think i'll hack up my adapter and see if i can get the jtag working |
| 19:47.56 | *** join/#openjtag beewoolie-afk (n=beewooli@206.124.142.26) |
| 19:48.03 | prpplague | bee in the house |
| 19:48.33 | beewoolie-afk | yo yo yo. Let's pump up the volume |
| 19:48.35 | beewoolie-afk | [g2]: guess what I got in the mail toda... |
| 19:48.37 | beewoolie-afk | O |
| 19:48.49 | beewoolie-afk | I've not got 1,2,3,4,5,6 jtag emulators. |
| 19:48.57 | beewoolie-afk | s/not/now/ |
| 19:49.47 | beewoolie-afk | [g2]: Do we have documentation on this one from gateworks? |
| 19:50.04 | prpplague | beewoolie-afk: hehe and you have one more on the way |
| 19:50.14 | beewoolie-afk | Yipee! |
| 19:50.28 | beewoolie-afk | The cool things is that I'll be able to use all of them. |
| 19:50.29 | prpplague | beewoolie-afk: iirc there are shipping next week |
| 19:50.36 | beewoolie-afk | Excellent. |
| 19:51.14 | beewoolie-afk | There's bupkas on the Gateworks site about this device even though it claims to be REV C. |
| 19:52.43 | prpplague | gateworks? |
| 19:52.47 | prpplague | not heard of them |
| 19:52.49 | prpplague | url? |
| 19:52.56 | beewoolie-afk | http://www.gateworks.com/ |
| 19:54.04 | prpplague | is this a board or a jtag device? |
| 19:56.06 | vmaster | they're working on a parallel port dongle that comes with a S3, iirc |
| 19:56.26 | vmaster | might even have jtag on it |
| 19:56.33 | vmaster | g2_lap_ mentioned something about it |
| 19:56.40 | prpplague | ahh interesting |
| 19:58.57 | prpplague | vmaster: http://www.cs.indiana.edu/~bhimebau/ftdi_msp/ftdi_msp.html |
| 19:59.31 | vmaster | yeah, saw that |
| 20:00.07 | prpplague | vmaster: ahh day late and a dollar short |
| 20:42.44 | beewoolie-afk | prpplague: that hurts. |
| 20:43.45 | beewoolie-afk | It's an interesting solution because it shrinks everything down to the two chips. |
| 21:05.01 | [g2] | beewoolie-afk Not yet. There's a little bad new on the device |
| 21:05.20 | [g2] | It's got a CPLD on board not a S3 |
| 21:07.48 | beewoolie-afk | [g2]: How is that bad? |
| 21:08.47 | [g2] | beewoolie-afk as far as adding lots of logic the CPLD doesn't have anywhere near the gate count as the S3 right ? |
| 21:08.55 | beewoolie-afk | Sure. |
| 21:09.02 | [g2] | well I think that's bad |
| 21:09.49 | [g2] | clearly, I think these devices will be a nice set up from the parallel port and possibly the USB intf |
| 21:10.12 | [g2] | but I'm not sure it'll work very long-term |
| 21:12.40 | beewoolie-afk | I'm game to make it work. |
| 21:13.00 | beewoolie-afk | If you can get documentation about their interface, I'll add a driver to my existing codebase. |
| 21:13.22 | vmaster | you might want to have a look at this: http://www.amontec.com/jtag_accelerator.shtml |
| 21:13.49 | vmaster | the documentation is quite good |
| 21:14.59 | beewoolie-afk | vmaster: they mention you. Very nice of them. |
| 21:15.45 | vmaster | well, my openocd is the only software supporting it right now (others are working on it), and some of the ideas come from me |
| 21:16.07 | [g2] | beewoolie-afk cool... I pinged them earlier today |
| 21:17.56 | beewoolie-afk | vmaster: looks to be very well documented. |
| 21:18.43 | vmaster | effective tck rate is about 2mhz on my system, but it all depends on the speed of your parallel port |
| 21:18.54 | beewoolie-afk | vmaster: I like the fact that it tells you the TAP state instantaneously. |
| 21:20.16 | beewoolie-afk | vmaster: are you using EPP? |
| 21:21.06 | vmaster | yeah |
| 21:21.13 | vmaster | an access takes about 1.5us |
| 21:22.31 | [g2] | beewoolie-afk can you point me to where you code is hiding ? |
| 21:22.56 | beewoolie-afk | [g2]: It hasn't been released. I'm still tinkering with it. |
| 21:23.16 | beewoolie-afk | When I get the Xscale driver working, I'll publish what I've got. |
| 21:23.35 | beewoolie-afk | It still needs an interface which will probably look a lot like apex. :-) |
| 21:37.00 | [g2] | beewoolie-afk I should get docs later today |
| 21:37.49 | beewoolie-afk | Excellent. |
| 21:38.52 | [g2] | beewoolie-afk did you open it up yet ? |
| 21:39.01 | beewoolie-afk | No. I can. |
| 21:39.07 | beewoolie-afk | Should I/ |
| 21:39.09 | beewoolie-afk | ? |
| 21:39.14 | [g2] | no need, just wondering |
| 21:39.17 | beewoolie-afk | :-) |
| 21:39.28 | beewoolie-afk | I considered it when it came in, but I've been doing other things. |
| 21:39.40 | [g2] | There's a Xilinx standard header on there so we could reprogram the CPLD |
| 21:39.58 | beewoolie-afk | Hey, that could be cool. |
| 21:40.03 | beewoolie-afk | Is that a jtag header? |
| 21:40.10 | [g2] | but of course :) |
| 21:40.13 | beewoolie-afk | How recursive. |
| 21:40.28 | [g2] | nod |
| 21:40.41 | [g2] | self modify FPGA code :) |
| 21:40.52 | beewoolie-afk | Can the existing code be downloaded? |
| 21:41.26 | [g2] | no I don't think so |
| 21:42.13 | [g2] | They are a little concerned that if everything is available, then ppl could just clone the device etc... |
| 21:42.34 | beewoolie-afk | that's what I'd expect from the CPLD manufacturer. |
| 21:42.38 | beewoolie-afk | It's really a good idea. |
| 21:42.49 | beewoolie-afk | No one would want to put their devices out there. |
| 21:43.18 | beewoolie-afk | I'm no FPGA hacker, so I'm comfortable just using their firmware for the time being. |
| 21:43.40 | [g2] | nod. I'm with you for now |
| 21:43.58 | [g2] | having a higher speed device is what I've been looking for |
| 21:44.21 | [g2] | I'm gonna put JTAG / Serial on some other XScale devices :) |
| 21:49.42 | vmaster | what cpld is on these devices? |
| 21:50.09 | beewoolie-afk | Guess that means I need to open it up... |
| 21:50.17 | vmaster | heh, maybe g2 already knows |
| 21:51.01 | [g2] | XC9572XL |
| 21:51.20 | vmaster | thx |
| 21:51.43 | vmaster | the amontec chameleon has a xcr3128 |
| 23:23.32 | beewoolie-afk | This indiana U schmatic isn't as interesting as I though. |
| 23:24.23 | beewoolie-afk | s/though/thought/ |