**** BEGIN LOGGING AT Thu Feb 23 10:59:56 2006 Feb 23 11:02:53 Ok the intel document number is 29066721 Feb 23 11:02:57 I can mail it to you too.... Feb 23 11:02:58 but from the ep9301 data it can boot from ether serial , eeprom , async flash or sram/rom , sync flash or sdram or spi Feb 23 11:03:38 no point , theres better devices arround Feb 23 11:05:12 like why have a 32mb device when we could have a 512mb Feb 23 11:06:58 debian arm fits in a 256mb c/f , could have it loaded onto the onboard flash Feb 23 11:07:14 really? Feb 23 11:07:25 I guess for the basic package set. Feb 23 11:09:21 they do 1G devices also , so with 2 chips thats a 2G block although looks like there may be an address limit of 256mb per device on the ep9302 Feb 23 11:10:43 64mb if async mode is used Feb 23 11:14:12 AchiestDragon: my ts7250 has an 8bit NAND flash part Feb 23 11:14:44 AchiestDragon: okay, i don't know how else they do that then, they must be using either the CSx, the spi rom or the serial method for booting Feb 23 11:14:53 AchiestDragon: and the first block of NAND flash does have the CRUS header in it Feb 23 11:15:29 then there booting it direct Feb 23 11:15:46 its just read as a memory location and booted Feb 23 11:15:48 but the cpld doesn't connect to all the address lines, iirc Feb 23 11:17:18 but the ep9301/2 devides the memory up into 16 blocks that some are provided as ce outputs for a unique address space Feb 23 11:18:12 so all the 64mb address space is avalable to each device Feb 23 11:19:06 the cpld only connects to A25, A24, A23, A22 Feb 23 11:19:09 the sdram is mux'ed ant theres provision for 4 sync devices so 256mb per sync device is posible Feb 23 11:19:14 and you can't boot directly off the nand flash Feb 23 11:19:23 so there must be some other way they're doing it Feb 23 11:20:35 z3ro: ok, let me find those files Feb 23 11:21:00 lennert: do you want me to upload example files? Feb 23 11:21:13 z3ro: example files for what? Feb 23 11:21:35 a ngc file, and the decrypted version. Feb 23 11:21:54 i'm not sure if i want to go there :) Feb 23 11:22:02 or were you talking about different files? Feb 23 11:22:08 * z3ro isn't sure what you meant. Feb 23 11:22:21 the files that describe the .bit bits i found so far Feb 23 11:22:28 oh, okay Feb 23 11:22:35 i should have them somewhere here.. *looks* Feb 23 11:24:42 i used the tools at http://svn.wantstofly.org/vhdl/utils/ for diffing .bit files Feb 23 11:25:01 there's a file that'll decode the bitstream and a simple shell script that can diff two decoded dumps Feb 23 11:25:28 okay :) Feb 23 11:25:35 can't find the other files yet Feb 23 11:25:44 let me know when you do. Feb 23 11:27:19 ok before i get any further with the ep board theres a few points that i could do with some feedback on Feb 23 11:28:30 the spec so far is 4 sdram devices ( 64mb seems to be the biggest type i can get ) so 256mb ram Feb 23 11:30:37 although should support 256mb devices and 4 would give 1Gb , need to study the ep9302 data a bit more to see the size limit of them Feb 23 11:30:48 z3ro: i put some stuff in http://www.wantstofly.org/~buytenh/xilinx/ Feb 23 11:31:00 z3ro: i think there was more but no idea where i put it.. Feb 23 11:31:09 lennert: okay Feb 23 11:31:28 z3ro: for the xc3s50 it has which frames contain the keeper/driveattrbox/ioattrbox/slew bits Feb 23 11:31:41 z3ro: pull up/pull down/io standard/drive strength Feb 23 11:31:50 lennert: 403 on http://www.wantstofly.org/~buytenh/xilinx/ Feb 23 11:31:55 z3ro: the bram as well Feb 23 11:32:01 oh, a sec Feb 23 11:32:08 fixed Feb 23 11:32:08 2 flash chips at 64mb each , would like to suport 512mb total but that may not be posible Feb 23 11:32:22 lennert: okay, thanks. Feb 23 11:32:23 all the iob bits are in there Feb 23 11:32:44 i'm sure i had the clb bits decoded too Feb 23 11:32:58 and bits of the routing Feb 23 11:33:13 the easiest way to go about it is via xdl files Feb 23 11:33:22 write an xdl file, convert it to ncd, and bitgen it Feb 23 11:33:30 change the xdl slightly, same thing again, and bit-diff it Feb 23 11:33:46 hmm, I see. Feb 23 11:34:20 i haven't had time to work on this lately at all Feb 23 11:34:33 okay, well, this information should be helpful. :) Feb 23 11:34:47 the expantion bus is a 50mhz bus , and suport is provided for etherent , 1 rs232 , 2 usb , and a 16 bit dio port , theres a full 14 pin arm jtag for the ep9302 and the other ports are brought to a connector so if the user wants he can get to them Feb 23 11:34:55 good luck Feb 23 11:35:02 btw, do I have permission to reference this work or include information from the files, and do you want credit or to remain anonymous? Feb 23 11:35:46 any requests for extras or thougts on the spec Feb 23 11:36:08 z3ro: feel free to incorporate everything, but please keep me anonymous Feb 23 11:36:25 lennert: okay, absolutly. Feb 23 11:36:26 z3ro: the files are under the do whatever you want with them license Feb 23 11:36:35 okay Feb 23 11:36:43 AchiestDragon: sounds good, not sure what else to add Feb 23 11:38:24 the fpga board for is next , and an expation board that will give CF, ide and video, audio and ps2 is planned later Feb 23 11:38:55 fpga board for the jtag /anlizer Feb 23 11:41:29 is wondering atm if i can use a cpld to frig the ep9302 into using the sdram timings to let us use ddr rams then may be able to get more ram using less chips Feb 23 11:56:33 :S looks like the ram is limited to 4 64mbyte blocks so 256mb ram is going to be the max Feb 23 11:57:16 and 128mbytes of flash with the 2 chips Feb 23 11:57:22 that enough ? Feb 23 11:58:49 theres also 128mb of ddr ram on each fpga card and alowance for 64mb of video ram Feb 23 12:00:29 btw we have 2 external dma chans on this that the ts72xx did not impliminet Feb 23 12:17:52 lennert: are you interested in the xilinx-xdb v1.4e format, used in the ngo and ngd files? I put some example files and diff's online, but I'm having a hard time deciphering the format of the decrypted content. thought maybe you might want to look at it. Feb 23 12:18:15 the string table is easy to figure out, but the rest is less obvious. Feb 23 12:20:05 z3ro: no, please -- i'd like to stay out of that Feb 23 12:21:06 lennert: hmm. I thought you would want to crack these formats, since you're already working on the 'configuration bits' problem... :( Feb 23 12:21:20 lennert: is it some legality thing? Feb 23 12:21:37 yeah Feb 23 12:21:50 hmm :( Feb 23 12:22:03 i believe that i'm in the clear as long as i just try to find out what bits do what in the .bit file, but i'm not so sure about de_crypt_ing their file formats Feb 23 12:22:21 decrypting = circumventing protection measure, yadda yadda Feb 23 12:22:56 afaik, it's legal to reverse engineer a file format in order to find out what information it contains. Feb 23 12:23:25 I think it's only illegal if it's patented. Feb 23 12:23:41 but I'm not a lawyer, I've just read this online somewhere. Feb 23 12:25:22 guess if you read the data from the chip and store it in your own format then theres not much they can do about it Feb 23 12:25:47 unless the data beeing read is theres Feb 23 12:26:47 well, I'm just stuck on figuring out one of their file formats. The file is "encrypted", but it's just xor encryption, so it's not like they wanted it to be secure... Feb 23 12:26:54 the decoded content is what I'm stuck on. Feb 23 12:27:09 first part is a string table, but the rest is very strange. Feb 23 12:27:22 it's all ascii (the decoded content) Feb 23 12:27:55 i would concentrate on the xdl file format Feb 23 12:28:05 xdl is all ascii representation of an ncd file Feb 23 12:28:11 xilinx provide tools to convert to/from ascii Feb 23 12:28:15 and ascii is human-readable Feb 23 12:28:20 so you're in the clear there Feb 23 12:28:30 I'd like to avoid their tools... Feb 23 12:28:56 i think you'll have to use their tools at least initially Feb 23 12:29:08 unless you can have someone else do the 'cleanroom' work Feb 23 12:29:15 take a look at z3ro.geek.nz/fpga/xdm/0_foobar.ngo.indent (an indented version of the decrypted content) Feb 23 12:29:51 that's the data inside a .ngo file. Feb 23 12:30:21 it's the text after the first 4 lines that I'm stuck on Feb 23 13:48:03 lennert: Do you know where I can add the static partitions? Feb 23 13:50:42 NAiL: either command line or in the mtd map driver Feb 23 13:50:48 NAiL: not 100% sure what the command line format is Feb 23 13:51:17 mtdparts=flash00:128k(u-boot),128k(env),128k(spare),7808k(jffs2) Feb 23 13:51:29 that's what my cmdline looks like Feb 23 13:52:20 lennert: are you going to be at the pre-fosdem beerdrinking? Feb 23 13:54:25 hmmm beer Feb 23 13:54:39 Beer Is Good (tm) Feb 23 13:55:20 vmaster: yes, in roy de l'espagne on friday Feb 23 13:55:30 Sleep is good 9tm) Feb 23 13:55:39 dyoung-zzzz: nite Feb 23 13:59:29 nite Feb 23 13:59:37 * prpplague needs beer Feb 23 13:59:48 prpplague: come to fosdem Feb 23 14:00:24 wish i could, busy on too many projects and short on $$ this month Feb 23 14:00:37 too bad.. Feb 23 14:00:44 yea Feb 23 14:00:55 lennert: i used to have a large budget for conferences Feb 23 14:01:00 lennert: at my last company Feb 23 14:01:19 lennert: then they go some new management and they deemed it a waste of money Feb 23 14:01:33 lennert: no training, no conferences, no certs Feb 23 14:01:47 lennert: they had 20+ people quit on the same day Feb 23 14:01:50 haha Feb 23 14:01:55 out of how many? Feb 23 14:02:03 300 Feb 23 14:02:24 probably the smartest 20 as well Feb 23 14:02:25 over the next month they lost a total of about 75 more ppl Feb 23 14:02:43 including myself, hehe Feb 23 14:03:18 i'm sure the management proudly advertised their 33% staff reduction Feb 23 14:03:33 to show their cost-cutting methods were effective Feb 23 14:03:35 lennert: yea thats how they played it up Feb 23 14:03:53 to CEO: "Oh, you didn't want us to cut productivity by 100%? You didn't tell us.." Feb 23 14:04:07 lennert: i found out later that the vp that made the decision about the conferences got a giant bonus for reducing costs Feb 23 14:04:27 company went down the crapper yet? Feb 23 14:04:34 lennert: yea they are still in business, but the moral is horrible Feb 23 14:04:42 serves them right Feb 23 14:04:46 lennert: most of the ppl working there are "rejects" Feb 23 14:04:56 that sucks for them Feb 23 14:05:10 the central IT institution of Leiden University is similar Feb 23 14:05:21 management was replaced by idiots one day Feb 23 14:05:29 all the people that had no trouble finding something different quit Feb 23 14:05:36 hehe Feb 23 14:05:39 so now they only have people that are stuck there because they are incompetent Feb 23 14:05:59 so whatever you want to do here, you have to deal with incompetent idiots Feb 23 14:05:59 lennert: yea the management was replace all with sales people Feb 23 14:06:08 jeeze Feb 23 14:06:42 (OT, www.stfu.se/stfuse.swf is fun :) Feb 23 14:06:44 lennert: the company i work for now has never really had a budget for conferences Feb 23 14:07:01 lennert: i'm trying to get them to see the benefit of them Feb 23 14:07:51 lennert: she singing something? Feb 23 14:07:57 yeah Feb 23 14:08:14 it's a lot less fun without audio Feb 23 14:08:35 lennert: i'll have to play it later i have my audio card disable right now due to some testing i was doing Feb 23 14:08:42 kay Feb 23 14:22:09 lennert: hehe you'll have to drink some good belgian abbey ale for me Feb 23 14:22:16 lennert: at fosdem Feb 23 14:24:00 okay :) Feb 23 14:25:41 lennert: maybe a Duvel or Chimay Feb 23 14:25:54 mmm duvel Feb 23 14:26:58 lennert: i got this nice set of 4 tulip duvel glasses with a large 1500ml bottle once Feb 23 14:27:18 lennert: my #$%#%^$#%^ ex-wife put them in the dishwasher and they all got broken Feb 23 14:27:54 whoops Feb 23 14:28:11 i can see why you divorced her ;) Feb 23 14:28:14 hehe Feb 23 14:28:16 inexcusable Feb 23 14:28:35 lennert: hehe no i divorced her because she was a bitch, hehe Feb 23 14:43:57 thinks the 2 go together normaly Feb 23 14:45:19 AchiestDragon: 'ex-wife' and 'bitch'? surely :) Feb 23 14:45:29 AchiestDragon: 'ex-wife' and 'sweetheart' seems unlikely Feb 23 14:45:36 she and bitch Feb 23 14:45:40 oh :) Feb 23 14:45:54 well, 'he' and 'bitch' don't Feb 23 16:06:14 <[g2]> execept for transgenders Feb 23 16:06:30 <[g2]> lmao Feb 23 16:08:23 hehe Feb 23 16:23:18 yay, fixed the bugs in xprog.c :) Feb 23 16:23:31 (actually reading the state diagram helps. :p) Feb 23 21:43:21 ~seen beewoolie-afk Feb 23 21:43:26 beewoolie-afk was last seen on IRC in channel #openjtag, 1d 16h 12m 30s ago, saying: 'On one device I've looked at, the largest packet is 128 bytes.'. Feb 24 08:14:34 lennert: ping? Feb 24 09:55:39 pong Feb 24 09:55:49 NAiL: pong, that is Feb 24 09:55:55 yesh Feb 24 09:56:33 I can't figure out the darn partitions on the DS101. There seems to be some "stuff" at the start of the flash, so redboot doesn't start at offset 0 Feb 24 09:56:53 right, that's possible Feb 24 09:57:02 there's probably some small pre-bootloader there Feb 24 09:57:11 but the partition table should be at the end of the flash Feb 24 09:57:13 in the last block or so Feb 24 09:57:30 FLASH: 0x90000000 - 0x92000000, 256 blocks of 0x00020000 bytes each. Feb 24 09:57:34 FIS directory 0x91FE0000 0x91FE0000 0x00020000 0x00000000 Feb 24 09:57:52 right, it doesn't have a redboot partition table, you said earlier Feb 24 09:58:59 I'm not sure that raw flash access is safe ;)' Feb 24 09:59:34 the (i)NFTL stuff puts some partition stuff on the flash, if I understand it correctly. Feb 24 09:59:59 but that is apparently not read by the 2.6-kernel Feb 24 10:00:34 isn't there an #mtd channel on freenode? Feb 24 10:00:53 yeah, I'm there Feb 24 10:01:01 right :) Feb 24 10:01:17 It's just really, really slow ;) Feb 24 10:01:38 last time I asked, it took three days before anyone said anything :-P Feb 24 10:01:59 i'm still waiting for comments on my physmap rewrite Feb 24 10:02:23 slow indeed Feb 24 10:02:29 what was that? I saw that mail go past, I think Feb 24 10:02:42 it rewrote the physmap mtd map driver as a platform device Feb 24 10:02:59 the current problem with physmap is that if you compile it in, you have to use it, and you can't instantiate two or three Feb 24 10:03:14 so, if it's a platform device, you can either instantiate 0, 1, 2 or as many as you like Feb 24 10:04:28 ah Feb 24 10:04:33 (I think) Feb 24 10:04:35 ;) Feb 24 10:04:42 hehe :) Feb 24 10:05:16 They're very interested in those bit error rates though :-P Feb 24 10:05:25 yeah Feb 24 10:05:28 and to OOB or not to OOB Feb 24 10:05:40 and whether yaffs++ or jffs++ **** ENDING LOGGING AT Fri Feb 24 10:59:56 2006