**** BEGIN LOGGING AT Tue Jun 26 02:59:57 2007 Jun 26 04:45:41 hi, rwhitby. gonna try some slug testing now (that I have time) Jun 26 04:46:36 cool - I'm around at the moment - got a meeting in 15 minutes for up to an hour, but free after that too for a couple of hours Jun 26 04:47:13 ok thanks... Jun 26 04:54:27 when you get a chance, could you recommend an 8MB or 16MB image for booting (slugimage). Jun 26 04:57:51 RobNC: what do you have built there now? Jun 26 05:00:13 rwhitby: I still have the build env so I could make the 16mb image I guess. Jun 26 05:00:58 did we determine if redboot looks at 8MB for trailer, or at end of flash? that's the first question. Jun 26 05:02:02 I thought it was more a question as to whether linux (slugos) looks at 8MB for trailer, is that right? is the intent to test flash or ram? Jun 26 05:02:20 linux is not a problem. the problem is the redboot test for the secomm trailer Jun 26 05:02:47 just flash a normal 8MB image, and see what happens :-) Jun 26 05:02:54 w/o mods to flash, it looks as though redboot finds the 16MB just fine but as to RAM, that would be the best question Jun 26 05:03:05 if that fails to boot, replicate the same image in the upper 8MB and boot that Jun 26 05:03:26 I think I did that already, but I gotta get rid of 8MB trailer Jun 26 05:03:26 RedBoot finds 16MB, but may not boot linux due to not finding the trailer signature Jun 26 05:03:43 you won't tell anything until you run the "boot" command in redboot Jun 26 05:03:56 just booting redboot itself doesn't say anything about the trailer Jun 26 05:04:29 (since it's the SerComm-added 'boot' command in RedBoot which looks for the trailer, and refuses to load the kernel if it doesn't find the trailer) Jun 26 05:04:32 gotcha. so I should be able to get just a generic image and slugimage it Jun 26 05:04:36 yep Jun 26 05:04:41 and try and boot linux Jun 26 05:05:05 if the 'boot' command fails and puts you in upgrade mode, then flash two 8MB images one after each other and try again. Jun 26 05:05:31 back in about an hour - if you can pastebin the results of that it would be a good step forward. Jun 26 05:05:38 ok I see. I can do that quite easily but the JTAG is a pain but not impossible. Thanks see you then Jun 26 05:06:09 you should be able to do all this with load over serial or tftp Jun 26 05:06:12 no jtag required Jun 26 05:06:18 (since you've already got redboot on it) Jun 26 05:06:29 gotcha (yes no overwriting redboot) Jun 26 05:06:34 bbiab Jun 26 05:09:29 * rwhitby-n800 watches RobNC's progress ... Jun 26 06:23:03 rwhitby: need help with redboot param "mfill"... Jun 26 06:25:21 wait - lemme try something else Jun 26 06:29:34 dunno about mfill Jun 26 06:30:05 its ok - think I figured it out. gotta fill ram, then fis write Jun 26 06:42:28 rwhitby why would serial trailer from di-nslu2 be different than one from JTAG capture of OEM flash? Jun 26 06:44:05 dump -b 0x507ffff0 -l 0x10 => 0001 0000 2363 90f7 6552 634f 6d4d 00b4 (this is from di-nslu2.bin) Jun 26 06:44:21 same command on JTAG capture of OEM flash from Linksys shows: Jun 26 06:45:27 0001 0000 2373 90f7 6552 634f 6d4d 00ea Jun 26 06:48:08 different product version Jun 26 06:48:13 normal Jun 26 06:48:26 ok good to be sure Jun 26 06:48:40 did it boot linux? Jun 26 06:50:03 yeah it booted di-nslu2 (debian) with 8mb flash in lower and 16mb flash duplicate of 8mb. erasing serial trailer now - about to try to boot as soon as flashing done Jun 26 06:51:31 nope.... says "not have eRcOmM" Jun 26 06:52:28 I just filled 0x507ffff0-> 0x507fffff with FFFF Jun 26 06:52:42 I meant 0xffffffff Jun 26 06:53:09 ok, that means it's looking at end of flash for trailer Jun 26 06:53:38 should be at end of 16MB right? I verified it is there Jun 26 06:53:50 yes Jun 26 06:54:18 which means we can have a big rootfs in the middle Jun 26 06:54:35 true, so it won't look beyond 8MB Jun 26 06:54:47 ? Jun 26 06:54:49 ok... gonna copy back the trailer Jun 26 06:55:00 I mean it won't look beyond 8MB for flash trailer (sercomm) Jun 26 06:55:13 oh, that's bad Jun 26 06:56:20 where is it looking for the trailer? 8mb or 16mb? Jun 26 06:57:11 it does recognize size correctly tho... Jun 26 06:57:46 it doesn't say where it is looking for trailer Jun 26 07:00:01 I verified flash at 0x507ffff0 -l 0x10 has all FFs and 0x50fffff0 -l 0x10 has the eRcOmM Jun 26 07:05:24 and it boots linux with 507ffff0 with ffs and 50fffff0 with eRcOmM ? Jun 26 07:05:27 RobNC: ^ Jun 26 07:06:07 I'm trying to verify whether the 'boot' command in RedBoot (which is a SerComm-added custom command) is looking at 8MB-16 or 16MB-16 for the trailer Jun 26 07:06:57 looks to me like if 0x507ffff0 has eRcOmM then it boots. If it has FFs then no boot Jun 26 07:08:14 ah, that's the worse of the two alternatives then Jun 26 07:08:39 I was hoping that since RedBoot recognised the whole of ram, then the sercomm boot command would use that to find the trailer. Jun 26 07:09:02 but it looks like the sercomm boot command is just looking at a fixed offset (which doesn't surprise me at all) Jun 26 07:09:22 wait - redboot detects whole of flash not ram :-( Jun 26 07:09:34 sorry, flash not ram is what I meant Jun 26 07:10:13 redboot recognises all of flash, but sercomm looks for the trailer just before the 8MB mark. is that correct? Jun 26 07:11:00 seems like you are right rwhitby Jun 26 07:11:13 trying again just to be sure. I have a teraterm log of the whole thing Jun 26 07:11:41 we can probably just do a binary edit on redboot to fix it. Jun 26 07:12:10 yeah that is true - shouldn't be so hard to find if just changing address it is looking at - binary diff Jun 26 07:16:33 yup just verified it again... Jun 26 07:16:38 just as I said above. Jun 26 07:17:58 ok, not suprising Jun 26 07:18:15 to be otherwise would have required forethought on the part of sercomm engineers. Jun 26 07:27:07 sorry I stepped away. gotta zz now time is 330 Jun 26 07:27:27 what other tests can u think of Jun 26 07:28:23 at least Apex sees the 16MB of memory too Jun 26 07:36:22 http://pastebin.ca/589872 Jun 26 07:46:08 http://pastebin.ca/589879 - updated (backspaces were shown as spaces) Jun 26 07:46:20 ttyl.... Jun 26 13:36:49 03oleo * r6304 10optware/trunk/sources/uclibcnotimpl/ (. math.c): uclibcnotimpl: add math.c Jun 26 13:38:00 03oleo * r6305 10optware/trunk/ (3 files in 2 dirs): transmission: r1874->r2199 - now sequential file checking Jun 27 01:02:02 I know this isnt the place to ask questions but the general is dead. I plugged my flash disk (to be unslung) and formatted it in Port 2 but it is now recognized as Disk 1 if I unsling to this drive even though its in Port 2 but considered Disk 1 will my actual Port 1 still have NTFS capabilities **** ENDING LOGGING AT Wed Jun 27 02:59:56 2007