00:31:50 | Sammy | morning all... |
04:07:59 | Russ | morning |
08:17:21 | Sammy | evering ... |
08:17:40 | Sammy | erikm :-) |
08:17:43 | erikm | hi all |
08:24:42 | Sammy | My God, another typhoon come again ...;-( |
08:28:11 | erikm | oops, storm season over there? |
08:31:22 | erikm | gets a disk |
08:42:29 | Sammy | I don't know , maybe the nature are punish us too give so many thing's and waste so many things from him ... |
08:43:00 | Russ | its just the weather |
08:43:24 | Russ | not the first, nor the last time |
08:44:19 | Sammy | but the storm always come in Jul and Aug and now ...is october ... still have typhoon so strange ... |
08:47:39 | Russ | you just need to look at the slightly more distant past |
08:56:38 | Sammy | hope not became another venice ....;-| |
08:57:26 | erikm | Sammy: nothing unusual. just normal statistical variations |
11:26:49 | sammy_ | need to come home and hide the typhoon ... |
13:10:06 | prpplague | morning |
14:52:24 | prpplague | erikm: you alive, got a quick question... |
15:05:47 | erikm | wakes up |
15:09:35 | prpplague | erikm: i guess i'm blind, but i can't seem to locate where the struct tag is defined |
15:14:21 | erikm | in kernel |
15:14:28 | erikm | include/asm-arm/setup.h |
15:15:57 | prpplague | thanks |
15:28:17 | erikm | I even documented the kernel sources nicely |
15:28:52 | prpplague | erikm: i noticed |
15:31:56 | prpplague | erikm: as i understand it, the boot params "could be" hardcoded into the kernel, however the ultimate goal is not to have hardcoded arch/platform specific infor in the kernel, and that the bootload should be responsible for passing the correct boot params along to the kernel |
15:32:23 | erikm | exactly |
15:34:01 | erikm | the only thing that the kernel still has hardcoded is the default command line, but that gets overwritten as soon as it finds an ATAG_CMDLINE structure |
15:35:03 | erikm | btw, this is exactly how the LART worked before blob-2.0.3: everything was hardcoded in the fixup() function, but as soon as blob-2.0.3 got proper command line support I completely removed the fixup() function |
15:35:15 | erikm | (fixup() function in kernel) |
15:35:58 | prpplague | gotta ya |
15:37:29 | prpplague | i see that in some of the other platforms that apparently are still using bootloaders that don't support commandline |
15:38:50 | prpplague | now theorecticaly i could write a set of tags to a flash location have have the kernel look at the flash location for the tags? |
15:41:29 | erikm | no |
15:41:34 | erikm | ehm, yes |
15:42:35 | erikm | but that wouldn't give you much flexibility |
15:43:17 | prpplague | i.e. bad bootload, again the real goal would be a quality bootloader that handled the correct commandline |
15:43:27 | erikm | for example: if a plug a 64MB memory board to my LART, blob detects it and issues an extra ATAG_MEM node so the kernel knows there is 96MB available |
15:45:58 | prpplague | erikm: for me, its sometimes good to understand whats bad about a way of doing things, to really appreciate the direction of development |
16:13:31 | prpplague | is away: lunch |
17:03:24 | prpplague | is back (gone 00:49:52) |
18:59:52 | prpplague | welcome back |
19:00:20 | prpplague | hope you've been hard at work shipping tuxscreens BZFlag |
19:00:46 | BZFlag | heh. didn't ship any today. Though there are already orders for tomorrow. |
19:02:34 | prpplague | BZFlag: get an order from "Dave Anders" ? |
19:05:18 | BZFlag | what method and what date? |
19:05:59 | prpplague | i sent the money order off last wed. |
19:06:03 | prpplague | via usmail |
19:07:10 | BZFlag | I don't have the mailed ones on record here. I'd have to look at home. |
19:07:34 | BZFlag | send me an email with the date you sent it and the name and shipping address. I'll look it up. |
19:07:52 | prpplague | BZFlag: ok thanks |
19:09:31 | prpplague | sales@tuxscreen.net ?? |
19:10:01 | BZFlag | or tim@rikers.org, either is fine. |
19:14:15 | prpplague | on the way |
20:07:16 | BZFlag | are the pins for the jtag dongle labeled anywhere? |
20:10:53 | erikm | BZFlag: the LART jtag dongle schematics are available on the LART site |
20:20:48 | BZFlag | yes, but they do not include pin labels, just pure wiring. |
20:21:29 | BZFlag | I found the jtag connector info in the lart schematics, but I'm not sure which end is pin 1 |
20:31:53 | erikm | lemme check on a real lart |
20:33:32 | Russ | isn't there a key on pin 2 or something |
20:33:37 | erikm | pin1 = pin8 = GND |
20:33:49 | erikm | so at that point nothing will go wrong :) |
20:34:29 | erikm | ok, put the dongle on the table, D25 connector pointing to you |
20:34:56 | erikm | and the largest number of pins on the D25 on the upper side |
20:35:13 | erikm | (which should also mean that the components are on the upper side) |
20:35:35 | erikm | then and only then, pin 1 of the JTAG connector is the left most pin |
21:33:20 | prpplague | Russ: you around for a question? |
21:53:14 | Russ | yah |
21:57:19 | prpplague | when allocating the ram location for the initrd image, how are the locations determined, i see that the lart is loading at 0xc080000 while the shannon is being loaded at 0xc800000, while the default is 0xc040000 |
21:59:12 | Russ | a place where it fits |
22:05:04 | prpplague | so there is no major logic in deciding, just alocation that doesn't stradle banks, and has room? |
22:34:54 | Russ | prpplague: right |
22:46:03 | prpplague | Russ: one last question for today, can you give me some idea of the role KERNEL_RAM_OFFSET and RAMDISK_RAM_OFFSET play? |
23:30:33 | prpplague | later |