**** BEGIN LOGGING AT Tue Mar 13 02:59:56 2007
Mar 13 03:17:01 <AchiestDragon> burst speeds of 6 to 10 mhz maybe posible but to get it sustained at that would not be ,, , and even shourt bursts could have interuptions that delay it
Mar 13 03:19:26 <AchiestDragon> i have the other problem with the geep-b , in that the ram bandwith is 100mhz for burst access thats 1 16bit read ,, so in actual fact it would be posible to achive a 1.6Ghz but the fpga is not fast enough to cope with that 
Mar 13 03:21:47 <AchiestDragon> it will do 125mhz ,, and by using a cpld it could be posible to dubble that and get 250mhz but as long as it can do 60mhz jtag clock  that should be fine 
Mar 13 03:24:13 <AchiestDragon> at the moment its down to coding ,, im working on the avr supervisor program for it atm and still the host code to start on , as well as the vhdl and code for any softcore 
Mar 13 03:24:57 <AchiestDragon> the avr code is beeing written in asembly 
Mar 13 10:37:27 <aproemel> hey vmaster :-)
Mar 13 10:40:38 <vmaster> hey aproemel 
Mar 13 10:41:16 <aproemel> Do you think 2 weeks u-boot 5 weeks kernel to start init is to less?
Mar 13 10:41:47 <vmaster> it could work - of course you're on a tight schedule, but you already know that i guess
Mar 13 10:42:04 <vmaster> most of the porting is already done, and the board specific stuff is usually not that much
Mar 13 10:42:30 <aproemel> yes, but i have to tell my boss in a few hours, how i plan to do my work
Mar 13 10:42:44 <vmaster> especially since you don't need to support anything but memory + serial
Mar 13 10:43:26 <aproemel> is it possible to boot a kernel like on a x86 arch ?
Mar 13 10:43:34 <aproemel> so that the kernel boot itself ?
Mar 13 10:45:09 <vmaster> you mean once you have a running kernel, you boot a new one?
Mar 13 10:45:35 <vmaster> i've heard that this works on arm, too, but never tried it, and don't know what's necessary for it to work
Mar 13 10:46:33 <aproemel> i mean, that i write the kernel at an address an boot it, like you write it at the start of a harddisk and then restart the x86 arch
Mar 13 10:46:45 <aproemel> then the kernel boots itself, without any bootloader
Mar 13 10:47:13 <vmaster> well, on x86 you have the BIOS, which does everything a bootloader does
Mar 13 10:47:24 <aproemel> i know
Mar 13 10:47:53 <vmaster> heh, yeah, then if you have a BIOS on your target, you can do without a bootloader ;)
Mar 13 10:47:59 <aproemel> if i upload it with my loader thats inits ram, serial and the clocks, and jump to my kernel, do you think that'll work ?
Mar 13 10:48:12 <vmaster> the kernel needs to be passed some arguments
Mar 13 10:48:27 <vmaster> machine id, board id, and ATAG parameters with memory layout and cmdline
Mar 13 10:48:46 <aproemel> what does the kernel without these informations?
Mar 13 10:48:58 <vmaster> it fails booting
Mar 13 10:49:07 <aproemel> does an output appear?
Mar 13 10:49:17 <vmaster> no - at that time, there is /nothing/
Mar 13 10:49:30 <vmaster> you can add a "shim" that initializes this information
Mar 13 10:49:43 <vmaster> people do that when their target has a bootloader that fails to pass the correct information
Mar 13 10:50:15 <aproemel> :-( it was only an idea, to test my kernel build on my test board, without flash
Mar 13 10:50:51 <vmaster> you could use the embest u-boot - you already have that working, don't you?
Mar 13 10:51:49 <aproemel> mhh, i have to build the kernel new, cause board ID
Mar 13 10:52:40 <vmaster> yeah, you can easily fake that in the kernel
Mar 13 10:54:07 <aproemel> i'm currently run from one into another meeting :-(
Mar 13 10:55:51 <aproemel> today 2 meetings, tomorrow 1, yesterday 1
Mar 13 10:56:09 <vmaster> ouch
Mar 13 10:56:13 <aproemel> I spent more time to work for these meetings, that on my work :-(
Mar 13 10:56:17 <vmaster> hard to get any work done this way
Mar 13 10:56:19 <vmaster> yeah
Mar 13 10:57:10 <aproemel> i'm working 7-15:30 without any break, then drive home, eat something and have to correct my written text of the morning ^^
Mar 13 10:57:52 <vmaster> heh, during the final phase of my diploma thesis i worked from 10:00 to 24:00+ ;)
Mar 13 10:58:10 <aproemel> 10+ ^^ i get up at 5:30 every day
Mar 13 10:58:17 <aproemel> drive 1hour
Mar 13 10:58:25 <aproemel> and 1 hour back home
Mar 13 10:59:27 <aproemel> i hope, that if i get my hardware, that it works correctly
Mar 13 11:00:28 <aproemel> because i don't want to debug hardware a few weeks first, if i had to do so, everything isn't possible to get realized :-(
Mar 13 11:05:40 <aproemel> in 45min i have to go to the meeting
Mar 13 11:14:36 <rwhitby> a good arm shim can be found on marc singer's site
Mar 13 11:15:27 <rwhitby> ftp://ftp.buici.com/pub/arm/arm-kernel-shim/
Mar 13 11:15:45 <aproemel> thnx rwhitby 
Mar 13 11:16:06 <rwhitby> (we use that for the ixp4xx boards that we support in nslu2-linux.org)
Mar 13 11:16:18 <rwhitby> (where the vendor's bootloader is broken)
**** ENDING LOGGING AT Wed Mar 14 03:00:00 2007