**** BEGIN LOGGING AT Tue Mar 05 02:59:58 2013 Mar 05 12:40:07 hi guys, about kexec purgatory, is it just necessary/used in kdump or it is used in a reboot withoug kdump thing? Mar 05 15:51:00 kirotawa: ant_work may know this :) Mar 05 15:51:07 (or may not) Mar 05 15:51:24 hehe, Jay7 tks. Mar 05 15:51:37 this channel is about kexecboot not just kexec (but we are using and building it) Mar 05 15:52:51 Jay7: yep, I first thought here is about kexec only, but since you are working with it I supossed that you have that knowledge about kexec, :) Mar 05 15:58:07 hi Mar 05 15:58:31 afaik the purgatory is used in both kexec-only and kdump cases Mar 05 15:58:47 is there you'll store the new kernel, there are checks Mar 05 15:59:21 so is correct to say that purgaroty is the last phase before it jumps to the new kernel, right? Mar 05 16:01:22 yes, in some cases purgatory reinitializes the hardware if necessary then jumps to start kernel Mar 05 16:01:39 hum, nice to know. Mar 05 16:01:48 thanks a lot ant_work Mar 05 16:01:51 kexec_load does specify the address Mar 05 16:02:21 afik the relocation is automatic Mar 05 16:02:36 if th ekernel doesn't like those addresses Mar 05 16:02:38 hum, now that you talked about address. that kernel_load_address in do_bzImage_load, Mar 05 16:03:12 that address is in that point where is the entry point, kernel entry point? Mar 05 16:04:16 I've seen here that it is constant, always 0x1000000. Mar 05 16:05:18 that it pass to regs32.eip as kernel_load_address (entry point). So, I supposed that it load the kernel always there and after kernel relocate itself. Mar 05 16:08:00 found the easy thread: http://lists.infradead.org/pipermail/kexec/2012-June/006406.html Mar 05 16:11:19 nice. Mar 05 16:19:20 so, is correct to afirm that in purgatory I can grab that address, that entry point address. Mar 05 17:01:55 bluelightning: I have had an enlightement about colie Mar 05 17:02:00 http://www.tekprosystems.com/images/zaurus/zguts1.jpg Mar 05 17:02:21 ant_work: hi Mar 05 17:02:26 "presumably have four 8-bit flash chips laid out on a 32-bit bus" Mar 05 17:02:26 ant_work: what's that? Mar 05 17:02:30 collie Mar 05 17:02:35 I see only 2 chips Mar 05 17:02:39 hmm... Mar 05 17:03:28 and the tech sheet says Page Mode Dual Work Flash MEMORY - Eight 4K-word Parameter Blocks , One-hundred and twenty-seven 32K-word Main Blocks, Top Parameter Location Mar 05 17:04:07 64M,Smart,Top Boot Block, 90ns Mar 05 17:04:08 organization: x16 Mar 05 17:04:59 so it should be 2 16-bit chips interleaved on a 32 bit bus Mar 05 17:06:06 and well, the LH28F640BFHE-PTTL90 pdf shows at least two erase regions Mar 05 17:07:26 so it seems what pawel did is wrong :/ Mar 05 17:07:29 http://lists.infradead.org/pipermail/linux-mtd/2005-November/014273.html Mar 05 17:10:14 well, I have first to 1) fix meta-initramfs 2) send linux-yocto 3.8 for meta-hh then I'll come back to collie Mar 05 17:20:02 bbl **** ENDING LOGGING AT Wed Mar 06 02:59:58 2013