**** BEGIN LOGGING AT Fri May 06 02:59:58 2016 May 06 07:29:38 good morning May 06 07:35:16 daal, did you compile with CPU_FREQ ? May 06 07:45:26 Hi ant_work May 06 07:46:43 yes, my kernel had CPU_FREQ set but it is also broken, it is with 416mhz as default, but to overclock you need to hardcode 624mhz May 06 07:48:02 you can not set the Mhz via cpufreq tools May 06 07:48:31 works fine if hardcoded in cpufreq pxa May 06 07:51:24 there was a thread on oesf.org, I send a overclocked kernel to danboid, it worked fine May 06 07:52:27 ok, I have noticed one Zaurus (5600) crashes if I enable it May 06 07:52:36 in the minimal kexecboot kernel May 06 07:52:48 it is a pxa250 May 06 07:52:54 ok, kexecboot I did not include CPU_FREQ May 06 07:53:01 may be a pxa250 issue May 06 07:53:22 seems ok on C860 (pxa255) May 06 07:53:28 tested May 06 07:53:38 can you overclock from commandline ? May 06 07:53:48 anyway, I'll let it disabled in kexecboot May 06 07:54:01 sure, no need for it I guess May 06 07:54:21 I am standardizing all features between models May 06 07:54:35 I have added F2FS (but not on collie) May 06 07:54:46 dunno about ROMFS, maybe for experiments May 06 07:54:56 * daal had no time to test the new kexecboot yet May 06 07:55:09 maybe Varthall gave it a go ? May 06 07:55:24 otherwise it is ms-dos, vfat, ext2, ext3, ext4, jffs2, jffs.summary, ubifs, cramfs, f2fs May 06 07:56:01 there seems to be an issue with MMC for PXA in kernel 4.4 May 06 07:56:10 sometimes the SD card is not detected May 06 07:56:31 or it is detected if you also have a CF inserted May 06 07:57:13 not happening on collie/5500 with the same cards May 06 07:57:37 but mmc on collie is over SPI, is slower May 06 07:57:41 hm, may be related to the MMC issue I found in 4.4 May 06 07:57:56 was that fixed ? May 06 07:58:12 the read only? yes May 06 07:58:18 ah, ok May 06 07:58:24 I still have doubts, though May 06 07:58:49 it seems Zaurus uses gpio-inverted and the code was working by pure luck May 06 07:59:05 I have asked the maintainer, maybe he's working on it May 06 07:59:10 using the 4.3 mmc code works fine May 06 07:59:17 ok May 06 07:59:28 I guess there is some more patching to do May 06 07:59:40 I am at 4.4.8 (last stable) May 06 08:00:03 https://github.com/greguu/linux-4.4-cxx00/commit/72ef5fe2bfca9442ec78d7a333756753aeb9941f May 06 08:00:19 this is what I did to my 4.4.0 to get mmc work on C3x00 May 06 08:01:09 and to be correct I used 4.2.3 code for mmc May 06 08:07:12 it seems there is also something wrong with the framebuffer pxafb on the 4.4 series May 06 08:08:00 it blanks white first, and not only that, if I try to access the framebuffer via directfb it also gets a white screen, memory overlap or something May 06 08:08:17 was working fine under 3.x kernel May 06 08:08:47 hm, I have to pull out spitz May 06 08:08:56 I don't see these problem on other models May 06 08:09:00 anymore May 06 08:09:19 white screen on first blank ? May 06 08:09:32 yes, it did happen to collie/poodle as well May 06 08:09:45 poodle uses pxa_fb May 06 08:10:28 the issue is fixed for other models, but not spitz ? May 06 08:11:00 last I tried to boot on spitz I got console May 06 08:11:10 but it was long ago, 4.2 probably May 06 08:11:42 on collie it was a clock issue (clock 0) May 06 08:12:12 then on both it was the backlight part (locomo) May 06 08:12:28 there are 20-30 pending patches but it works ;) May 06 08:12:53 on corgi/c860 all is surprisingly ok with 4.4 May 06 08:46:25 morning all May 06 08:47:27 daal: the latest kexecboot doesn't boot on my 1000, me and ant suspect it might be as there's a Cacko install on my /home partition on NAND May 06 08:48:32 I'll test on 3200 and push the changes for meta-handheld today May 06 08:48:41 bought a new USB-SD reader (I had other 2 which seem to be broken), will try to erase that partition as soon as I have some free time, I hope this weekend May 06 08:48:50 I think I'll use the same kernel config for akita and spitz May 06 08:49:21 did you maybe repartition the device? May 06 08:50:42 I followed the cacko install info, I have resized the root partition to 32 MB: http://my-zaurus.narod.ru/release-notes-1.23.txt May 06 08:50:56 ah, ok, maybe it's that May 06 08:51:00 (sorry, meant / before, not /home) May 06 08:51:17 kexecboot can read and pass the partition info May 06 08:51:22 does kexec require a specific size? May 06 08:51:30 but must be taught about repartitioning... May 06 08:51:43 the kernel itself hardcodes mtd values May 06 08:51:47 aha May 06 08:52:11 I promised long ago to rewrite in kernel space the routine of kexecboot May 06 08:52:15 as module May 06 08:52:22 no time yet :/ May 06 08:53:05 the old 2.x boots on my system though, so it seems to support dynamic partition sizes May 06 08:53:20 yes, it was passed by atags with kernel 2.4 May 06 08:53:32 what are the partition sizes expected by the latest kexec? May 06 08:53:34 I see May 06 08:53:49 do you have 32MB as root? May 06 08:53:59 yes May 06 08:55:49 ok, then I can recompile one kernel May 06 08:55:51 http://lxr.free-electrons.com/source/arch/arm/mach-pxa/spitz.c#L800 May 06 08:58:24 standard kernel expects mtdparts=sharpsl-nand:7168k at 0k(smf),59392k at 7168k(root),-(home) May 06 08:58:51 still, kexecboot should just fail to detect the fs and move on... :/ May 06 08:59:24 I see May 06 09:00:11 I can also repartition root to 58MB May 06 09:00:21 and check that it works May 06 09:00:31 I'll let you know May 06 09:01:16 have to go now, later May 06 16:11:21 latest builds May 06 16:11:23 https://github.com/LinuxPDA May 06 21:36:17 Varthall, hi May 06 21:36:19 May 06 16:11:21 latest builds May 06 21:36:19 May 06 16:11:23 https://github.com/LinuxPDA May 06 21:36:29 tested on SL-C3200 **** ENDING LOGGING AT Sat May 07 02:59:58 2016