**** BEGIN LOGGING AT Fri Jun 09 03:00:04 2017 Jun 09 09:00:50 ant_work: good progress on 3.2 , so what is the issue between 3.2 and 3.5+ ? Jun 09 09:03:28 bluelightning: next time you are back in NZ, give me a price on the C1000 and ship it to the south island ;) Jun 09 09:05:04 ant_work: mind you that the f2fs port for <3.10 is outdated Jun 09 09:05:23 greguu, I have the setup for building 3.2, 3.4, 3.8, 3.10 Jun 09 09:05:36 yes, I have read :/ Jun 09 09:05:47 greguu: I think that one belongs to my work, so I couldn't sell it unfortunately Jun 09 09:06:01 greguu: the one I own is the SL-C3000 (spitz) Jun 09 09:06:36 bluelightning: no worries, I will ask you then to do some tesing intead :) Jun 09 09:08:14 greguu: yep, should be able to do that :) Jun 09 09:08:48 ant_work: found the issue why 3.2+ kernel do not boot ? Jun 09 09:09:51 ant_work: hi Jun 09 09:10:05 ant_work: http://logs.nslu2-linux.org/livelogs/kexecboot/kexecboot.20170609.txt (last few lines) Jun 09 09:10:39 btw realised now that 3.2 can't boot ext4, I'll try with ext3 Jun 09 09:11:59 also about the slow menu in kexecboot, this can be fixed - ant_work can make that happen for sure Jun 09 09:12:24 or just inject the 0.6 initramfs from the latest arch kexecboot kernel :) Jun 09 09:13:33 might be because of the unrecognised mtdblock partitions? Jun 09 09:13:52 as if it's constantyl trying to mount them? Jun 09 09:14:50 no, if I remember correctly , this can be tweaked in 0.6 to be more smooth Jun 09 09:15:06 ant_work should know all about this Jun 09 09:15:39 ok Jun 09 09:17:40 I did disable some of the zaurus specific parsing in the arch kexecboot, as it would not fit into NAND anyway Jun 09 09:18:16 also removing JFFS2 from NAND stops kexecboot from parsing as it should not recoginize it.. in theory. Jun 09 09:22:30 sorry, was on phone call Jun 09 09:22:41 there was a bug with menu in 0.5.9 iirc Jun 09 09:22:57 but on poodle and corgi is not so slow Jun 09 09:23:11 yes some refresh thing,it was slow as hell Jun 09 09:23:36 the mtd are not recognised...you erased them, isn't? Jun 09 09:24:17 I'll rebuild a serie of 3.2 and 3.10 later today Jun 09 09:26:29 about ext4, it is compiled..I'll recheck Jun 09 09:27:22 greguu, so that backport for 3.2 is unfixable? Jun 09 09:27:51 I erased mtd2 and 3, what is strange is that it is not recognising even mtd0 and 1 Jun 09 09:28:09 mtd0 is a prom Jun 09 09:28:12 well, I tried the "unleashed" port on 3.5 and it did not boot Jun 09 09:28:18 ah so ext4 should work... strange Jun 09 09:28:21 mtd1 is 'nandlogical' so no filesystem Jun 09 09:28:25 the proper port works on 3.10.y Jun 09 09:28:34 ah ok Jun 09 09:29:24 let's hope we understanf what's wrong in that 3.x serie Jun 09 09:29:46 dromede wrote about issues with 3.7 back then Jun 09 09:30:02 but I see the supposed fixes in 3.8 Jun 09 09:30:16 ..so 3.8 and 3.10 should boot on akita as well.... Jun 09 09:30:26 ah, so were they accepted at the end? Jun 09 09:30:55 yes somehow Jun 09 09:33:19 May 26 07:10:00 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/arch/arm/include/asm/assembler.h?h=v3.8&id=1ecec696c8bb9b4cefb09495d81d081d1c81b578 Jun 09 09:33:19 May 26 07:11:36 https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/drivers/video/backlight/corgi_lcd.c?h=v3.8&id=1581b088fc91cbb974ad2b98431a8ecabb4852ee Jun 09 10:25:03 ant_work we should dissect the changes and find out what the issue is, I am sure we can get a 3.10.y going on akita Jun 09 10:26:31 VartiWork reported notification issue with corgi backlight on Akita, this is driven by systemd on Archlinux.. any hint what may be different between akita and spitz/borzio ? Jun 09 10:26:52 I thought the display is the same for all Cxxxx series Jun 09 10:27:32 AFAIR the Akita has a different chip for the backlight, let me check Jun 09 10:29:57 http://lists.infradead.org/pipermail/linux-arm-kernel/2012-December/136073.html Jun 09 10:30:26 >> Akita machines have backlight controls hooked to a gpio expander chip, max7310. >> The driver that controls the max7310 sets the cansleep flag and the corgi_lcd >> driver only uses plain gpio_set_value calls for changing backlight controls. >> This triggers the WARN_ON on akita machines. Jun 09 10:30:45 >> Akita is the only exception in this case since other users of corgi_bl access >> backlight gpio controls through a different gpio expander which does not set the cansleep flag. Jun 09 10:30:51 lunch break, later Jun 09 10:31:34 discussions here (near the end): http://logs.nslu2-linux.org/livelogs/kexecboot/kexecboot.20170526.txt Jun 09 10:31:43 and http://logs.nslu2-linux.org/livelogs/kexecboot/kexecboot.20170527.txt Jun 09 10:32:43 ok, thanks, maybe this needs to be considered in the 4.12 kernel Jun 09 10:32:55 max7310.. Jun 09 12:01:40 greguu, VartiWork: this backlight issue is considered in the second patch above Jun 09 12:01:50 in 3.8 Jun 09 12:02:32 so in theory is fixed... Jun 09 12:02:47 hmm Jun 09 12:36:18 heh, the MTD telenovela is following... Jun 09 12:36:55 neverthless the squashed patches me and greguu use are okù Jun 09 12:37:11 new objections raised to accept the patch? Jun 09 12:37:28 finally the big maintainer reviewed them Jun 09 12:37:56 some changes requested ofc Jun 09 12:38:25 mostly error-paths..but I'd say superfluous because the detection cannot fail... Jun 09 12:38:44 the structure is always the same after nand restore Jun 09 12:39:15 now they ask about endianness :) Zaurus is LE only ;) Jun 09 12:39:39 I hope lumag will now give the final push Jun 09 12:41:48 ok, things are progressing anyway Jun 09 13:01:34 seem so Jun 09 13:01:36 bbl **** BEGIN LOGGING AT Fri Jun 09 23:33:10 2017 **** ENDING LOGGING AT Sat Jun 10 03:00:01 2017