**** BEGIN LOGGING AT Wed Feb 14 03:00:02 2018
Feb 14 08:26:59 <ant_work> hi greguu
Feb 14 08:27:13 <greguu> hi ant_work
Feb 14 08:27:18 <VartiWork> hi
Feb 14 08:27:33 <ant_work> I have built 4.16-rc1 for c860 but have to run-test it
Feb 14 08:27:56 <ant_work> let's try to find th enew breakage :/
Feb 14 08:28:02 <greguu> sure
Feb 14 08:28:31 <greguu> I did a 4.16-rc1 vanilla build yesterday for Cxx00
Feb 14 08:29:09 <ant_work> I will log debug  on serial
Feb 14 08:29:21 <greguu> great, I have not JTAG cable
Feb 14 08:29:23 <ant_work> spi and kb
Feb 14 08:29:38 <ant_work> bout kb, you maybe can test with evtest
Feb 14 08:30:07 <ant_work> the driver seems unchanged so it's some mapping issue
Feb 14 08:30:11 <ant_work> modifiers?
Feb 14 08:30:42 <greguu> ok, I will look at evtest
Feb 14 08:31:01 <ant_work> you can check keypress in kexecboot as well if compiled for debug
Feb 14 08:31:17 <ant_work> but latest it is 4.4
Feb 14 08:31:20 <greguu> the issue showed up with 4.9, but it could have been introduced earlier, 4.4 was ok
Feb 14 08:31:42 <ant_work> about this, I think best is to keep the kexecboot-kernel 3.x version around
Feb 14 08:31:49 <ant_work> seems snappier
Feb 14 08:31:53 <greguu> yes
Feb 14 08:32:08 <ant_work> I am sure it is the deferred probing slowing down all
Feb 14 08:32:14 <greguu> the 3.10.y kernel seems to be ok so far for kexecboot - appart for the lack of C1000 and Cxx0 support
Feb 14 08:32:19 <ant_work> this impacts also in sd/cf detection
Feb 14 08:32:33 <ant_work> wait, I am sure we fixed akita...
Feb 14 08:32:45 <greguu> really ?
Feb 14 08:32:58 <greguu> 3.10.y would not boot akita.
Feb 14 08:33:07 <greguu> this was still outstanding
Feb 14 08:33:35 <ant_work> iirc it was fixed, problem was just Varti's f2fs rootfs
Feb 14 08:33:48 <greguu> ah, no
Feb 14 08:34:10 <greguu> I created a frankenstein kernel out of the 2.6 kernel and injected a new kexecboot
Feb 14 08:34:11 <ant_work> https://github.com/LinuxPDA/linux-kexecboot/commits/master
Feb 14 08:34:21 <ant_work> afair Varti tested these
Feb 14 08:34:28 <greguu> this did not allow ext4/f2fs /boot partitions
Feb 14 08:35:03 <ant_work> this might be a distro issue
Feb 14 08:35:17 <ant_work> I am sure I can kexec a kernel in any partition
Feb 14 08:35:22 <greguu> nah, the kernel did not boot at all VartiWork ?
Feb 14 08:35:25 <ant_work> from kexecboot
Feb 14 08:35:41 <ant_work> you're kidding me ;)
Feb 14 08:35:51 <ant_work> VartiWork, ping
Feb 14 08:36:14 <greguu> well. the "workaround" kernel (frankenstein) was 2.6 based and lacked f2fs/ext4 support
Feb 14 08:36:15 <ant_work> jeez I was about buying an akita when Varti said it booted
Feb 14 08:36:27 <greguu> with my frankenstein kernel maybe
Feb 14 08:36:55 <ant_work> I really hope you're wrong remembering the matter :)
Feb 14 08:36:57 <VartiWork> it did not boot due to the backlight problem
Feb 14 08:37:09 <greguu> black screen ?
Feb 14 08:37:09 <ant_work> I am wrong...
Feb 14 08:37:41 <greguu> or white screen ?
Feb 14 08:37:52 <ant_work> we have the logs of beg june
Feb 14 08:37:57 <VartiWork> I don't remember if the screen was black or white, but yes, it remained stuck
Feb 14 08:38:26 <ant_work> wel, please try again the ones at LinuxPDA
Feb 14 08:38:35 <ant_work> on akita and tosa
Feb 14 08:38:40 <VartiWork> ok
Feb 14 08:39:43 <ant_work> I still have the sources, will update to latest kexecboot 0.7 soon (with devicetree support, etc etc)
Feb 14 08:39:48 <VartiWork> my first success with the frankenstein kexec: http://logs.nslu2-linux.org/livelogs/kexecboot/kexecboot.20170526.txt
Feb 14 08:39:55 <VartiWork> I'm searching for my older posts
Feb 14 08:40:08 <greguu> devicetree ? ahhhhhhh.....
Feb 14 08:40:19 <VartiWork> ah, there's my thread on oesf too, let me check
Feb 14 08:41:27 <VartiWork> https://www.oesf.org/forum/index.php?showtopic=34502
Feb 14 08:41:48 <VartiWork> 4.4.8 did not boot
Feb 14 08:42:44 <ant_work> you see immediately if the kernel logo appears on cold bopot
Feb 14 08:43:41 <ant_work> please retry the 3.x first
Feb 14 08:46:03 <VartiWork> http://logs.nslu2-linux.org/livelogs/kexecboot/kexecboot.20170526.txt
Feb 14 08:46:11 <VartiWork> May 25 13:13:25 <ant_work> seems akita suffers after 3.7 May 25 13:13:36 <ant_work> that explains my working 3.2
Feb 14 08:46:50 <VartiWork> I'll test them again anyway
Feb 14 08:53:53 <ant_work> Jun 12 14:36:34 <VartiWork> I have tried both the 3.2.18 and 3.2.84-f2fs versions, both they behave the same
Feb 14 08:53:53 <ant_work> Jun 12 14:37:32 <VartiWork> i.e. they boot as expected, menu browsing is now fast, when trying to boot my alarmz install the screen remains black
Feb 14 08:54:05 <ant_work> so these are ok
Feb 14 08:55:40 <VartiWork> hem no... they don't work :)
Feb 14 08:56:23 <ant_work> this seems an after-kexec problem, aka yours ;)
Feb 14 08:57:07 <ant_work> anyway, there you have all 3.x variants with f2fs port
Feb 14 08:58:27 <greguu> I believe VartiWork also tried my 3.10.y version, even the "special one" that was compiled with only Akita support (eg removed borzoi/spitz) - but no luck
Feb 14 08:59:39 <greguu> not sure about Varti's statement on Jun 12th..
Feb 14 09:00:30 <VartiWork> ant_work: true, kexec itself worked, sorry to have mixed it up
Feb 14 09:00:52 <greguu> ah, I remember !
Feb 14 09:01:03 <greguu> 3.2.18 and 3.2.84 worked on Akita
Feb 14 09:01:10 <greguu> but the F2FS support was borked
Feb 14 09:01:22 <greguu> only 3.10.y has proper backport but does not boot on Akita
Feb 14 09:01:50 <VartiWork> yes, makes sense
Feb 14 09:02:15 <ant_work> seems that I found the 3.4 from commit logs
Feb 14 09:02:43 <greguu> also on 3.4 the F2FS backport is outdated
Feb 14 09:03:03 <greguu> we settled on 3.10.y for current offical F2FS backport
Feb 14 09:03:39 <greguu> so we can have a kexecboot for akita with ext4 support but not f2fs without a seperate /boot parition - same as with 2.6 kernel
Feb 14 09:04:06 <greguu> well, better than then 2.6 depending if you intend to use F2FS or not
Feb 14 09:05:50 <greguu> The idea was to have a unified kexec kernel for Cxx00 ...
Feb 14 09:06:43 <ant_work> the 3.4 lacks fscrypto
Feb 14 09:06:46 <ant_work> I see
Feb 14 09:11:11 <greguu> the 3.2 and 3.4 f2fs backports could not mount a f2fs fs that was created with a newer kernel
Feb 14 09:11:32 <greguu> 3.10 raised it to the new standard
Feb 14 09:12:59 <greguu> ant_work : but looking back to the irc logs, your assumtion the device detection and ASM code may be the cause for Akita failing to boot on 3.7+
Feb 14 09:16:08 <ant_work> I remember I was about buying one akita last spring ;)
Feb 14 09:16:25 <greguu> me too, I looked on rakuten...
Feb 14 09:16:43 <greguu> but never got one.. It would help if we had one to troubleshoot the issue
Feb 14 09:17:39 <ant_work> I can't find the log but I did not buy it because apparently the issue was solved
Feb 14 09:18:16 <greguu> hehe
Feb 14 09:19:55 <greguu> the 2.6 frankenstein kernel does include kexecboot 0.6 so the newer boot.cfg standard is there. some just have to live with having a seperate /boot partition if they like ext4 or f2fs on /
Feb 14 09:20:04 <ant_work> <ant_work> visto che akita ora ha kexecboot direi che non ne ho bisogno
Feb 14 09:20:06 <greguu> but still we should be looking into the issue
Feb 14 09:20:07 <ant_work> ^ VartiWork
Feb 14 09:20:26 <greguu> hey ant_work : no secret talk :)
Feb 14 09:20:30 <ant_work> some tracky in the PV log
Feb 14 09:20:32 <ant_work> ;)
Feb 14 09:23:12 <ant_work> nope, can't find more
Feb 14 09:28:18 <VartiWork> I'm still hoping to receive the serial cable from dromede btw, I hope that it would be helpful
Feb 14 09:29:12 <VartiWork> I remember that comment from you :)
Feb 14 09:29:51 <VartiWork> it was at the time when I was mass-buying Z's from Japan
Feb 14 09:30:16 <greguu> how many did you get :)
Feb 14 09:30:56 <VartiWork> (it would have been anyway useless to combine all the Z's into one package, you can't do that anymore because of the Lithium batteries shipping limitations)
Feb 14 09:31:22 <greguu> the JTAG cable seems to be more rare than the zaurus units
Feb 14 09:32:37 <VartiWork> greguu: 14
Feb 14 09:33:15 <greguu> always good to read http://www.penguin.cz/~utx/zaurus/hardware
Feb 14 09:33:40 <greguu> VartiWork: Any spare borzoi ?
Feb 14 09:33:44 <VartiWork> greguu: but I have a batch of 8 Tosas, which were sold together
Feb 14 09:34:02 <VartiWork> yes, I have a Borzoi too
Feb 14 09:34:34 <greguu> ok great, I need a backup unit. consider it sold if mine fails
Feb 14 09:35:27 <VartiWork> I wouldn't like to part it for now... I wanted to make tests on it too, and wanted to make photos of it to put in the wiki
Feb 14 09:35:42 <greguu> ah, you got only one ?
Feb 14 09:37:41 <greguu> it was the best unit produced afaik
Feb 14 09:37:56 <greguu> the c3200 could not be overclocked
Feb 14 09:41:10 <greguu> anyway, back to the inital topic. ant_work: please confirm 4.16-rc1 issues on spitz/borzoi
Feb 14 09:41:22 <greguu> I will troubleshoot the kb issue
Feb 14 09:41:41 <ant_work> ok
Feb 14 09:43:48 <greguu> I tried with gcc version 7.3.0
Feb 14 09:44:23 <greguu> and 6.3.1
Feb 14 09:45:08 <greguu> both ArchLinuxARM prebuild toolchains
Feb 14 09:45:50 <ant_work> 7.3 here as well
Feb 14 09:49:05 <greguu> I noticed a new config option CONFIG_PINCTRL_PXA27X in 4.15/4.16
Feb 14 09:50:00 <greguu> also for PXA25x
Feb 14 09:52:49 <greguu> it seems to have been there for a while.. but previous "oldconfig" did not bring it up , even with 4.12
Feb 14 10:01:41 <VartiWork> yes, only one...
Feb 14 10:06:31 <ant_work> it is there since 4.5-1
Feb 14 10:07:31 <ant_work> dor pxa25x and 27x
Feb 14 10:13:01 <greguu> hm, must have missed it
Feb 14 10:13:07 * greguu is gone for the day
Feb 14 10:15:35 <ant_work> cu later
Feb 14 10:18:53 <VartiWork> night greguu
**** ENDING LOGGING AT Thu Feb 15 03:00:01 2018