**** BEGIN LOGGING AT Sun Aug 31 02:59:59 2014
Aug 31 06:35:12 <drathir> mornin...
Aug 31 08:33:49 <Pali> DocScrutinizer05: you had some objections about lis3lv02d driver in upstream kernel
Aug 31 08:34:09 <Pali> can you write somewhere notes about it?
Aug 31 08:40:38 <Pali> (btw, I found lis3 accelerometer in my notebook connected via smbus intel Lynx Point controller and lis3lv02d_i2c driver can detect and use it)
Aug 31 09:02:24 <wizbit> is it possible to convert mydocs to ext3?
Aug 31 09:02:35 <wizbit> and let it automount
Aug 31 09:32:03 <wizbit> figured it out :D
Aug 31 09:32:14 <wizbit> ext3 kicks the arse out of fat
Aug 31 09:32:23 <wizbit> what keeps on getting corrupted
Aug 31 14:43:14 <sixwheeledbeast> wizbit: I believe some applications play up with non-FAT MyDocs
Aug 31 14:48:38 <DocScrutinizer05> camera had an issue since they created dcim/ but tried to write to DCIM/* or sth like that
Aug 31 14:49:39 <DocScrutinizer05> other than that, MyDocs *should* be fs-neutral. Of course except mass storage mode
Aug 31 14:54:07 <DocScrutinizer05> but I'm generally doubting ext3 (or any journalling fs in that respect) is really the ideal fs for flash. Journal is optimized for storage where sequencial writes take very little time and "come for free". On flash *every* write is something you better avoid if you can
Aug 31 14:55:22 <Pali> DocScrutinizer05: see ^^^ about lis driver
Aug 31 14:55:36 <DocScrutinizer05> yeah. I pondered what to say
Aug 31 14:57:40 <DocScrutinizer05> maybe just so much: the driver MUST be based on IRQ driven concept. Poll is unacceptable. And it also must support a way to configure the filters of lis302 - a consequence from mandatory IRQ usage, since the filters determine when a IRQ will fire
Aug 31 14:58:55 <DocScrutinizer05> the 100Hz clocked mode is deprecated, even though it uses IRQ
Aug 31 14:59:10 <sixwheeledbeast> It was to do with file permissions for the applications IIRC.
Aug 31 14:59:24 <DocScrutinizer05> generaly an IRQ must happen only when an event is pending
Aug 31 14:59:35 <DocScrutinizer05> sixwheeledbeast: that too
Aug 31 14:59:58 <DocScrutinizer05> rather: owner
Aug 31 15:00:17 <DocScrutinizer05> FAT doesn't know about file owners
Aug 31 15:00:22 <sixwheeledbeast> :nod: 
Aug 31 15:01:57 <DocScrutinizer05> sth like: when DCIM is missing, then a root process creates DCIM root:root 644. Then a user process tries to write to dcim/1st-photo.jpg
Aug 31 15:03:12 <DocScrutinizer05> on FAT that will work since no file owners and no 644 permissions either, and no distinguishing between upper and lower case filenames
Aug 31 15:03:49 <DocScrutinizer05> but I think that's been fixed by Nokia somewhere in PR1.2 or 1.3
Aug 31 15:09:39 <wizbit> could using the wrong charger damage the n900 battery / msd card?
Aug 31 15:10:42 <DocScrutinizer05> no
Aug 31 15:10:58 <wizbit> my mssd card keeps on getting corrupted
Aug 31 15:11:33 <wizbit> maybe cfdisk is corrupting it
Aug 31 15:11:35 <DocScrutinizer05> a *defect* charger however could fry any device you plug to it
Aug 31 15:12:03 <DocScrutinizer05> mssd?
Aug 31 15:12:38 <wizbit> yep
Aug 31 15:13:09 <wizbit> i will mkfs.ext3 again
Aug 31 15:13:19 <wizbit> could a defect battery corrupt it?
Aug 31 15:13:46 <DocScrutinizer05> what's mssd?
Aug 31 15:14:09 <wizbit> the extra ssd card you install on a n900
Aug 31 15:14:25 <DocScrutinizer05> commonly used terms are "eMMC" and micro-SD aka uSD or µSD
Aug 31 15:14:57 <DocScrutinizer05> "eMMC" is embedded MultiMediaCard
Aug 31 15:15:11 <DocScrutinizer05> so MyDocs basically
Aug 31 15:16:32 <DocScrutinizer05> a defect battery that causes device to shut down hard, without an opportunity to flush/sync storage buffers, will definitely damage eMMC and uSD likewise
Aug 31 15:17:34 <wizbit> aye ok
Aug 31 15:17:57 <wizbit> shutting down the phone from shutdown menu should be a clean to shut it down?
Aug 31 15:19:18 <DocScrutinizer05> yes
Aug 31 15:22:20 <wizbit> what about if i did a shutdown from the terminal on the n900?
Aug 31 15:22:23 <wizbit> or reboot
Aug 31 15:27:39 <DocScrutinizer05> should all be safe
Aug 31 15:27:58 <DocScrutinizer05> more or less
Aug 31 15:28:34 <sixwheeledbeast> wizbit: You on CSSU-T?
Aug 31 15:28:55 <wizbit> cssu stable
Aug 31 15:29:26 <wizbit> eMMC seems ok that moment
Aug 31 15:29:44 <wizbit> i make sure i logout of ssh sessions and im not cd into a eMMC partition
Aug 31 15:29:48 <wizbit> then reboot using phone only
Aug 31 15:31:04 <wizbit> umount wont work if cd is on eMMC
Aug 31 17:38:28 <DocScrutinizer05> you shouldn't umount anything before shutdown. initscripts do that for you
Aug 31 17:43:29 <DocScrutinizer05> for reboot in terminal the preferred methos is `dsmetool --reboot`
Aug 31 17:47:15 <ShadowJK> Question is if the uSD card is corrupted already before reboots
Aug 31 17:50:24 <DocScrutinizer05> yup, sure
Aug 31 17:50:59 <DocScrutinizer05> poor contacts may do havoc to uSD
Aug 31 17:51:39 <DocScrutinizer05> also there are bad cards that need no further trigger to corrupt data
Aug 31 17:53:24 <ShadowJK> And don't run fsck on ext3 on sd
Aug 31 17:53:42 <ShadowJK> but if also vfat is getting damaged...
Aug 31 17:54:47 <DocScrutinizer05> otoh removing battery too fast after shutdown is probably also a nice way to mess up stuff. Device is not off already when screen goes black, it still does another few seconds of cleanup in act-dead mode before it actually shits down
Aug 31 17:55:39 <DocScrutinizer05> shuts*, but meh, fits
Aug 31 17:58:57 <dos1> I have modified power off LED pattern to not go completely off at the end of the fade out, but to still shine a bit, so when it goes off I know that the shutdown is completed
Aug 31 18:00:50 <ShadowJK> My samsung pro card has survived sudden powerloss many many times 
Aug 31 18:07:30 <kerio> yeah, fsck on ext3 on a sd is silly
Aug 31 18:07:35 <kerio> just mount it and use it
Aug 31 19:19:06 <joga> dos1, sounds like a good idea
Aug 31 19:19:26 <joga> sometimes it takes so long for it to have shut down
**** ENDING LOGGING AT Mon Sep 01 02:59:58 2014