[01:06:56] <[Saint]> TheSeven: moving to a more appropriate place
[01:07:17] <[Saint]> UMSboot is up and mounted - no onscreen instruction from the iPod.
[01:07:30] <[Saint]> Just half a black screen with a corrupt emCORE in the other half.
[01:08:14] <[Saint]> using "installer-ipodclassic-diskmode-beta"
[01:13:32] *** Quits: krnlyng (~liar@83.175.90.24) (*.net *.split)
[01:13:32] *** Quits: TheSeven (~quassel@rockbox/developer/TheSeven) (*.net *.split)
[01:18:46] *** Quits: [Saint] (~saint@rockbox/staff/saint) (Quit: Quit.)
[01:18:58] *** Joins: [Saint] (~saint@rockbox/staff/saint)
[01:19:42] *** Joins: krnlyng (~liar@83.175.90.24)
[01:19:42] *** Joins: TheSeven (~quassel@rockbox/developer/TheSeven)
[02:16:35] <[Saint]> TheSeven: you are a beautiful magical bastard
[02:16:54] <[Saint]> Boot time is about ~1s longer, but, I'll take it over ugly disk noises.
[02:17:46] <[Saint]> emCORE now actually spins down the disk before handing off to Rockbox. As intended, one assumes.
[02:19:43] <[Saint]> The only issue I have noticed is that it takes several seconds to shut emCORE down now regardless if the disk is spinning or not.
[02:20:03] <[Saint]> (also a non-issue, comparatively)
[02:20:24] <user890104> [Saint]: which version are you testing? diskmode-beta?
[02:20:36] <user890104> what revision of emCORE is this?
[02:20:47] <[Saint]> This emCORE build you've given me, unlike user890104's, doesn't have the weird black screen fuckup in UMSboot either.
[02:21:17] <[Saint]> user890104: 928M
[02:21:44] <[Saint]> (the M is some form of CE-ATA-compatibility-ness)
[02:22:11] <[Saint]> TheSeven fell off the internet in the middle of explaining exactly what.
[02:22:15] <user890104> r926 fixes "weird black screen fuckup in UMSboot"
[02:22:21] <user890104> :)
[02:22:35] <[Saint]> Ah. Awesome. :)
[02:23:07] <[Saint]> Its identical to the corruption one gets over top of the Apple logo during first install.
[02:23:16] <user890104> yes
[02:23:19] <[Saint]> I should say it *was*.
[02:23:26] <[Saint]> Now it doesn't exist anymore.
[02:23:37] <user890104> the fix is an one liner: http://websvn.freemyipod.org/comp.php?repname=freemyipod&compare[]=/@925&compare[]=/@926
[02:23:56] <[Saint]> Hah. Cute.
[02:24:09] <[Saint]> The little things, huh? :)
[02:24:18] <user890104> i just had to copy it over from emCORE's LCD driver
[02:25:06] <user890104> that's the downside of having the same drivers in two different apps with different code layouts
[02:25:35] * [Saint] has a suggestion for the installer
[02:25:52] <[Saint]> "Don't re-install themes if path/to/theme isn't present on upgrade"
[02:26:43] <user890104> well, it's a reinstall-slash-repair thing - if it's missing, we should restore it
[02:27:05] <[Saint]> Its not really repair, IMO. 
[02:27:14] <[Saint]> As all those files in the disk are irrelevant really.
[02:28:01] <user890104> well, it does its best to bring back the winter spirit, even if you don't like it
[02:28:27] <[Saint]> The Christmas easter egg that doesn't quit.
[02:28:31] <[Saint]> Ever.
[02:28:36] <[Saint]> For several years. ;)
[02:29:02] <user890104> since we have a RTC driver, we can choose wheter or not to display the winter theme, maybe only in december
[02:29:45] <user890104> well, i'd disable it at all in the next release, and offer it as a separate download
[02:29:50] <[Saint]> Can it be a persistant setting?
[02:30:01] <[Saint]> (if its not removed entirely)
[02:30:09] <user890104> but it doesn't sound like a good idea, because someone can mix up app versions (similar to fastboot)
[02:30:34] <[Saint]> The only thing that really bothers me is that I remove the theme, and, it puts it straight back.
[02:30:48] <user890104> the original theme looks for a custom theme. if this code is removed, nothing will ever try to run the custom one
[02:30:49] <[Saint]> I wouldn't mind it putting the file back, but leaving my selection.
[02:31:17] <[Saint]> I just find it slightly annoying having to go back in and remove it again if I don't want it.
[02:31:40] <user890104> we can also do some temporary makefile magic, to skip copying the winter theme to fsfiles
[02:31:51] <user890104> when building an installer
[02:32:32] <[Saint]> Would saving the path to the user selected theme in the settings be an option?
[02:34:16] <user890104> we need to implement a file browser then
[02:34:27] <[Saint]> Ah, yes. True.
[02:34:52] <user890104> i have an attempt to do that, as an emcore app
[02:35:06] <[Saint]> 24MB/s write - 11MB/s read with Disk Mode, btw.
[02:35:26] <[Saint]> vs. 6MB/s/2MB/s in Rockbox.
[02:35:34] <user890104> but it's on the hdd of an inaccessible VM, so i need to extract it somehow
[02:35:34] <[Saint]> Well done, you majestic fuckers.
[02:35:46] <[Saint]> Well done.
[02:35:49] <user890104> double buffering FTW
[02:35:52] <user890104> thanks to TheSeven
[02:36:10] <user890104> can you test OF's diskmode and OF?
[02:36:15] <[Saint]> its like a ~400% improvement
[02:36:22] <user890104> since you started comparing the speeds...
[02:36:38] <[Saint]> I need to restore for that - but, yes. I can.
[02:36:56] <[Saint]> Not right now though. But, yes. Yes I can.
[02:38:48] <user890104> if you have an emcore bootstrapper (a thing that i made for my tests, which consists of bootsub+emcore instead of bootstub+umsboot), you can switch back and forth by restoring a NOR backup
[02:39:06] <user890104> and you'll need to use ipodscsi for repartitioning
[02:39:12] <user890104> (or itunes)
[02:39:27] <user890104> this should work in theory, unless i'm missing something
[02:39:28] <[Saint]> I'd need to restore first to *get* that backup, though. So...yeah. ;)
[02:39:38] <[Saint]> Seeing as I'm apparently "Mr. CE-ATA Guy"
[02:40:10] <[Saint]> AFAIK neither of you guys have the CE-ATA variant.
[02:40:15] <user890104> actually, this is a pretty hacky way for early dual-boot support
[02:40:46] <user890104> bit it will need to boot emcore "tethered"
[02:40:55] <[Saint]> Huh - clever.
[02:40:56] * [Saint] nods
[02:42:00] <[Saint]> Can the installer automagically grab diagmode from a new install and embed it in the emCORE install for future use?
[02:42:02] <[Saint]> I may have asked this before.
[02:42:08] <[Saint]> I know it can't be distributed.
[02:43:28] <user890104> the installer can read diagmode from OF's NOR images, then put it back in place (compressed, because we're alsmost out of space already)
[02:43:42] <user890104> then the bootmenu can be made clever enough to boot it
[02:44:05] <user890104> a-la-nano2g
[02:44:34] <[Saint]> Thinking about the tethered dual boot - would it actually need to be tethered? Could the necessary magic not be stored on disk and a RAMdisk created to do the juggling around, and then jump into the OF?
[02:45:28] <user890104> once we run our bootstub/emcoreldr, OF and OF disk mode can't boot anymore
[02:45:49] <user890104> so we need a fresh norloader from apple to be loaded first
[02:46:38] <user890104> but it won't be kind enough to run our code, unless it's signed by the cupertino guys
[02:47:31] <user890104> we should be able to decrypt it patch it then encrypt it
[02:47:44] <user890104> ...or, fix our code so it doesn't break apple's
[02:48:10] <[Saint]> Which would require an understanding of why it breaks.
[02:48:56] <user890104> yes, and we already have an idea about that
[02:49:08] <[Saint]> Oh? I must've missed that.
[02:49:56] <user890104> IIRC there's something wrong in the state which we set to the i2c interface
[03:26:10] *** Quits: [Saint] (~saint@rockbox/staff/saint) (Remote host closed the connection)
[03:27:08] *** Joins: [Saint] (~saint@rockbox/staff/saint)
[06:41:05] *** Quits: TheSeven (~quassel@rockbox/developer/TheSeven) (Ping timeout: 264 seconds)
[06:42:14] *** Joins: TheSeven (~quassel@rockbox/developer/TheSeven)
[08:01:59] *** Quits: [Saint] (~saint@rockbox/staff/saint) (Remote host closed the connection)
[08:02:58] *** Joins: [Saint] (~saint@rockbox/staff/saint)
[10:19:14] <TheSeven> [Saint]: ok, so the spindown itself is working fine? let me make a build without the delay then
[10:19:57] <TheSeven> http://theseven.homenet.org/~theseven/tmp/installer-ipodclassic.ubi
[10:20:01] <TheSeven> is that one fine as well?
[10:21:29] <TheSeven> [Saint]: feel free to make a summer theme ;)
[10:22:55] <TheSeven> [Saint]: the downside of that improvement is that write errors may be misattributed to wrong sector numbers or get lost completely
[10:23:43] <TheSeven> the drive does something very similar internally already though, so even with rockbox's code write errors aren't accurate, so I don't really bother if they are more inaccurate now :P
[10:25:47] <TheSeven> user890104: hm... has anyone ever tried to chainload apple's norloader from emcore?
[10:26:04] <TheSeven> I mean it would get totally confused about the flash contents, but maybe it doesn't even look at those before booting the OF? :P
[10:26:33] <TheSeven> at the very least it would be interesting to know if the i2c driver of that code locks up as well
[10:37:48] <TheSeven> hm, my disk is showing signs of head trembling in certain areas...
[10:38:10] <TheSeven> I really wonder what the root cause of all these drives failing is
[10:38:32] <TheSeven> the behavior doesn't quite align with a headcrash
[10:38:42] <TheSeven> more with servo info degrading over time or something
[10:38:51] <[Saint]> Its curious that both drive variants appear to be equally quirky and failure prone.
[10:39:12] <copper> mine has survived a drop on the sidewalk
[10:39:19] <TheSeven> while spinning?
[10:39:28] <copper> while playing
[10:39:33] <copper> so, maybe not while spinning
[10:39:51] <[Saint]> Odds are heads were parked.
[10:40:02] <[Saint]> Unless it was hilariously bad timing.
[10:40:08] <TheSeven> this damage could possibly be related to writing while being shaken or something, if the head steering isn't very good
[10:40:57] <TheSeven> from what it looks like, the drive has trouble *finding* the sectors, not reading them
[10:41:03] <[Saint]> With the heads parked they can withstand around 6G of shock force iirc.
[10:41:12] <TheSeven> WAY more
[10:41:21] <[Saint]> Ah. OK.
[10:41:56] <TheSeven> desktop HDDs are often rated 200 to 500G
[10:42:12] <copper> what
[10:42:22] <[Saint]> Oh. Whoops. I see what happened there
[10:42:28] <[Saint]> I meant to type 60
[10:42:35] <copper> yeah right :P
[10:42:40] <[Saint]> But it seems I'm still off by a fair margin.
[10:42:58] <copper> sitty geeez!
[10:43:20] <[Saint]> But, yeah, 200G sounds like heaps.
[10:43:29] <[Saint]> But for impact force its really not.
[10:43:30] <TheSeven> 60G is what they can typically take while operating
[10:43:38] <[Saint]> Ah.
[10:44:32] <[Saint]> Do these drives have free fall detection?
[10:44:53] <[Saint]> I know some drives can attempt to park the heads during free fall.
[10:44:54] <user890104> [10:25:47] <TheSeven> user890104: hm... has anyone ever tried to chainload apple's norloader from emcore? << IIRC no, you can be the first one to try it, if you have a decrypted norloader image at hand
[10:45:38] <user890104> it should be able to read syscfg fine
[10:45:53] <user890104> but maybe fail at the apple logo?
[10:46:10] <TheSeven> hm, we need to have a look at what the handover conditions for that one are
[10:46:46] <user890104> unless they made their code ship showing the logo if the file doesn't exist
[10:46:49] <TheSeven> IIRC it's running from SRAM address 0, with just about nothing initialized but the SoC itself
[10:47:06] <TheSeven> i.e. the first thing it will do is initialize the PMU (I2C!) and then the DRAM (again)
[10:47:21] *** Joins: slenselink__ (~STeeF@office.hostnetbv.nl)
[10:47:32] *** slenselink__ is now known as STeeF
[10:47:40] <TheSeven> IIRC it can handle a missing logo
[10:48:06] <TheSeven> I'd think it's likely that it ends up showing the "bad battery" icon though if it doesn't like the flash file table
[10:48:34] <user890104> only if it can read the bad battery icon
[10:50:55] <TheSeven> that's embedded
[10:51:32] <user890104> ah, i see
[10:51:47] <TheSeven> I'm mostly curious if it manages to access the PMU at all though
[10:52:28] <[Saint]> It sure is. I figured that out when I couldn't dump it from the OF image when I was trying to dump the OF bitnaos for the iLike thene
[10:52:29] <TheSeven> in that case we could try to break out of it after the first successful I2C transfer and then hand over to the OF, hoping that it likes that state better
[10:52:41] <TheSeven> (and then try to narrow down the root cause)
[11:32:13] *** Quits: copper (~copper@unaffiliated/copper) (Quit: ZNC - http://znc.in)
[11:32:43] *** Joins: copper_ (~copper@unaffiliated/copper)
[11:32:52] *** copper_ is now known as copper
[12:50:38] <copper> TheSeven: the adapter guy replied to my e-mail: https://outpost.fr/tmp/qcq.txt
[13:01:55] <copper> (for future reference, his name is Tarkan)
[15:16:21] <TheSeven> just for your information: switching off the BBT function will only move the error around - it's just the first thing that attempts to access sectors
[15:16:51] <TheSeven> it's not what he thinks and in no way specific to SSDs, it's mostly intended to work around broken HDDs
[15:17:32] <TheSeven> it works transparently on top of the disk and doesn't interfere with any of the disk/SSD's internals
[15:18:28] <TheSeven> a thing that's possibly worth trying is forcing PIO mode
[15:18:58] <TheSeven> but it's highly unlikely to help if the same adapter works with one card but not another
[16:03:49] <copper> TheSeven: I have time to test it, if you have time to build it :)
[16:11:14] <copper> I've temporarily removed my SDXC card from eBay, just in case
[16:12:17] <copper> TheSeven: apparently it's mostly this one guy with this one card that has it working
[16:12:43] <copper> everyone else, including me and Tarkan with the two same cards, is having that problem
[20:45:41] <TheSeven> i'm totally confused why it would work with emcore with one card but not another, but with both cards with the OF
[21:08:42] <copper> I have no clue
[21:08:50] <copper> I know nothing about such things
[21:09:29] <copper> though I alway get the sense that no-one is ever following specs religiously
[21:10:03] <copper> (if specs are helpful at all)
[21:10:32] <copper> I generally distrust consumer electronics companies
[21:11:00] <copper> because their prices are almost always driven down
[21:11:14] <copper> and the big juicy markups are in the enterprise market
[21:11:40] <copper> I don't think it's a coincidence that my 128 GB SDXC card from Kingston is 50% cheaper than the competition
[21:12:00] <copper> well, I always exagerate
[21:12:02] <copper> maybe 30%
[21:13:27] <copper> I doubt anyone is spending much money on testing before release
[21:14:01] <copper> or on R&D
[21:15:05] <copper> mmm-kay, I'm going to shut up now
[22:30:02] <TheSeven> kingston is an interesting phenomenon
[22:39:48] <copper> I thought you were about to elaborate!
[22:56:55] <TheSeven> they basically resell chips from other manufacturers, yet at very competitive prices
[22:57:05] <TheSeven> not quite sure what's going on there...