**** BEGIN LOGGING AT Fri Nov 14 03:00:00 2014 Nov 14 08:08:07 lumag_: gm Nov 14 08:08:10 http://www.circlemud.org/jelson/sdcard/SDCardStandardv1.9.pdf Nov 14 08:08:18 sandisk 2003 Nov 14 08:08:31 almost contemporary to collie Nov 14 08:21:04 ah..RCA register is not available in SPI mode... Nov 14 08:21:15 (Relative card address) Nov 14 08:51:50 ant_work, hello Nov 14 08:53:24 It seems like I somehow has lost the LCD initialization. Nov 14 08:54:43 hi Nov 14 08:54:50 I'm reading ^manual page 95 Nov 14 08:54:58 is 38 error hex or decimal? Nov 14 08:55:19 decimal Nov 14 08:55:22 It's -errno Nov 14 08:55:39 so it is not R1? Nov 14 08:55:46 Enable MMC_DEBUG Nov 14 08:55:58 You will see information about actual commands and their statuses Nov 14 08:56:13 ok, is a bit like mtd Nov 14 08:57:28 about lcd, after boot I see the OE logo vanishing and fade white Nov 14 09:01:00 Yes. Nov 14 09:01:31 in drivers/video/backlight/locomo_lcd.c changing FB_BLANK_NORMAL to FB_BLANK_UNBLANK should help Nov 14 09:04:33 your patchset didn't touch locomo_lcd and locomo_kb yet Nov 14 09:04:57 is the suspend/rersume still necesary? I see it mentions no support but now you've added pm ops Nov 14 09:05:43 Hmm. Nov 14 09:06:07 Patch 4 touches locomo_kbd and 6 adds lcd Nov 14 09:06:25 ah, ok then Nov 14 09:06:41 is about the 3 stale patches in oe Nov 14 09:07:03 meta-hh? Nov 14 09:07:10 yes Nov 14 09:07:41 for 3.18 only I've rebased collie-gpio-charger and put it as last patch Nov 14 09:07:53 (defconfig is already gpio-charger) Nov 14 09:08:17 input is necessary, but it is 'not for upstream' Nov 14 09:08:18 then adding your 16 patches I've removed locomo-fix and locomo_lcd patches Nov 14 09:08:36 (and added the 3 drivers in defconfig) Nov 14 09:09:00 now the pending fix we talked last night + thi slcd unblank Nov 14 09:09:12 is there more to add? Nov 14 09:11:22 pending fix? Ah, that is for pxa and it is not important till Robert merges his clock patches. Nov 14 09:11:59 the cpu clock patches are necessary, but I think you have them in. Nov 14 09:12:26 I mean the writew (pre-regmap) Nov 14 09:12:56 The DAC resume? Nov 14 09:13:02 Yes, it will be good to have. Nov 14 09:14:31 finally the locomo_kb2 hack...why does the mapping differs in kernel with collie/podle? Nov 14 09:14:40 iirc service man has th ekeymap Nov 14 09:15:12 why if only these 2 machines are using it? Nov 14 09:17:15 Maybe collie and poodle have different keymaps. Nov 14 09:17:40 Also kbd tweak does some more changes - like changing scancodes of some keys Nov 14 09:17:54 http://pages.cs.wisc.edu/~lenz/zaurus/ Nov 14 09:18:03 seems the initial source Nov 14 09:20:47 I'm not going to read through 2.4 source to find the keymap. It is ugly there. Nov 14 09:21:15 Haojan said once I should send it upstream Nov 14 09:21:37 we carry around this patch since 2.6 Nov 14 09:23:08 maybe putting it in your locomo patchset? Nov 14 09:29:18 I will think about it. Nov 14 09:29:29 It contains at least one mistake if I'm not mistaken. Nov 14 09:30:11 And mail/menu keys are switched. Nov 14 09:39:14 I will check later. Nov 14 09:42:40 ant_work, there might be a quick fix for gpio-keys configuration. Nov 14 09:48:06 Hmm. Nov 14 09:48:07 Or not. Nov 14 09:48:12 It is defined correctly. Nov 14 09:48:43 what is the issue? random crashes at boot? Nov 14 09:48:59 No. Nov 14 09:49:21 When testing new locomo driver I got stalls if I had gpio-keys enabled. Nov 14 09:49:31 This might be related to irq problems. Nov 14 09:49:35 Or might be not. Nov 14 09:51:12 I have it enabled ... input: gpio-keys as /devices/platform/gpio-keys/input/input2 Nov 14 09:51:49 If you press menu or mail keys, does the unit survive? Nov 14 09:52:45 I'll test Nov 14 09:54:08 we have it since 3.10 Nov 14 09:55:38 afair I couldn't trigger syreq Nov 14 10:21:02 lumag: I've found a thread where they say solution to er 38 was reformat dthe card + redo partitons Nov 14 10:23:02 Disabling gpio-keys helps to resolve that hang. Nov 14 10:24:18 link to the thread/ Nov 14 10:25:02 https://community.freescale.com/thread/91654 Nov 14 10:25:11 There might be an issue with SDHC/SDXC cards, since (if I understood correctly) SPI and native modes threat sector size differently for big card.s Nov 14 10:42:02 hm... I understand it returns R2... Nov 14 10:42:28 if that's true it would be 3 error-bits set Nov 14 10:43:05 bit 1 wp erase skip / lock-unlock failed Nov 14 10:43:23 bit 2 error Nov 14 10:43:32 bit 3 wp violation Nov 14 10:44:49 wp = write protect Nov 14 10:45:15 38 = 00100110 Nov 14 10:46:18 bit 0 means Card is locked Nov 14 10:46:18 —Supported by the SanDisk SD Card. Nov 14 10:46:26 but is unset... Nov 14 10:48:27 (I could have reversed lsb and msb :p ) Nov 14 10:49:09 * MMC/SD in SPI mode reports R1 status always, and R2 for SEND_STATUS Nov 14 10:49:32 ...linux.git/tree/include/linux/mmc/mmc.h Nov 14 10:50:14 need to switch on more debug Nov 14 10:57:38 reading now physical specs..http://users.ece.utexas.edu/~valvano/EE345M/SD_Physical_Layer_Spec.pdf Nov 14 10:58:20 new: https://www.sdcard.org/downloads/pls/simplified_specs/part1_410.pdf Nov 14 11:00:41 and that's sdio http://www.google.it/url?sa=t&rct=j&q=&esrc=s&source=web&cd=7&cad=rja&uact=8&ved=0CEcQFjAG&url=http%3A%2F%2Fwww.sandisk.com%2Fmedia%2FFile%2FOEM%2FManuals%2FSD_SDIO_specsv1.pdf&ei=1N5lVNmLDoK8ygPY74LoDQ&usg=AFQjCNEmERWNqBnMJzJVRX6qi_ItlIv6bg&sig2=V8g2I9C0fIs72VVSPBYeeQ&bvm=bv.79142246,d.bGQ Nov 14 11:00:45 argh ..sorry Nov 14 11:01:00 www.sandisk.com/media/File/OEM/Manuals/SD_SDIO_specsv1.pdf Nov 14 11:19:42 ant_work, -38 is -errno. Nov 14 11:20:00 :) Nov 14 11:49:00 http://lists.infradead.org/pipermail/linux-arm-kernel/2011-June/053833.html Nov 14 11:49:01 ^ Nov 14 11:49:24 just reading an old patch ;) err = mmc_wait_for_cmd(card->host, &cmd, retries); Nov 14 11:50:15 I really have to enable mmc debug to see Nov 14 11:51:28 it is indeed similar to mtd Nov 14 11:52:13 ...so I'll need 6 months to grasp it ;) Nov 14 11:52:47 the SanDisk syndrome is strange, though Nov 14 12:30:22 ant_work, I don't know. Maybe it's just a misfeature of SanDisk firmware. Nov 14 12:30:50 It might be a bug in locomo spi controller. Nov 14 13:00:48 today I'll try on old collie Nov 14 13:00:51 2.4 Nov 14 13:01:10 I bet the card will work Nov 14 13:01:28 the <1GB at least Nov 14 13:02:03 btw if you flash the old kexecboot kernel you get the detection done by 3.10+old patc Nov 14 13:02:15 err.. the old is 2.6.31 Nov 14 13:02:25 3.10 has no SD Nov 14 13:17:53 Jay7: btw can you recover the kernels on kexecboot.org? Nov 14 13:18:53 lumag: heh...desk is free of patches.Now waiting RMK... Nov 14 13:19:10 well done Nov 14 13:40:32 yes. rmk seems to really (mostly) ignore sa1100 (or me). Nov 14 13:43:20 I have few more patches in my sleeve. Nov 14 13:43:39 But I'd really like to finish with locomo refactoring and pushing sa1100 bits through rmk. Nov 14 13:46:27 Thankfully Robert is much more responsive. Nov 14 13:46:39 So pxa patches usually can flow in. Nov 14 13:48:23 Sometimes I have the feeling that it would be much easier to restart sa1100 support. To start it from scratch. Nov 14 13:48:46 But I think the ARM maintainer will not approve that. Nov 14 13:56:23 ah, linusw swnt 3.17 for h3600 Nov 14 13:56:26 *sent Nov 14 14:01:45 pointer? Nov 14 14:02:04 oe-core ML? Nov 14 14:03:04 oe-dev Nov 14 14:03:11 for meta-handheld Nov 14 14:03:49 Found it. Nov 14 14:03:56 Good. Nov 14 14:04:01 we have an ally Nov 14 14:08:20 bluelightning: pls try to keep up with the patches this weekend Nov 14 14:08:36 we are preparing for 3.18 Nov 14 14:09:15 klibc is broken in master (sigh..again) Nov 14 14:11:00 bluelightning: btw, seen the thread about sato/broken calibration? Nov 14 14:44:26 bbl Nov 14 17:58:49 ant_home: ok will try to do that Nov 14 17:58:59 ant_home: no haven't seen that yet Nov 14 17:59:01 bbl Nov 14 18:03:54 lumag, on Sharp Rom only the Lexar 1G and SanDisk 64M are recognized Nov 14 18:04:21 the 64M still gives -38 error now Nov 14 18:05:01 as the SanDisk 1G Nov 14 18:05:14 (and ofc larger) **** ENDING LOGGING AT Sat Nov 15 03:00:00 2014