| 00:03 | MikeSeth | ok time to hack |
| 00:05 | MikeSeth | boot |
| 00:05 | MikeSeth | oops |
| 00:58 | Coburn | _rmk_: are you there? |
| 01:19 | SoCX2 | how exactly do i go about setting up console on uboot? i know to put console=ttymxc0,115200n8 somewhere but the wiki doesn't really say where.. |
| 01:19 | _rmk_ | Coburn: yep |
| 02:22 | Coburn | oops |
| 02:22 | Coburn | didn't see _rmk_'s reply |
| 02:22 | Coburn | is he still thar? |
| 08:10 | aplund | I just posted something to the forums, but just discovered this channel, so I might ask here too: Where do you get the libvpu library? |
| 08:29 | jnettlet | aplund, the easiest place to get them is from the Yocto mirrors. http://download.ossystems.com.br/bsp/freescale/source/ |
| 08:48 | aplund | jnettlet: thanks |
| 09:54 | _Adik_ | jnettlet: yesterday I recieved two more SD cards and made some more tests |
| 09:56 | _Adik_ | funny thing is that both new cards (the one without and the one with UHS-1 but from different vendor) worked without problems, even on unpatched u-boot, kernel also works fine on both |
| 09:56 | _Adik_ | bo there is defineatly something specific to the Sandisk Extreme Pro cards |
| 09:56 | _Adik_ | *so |
| 10:01 | _Adik_ | the card itself is fine as I am able to read and write it on both my pc and mac without any issues, it is freakin fast (95/65MB/s) so would be nice if I could use it on cubox-i |
| 11:23 | jnettlet_ | _Adik_, the problem is actually an SOC errata and won't happen all the time. Are you seeing other problems with the card besides the boot problem? |
| 11:40 | _Adik_ | jnettlet_: yep, after it sucesfully boots there are read problems |
| 11:40 | jnettlet_ | _Adik_, what kernel are you using? |
| 11:40 | _Adik_ | 3.0 |
| 11:41 | _Adik_ | BUT as I mentioned, it seems that it is not a kernel problem as second SD card, which is also UHS-1 enabled (toshiba) works fine on the same kernel |
| 11:41 | jnettlet_ | It is probably trying to switch to voltage switch 1.8v signaling in order to do uhs speeds failing and not recovering. I have that fixed in the newer kernels but didn't know it was a problem in the 3.0.35 kernel |
| 11:42 | jnettlet_ | _Adik_, I bet if you look at the card properties in /sysfs you will see it is only running at 50Mhz |
| 11:43 | _Adik_ | jnettlet_: the question is, why Sandisk one does not work the same way as Toshiba |
| 11:43 | _Adik_ | i.e. does not fallback into standard slower class 10 w/o uhs-1 |
| 11:44 | jnettlet_ | _Adik_, because all the SDHC cards have their own ARM processors that run code to help negotiate this. |
| 11:44 | _Adik_ | hmm, you mean it can behave differently? |
| 11:45 | jnettlet_ | absolutely. The SD protocol is just a spec and both host and cards need to negotiate what can and can't be supported |
| 11:46 | MikeSeth | jnettlet_: there's a bunch of reports on the forums about images refusing to boot |
| 11:46 | MikeSeth | I wonder if this is related |
| 11:46 | jnettlet_ | possibly. Are any of them posting logs? |
| 11:46 | MikeSeth | I will go around and ask |
| 11:46 | jnettlet_ | thanks |
| 11:47 | _Adik_ | jnettlet_: do you have a image of any linux that contains your kernel with fixes for me to try? |
| 11:47 | jnettlet | 11:47 * jnettlet_ didn't think the 3.0.35 image was supporting uhs modes at all |
| 11:48 | MikeSeth | from a cursory glance, the two main problems on user side are the card problems and blank screen after kernel boot |
| 11:48 | jnettlet_ | _Adik_, nope I am still working on the patches. My work was derailed by my CBi having problems. Haven't been able to get a new dev CBi to finish it yet because demand is so high |
| 11:49 | jnettlet_ | MikeSeth, which images? |
| 11:49 | _Adik_ | jnettlet_: I see, but still - if you have something to test, feel free to ping me |
| 11:49 | _Adik_ | in addition, I can even send the 'problematic' card to someone to debug :) |
| 11:50 | jnettlet_ | _Adik_, unfortunately debugging mmc remotely is just useless. |
| 11:50 | MikeSeth | jnettlet_: fsl ubuntu, android, I think one report about debian too |
| 11:50 | jnettlet_ | _Adik_, I have one, which is why I was working on this. |
| 11:50 | MikeSeth | I don't want to mislead you so let me get back to you |
| 11:50 | MikeSeth | also I've spoken to rabeeh about a community issue tracker, maybe that will get done |
| 11:50 | jnettlet_ | MikeSeth, well really we need to know, TV's type of video connection, and resolution they are trying to run at. |
| 11:50 | _Adik | 11:50 * _Adik_ wonders if the one jnettlet_ have is a Sandisk Extreme Pro one |
| 11:51 | _Adik_ | this is 'top of the line' microsd card for now |
| 11:52 | jnettlet_ | _Adik_, mine are Sandisk Ultra ush1 cards. They are faster at random 4k writes :-) |
| 11:52 | jnettlet_ | which makes running desktop linux on them a much nicer experience |
| 11:52 | _Adik_ | :-) |
| 11:54 | _Adik_ | jnettlet_: the question is if i.mx6 can use the full potential of these cards |
| 11:55 | jnettlet_ | _Adik_, yes. The Cubox-i has the circuitry to support uhs 1.8v signal speeds. |
| 11:56 | jnettlet_ | although I have yet to actually successfully report using it from any of the manufacturers. All the code I have seen submitted has been broken in one way or another so even though they got the cards "working" they weren't actually doing the switch over to UHS |
| 11:57 | _Adik_ | hmm |
| 11:57 | _Adik_ | so I have not better choice like just sit and wait for now |
| 11:59 | jnettlet_ | well you have two other cards that work, so it isn't like you are dead in the water. Also what are you doing that really requires the extra throughput? |
| 12:01 | _Adik_ | jnettlet_: well you know how it works, I just spent some $ on cubox-i so I just wanted to put the fastest sd card inside, unfortunately the card that should be that fast is not working at all :) |
| 12:02 | _Adik_ | so I bought two more cards that were both chesper than the Sandisk and both are working well... |
| 12:02 | _Adik_ | life is tough :) |
| 12:03 | _Adik_ | but replying for your question, I ordered two cubox-i'es, one for media-center, second one will be my personal server for services like http, mail etc |
| 13:24 | _dab_ | the plural of cubox-i is cubox-ii ? or cubox-ies? |
| 13:35 | _rmk_ | and how do people pronounce cubox anyway :) |
| 13:49 | _dab_ | qbox |
| 13:59 | jnettlet_ | cube ox? |
| 14:00 | jnettlet_ | see you box |
| 14:00 | jnettlet_ | cue box |
| 14:01 | _dab_ | given my load averages on my i4pro maybe queue box |
| 14:02 | jnettlet_ | _dab_, you are probably lacking the ipu patch, or booting with cpufreq default of ondemand |
| 14:03 | jnettlet_ | both cause a load average of right around 1 all the time |
| 14:03 | _dab_ | aha. I am lacking a recent kernel. On my list is to build the latest stable and see what does not work |
| 14:04 | jnettlet_ | _dab_, are you using 3.0.35 or upstream? |
| 14:05 | _dab_ | 3.0.35 |
| 14:05 | jnettlet_ | you are probably lacking the ipu patches then |
| 14:05 | _dab_ | It is time to to move to 3.13 |
| 14:06 | _dab_ | I need to get wifi and bt stable. That should be easy at 3.13 |
| 14:06 | jnettlet_ | hahahaha |
| 14:06 | _dab_ | hoho |
| 14:07 | jnettlet_ | we are stilling sorting out the device-tree needed to make that seamless |
| 14:07 | jnettlet_ | and there will be no gpu or vpu acceleration. |
| 14:09 | _dab_ | I really dont care too much about the accel, As long I has have something displays graphics. Not fussed about video *yet*. |
| 14:10 | _dab_ | How far away is the DT? |
| 14:10 | jnettlet_ | _dab_, you will need _rmk_'s patchset on top of 3.13 to get hdmi output and 3.13 support. Allthough the device-tree half has been accepted by Linux now for 3.14 inclusion |
| 14:11 | jnettlet_ | my 3.10 kernel is progressing, down to the last couple of issues |
| 14:11 | _dab_ | Where do I get rmk's patchset? |
| 14:11 | _dab_ | and where do I get your 3.10 (longterm?) |
| 14:18 | _dab_ | was it the wrong question? |
| 15:08 | kuguar-tsk | try use geexbox&youtube, found my favorite channel, open stream... |
| 15:11 | kuguar-tsk | and nothing, black window |
| 15:16 | diget | hi |
| 15:17 | kuguar-tsk | hi |
| 15:18 | dv_ | kuguar-tsk: I am unfamiliar with geexbox. does it start xbmc ? |
| 15:19 | kuguar-tsk | yes |
| 15:19 | kuguar-tsk | http://yadi.sk/d/KivLUPNvH7kvJ |
| 15:21 | dv_ | maybe something's wrong with its VPU support |
| 15:23 | dv_ | I recommend talking to the geexbox guys about this. without any logs etc. it is not possible to find out what is wrong |
| 15:23 | xraxor | im playinf YT videos in geexbox and fine |
| 15:24 | xraxor | but sometimes it acts funny |
| 15:24 | kuguar-tsk | hm |
| 15:24 | dv_ | youtube videos can come in several containers and bitstream formats |
| 15:24 | dv_ | so no wonder some can cause problems |
| 15:25 | dv_ | try out other youtube videos, also very old ones |
| 15:25 | xraxor | try this image from emmveepee http://imx.solid-run.com/forums/viewtopic.php?f=7&t=335&start=30#p2887 |
| 15:26 | xraxor | yep, sometimes it works but sometimes wont |
| 15:26 | xraxor | playlists seems to cause probs as well |
| 15:34 | crypt0s | anyone here make a uImage kernel for Uboot before? |
| 16:01 | MikeSeth | crypt0s: what's the question? |
| 16:07 | crypt0s | MikeSeth: was having issues getting uImage create |
| 16:08 | crypt0s | MikeSeth: but I fixed it by looking up the loadaddress after compile of imx6 uboot tools |
| 16:08 | crypt0s | obstensibly all I have to do right now is write the SPL and uBoot into the SD, create the EXT4 partition, then unpack my stage3 tarball |
| 16:53 | MikeSeth | dd if=binary/SPL of=$DEVICE bs=512 seek=2 |
| 16:53 | MikeSeth | dd if=binary/u-boot.img of=$DEVICE bs=1K seek=42 |
| 16:53 | MikeSeth | crypt0s: ^^^ |
| 17:05 | crypt0s | MikeSeth: I did that I was talking about getting the kernel image compiled in uboot format |
| 17:05 | crypt0s | MikeSeth: although - when I did those steps, my Cubox-i did not bring up the SPL or uBoot...I don't know if I messed something up and I dont know where to put the uEnv file |
| 17:07 | _rmk_ | crypt0s: when you created the EXT4 partition, you did avoid the first 1MB of the SD card? |
| 17:08 | _rmk_ | if the ext4 partition overlaps uboot, it won't work very well |
| 17:09 | _rmk_ | right, fingers crossed, I should have wifi working entirely from dt now |
| 17:09 | _rmk_ | still a bit of a hack (for the oscillator enable) but... |
| 17:10 | _rmk_ | 4: wlan0: mtu 1500 qdisc noop state DOWN qlen 1000 link/ether 40:2c:f4:ae:27:0b brd ff:ff:ff:ff:ff:ff |
| 17:11 | _rmk_ | hci0: Type: BR/EDR Bus: UART BD Address: 43:29:B0:00:2F:DD ACL MTU: 1021:6 SCO MTU: 64:1 UP RUNNING PSCAN |
| 17:15 | MikeSeth | crypt0s: uenv goes to / of the first partition |
| 17:54 | MikeSeth | hrm, what am I doing wrong |
| 17:54 | MikeSeth | uboot says no partition table on mmc0 |
| 18:03 | MikeSeth | mmc read 0 1 1 returns error |
| 18:03 | MikeSeth | mmc part says unknown partition table |
| 18:03 | MikeSeth | hints anyone? |
| 18:12 | jas-hacks | MikeSeth: is it the correct mmc device? |
| 18:12 | MikeSeth | jas-hacks: yeah, it shows as 0 in mmc.list |
| 18:13 | MikeSeth | but apparently I was using a wrongly configured uboot, when I try precompiled images from rabeeh this does not happen |
| 18:14 | jas-hacks | did you compile it for the correct board? |
| 18:15 | MikeSeth | No, I yanked them, quite possibly from wrong/old image |
| 18:15 | MikeSeth | I'll build my own ones and see what happens |
| 20:01 | MikeSeth | jnettlet_: just imaged the first 1Mb from a working rootfs, custom built uboot, wrote it over to the SD, no partition table found |
| 20:01 | MikeSeth | I'll go dig in hste's scripts, maybe he figured out some magic I'm missing |
| 20:11 | MikeSeth | ok, zeroed out the card, rebuilt uboot, wrote uboot and created a single partition, uboot says no partition table |
| 20:11 | MikeSeth | something is screwy |
| 20:12 | MikeSeth | can it be windows/virtualbox corrupting the card? |
| 20:15 | _rmk_ | What does fdisk -l /dev/whatever-your-sd-card-is-called say? |
| 20:15 | hste | MikeSeth: if u put the card back in the pc, do u have a partitiontable then? |
| 20:16 | MikeSeth | hste: yes |
| 20:16 | MikeSeth | _rmk_: it shows the partition table normally |
| 20:16 | MikeSeth | but I think I found the probmel |
| 20:16 | _rmk_ | it should look like this: |
| 20:16 | _rmk_ | Device Boot Start End Blocks Id System |
| 20:16 | _rmk_ | /dev/mmcblk0p1 2048 262144 130048+ c W95 FAT32 (LBA) |
| 20:17 | MikeSeth | uh, wait |
| 20:17 | _rmk_ | note the start... |
| 20:17 | MikeSeth | first partition HAS to be fat? |
| 20:17 | _rmk_ | I don't believe so |
| 20:17 | hste | MikeSeth: I use a single ext4 and that works for me |
| 20:17 | _rmk_ | the important thing is the start doesn't overlap where uboot gets stored |
| 20:17 | MikeSeth | _rmk_: fdisk insists the first partition be on offset 2048 anyway |
| 20:18 | _rmk_ | good, so that's not the problem |
| 20:18 | _rmk_ | are you using a SD card reader? |
| 20:19 | MikeSeth | _rmk_: yes, in windows -> vagrant -> virtualbox -> x86 debian |
| 20:19 | MikeSeth | but I suspect I am writing the wrong uboot image |
| 20:19 | MikeSeth | I am writing uboot.img |
| 20:19 | hste | MikeSeth: have u tried dd the uboot again? |
| 20:19 | _rmk_ | I have seen some media card readers which try to be too clever and do stuff like automatically reformat the card |
| 20:19 | MikeSeth | but I just realized that the rest of instructions say write .imx |
| 20:19 | MikeSeth | hste: yeah, in several combinations |
| 20:20 | MikeSeth | am I supposed to post-process the built uboot image somehow? |
| 20:21 | MikeSeth | _rmk_: entirely possible, but the card reads back fine |
| 20:22 | MikeSeth | hste: the uboot images cloned from your jessie root fs work; those that I built don't |
| 20:23 | MikeSeth | also, mmc read 0 1 1 returns error |
| 20:23 | hste | MikeSeth: I have only the old one. do u try use the new with SPL? |
| 20:24 | MikeSeth | yeah |
| 20:25 | MikeSeth | uboot is basically unable to read the card, can this be caused by newer commits in rabeeh's tree? hmmm |
| 20:25 | MikeSeth | hste: how do you generate uboot.imx? |
| 20:28 | hste | MikeSeth:it get built when doing the make |
| 20:30 | MikeSeth | apparently not |
| 20:30 | hste | MikeSeth:the old uboot, not the new |
| 20:30 | MikeSeth | oh |
| 20:30 | MikeSeth | okay there is a major confusion in the wiki then |
| 20:31 | hste | looks right for me for the new uboot: sudo dd if=SPL of=/dev/sdX bs=512 seek=2 and sudo dd if=u-boot.img of=/dev/sdX bs=1K seek=42 |
| 20:32 | MikeSeth | yeah that's what I do |
| 20:32 | MikeSeth | /dev/sdf1 2048 7744511 3871232 83 Linux |
| 20:32 | hste | and on the new with SPL its u-boot.img |
| 20:32 | MikeSeth | the partition table is correct |
| 20:33 | hste | try only dd the SPL |
| 20:34 | MikeSeth | then I won't get the console |
| 20:34 | hste | yes but uboot-img shouldn't be erased doing that dd |
| 20:34 | MikeSeth | oh |
| 20:34 | MikeSeth | okay |
| 20:37 | MikeSeth | ARGH |
| 20:37 | MikeSeth | unbelievable |
| 20:37 | MikeSeth | it was the card itself! |
| 20:37 | MikeSeth | Just opened a blank card, ran the script and it sees the partition just fine |
| 20:37 | MikeSeth | something is very screwy with that card |
| 20:38 | MikeSeth | in fdisk you can not create a partition (you press new, it asks you for size, you press enter, no new partition) |
| 20:38 | MikeSeth | so I did that in cfdisk |
| 20:38 | MikeSeth | I thought that's just local fdisk weirdness but that didnt happen on the new card |
| 20:38 | MikeSeth | so I am looking at something extra curious here |
| 20:40 | MikeSeth | can it be some sort of DRM crap on the card? |
| 20:46 | MikeSeth | and it boots! |
| 20:47 | MikeSeth | the faulty card is Transcend Micro |
| 20:47 | MikeSeth | the good one is Sandisk |
| 20:48 | MikeSeth | now I have working chain to build debian images |
| 20:50 | MikeSeth | jnettlet_: beside Sandisk Ultra, are there other cards that have issues w/current uboot/kernel? I just discovered something funny with a Transcend Micro card, uboot is unable to read the first sectors from it (though it can and will load itself), so I want to post about this on the forums |
| 20:51 | MikeSeth | jnettlet_: s/Sandisk Ultra/Sandisk Extreme Pro/ |
| 21:14 | jnettlet_ | MikeSeth, not that I know of. I haven't been paying too close attention to brands. My classification has been uhs/sdhc |
| 21:17 | MikeSeth | alright, thanks |
| 21:17 | MikeSeth | any guess what would make a Transcend Micro card unreadable to uboot? |
| 21:53 | _dab_ | MikeSeth: I have had no problems with Transcend cards(class 10) or Integral cards(class 10). |
| 21:55 | _dab_ | However, I have had problems with cards placed in a windows SD drive. Linux, no problem. Let a card near a windows 7 box then that is a problem. |
| 21:56 | _dab_ | occasionally mashes the partition table |
| 21:56 | _dab_ | i.e. MBR |
| 21:58 | MikeSeth | _dab_: huh, that might very well be the case here |
| 21:58 | MikeSeth | though it only does this for one card but not the other |
| 22:02 | _dab_ | I use: echo -e "n\np\n1\n\n\nw\n" | fdisk $DEVICE |
| 22:02 | _dab_ | then ext4 |
| 22:04 | MikeSeth | _dab_: I use sfdisk to automate, but in fdisk what happens with the broken card is that after entering 1 it drops back to the menu and the partition is not created; no error displayed |
| 22:20 | _dab_ | opps: first echo -e "o\nw\n" | fdisk $DEVICE |
| 22:21 | _dab_ | That is two echos, and afterwards sync; hdparm -z $DEVICE |
| 22:22 | _dab_ | s/opps/oops/ |
| 22:26 | _dab_ | i.e. echo -e "o\nw\n" | fdisk $DEVICE; echo -e "n\np\n1\n\n\nw\n" | fdisk $DEVICE |