| 00:18 | mk01 | msalerno: yes. probably you don;t have correct MMC/SDIO drivers compiled ? |
| 00:19 | mk01 | (if you have custom kernel) |
| 00:19 | msalerno | I do not. It's the arch kernel package |
| 00:19 | msalerno | was booting fine |
| 00:20 | mk01 | ok. and what is the cmdline for kernel ? |
| 00:23 | msalerno | At the moment, it's: mmcroot=/dev/mmcblk0p2 rootwait rw info console=ttymxc0,115200 console=tty1 consoleblank=0 video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24,bpp=16 dmfc=3 |
| 00:23 | msalerno | and now it's working |
| 00:23 | msalerno | all I did was change debug to info |
| 00:24 | msalerno | When I first setup Arch, I was getting: lirc_serial lirc_serial.0: Bad irq number or handler |
| 00:25 | msalerno | Then I changed the uEnv to include debug, and it started working |
| 00:27 | mk01 | I have lirc_serial as module but after load have the same |
| 00:27 | mk01 | [ 1430.937846] lirc_serial: module is from the staging directory, the quality is unknown, you have been warned. |
| 00:27 | mk01 | [ 1430.938753] lirc_serial lirc_serial.0: Bad irq number or handler |
| 00:27 | mk01 | [ 1430.938774] lirc_serial: probe of lirc_serial.0 failed with error -22 |
| 00:27 | mk01 | [ 1430.938918] platform lirc_serial.0: lirc_dev: driver lirc_serial registered at minor = 0 |
| 00:27 | mk01 | so don't worry about that one |
| 00:27 | msalerno | Right, i'm not worried about the error, but my system would freeze on that line during bootup |
| 00:28 | msalerno | systemctl disable bluetooth brcm43xx |
| 00:28 | mk01 | I got the point, yes |
| 00:28 | msalerno | Wrong window |
| 00:29 | mk01 | what you have around MMC init ? |
| 00:29 | mk01 | [ 2.146634] sdhci: Secure Digital Host Controller Interface driver |
| 00:29 | mk01 | [ 2.146636] sdhci: Copyright(c) Pierre Ossman |
| 00:29 | mk01 | [ 2.146638] sdhci-pltfm: SDHCI platform and OF driver helper |
| 00:29 | mk01 | [ 2.147430] sdhci-esdhc-imx 2190000.usdhc: could not get ultra high speed state, work on normal mode |
| 00:29 | mk01 | [ 2.169544] mmc0: no vqmmc regulator found |
| 00:29 | mk01 | [ 2.170877] sdhci-esdhc-imx 2190000.usdhc: lookup for GPIO card-reset failed |
| 00:29 | mk01 | [ 2.462976] mmc0: SDHCI controller on 2190000.usdhc [2190000.usdhc] using ADMA |
| 00:29 | mk01 | [ 2.465833] mmc1: no vqmmc regulator found |
| 00:29 | mk01 | [ 2.483004] ata1: SATA link down (SStatus 0 SControl 300) |
| 00:29 | mk01 | [ 2.497733] mmc0: queuing unknown CIS tuple 0x80 (50 bytes) |
| 00:29 | mk01 | [ 2.506480] mmc0: queuing unknown CIS tuple 0x80 (7 bytes) |
| 00:29 | mk01 | [ 2.508660] mmc0: queuing unknown CIS tuple 0x80 (4 bytes) |
| 00:29 | mk01 | [ 2.538431] mmc0: queuing unknown CIS tuple 0x02 (1 bytes) |
| 00:29 | mk01 | [ 2.545654] mmc0: new SDIO card at address 0001 |
| 00:29 | mk01 | [ 2.562987] mmc1: SDHCI controller on 2194000.usdhc [2194000.usdhc] using ADMA |
| 00:29 | mk01 | [ 2.614608] mmc1: host does not support reading read-only switch. assuming write-enable. |
| 00:29 | mk01 | [ 2.640583] mmc1: new high speed SDHC card at address e624 |
| 00:29 | mk01 | [ 2.649845] mmcblk0: mmc1:e624 SU04G 3.69 GiB |
| 00:29 | mk01 | [ 2.657838] mmcblk0: p1 p2 |
| 00:29 | mk01 | mmc0 is BT |
| 00:30 | mk01 | mmc1 is actual SDcard |
| 00:34 | msalerno | here's the current init |
| 00:34 | msalerno | kernel: sdhci: Secure Digital Host Controller Interface driver |
| 00:34 | msalerno | kernel: sdhci: Copyright(c) Pierre Ossman |
| 00:34 | msalerno | kernel: sdhci-pltfm: SDHCI platform and OF driver helper |
| 00:34 | msalerno | kernel: sdhci-esdhc-imx 2190000.usdhc: could not get ultra high speed state, work on normal mode |
| 00:36 | mk01 | copy&paste one line at the time |
| 00:38 | msalerno | sorry |
| 00:38 | msalerno | kernel: sdhci: Secure Digital Host Controller Interface driver |
| 00:38 | msalerno | kernel: sdhci: Copyright(c) Pierre Ossman |
| 00:39 | msalerno | kernel: sdhci-esdhc-imx 2190000.usdhc: could not get ultra high speed state, work on normal mode |
| 00:39 | msalerno | kernel: mmc0: no vqmmc regulator found |
| 00:39 | msalerno | kernel: sdhci-esdhc-imx 2190000.usdhc: lookup for GPIO card-reset failed |
| 00:39 | msalerno | kernel: mmc0: SDHCI controller on 2190000.usdhc [2190000.usdhc] using ADMA |
| 00:39 | msalerno | kernel: mmc1: no vqmmc regulator found |
| 00:40 | mk01 | ok just the mmc1 is enough |
| 00:40 | mk01 | where it stopps ? |
| 00:40 | msalerno | ok |
| 00:40 | msalerno | kernel: mmc1: host does not support reading read-only switch. assuming write-enable. |
| 00:40 | msalerno | kernel: mmc1: new high speed SDHC card at address 0001 |
| 00:40 | msalerno | kernel: mmcblk0: mmc1:0001 00000 14.6 GiB |
| 00:40 | msalerno | kernel: mmcblk0: p1 p |
| 00:41 | mk01 | can you check the partition table somehow ? booting into ramdisk, or putting SD into other machine ? |
| 00:42 | msalerno | unfortunately or fortunately it's now working |
| 00:42 | mk01 | if there is "p1 p" for real (and not just missing from your copy) |
| 00:42 | mk01 | then p2 is not properly recognized |
| 00:42 | msalerno | I manually created all the partitions when I setup the OS |
| 00:42 | msalerno | the odd part is that sometimes it works and then sometimes it doesn't |
| 00:43 | mk01 | what is the FS ? |
| 00:43 | mk01 | btrfs / xfs / fs2 ? |
| 00:44 | msalerno | 01/fat and 02/ext4 |
| 00:45 | mk01 | still - once booted can you recheck the parts ? (parted or similar) |
| 00:46 | mk01 | and one remark - I have one microsd card 16gb big - it works on cubox but thaaaaat waaaaay SLOW |
| 00:46 | msalerno | Device Boot Start End Blocks Id System |
| 00:46 | msalerno | /dev/mmcblk0p1 8192 49151 20480 c W95 FAT32 (LBA) |
| 00:46 | msalerno | /dev/mmcblk0p2 49152 30702591 15326720 83 Linux |
| 00:46 | mk01 | cpu load, interrupts all at zero |
| 00:47 | mk01 | and copy of 1mb file takes like 30s |
| 00:48 | msalerno | My main card that I have been working with is 16gb |
| 00:48 | msalerno | the others are 4 |
| 00:49 | mk01 | ok the partitions are not aligned but it should not cause THAT effect. try asking Jnettlet tomorrow |
| 00:50 | msalerno | will do |
| 00:50 | msalerno | can you elaborate on not aligned? |
| 00:55 | mk01 | what is showing "fs2tune -l /dev/mmcblk0p2" ? |
| 00:55 | mk01 | and sorry p1 is aligned |
| 00:55 | mk01 | p2 start also |
| 00:57 | msalerno | tune2fs 1.42.9 (28-Dec-2013) |
| 00:57 | msalerno | Filesystem volume name: |
| 00:57 | msalerno | Last mounted on: / |
| 00:57 | msalerno | Filesystem UUID: 38107a73-17cf-4aa6-a978-3429f23d15f2 |
| 00:57 | msalerno | Filesystem magic number: 0xEF53 |
| 00:57 | msalerno | Filesystem revision #: 1 (dynamic) |
| 00:57 | msalerno | Filesystem features: has_journal ext_attr resize_inode dir_index filetype needs_recovery extent flex_bg sparse_super large_file huge_file uninit_bg dir_nlink extra_isize |
| 00:57 | msalerno | Filesystem flags: signed_directory_hash |
| 00:57 | msalerno | Default mount options: user_xattr acl |
| 00:57 | msalerno | Filesystem state: clean |
| 00:57 | msalerno | Errors behavior: Continue |
| 00:57 | msalerno | Filesystem OS type: Linux |
| 00:57 | msalerno | Inode count: 958464 |
| 00:57 | msalerno | Block count: 3831680 |
| 00:57 | msalerno | ahh... Need to stop copy/paste |
| 01:06 | mk01 | msalerno: leave it just tell me block size if is 4K. anyhow it was just curiosity how will ext4 handle the end not being 4k block - it will handle for sure just wanted to see tune2fs reports around the block count & block size |
| 01:36 | msalerno | Block size: 4096 |
| 01:37 | msalerno | Block count: 3831680 |
| 03:19 | cbxbiker61 | i just started testing my new San Digital JBOD with cuboxi, so far so good, the read/write speeds are normal using iozone on a 3TB btrfs partition |
| 03:20 | cbxbiker61 | i'm going to use btrfs to raid 0 two 3TB just for kicks and see how well the data rate holds up |
| 03:21 | mk01_ | out of curiosity. what are the speeds ? seq read? |
| 03:21 | mk01_ | with raid0 on USB hub I get cumulated 40-50Mb/s |
| 03:22 | cbxbiker61 | random random |
| 03:22 | cbxbiker61 | KB reclen write rewrite read reread read write |
| 03:22 | cbxbiker61 | 51200 4 9481 7456 17421 16432 3770 6694 |
| 03:22 | cbxbiker61 | 51200 512 52459 54240 44427 47449 40465 53754 |
| 03:22 | cbxbiker61 | 51200 16384 66333 66505 89935 92603 104164 64988 |
| 03:22 | mk01_ | on eSata it should do over 200-250 |
| 03:22 | mk01_ | and this is single disk, right ? |
| 03:22 | cbxbiker61 | the drive is a 3 year old WD, so a single drive max is pretty much 130MB |
| 03:23 | mk01_ | yep, ... |
| 03:24 | cbxbiker61 | i'm installing gsmartcontrol, i want to see if the smart behaves properly first, then i'll raid0 the two 3TB |
| 03:27 | mk01_ | whats gsmartXXX? like smartctl / smartd ? |
| 03:28 | mk01_ | dumping smart registers ? |
| 03:28 | cbxbiker61 | gtk gui into smartctl, nice interface |
| 03:28 | mk01_ | ah ok |
| 07:09 | zombiehoffa | wow, haven't updated geexbox since early january and the improvements to the latest release are huge. |
| 07:20 | cbxbiker61 | i just did a copy test from my main server to my cubox-i with jbod, it does pretty much 40MB/sec, so no complaints there, it should work very nicely as a NAS. |
| 07:23 | cbxbiker61 | btrfs/raid performance pretty much stays the same due to the data all being carried over a single sata cable, so other than data duplicity there's nothing to gain there |
| 07:28 | cbxbiker61 | the smart functionality seems to function normally in the jbod, so that's good |
| 07:58 | Juggie | not sure why you would want to use a cubox as a nas |
| 07:59 | Juggie | certainly there is better suited hardware |
| 08:05 | jnettlet | depends what you want to do on the NAS. As a dlna content provider you can actually do realtime transcoding in hardware. |
| 08:06 | jnettlet | or run some node.js services on the box |
| 08:07 | Juggie | cubox has encoding hardware? |
| 08:07 | Juggie | decoding yes i agree but encoding? |
| 08:20 | cbxbiker61 | Juggie, i prefer my devices to be easily changeable, with dedicated nas what happens when a board fries, or you want to keep the kernel/utilities up to date? In my case I have interchangeable boxes that can be swapped out easily/upgraded |
| 08:21 | cbxbiker61 | interchangeable cubox's and easily replaced jbods, that's the ticket for me |
| 08:21 | cbxbiker61 | and like jnettlet said, yes they do hardware encoding |
| 08:25 | cbxbiker61 | and there is a nice use case...tvheadend/xbmc where all of those features would be useful |
| 08:36 | cbxbiker61 | now this has me thinking... i've been toying with rolling out a tvheadend/xbmc setup for my aunt, but i didn't want to have to do a "big" system... |
| 08:37 | cbxbiker61 | with a cubox/jbod acting as the server i could roll out a "cheap" video system |
| 08:38 | cbxbiker61 | i've rolled out two big systems...i should test a smaller approach |
| 08:41 | jnettlet | Juggie, yes the VPU can encode and decode. |
| 08:41 | jnettlet | encode is limited to h264 |
| 08:42 | cbxbiker61 | yep, and i think that would be fine for accumulating compressed tv shows |
| 08:44 | rabeeh | Juggie: dv__ has benchmarked that a while a go on C1 / gst-1.0 plugins; he was able to trasncode high bitrate advanced main profile h.264 to base h.264 almost in real time |
| 08:45 | rabeeh | C1 is like CuBox-i1 hardware wise |
| 08:45 | rabeeh | i'v seen yesterday a thread on plex server about that actually |
| 08:45 | cbxbiker61 | i'm really hard pressed to think of a use-case that requires the bandwidth of the sata, other than storing video files and such, when you get right down to it hardly any normal use doesn't do more than 5-10MB/sec |
| 08:46 | cbxbiker61 | rabeeh, hi, plex doing hardware transcode? |
| 08:46 | rabeeh | cbxbiker61: hi... nice work btw :) |
| 08:46 | cbxbiker61 | on what? |
| 08:46 | rabeeh | they use ffmpeg; so it would only apply to vdpau |
| 08:47 | cbxbiker61 | ok, well in my case i could do gstreamer just as easily |
| 08:47 | Juggie | jnettlet, nice... had no idea it did encoding too |
| 08:47 | Juggie | now if only there was a upnp/dlna server that took advantage of it! |
| 08:47 | rabeeh | Juggie: encoding is pretty much a standard today in every SoC that has camera-in |
| 08:47 | jnettlet | Juggie, rygel can, it is based on gstreamer |
| 08:47 | jnettlet | you just need to setup the transcoding pipeline |
| 08:48 | Juggie | cool, good to know |
| 08:48 | rabeeh | Juggie: which gst plugins are you looking at? |
| 08:48 | Juggie | nil |
| 08:49 | Juggie | i just didnt know the soc did h264 encoding |
| 08:49 | cbxbiker61 | and gstreamer 1.x is supported, or am i mistaken |
| 08:49 | rabeeh | https://github.com/Freescale/gstreamer-imx |
| 08:49 | rabeeh | gst 1.0 is mostly dv__ work |
| 08:50 | rabee | 08:50 * rabeeh gathers a big amount of credit and throws towards dv__ |
| 08:50 | cbxbiker61 | yeah, i'll test all of that out, i think it'll be a big win |
| 08:51 | cbxbiker61 | i didn't want to have to roll out fire-breathing hardware for non-techies |
| 08:54 | Juggie | cbxbiker61, hopefully though there would be no transcoding of media required, though that is not always possible |
| 08:55 | jnettlet | if you intend to support the chromecast you will need to transcode. Only supports html5 standards, mp4 and vp8 |
| 08:55 | jnettlet | oh right, I believe that the VPU can encode to vp8 |
| 08:56 | jnettlet | nope, mp4 and mjpg |
| 08:57 | cbxbiker61 | for short term storage, i don't need to worry about transcode, since both r-pi and cubox can handle that, for longer term storage it's more useful to cut down on disk space |
| 08:58 | rabeeh | i think the transcoding will be mainly used in those cases to a real time transcoding for devices that can play that HD content |
| 08:59 | Juggie | only device that would need transcoding would be chromecast and potentially mobile |
| 08:59 | Juggie | mobile/tablet |
| 08:59 | cbxbiker61 | i'm sure that was the intent, but it should work fine for my case also, we'll see |
| 08:59 | cbxbiker61 | mpeg is just sooo big |
| 09:00 | Juggie | the intent of transcoding though is not to lower the bitrate for streaming |
| 09:00 | Juggie | the negociation is put in place to transcode when the device cannot handle the format of the media |
| 09:00 | Juggie | if both ends do mpeg, it would most likely unless you can override it, stream the mpeg |
| 09:01 | cbxbiker61 | mpeg will always encode to smaller h264 |
| 09:01 | Juggie | of course. |
| 09:02 | cbxbiker61 | it's all a lot less of an issue now that 6TB drives are out, and pi is handling mpeg fine |
| 09:02 | Juggie | i am saying that if the server has mpeg2 and the client supports mpeg2 then standard behaviour would be to stream it as is |
| 09:02 | cbxbiker61 | sure, but we're storing video for later use, not instantaneous |
| 09:02 | Juggie | that said, there is 0 reason to store files in mpeg2 anymore with the exception of say OTA recordings |
| 09:03 | Juggie | which you may want to re-encode later if you intend to keep them |
| 09:04 | cbxbiker61 | i only watch live maybe 5% |
| 09:04 | cbxbiker61 | with my big server box i have enough cpu to do comskip |
| 09:05 | cbxbiker61 | seeking has improved greatly in xbmc, so commercial skip is not critical |
| 09:10 | cbxbiker61 | interesting...xbmc - EGLNativeTypeIMX.cpp is including linux/mxcfb.h, even though it doesn't seem to use it, which creates a problem since mxcfb.h has kernel only defs |
| 09:13 | mk01_ | cbxbiker61: if I remember correctly mxcfb.h is used at autoconfigure stage as "prove" we have imx6 hw |
| 09:14 | cbxbiker61 | true, probably should check something else, otherwise native compile is a pain |
| 09:14 | cbxbiker61 | for now i'm just going to kludge it to work |
| 09:15 | cbxbiker61 | i don't see any reason not to use the same test as vpu_wrapper, at least for now |
| 18:12 | vivek_ | has anyone tried gstreamer-imx on debian jessie ? |