IRC log of #cubox of Thu 17 Apr 2014. All times are in CEST < Back to index

01:38 _dab_ cbxbiker61: which sandisk cards are you using on the i4/i2? extreme or ultra?
02:27 cbxbiker61 32G - extreme plus is the highest performance sd, i've used, also eht 32G 30M card works well too
02:50 Forrest-- hello - has anyone gotten any of the last 3 versions of geexbox to boot on a cubox-i4pro ?
03:26 _dab_ Forrest--: I cant get the last two to boot on i4 or i2u.
03:27 _dab_ thanks biker
03:29 Forrest-- hey dab
03:29 _dab_ hey
03:29 Forrest-- actually, funny, i j ust found out it's booting, i just didn't know because the display stopped working
03:29 Forrest-- but i can SSH in
03:29 Forrest-- i mean it definitely was having problems, it was auto-rebooting
03:29 Forrest-- but at this very second it seems to have been booted up
03:31 _dab_ I am just sitting around trying to guess what I might try next.
03:32 _dab_ out of ideas
03:34 _dab_ Forrest--: I dont think the SD cards are the problem either. I have some the work on the i4 but not on the i2 and vice versa
04:37 _dab_ geexbox apr16 nightly on transcend 8GB boots on i4pro, will not boot (no red light) on i2ultra
06:00 _dab_ jnettlet: u there?
09:08 rabeeh _dab_: i'm here
09:08 jnettlet _dab_, I am as well
09:10 mk01 jnettlet: yesterday evening I tested same kernel with previous u-boot (at commit 3fd848b597804f6a5b000136cd31ecf1fda00430) and all was well
09:13 rabeeh _dab_: https://github.com/SolidRun/u-boot-imx6/commit/dfb384fd31a103078b1e35dc83e53267367d4cf2
09:14 rabeeh that one fixed the u-boot sd card issues
09:18 _dab_ yes rabeeh that commit is in the uboot I am using via git pull.
09:18 jnettlet mk01, okay sounds like you are seeing problems with the pfd fix. :-\
09:18 mk01 jnettlet: although wlan reports "brcmfmac: _brcmf_set_multicast_list: Setting BRCMF_C_SET_PROMISC failed, -23" on DHCP (but gets IP without issues) - even though I have the feeling Russel was telling here on IRC in a discussion that -23 is not a real error. i don't remember exactly.
09:19 _dab_ jnettlet: do you have an i2
09:19 rabeeh jnettlet: the DDR QOS improvement that was on LK 3.0.35 is not in LK 3.10.30
09:19 jnettlet _dab_, dl not d
09:19 rabeeh where do you think the proper way to have that? in u-boot or LK?
09:20 rabeeh this is probably the issue that Forrest-- previously reported about
09:20 mk01 jnettlet: but the wlan is not related to the other problem, just started to seeing it recently but don't remember when exactly (because again wlan is working fine anyhow)
09:21 mk01 jnettlet: if you need some help with testing later regatrding PFD, let me know
09:21 jnettlet rabeeh, so far in your geexbox testing we haven't needed it since they bumped up the ipu burst size
09:21 jnettlet mk01, I need to spend some serious time on the wlan driver
09:22 mk01 jnettlet: seriously, that wlan was just 'btw'
09:22 jnettlet tomlohave, has also pointed out a patch floating around so you don't need to run patchram in userspace
09:22 jnettlet that is for bluetooth of course
09:23 jnettlet rabeeh, ultimately I think I would like to expose the QOS settings to userspace via sysfs.
09:24 rabeeh jnettlet: there were reports for lower resolutions that it gets blanking effect (which is the IPU underruns)
09:24 rabeeh i recall that from sraue too
09:24 rabeeh but regardless; the i1 with 32bit would need that fix
09:24 jnettlet at lower resolutions?
09:25 rabeeh yes; the issue would happen only on lower resolutions since upscaling is a mem2mem via IC that puts more load on the ddr
09:25 mk01 jnettlet: lower resolution video files
09:25 jnettlet I have run my HB1 at 1080p 32-bit depth without the blanking effect showing up.
09:25 mk01 I was having it too
09:25 jnettle 09:25 * jnettlet will test
09:25 rabeeh even 480p would blank
09:26 jnettlet xbmc is doing direct-textures now. Seems as though the GPU should be doing the scaling
09:26 rabeeh jnettlet: imho that should be part of the dtb
09:26 mk01 need to go guys, see you around
09:26 mk01 bye
09:26 rabeeh mk01: ttyl
09:28 rabeeh _dab_: just sent you email; do you see anything on the serial console on i2u with latest u-boot?
09:30 _dab_ rabeeh: no
09:31 rabeeh and how stable is the older u-boot?
09:32 rabeeh the older u-boot with the Debian image is a pre SPL one
09:32 rabeeh it was only for i4pro; so i doubt it will fully boot to Debian
09:32 rabeeh (because of 2GByte memory vs. 1GByte)
09:33 _dab_ No, it is SPL and it is stable apart from having to cold boot. warm boot a bit shaky
09:35 _dab_ the debian image is my own debootstrap and the uboot I built from the solidrun git
09:36 _dab_ I took the uSD out of the i4pro and it worked on the i2u
09:36 jnettlet which it should
09:37 jnettlet they are both very similar
09:37 jnettlet minus memory
09:40 _dab_ jnettlet: that was true for me, up until Mar24. Now, not so.
09:43 _dab_ U-Boot 2013.10-rc4-g3fd848b (Mar 09 2014 - 10:33:11) worked ok
09:43 _dab_ 09 not 24
16:32 wbx yeah, my cubox-i arrived!
16:43 rabeeh wbx: congrats
16:45 wbx hmm. trying to boot a self compiled kernel. is make uImage always required or does u-boot can boot a zImage?
16:46 davorin zImage it can
16:47 wbx the wiki does not contain any hints what to do with the zImage and dtb files.
16:47 davorin you just copy them over to the win partition
16:48 rabeeh wbx: bootz
16:48 wbx ah, I have to use bootz instead of bootm.
16:49 davorin can i haz some bootz?
16:53 wbx rabeeh: do you have an example for uEnv.txt?
16:59 rabeeh CuBox-i U-Boot > printenv bootargs
16:59 rabeeh bootargs=console=ttymxc0,115200 root=/dev/mmcblk0p2 rootfstype=ext4 rootwait rw consoleblank=0 video=mxcfb0:dev=hdmi,1920x1080M@60,if=RGB24,bpp=16 dmfc=3
16:59 rabeeh CuBox-i U-Boot > printenv bootcmd
16:59 rabeeh bootcmd=tftp 0x10800000 zImage; tftp 0x18000000 imx6q-cubox-i.dtb; bootz 0x10800000 - 0x18000000
16:59 rabeeh those are for instance to boot GB
17:23 wbx rabeeh: thx. kernel now is trying to start, but I get:
17:23 wbx did not find a cmdline flattened device tree
17:24 wbx ah, i think I found it.
17:26 wbx shouldn't use echo "" with $var