| 01:35 | Coburn | Well, gotta admit |
| 01:35 | Coburn | if you're a console user, irssi can't be beat |
| 02:09 | dmitrijus | Coburn: yes. |
| 02:47 | Coburn | If anyone needs me, give me a ping |
| 02:48 | Coburn | I'm using irssi and it doesn't have native sounds |
| 02:48 | Coburn | (ping dbsx) |
| 02:48 | Coburn | I'll be here if you need me |
| 13:10 | dbsx | tis very quiet here |
| 13:21 | michaelcubox | *beep* |
| 13:26 | purch | lets make some noice then |
| 13:28 | purch | topic is vmeta size and kirkwood drivers: kernel with no vmeta -> no sound card and crash |
| 13:29 | purch | headless with audio and max RAM |
| 13:30 | purch | after this is solved I will post the solution to the forum |
| 13:43 | dv505 | I tried the armhf debian build headless with sound. via hdmi=no sound, spdif=sound |
| 13:43 | dv505 | related? |
| 13:46 | michaelcubox | if you use a kernel without the dove gpu, why should there be sound? |
| 13:46 | michaelcubox | over hdmi,as the unit is completely turned off |
| 14:21 | dbsx | hi purch, I think its just a matter of having a kernel configured with the right options for spdif. WHich kernel are you using? |
| 14:32 | dbsx | purch: reading your question again. I solved my hassles by patching kernel so that I some RAM (not much) for vmeta. This was a long time ago and no one seemed interested. It was just a kernel config. |
| 16:41 | purch | spdif only |
| 16:46 | purch | vdorst 3.5.7 patched git repo, cubox_defconfig with EFI GUID support, audio works. |
| 16:48 | purch | to that .config I did 5 changes for 1GB RAM and audio card is not found and boot hangs |
| 16:49 | purch | I try with vmeta=64 |
| 18:54 | purch | nice, vmeta=64 and GPU=16 |
| 18:54 | purch | lets test sound |
| 18:56 | purch | works! |
| 18:57 | purch | free says: total MEM 951060 |
| 19:17 | rabeeh | purch: great :) |
| 20:25 | purch | too early, libmad0 crashes mpd, needs a rebuild |
| 20:25 | purch | https://bugs.launchpad.net/ubuntu/+source/libmad/+bug/989846 |
| 20:30 | rabeeh | purch: which bitrate are you trying to play btw? |
| 20:30 | rabeeh | on mpd? |
| 20:41 | purch | format 44100:16:2 |
| 20:42 | purch | rabeeh: ^ |
| 20:44 | purch | 48000 wont help either |
| 20:54 | purch | rabeeh: rebuilding it with --disable-aso |
| 20:55 | purch | I hate ubuntu, they fixed it but rolled it only to next release 13.04 |
| 21:06 | purch | libmad0 rebuild fixed mpd |
| 21:07 | purch | well, maybe tomorrow I have time to write this journey to the forum |
| 21:08 | rabeeh | purch: that would be great |
| 21:09 | rabeeh | the patch they have submitted seems to disable some assembly code in libmad |
| 21:09 | rabeeh | i wonder why - probably it's hardfp ABI changes related |
| 21:09 | purch | something like that |
| 21:13 | purch | next to install deluge and if cubox runs out of RAM then it is time to order that 2GB version :) |
| 23:16 | dbsx | methinks that Solidrun should change the boot.scr to contain mtdparts. The kernel parameter should be provided. |
| 23:16 | dbsx | mtdparts=spi0.0:768k(u-boot),64k(uboot_env),64k(macaddr)ro,-(kernel) |
| 23:19 | shesselba | dbsx, rabeeh: in the not so distant future we also need a page for fdt blob |
| 23:29 | dbsx | shesselba: is the fdt blob less than 64k ? |
| 23:31 | shesselba | currently it is slighly larger than 4k, so 64k as the minimal page size of newer cuboxes is enough |
| 23:38 | dbsx | ok. so mtdparts=spi0.0:448k(u-boot),64k(uboot_env),64k(sundry)ro,-(kernel) |
| 23:38 | dbsx | the sundry blob which currently is used to save the mac address (16 bytes reserved) could be used also for the FDT. |
| 23:38 | dbsx | every uboot I have seen will fit in 448k or even 384k |
| 23:39 | dbsx | note that the mtdparts/mtdids set by the current uboot are junk (wrong) |
| 23:41 | dbsx | shesselba: or does the FDT need to be on a page boundary? |
| 23:43 | shesselba | I don't think it has to be on any boundary, it will be copied to ram like uImage |
| 23:43 | shesselba | besides, when having a fdt blob, mtdparts can be removed completely as it will also go into it |
| 23:46 | dbsx | ok then mtdparts=spi0.0:384(u-boot),64k(uboot_env),64k(sundry)ro,-(uimagel) |
| 23:46 | dbsx | and we can put the FDT at the start of the sundry page and the mac address (which is only accessed by code) at the end of the sundry page. I presume the FDT does not change unless the hardware changes. |
| 23:47 | shesselba | fdt will be static for each cubox revision, except cmdline, local-mac-address, and some minor stuff maybe. but that should be handled by u-boot |
| 23:48 | shesselba | it will read the fdt blob and stored mac address and insert it in the fdt prior boot |
| 23:49 | shesselba | u-boot patches for dove and cubox have been posted and awaiting review. I _hope_ they will get into next release |
| 23:56 | dbsx | well if you into posting uboot patches then |
| 23:56 | dbsx | + #if defined(CUBOX) |
| 23:56 | dbsx | + setenv("mtdids","nor0=spi0.0"); |
| 23:56 | dbsx | + setenv("mtdparts","mtdparts=spi0.0:512k(u-boot),64k(uboot_env),64k(macaddr),-(uimage)"); |
| 23:56 | dbsx | + setenv("console","console=ttyS0,115200"); |
| 23:56 | dbsx | These 3 lines are wrong in rabeehs/marvells u-boot |
| 23:57 | dbsx | err... make that console=ttyS0,115200n8 |
| 23:58 | shesselba | please prepare a patch as soon as u-boot patches are in. or remind me of it |