| 00:08 | dbsx | rabeeh: re the multiple DT maintenance. With Dreamplugs (which subtley change their hardware over time) we have many versions of the plug that do not work with a DT but work with ONE simple startup module. So you end up with a choice of 1 kernel and many DTs or just one kernel with a patched startup program. |
| 00:57 | ojn | _rmk_: hummingboard is added to the boot farm, it should start doing boots as soon as Shawn's branch is published and included in -next |
| 00:58 | _rmk_ | ojn: cool, thanks. |
| 00:59 | ojn | I ran out of ethernet ports and had to go buy another 24-port switch. :) |
| 00:59 | _rmk_ | heh. |
| 00:59 | _rmk_ | well, I have a digi portserver coming at some point |
| 01:00 | _rmk_ | so I should be able to add more stuff to my farm |
| 01:01 | ojn | I haven't found a cost-efficient way of using multi-port serial devices, so I have a large tree of usb-to-serial adapters and a large usb hub (and a few small ones) |
| 01:02 | _rmk_ | the reason I don't like usb-to-serial is some of them have weirdo effects (ever watched a board boot at 115200 baud and wondered why the kernel messages come through soo slowly) |
| 01:03 | ojn | i've had reasonable luck so far. some of the cheap allwinner hardware will backpower over the cheap pl2303 serial adapters, but it could be argued that it's partially the fault of the allwinner boards (cubie 1/2) |
| 01:03 | _rmk_ | I've more or less steered clear of usb so far and I intend to carry on doing so... have you seen how expensive usb is for hid devices? |
| 01:04 | _rmk_ | 2ms of cpu time is not unknown... |
| 01:17 | dbsx | rabeeh: any progress on certifications on cubox-i ? |
| 07:25 | jnettlet | wumpus, ping |
| 07:34 | wumpus | jnettlet: pong |
| 07:36 | jnettlet | wumpus, hey so I am trying out etnaviv on the C1 with my kernel changes. I am getting a segfault when etnaviv is trying to run etna_fb_bind_resource |
| 07:36 | jnettlet | specifically in etna_bo_gpu_address |
| 07:37 | wumpus | any other output on the console? does it manage to open /dev/fb0 ? |
| 07:37 | jnettlet | figured I would ask if you had seen it before, before digging in myself |
| 07:37 | ojn | rabeeh: I didn't see you here earlier so I emailed. My carrier-1 is behaving a little oddly when power cycling it. Unless I remove and re-insert the SD card, it doesn't boot up when power is reapplied. Known issue? |
| 07:38 | jnettlet | a strace shows no problem opening /dev/fb0 and it properly sends the pan ioctl |
| 07:38 | wumpus | ok, does seem that it somehow ends up with a zero bo for the frame buffer, I heard of this problem before in #etnaviv but am not sure what the solution was |
| 07:39 | jnettlet | sounds like I should dig into it when I have time then. |
| 07:41 | wumpus | maybe that there is some bug there, the bo stuff is relatively new (an attempt to offer a drm-ish API over the vivante kernel driver), though I'm sure I've seen it working on gcwzero and normal cubox |
| 07:42 | wumpus | one of the bo allocate functions must be returning a NULL pointer for some reason, which goes unchecked |
| 07:42 | jnettlet | I was figuring something like that. It will go on my list. |
| 07:43 | jnettlet | I think I have the latest MX6 vivante driver as stable as possible without too much more work. |
| 07:43 | jnettlet | I found another race, that effects powermanagement but only happens when there is lots of debug output happening so I am letting that go. |
| 07:44 | wumpus | I did get etnaviv demos to work on the C1, but that was with a pretty old kernel driver, and might be a pre-bo version of etnaviv |
| 07:45 | wumpus | great! |
| 07:46 | jnettlet | wumpus, and I have all _rmk_'s patches updated to the v4 driver, so the mips kernel should be able to apply them relatively easily. |
| 07:47 | wumpus | interesting, do you have a link for those? |
| 07:47 | jnettlet | not yet. just getting things ready to publish now |
| 07:48 | jnettlet | I will notify you when I push out the kernel for consumption |
| 07:49 | wumpus | thanks :) shouldn't be too difficult to merge if it's already based on v4, most we did is move files around and some code movement/deletion |
| 07:51 | jnettlet | wumpus, yeah I replicated that in another branch on the newer codebase and will merge the patches to that. |
| 07:51 | jnettlet | trying to unify the code a bit, but the latest code's user interface changes are just too much of a nightmare to manage. |
| 07:52 | wumpus | user interface changes? |
| 07:52 | wumpus | oh no |
| 07:52 | jnettlet | everything is changed to u32's and u64's instead of pointers and then cast back to pointers in the kernel. |
| 07:53 | jnettlet | I assume this is all in preparation for 64-bit ARM processors |
| 07:53 | wumpus | hm it makes a sense to do that for 32-64 bit compatibility |
| 07:53 | jnettlet | yep |
| 07:53 | wumpus | it's less random and strange as most of their user-space interface changes ,but still, frustrating :) |
| 08:00 | jnettlet | gc880 is doing 270fps under es2gears on the gc880 in the C1. That is not bad at all. |
| 08:01 | jnettlet | wow, I need more coffee. |
| 08:01 | jnettlet | too many gc880's in that last sentence :-) |
| 08:47 | rabeeh | ojn: jnettlet previously reported that DT kernels can't do 'reboot' |
| 08:48 | ojn | rabeeh: I just do a 'halt', then power off. when i come back to power it on it won't boot |
| 08:48 | ojn | I did notice that if I let it sit for a sufficient amount of time, it will boot up. Seems to be quite a while though (I haven't tried to narrow it down). 30+ minutes. |
| 08:48 | rabeeh | ojn: your power off in the farm is really unplugging the micro USB power cable? |
| 08:48 | ojn | rabeeh: it's powering down the outlet that the usb power adapter is plugged into. |
| 08:49 | rabeeh | so there is still a charge in the power adapter that drains out to certain level |
| 08:51 | rabeeh | is it far away from you? can u try unplugging the DC-in? |
| 08:52 | rabeeh | my guess is that the power supply internal capacitors are drained to certain level; which keeps the Vsnvs LDO alive (drains few tens of uA) |
| 08:53 | rabeeh | but 30 minutes is way too high for capacitors to self discharge; unless you have a super power supply from rare materials :) |
| 08:55 | jnettlet | rabeeh, dsd has seen this same behaviour |
| 08:56 | jnettlet | sorry dv has seen this same behaviour |
| 08:56 | jnettle | 08:56 * jnettlet has too many discussions going on |
| 09:35 | rabeeh | jnettlet: you need to fork() |
| 10:02 | jnettle | 10:02 * jnettlet would be too afraid of zombies |
| 10:02 | jnettlet | rabeeh, sorted the last bugs. 3.10 looks good to go. |
| 10:05 | jnettlet | Do we have an xbmc port for a debian rootfs yet? |
| 10:21 | _rmk_ | rabeeh: DT kernels (recent ones) can do reboot - that got fixed |
| 10:25 | jnettlet | _rmk_, does a reset from uboot work for you? |
| 10:28 | _rmk_ | don't know |
| 10:43 | ojn | rabeeh: sorry, my kid woke up, and i fell asleep while I was putting him back down. :) |
| 10:43 | bencoh | :-) |
| 10:43 | ojn | I can reproduce by disconnecting the micro USB at the board, so the remaining charge-in-adapter theory fails on that. |
| 11:42 | _rmk_ | there are times when some people's web browsers should be shot... |
| 11:42 | _rmk_ | /~rmk/cubox/carrier-1-v3.13-rc1/%7B0001-AR ... o-read.patch%7D: 1 Time(s) |
| 11:42 | _rmk_ | notice the extra %7B and %7D characters in there... |
| 11:45 | _rmk_ | ... from Singapore according to whois |
| 12:11 | _rmk_ | okay, Greg's pushed my -rc imx-drm fixes upstream :) |
| 12:26 | rabeeh | hste_ hste: do you have a working ubuntu 12.04 with gstreamer 0.10? |
| 12:40 | hste_ | rabeeh: no. I only have jas_hacks_ for 3.0.35:4.0.0. |
| 12:41 | dv | jnettlet: correct |
| 12:41 | dv | except for me it is like ~3-5 minutes |
| 12:44 | jnettlet | dv, I just need to test video decoding against 3.10 and I think we are ready to go |
| 12:45 | _rmk_ | jnettlet: did you want me to check whether reset worked in uboot? |
| 12:45 | jnettlet | _rmk_, if you could it would be helpful. |
| 12:46 | dv | jnettlet: ah cool |
| 12:46 | dv | jnettlet: how do you do that? |
| 12:46 | _rmk_ | Hit any key to stop autoboot: 0 |
| 12:46 | _rmk_ | C1 U-Boot > reset |
| 12:46 | _rmk_ | resetting ... |
| 12:46 | _rmk_ | U-Boot 2013.10-rc4-g61ecf63 (Nov 17 2013 - 12:10:02) |
| 12:46 | jnettlet | ah yours works |
| 12:46 | _rmk_ | and you have the version there :) |
| 12:47 | jnettlet | yep, although my boards have always had this hang problem. Both of them. |
| 12:47 | _rmk_ | I think this uboot may be rabeeh's tweaked version of yours |
| 12:47 | jnettlet | yeah, I have since pulled his patches into my tree |
| 12:48 | jnettlet | so they should be on par. |
| 12:49 | jnettlet | what would be differences? sdcard, power adapter |
| 12:51 | hste_ | rabeeh: I could try making it work on c1, but then it will be using a kernel that only works for that rootfs and not the same as default. Its also only armel |
| 12:58 | jnettlet | _rmk_, if you have a chance could you dump the extended info from u-boot with the version command? I wonder if it is a compiler difference that is my problem. |
| 12:59 | _rmk_ | jnettlet: in a moment, it's jsut rebooting again :) |
| 13:00 | _rmk_ | U-Boot 2013.10-rc4-g61ecf63 (Nov 17 2013 - 12:10:02) |
| 13:00 | _rmk_ | arm-linux-gnueabihf-gcc (crosstool-NG linaro-1.13.1-4.7-2013.03-20130313 - Linaro GCC 2013.03) 4.7.3 20130226 (prerelease) |
| 13:00 | _rmk_ | GNU ld (crosstool-NG linaro-1.13.1-4.7-2013.03-20130313 - Linaro GCC 2013.03) 2.23.1 |
| 13:00 | _rmk_ | anything else? |
| 13:05 | jnettlet | nope that is it. thanks |
| 13:27 | rabeeh | hste_: any idea of any Linux distro that works with BSP 4.1.0 that supports gstreamer 0.10? |
| 13:37 | hste_ | rabeeh: not any newer ones. Do u know if there has been some more info from freescale on gst 1.0 support? |
| 13:56 | dv | hste_: only my work :) |
| 14:40 | hste_ | dv: :) have u tried using totem with it? |
| 14:45 | dv | oh, right |
| 14:45 | dv | the cluttersink problem |
| 14:45 | dv | I unfortunately didnt have time to do anything the last few days |
| 14:54 | jnettlet | what is the problem with cluttersink? Is clutter on your platform compiled to use gles as a backend? |
| 15:13 | dv | it uses something else |
| 15:13 | dv | as in, not my sink |
| 15:13 | dv | so it does deep copies of frames, even though they are stored in dma buffers |
| 15:23 | jnettlet | oh this is beautiful. The chromecast is using yet another memory management driver created by Marvell. https://code.google.com/p/chromecast-mirrored-source/source/browse/?repo=kernel#git%2Flinux-3.8%2Farch%2Farm%2Fmach-berlin%2Fmodules%2Fshm |
| 15:23 | dv | /o\ |
| 15:24 | jnettlet | The Galois Memory Engine |
| 15:27 | jnettlet | ah that is for DRM |
| 15:28 | jnettlet | that makes sense |
| 16:12 | schnuffle | Hi folks, I don't see any guidelines, are there any? |
| 16:13 | jnettlet | guidelines for which? |
| 16:14 | schnuffle | jnettlet: Rules may be the better word. Like: All languages are welcome. No politics. No religion. No swearing, no sexual/racist/warez/spam/flood |
| 16:15 | jnettlet | schnuffle, we tend to assume that all members are mature adults and are polite to one another. |
| 16:16 | jnettlet | you are welcome to talk in any language you want, but if you want help then you should probably stick with english :-) |
| 16:16 | schnuffle | jnettlet: :) that's rare, somebody believing that adults behave reasonable in the internet |
| 16:17 | jnettlet | well this is a relatively small dev oriented channel, we don't get many people here that aren't interested in helping, or being helped. |
| 16:19 | schnuffle | jnettlet: okay fine, I think I'll have some questions as soon as I get my hardware. Damned, on holiday and my gadgets haven't arrived :) |
| 16:19 | jnettlet | schnuffle, sounds great. welcome |
| 16:21 | schnuffle | jnettlet: yep, I'm really keen on starting to play around a bit |
| 16:23 | schnuffle | aah, if there're some official people in here. just a minor thing I want to mention. I got the shipping email with a link to the tracking site. As long as the tracking number is not in the system already you get redirected to a search which is totally in hebrew, the &L=en in the URL gets lost. Could be confusing for some people |
| 16:44 | jnettlet | Was somebody talking about this error last week? imx-ipuv3 2400000.ipu: IPU Warning - IPU_INT_STAT_10 = 0x00080000 |
| 16:57 | hste | jnettlet: I think it was on the openbricks-devel https://groups.google.com/forum/#!topic/openbricks-devel/Yv_C5Kk6zWA |
| 16:57 | jnettlet | hste, thanks |
| 17:24 | rabeeh | schnuffle: here. go ahead |
| 17:25 | schnuffle | rabeeh: ? |
| 17:25 | rabeeh | jnettlet: that's underrun of the IPU in resizing |
| 17:25 | rabeeh | schnuffle: you had questions? |
| 17:25 | rabeeh | oh; it's just only the shipping email then |
| 17:25 | rabeeh | jnettlet: try - |
| 17:25 | rabeeh | devmem 0x020e0018 32 0xffffffff |
| 17:26 | rabeeh | this tags the qos field of all AXI transactions from the IPU towards the DDR controller as high priority |
| 17:26 | rabeeh | this write made xbmc on c1 reality :) |
| 17:27 | jnettlet | okay. still trying to sort out xbmc on Ubuntu 13.04 to test the GPU changes. |
| 17:44 | jnettle | 17:44 * jnettlet has backported the zswap module to the 3.10 kernel to see how it helps with the 512MB's of RAM in the HB |
| 19:05 | hste | rabeeh: does the devmem 0x020e0018 32 0xffffffff have some downside on other things or is it just positive to have it added? |
| 19:16 | jnettlet | _rmk_, have you pushed patches for your cubox-i alsa driver yet? I am trying to decide if I want to bother diving into the freescale 3.10 asoc mess. |
| 20:08 | hste | jnettlet: do u have git where u push the patches? |
| 20:31 | _rmk_ | jnettlet: that's right at the end of the imx-drm stuff |
| 20:38 | rabeeh | hste: no downsides was reported |
| 20:44 | rabeeh | _rmk_: http://imx.solid-run.com/forums/viewtopic.php?f=2&t=282&p=1289#p1284 |
| 20:45 | rabeeh | tons of credit goes to you on this. btw - is there a public place for the patches? we should write a post about this !!! |
| 20:49 | _rmk_ | there's only mailing lists |
| 20:58 | jnettlet | hste, they will end up on my github account, but I have just recently beaten the 3.10 kernel into shape |
| 21:01 | jnettlet | still tracking down a couple of crashers for Xorg, and vpu testing. |
| 21:01 | jnettlet | Then I progress to drop Exa acceleration in favour of _rmk_'s straight X acceleration. |
| 21:02 | _rmk_ | jnettlet: I was planning to look at making the glyph stuff more efficient |
| 21:05 | jnettlet | _rmk_, I have to see how much X uses triangles at this point. Vivante's 2d hardware can accelerate triangles |
| 21:05 | jnettlet | very useful for drawing circles |
| 21:09 | jas-hacks_ | jnettlet: Do you get better results without Exa? |
| 21:10 | _rmk_ | jas-hacks: that's my theory |
| 21:11 | jas-hacks_ | _rmk_: any patches I could try with 3.0.35 kernel |
| 21:12 | jnettlet | well EXA was never about performance. It took EXA a long time to even catch up with XAA for most drawing operations. |
| 21:12 | jnettlet | EXA was mostly memory management and taking advantage of 3d engines for compositing |
| 21:13 | jnettlet | The big problem we see now, which Intel started addressing a while ago is that EXA is aimed at dealing with having to migrate pixmaps back and forth from system RAM to VRAM. |
| 21:13 | jnettlet | this is not really a problem that SOC's with integrated GPU's have |
| 21:16 | jnettlet | SDL2 is now compiled against Vivante's GLESv2 libraries. Time to build ioquake3 |
| 21:25 | jas-hacks_ | dv: If I make window full screen video freezes when its in the foreground |
| 21:31 | jas-hacks_ | dv: its a problem with the xrender compositing |
| 21:34 | dv | jas-hacks_: ask me again on monday, i will be busy until then |
| 21:37 | jas-hacks_ | no problems |
| 21:50 | rabeeh | jas-hacks_ : in your blog - "The mesa utilities glxgears,es2gears, glmark2, glmark2-es2 are pre-installed. glmark2 will end after the jellyfish demo with a segmentation fault." |
| 21:51 | rabeeh | it runs all the way for me |
| 21:51 | hste | rabeeh: the newest fsl-libs have fixed that |
| 21:52 | jas-hacks_ | rabeeh: yes it works all the way after I upgraded the viv libs, blog description is old |
| 21:52 | rabeeh | ok. |
| 21:52 | rabeeh | is there something that prevents drm.ko / vivante.ko being statically built with the kernel? |
| 21:53 | rabee | 21:53 * rabeeh has dejavu... again |
| 21:53 | rabeeh | i'd like to have it part of a .config file that i want to submit (that will have lots of kernel modules too). |
| 21:53 | jnettlet | rabeeh, there isn't. I run my kernels with them compiled in |
| 21:54 | _rmk_ | I have drm stuff built-in so the boot messages are on the screen |
| 21:54 | jas-hacks_ | There was some issues with it being in the kernel at some point. |
| 21:54 | hste | rabeeh: wasn't that some probs when u used cgroups? |
| 21:55 | rabeeh | i knew there i had dejavu |
| 21:55 | jnettle | 21:55 * jnettlet wants to break 300 for es2gears |
| 21:55 | rabeeh | c1? |
| 21:55 | rabeeh | jnettlet: c1? |
| 21:55 | jnettlet | yeah. I am at 280 max right now |
| 21:56 | hste | jnettlet: on 1080P? |
| 21:56 | jnettlet | 720p, but 1080p shouldn't make a difference |
| 21:56 | rabeeh | hste: it's a window inside it |
| 21:56 | jnettlet | it is rendering the same size demo window |
| 21:56 | rabeeh | jnettlet: it should matter |
| 21:57 | jnettlet | the demo uses the same surface size, just makes things smaller on the screen with higher resolutions |
| 21:58 | rabeeh | jnettlet: gpu in c1 is limited by ddr bandwidth; if you have a smaller resolution then you free half of the ddr bandwidth that goes only for refreshing |
| 21:58 | rabeeh | for instance 1920x1080x32bpp@60hz is ~480MB/sec |
| 21:59 | rabeeh | while 720p60 is ~200MB/sec |
| 21:59 | rabeeh | so you have extra 280MB/sec that the gpu can burn |
| 21:59 | jnettlet | true. |
| 22:00 | rabeeh | try even with lower resolution like 640x480; you will break the 300 easily with it. |
| 22:00 | jnettlet | will test later. I am eating a late dinner now, wife just got home. |
| 22:00 | rabeeh | bon apetit |
| 22:18 | rabeeh | jnettlet: what is the cpu utilization on your drivers when running es2gears? |
| 22:19 | jnettlet | rabeeh, 100% as it should be. one of the reasons glxgears/es2gears is not a great benchmark. |
| 22:19 | jnettlet | But I can give you breakdown of es2gears/xorg/galcore in a sec |
| 22:19 | rabeeh | on my quad it's 282fps |
| 22:20 | rabeeh | with ~97% on one processor (as it's not multithreaded application) |
| 22:21 | rabeeh | what i'm saying is that resolution might not change anything since it's cpu bound |
| 22:21 | jnettlet | So at 1080p I am doing 268fps with a bare X-server |
| 22:21 | jnettlet | just peaked at 270 |
| 22:23 | rabeeh | so it does make a difference |
| 22:23 | jnettlet | userspace is about 90% with the remaining cpu 10% system, es2gears 90% Xorg 7% galcore daemon the rest |
| 22:23 | hste | jnettlet: I had big difference on using lxde vs xfce |
| 22:24 | jnettlet | just peaked. 1390 frames in 5.0 seconds = 277.833 FPS |
| 22:24 | jnettlet | just hit 278 |
| 22:25 | rabeeh | jnettlet: want an overclocking patch :) |
| 22:25 | rabeeh | ? |
| 22:25 | jnettlet | sure I can give it a try. testing gtkperf now for 2d performance at 1080p haven't tested this yet. |
| 22:26 | jnettlet | 37.61 so about a 7% drop off there. |
| 22:26 | jnettlet | still not too bad |
| 22:27 | rabeeh | jnettlet: which window manager are you using? |
| 22:27 | rabeeh | hste: what sort of difference between lxde and xfce? you mean es2gears? |
| 22:29 | jnettlet | under xfce I am holding steady at about 267fps |
| 22:30 | jnettlet | glmark2-es2 is definitely going to be rough though. |
| 22:30 | hste | on glmark2-es2 it was in xfce 104 in lxde 197 |
| 22:31 | hste | thats on gk802 |
| 22:32 | jnettlet | damn actually it is holding up nicely. only seeing about 10-20% user 10-20% system |
| 22:32 | jas-hacks_ | as a comparison I get 340fps on xfce with 1.2Ghz quad with 4.1.0 |
| 22:32 | jnettlet | load average of 1.4 and dropping |
| 22:36 | jnettlet | not bad glmark2-es2 90 at 1080p-32bpp |
| 22:36 | jnettlet | the C1 is really humming now |
| 22:38 | jnettlet | 2 IPU warnings |
| 22:43 | jnettlet | neverball at 1080p not so playable, 720p it flies though |
| 22:45 | jnettlet | glmark2-es2 1080p fullscreen is definitely stressing the ddr. cpu usage is down to almost nothing because it is all about the GPU and DDR |
| 22:45 | jnettlet | running pretty smooth though |
| 22:49 | jnettlet | 1080p fullscreen is a glmark-es2 score of 41. definitely pushing the limits there. |
| 22:49 | hste | not bad for such a limited device |
| 22:50 | hste | I think you would get even better results in lxde |
| 22:51 | jnettlet | it shouldn't matter too much, as window drawing shouldn't be happening behind a fullscreen window. |
| 22:52 | jnettlet | I will test against a bare Xserver. ultimately I don't see hardly any cpu utilization which means we are hitting hardware limitations |
| 22:52 | jnettlet | I will finish compiling XBMC and see how it handles that at 1080p |
| 22:53 | jnettlet | but now I am going to finish watching an episode of Fringe |
| 23:25 | dbsx | i4pro just arrived. What should I try first? |
| 23:25 | rabeeh | geexbox xbmc :) |
| 23:26 | rabeeh | http://cubox-i.com/getting-started/ |
| 23:27 | _rmk | 23:27 * _rmk_ gets to his 5th kernel build and his Nth attempt to get an ARM Ltd supported platform to boot... |
| 23:27 | rabeeh | dbsx: do you like the new weight? |
| 23:27 | _rmk_ | after 2.5 hours, I've almost had enough of this DT crap |
| 23:27 | rabeeh | _rmk_: 64bit? :) |
| 23:27 | _rmk_ | Versatile Express with the CA15 tile |
| 23:28 | rabeeh | how many cores? |
| 23:28 | _rmk_ | 8 but only 4 should be active |
| 23:28 | rabeeh | big little crap? |
| 23:28 | dbsx | I like the packaging and I like the extra weight the HDMI cable no longer lifts the box off the desk. |
| 23:28 | _rmk_ | yea |
| 23:28 | _rmk_ | the more things I try the less stuff that works |
| 23:28 | _rmk_ | I've yet to see it get past the decompressor |
| 23:28 | rabeeh | dbsx: we had the weight in mind for a long time; jnettlet did the last push to get it in |
| 23:29 | sam_nazarko | rabeeh: |
| 23:29 | sam_nazarko | I want to test geexbox as I have not been able to boot carrier-1 with a kernel myself and want to just get a UART to see things are going |
| 23:29 | rabeeh | hi sam |
| 23:29 | _rmk_ | hacking on non-solidrun stuff is becoming very frustrating! |
| 23:30 | sam_nazarko | but there is no c1 image for geexbox, which cubox-i one should I use? |
| 23:30 | mijanek | I'd love to play with i4pro too, but my order (form 17.9.) seems to got lost... no shipping notification yet :-( |
| 23:31 | rabeeh | _rmk_: you saying this is like a love letter :) |
| 23:31 | _rmk_ | ... and I still need to rip apart my network and rebuild it post FTTC "downgrade" |
| 23:31 | rabeeh | sam_nazarko: there are two ways to do that - |
| 23:32 | rabeeh | 1. either use the geexbox mksdcard.sh script that should do the work for you |
| 23:32 | rabeeh | 2. or use cubox-i1 and try it as is; if it doesn't work as-is just replace the u-boot that it has with a u-boot for c1 |
| 23:33 | rabeeh | u-boot sources are from - https://github.com/rabeeh/u-boot-imx6 |
| 23:33 | rabeeh | if you want ready to use images then i can send you too. |
| 23:33 | sam_nazarko | cool |
| 23:33 | rabeeh | mijanek: what do you mean seems to got lost? |
| 23:34 | sam_nazarko | i've tried a manual C1 boot but had no luck so far, just red light. i did follow instructions for partitioning (although used parted, not fdisk) |
| 23:34 | sam_nazarko | so I'm interested in seeing if the geexbox version works |
| 23:34 | rabeeh | if you just want to see u-boot then even no partitioning is required. |
| 23:34 | mijanek | as I saw you are allerady shipping them out but I got only nwesletter but nothing regarding shipping.. |
| 23:34 | rabeeh | just dd u-boot to offset 1KByte from the start of the micro sd |
| 23:35 | rabeeh | sam_nazarko: http://imx.solid-run.com/wiki/index.php?title=Building_the_kernel_and_u-boot_for_the_CuBox-i_and_the_HummingBoard#U-Boot_Upstream |
| 23:35 | sam_nazarko | i did that |
| 23:35 | rabeeh | sudo dd if=u-boot.imx of=/dev/sdX bs=512 seek=2 |
| 23:35 | rabeeh | ? |
| 23:35 | sam_nazarko | didn't have much luck, but will try again |
| 23:35 | sam_nazarko | yep |
| 23:35 | rabeeh | notice that u-boot 2009.08 is different than the upstream one |
| 23:35 | sam_nazarko | I used the solidrun u-boot |
| 23:36 | rabeeh | btw we completely neglected u-boot 2009.08 since it's useless and the upstream is the focus |
| 23:36 | sam_nazarko | root@sam-wheezy:/home/sam# parted -s /dev/sdb print |
| 23:36 | sam_nazarko | Model: Generic STORAGE DEVICE (scsi) |
| 23:36 | sam_nazarko | Disk /dev/sdb: 4013MB |
| 23:36 | sam_nazarko | Sector size (logical/physical): 512B/512B |
| 23:36 | sam_nazarko | Partition Table: msdos |
| 23:36 | sam_nazarko | Number Start End Size Type File system Flags 1 4194kB 4013MB 4009MB primary fat32 |
| 23:36 | sam_nazarko | oh |
| 23:36 | sam_nazarko | that I did not know |
| 23:36 | sam_nazarko | I suspect I probably just followed out of date info |
| 23:36 | rabeeh | hmm... that's probably my mistake since i didn't really document it correctly |
| 23:37 | rabeeh | so choose if you want to build from scratch or i can provide you pre-built images |
| 23:37 | sam_nazarko | no problem |
| 23:37 | sam_nazarko | can you send me a pre-built for now? |
| 23:37 | sam_nazarko | any image, not too fussed (I can resize if necessary) |
| 23:37 | sam_nazarko | just want to test the HW |
| 23:37 | sam_nazarko | then I'll start building a custom kernel, rootfs, installers etc |
| 23:38 | rabeeh | sam_nazarko: http://dl.dropboxusercontent.com/u/72661517/tbr/u-boot-images.tar.xz |
| 23:38 | rabeeh | those should get you started with 8 images |
| 23:38 | rabeeh | just choose the c1 which will work for you |
| 23:38 | rabeeh | the others are for cubox-i * |
| 23:42 | sam_nazarko | I need to speed up with linXBMC development |
| 23:42 | sam_nazarko | i can start after around 15th jan |
| 23:42 | sam_nazarko | i think i'll make cubox-i the first target because it looks so promising |
| 23:55 | rabeeh | mijanek: i know it's frustrating not getting your shipping email |