[wolfgar]

Hi mijanek,

It is an issue in the kernel : Please check this commit to improve greatly the result : https://github.com/wolfgar/utilite/commi...3e0626aa37
Regards

Stephan

[wolfgar]

So what the final verdict for this Matrix? Must have? Anyone willing to upload some videofootage of this thing in action, switching channels etc.?

Hi,

The final result is better than what I feared at first sight when I saw the box and the fan for sure !
The dev team is committed to the task and they really want to deliver a device that works out of the box given the exchanges I had with them.
I have no time for now for a detailed video but I will post an article regarding this device (and even a whole comparison table of the imx6 devices I own by the end of the month)

Kind regards
Stephan

[mijanek]

@wolfgar Thx for your hints. Now I have changed the kernel with the mentioned patch from you. The artefacts on the bottom are now gone, but I get time to time some horizontal green bars over half screen on random possition (it is hard to take a picture of it, as it is really short there and not showing so frequently).

[wolfgar]

I have not said it solves the issue but it greatly improves it...

I have to figure out what the remaining (little) issue is !

[mijanek]

@wofgar yes for sure, It is an huge improvement and big thank to you, it is better to watch the video interlaced. (I mentoined it just whenewer you know this issue or not.. it could bee I missed again some patch :-) )

Currently I hove bad issue resulting in xbmc crash but I guess as I saw this could be fixed in your last commit (58105b4eec *Fix stupid redundant allocation) right?

Code:

ipu1_task: page allocation failure: order:4, mode:0x40d0
...
Out of memory: Kill process 3108 (xbmc.bin) score 8 or sacrifice child
Killed process 3108 (xbmc.bin) total-vm:451196kB, anon-rss:50880kB, file-rss:17756kB

[K405]

So what the final verdict for this Matrix? Must have? Anyone willing to upload some videofootage of this thing in action, switching channels etc.?

Hi,

The final result is better than what I feared at first sight when I saw the box and the fan for sure !
The dev team is committed to the task and they really want to deliver a device that works out of the box given the exchanges I had with them.
I have no time for now for a detailed video but I will post an article regarding this device (and even a whole comparison table of the imx6 devices I own by the end of the month)

Kind regards
Stephan

Looking forward to your article and thank you for taking the time to reply.

[mijanek]

@wolfgar, I have some problem with the memory allocation. So far I saw I applied all needed patches, but afer some time watching I get memory issue

Code:

imx-ipuv3 imx-ipuv3.1: [0xe9e6b400]Falied Alloc vditmpbuf[0]
ipu2_task: page allocation failure: order:4, mode:0x40d0
[<8004ae80>] (unwind_backtrace+0x0/0xf8) from [<800c7c60>] (warn_alloc_failed+0xcc/0x10c)
[<800c7c60>] (warn_alloc_failed+0xcc/0x10c) from [<800ca140>] (__alloc_pages_nodemask+0x4b0/0x704)
[<800ca140>] (__alloc_pages_nodemask+0x4b0/0x704) from [<800ca3a4>] (__get_free_pages+0x10/0x24)
[<800ca3a4>] (__get_free_pages+0x10/0x24) from [<803cde18>] (do_task+0xa84/0x16dc)
[<803cde18>] (do_task+0xa84/0x16dc) from [<803ceab8>] (get_res_do_task+0x48/0x5b8)
[<803ceab8>] (get_res_do_task+0x48/0x5b8) from [<803cf6e8>] (ipu_task_thread+0xe0/0xb30)
[<803cf6e8>] (ipu_task_thread+0xe0/0xb30) from [<80091db8>] (kthread+0x80/0x88)
[<80091db8>] (kthread+0x80/0x88) from [<80044b78>] (kernel_thread_exit+0x0/0x8)
Mem-info:
DMA per-cpu:
CPU    0: hi:   90, btch:  15 usd:   0
CPU    1: hi:   90, btch:  15 usd:   0
CPU    2: hi:   90, btch:  15 usd:   0
CPU    3: hi:   90, btch:  15 usd:   0
Normal per-cpu:
CPU    0: hi:  186, btch:  31 usd:   0
CPU    1: hi:  186, btch:  31 usd:  60
CPU    2: hi:  186, btch:  31 usd:   0
CPU    3: hi:  186, btch:  31 usd:   0
HighMem per-cpu:
CPU    0: hi:   90, btch:  15 usd:   0
CPU    1: hi:   90, btch:  15 usd:  15
CPU    2: hi:   90, btch:  15 usd:   0
CPU    3: hi:   90, btch:  15 usd:   0
active_anon:14712 inactive_anon:437 isolated_anon:0
active_file:1488 inactive_file:11425 isolated_file:0
unevictable:0 dirty:1 writeback:0 unstable:0
free:64915 slab_reclaimable:808 slab_unreclaimable:2088
mapped:6660 shmem:527 pagetables:341 bounce:0
DMA free:5464kB min:616kB low:768kB high:924kB active_anon:0kB inactive_anon:0kB active_file:0kB inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:186944kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? yes
lowmem_reserve[]: 0 1244 1593 1593
Normal free:7656kB min:4212kB low:5264kB high:6316kB active_anon:0kB inactive_anon:0kB active_file:104kB inactive_file:0kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:1274048kB mlocked:0kB dirty:0kB writeback:0kB mapped:0kB shmem:0kB slab_reclaimable:3232kB slab_unreclaimable:8352kB kernel_stack:984kB pagetables:1364kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? no
lowmem_reserve[]: 0 0 2794 2794
HighMem free:246540kB min:348kB low:640kB high:936kB active_anon:58848kB inactive_anon:1748kB active_file:5848kB inactive_file:45800kB unevictable:0kB isolated(anon):0kB isolated(file):0kB present:357632kB mlocked:0kB dirty:4kB writeback:0kB mapped:26640kB shmem:2108kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB unstable:0kB bounce:0kB writeback_tmp:0kB pages_scanned:0 all_unreclaimable? no
lowmem_reserve[]: 0 0 0 0
DMA: 4*4kB 19*8kB 21*16kB 21*32kB 1*64kB 1*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 1*4096kB 0*8192kB 0*16384kB 0*32768kB = 5464kB
Normal: 344*4kB 289*8kB 190*16kB 31*32kB 0*64kB 0*128kB 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB 0*8192kB 0*16384kB 0*32768kB = 7720kB
HighMem: 1293*4kB 671*8kB 420*16kB 227*32kB 107*64kB 51*128kB 17*256kB 11*512kB 2*1024kB 2*2048kB 1*4096kB 1*8192kB 1*16384kB 5*32768kB = 246540kB
13471 total pagecache pages
0 pages in swap cache
Swap cache stats: add 0, delete 0, find 0/0
Free swap  = 0kB
Total swap = 0kB
524288 pages of RAM
65219 free pages
73505 reserved pages
1615 slab pages
15361 pages shared
0 pages swap cached
imx-ipuv3 imx-ipuv3.1: [0xe9e6b200]Falied Alloc vditmpbuf[0]
mxc_ipu mxc_ipu: ERR: [0xe9e6be00] no-0x43120, timeout:4020ms!
mxc_ipu mxc_ipu: ERR: no-0x43120,ipu_queue_task err:-110
mxc_v4l2_output mxc_v4l2_output.0: display work fail ret = -110
mxc_ipu mxc_ipu: ERR:[0xe9e6be00] no-0x43120, split_task timeout,j:3,size:4.
mxc_ipu mxc_ipu: ERR:ipu[0] busy_vf:1, busy_pp:0.
mxc_ipu mxc_ipu: ERR:ipu[1] busy_vf:0, busy_pp:0.
mxc_ipu mxc_ipu: ERR: sp_task[0][0xe9e6b200] no-0x4312a done:0,state:ok,on_list:0, ipu:0x80d78960,timeout!
mxc_ipu mxc_ipu: ERR: sp_task[1][0xe9e6b400] no-0x43129 done:0,state:ok,on_list:0, ipu:0x80d78960,timeout!
mxc_ipu mxc_ipu: ERR: sp_task[2][0xe9e6ae00] no-0x43126 done:0,state:wait resource timeout,on_list:0, ipu:0x  (null),timeout!
mxc_ipu mxc_ipu: ERR: sp_task[3][0xe9e6a600] no-0x43125 done:0,state:ok,on_list:0, ipu:0x80d78960,timeout!
Physical memory allocation error!
Physical memory allocation error!
[root@matrixtv ~]# free
             total         used         free       shared      buffers
Mem:       1803132      1534668       268464            0          424
-/+ buffers:            1534244       268888
Swap:            0            0            0

there seems to be some memory leak (looks like with the ipu):

Code:

[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1049940       753192            0         4484
-/+ buffers:            1045456       757676
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1055956       747176            0         4484
-/+ buffers:            1051472       751660
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1057872       745260            0         4484
-/+ buffers:            1053388       749744
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1062928       740204            0         4484
-/+ buffers:            1058444       744688
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1064916       738216            0         4484
-/+ buffers:            1060432       742700
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1066772       736360            0         4484
-/+ buffers:            1062288       740844
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1068744       734388            0         4484
-/+ buffers:            1064260       738872
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1070804       732328            0         4484
-/+ buffers:            1066320       736812
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1072136       730996            0         4484
-/+ buffers:            1067652       735480
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1082252       720880            0         4484
-/+ buffers:            1077768       725364
Swap:            0            0            0
[root@matrixtv /]# free
             total         used         free       shared      buffers
Mem:       1803132      1084048       719084            0         4484
-/+ buffers:            1079564       723568
Swap:            0            0            0

edit: changed my wording, sorry

[wolfgar]

Hi mijanek

OK even with the patch you mentioned, you still have this behavior ?
I have a pending patch to push which should definitively help.
As soon as I have pushed it, I will track this memory leak if it is still there..

thanks for reporting
Can you tell me the duration between your free commands ?


Stephan

PS : If you are here to insult my work ("this looks very bad") , then you can pass your way...

[mijanek]

Its about per sec. Curioslily this happens only with the IPU patch, without I have fragments on the sreen but no memory leak, but I don't see any reason for it...

PS : No, sorry that wasn't pointed to your work or code. I do really appreciate what you have done! I ment this only, that it is really huge consumption of memory and starting after reboot with 1,4 G free they are exhousted in about 15-20 minutes of playing a video (on Idle is it fine).. And it could be caused by myself (mabe I am using wrong firmware version?)

[wolfgar]

Hi,

Thanks for this info !

PS : Sorry, it guess I clearly overreacted after a bad day
I doubt it is your fault but as soon as I have cleared my current work, I will play a few videos and check if I observe such a rapid memory leak...

[emveepee]

PS : If you are here to insult my work ("this looks very bad") , then you can pass your way...

Stephan, as a Wandboard and Cubox-i owner, your work getting XBMC running n the .MX6 platform is very much appreciated.

Un gros merci.

Martin

[wolfgar]

Hi mijanek

I have just push a big commit that is my current WIP state.
I will revert and push proper individual commits but it is available if you want to test that one...

Good night

---

Stephan, as a Wandboard and Cubox-i owner, your work getting XBMC running n the .MX6 platform is very much appreciated.

Un gros merci.

Martin

Thanks a lot Martin

[mijanek]

@wolfgar I have to apologise again, there is no memory issue on your patches nor xbmc, it was lying on the openpctv build environment I am using. After the kernel was build and new uImage was created they simple ignore this one and download some other prebuild from their project site (with obwiously massive issues inside) I took the build image now and hence everything is fine.

Really sorry about that, next time I'll double check before I post :-(

PS: The Interlaced video is now GREAT!

[slicemaster]

Hey, I just want to congratulate all involved with adding support for the i.MX6 line of chips to XBMC.
I remember when these chips first came to market and I thought they’d be a great match for xbmc.
In either case, I’m a huge fan of XBMC on ARM and I honestly believe that things will continue to
move this direction as time goes on.

Anyways, I’ve been a long time user of XBMC, and have been using it on three Raspberry Pis in my
home but have found that certain elements of the experience leave me wanting more than the
cheap little device can give (Although I have to give major cudos to Sam for his work on the raspbmc
distro, truly amazing work that has made xbmc on the pi truly usable, auto updates and easy setup,
something anyone can get working).

To get to the point, I was wondering how “usable” XBMC is on i.MX6 devices? I’m perhaps looking to
buy a Wand board Quad to play with and was hoping to get some feedback from current users and the
developers working on it and other i.MX6 devices.

I noticed earlier in the thread that pretty much all video codecs are being accelerated no problem and
that’s one major check but I just wanted to verify the rest to see if it would be a good fit for my usage pattern.

As of today Are these features supported on current i.MX6 builds…And to what extent are they supported?

1. DVD ISO playback over Ethernet (Menu Support?)
2. Blu-ray ISO playback over Ethernet
3. MP3 Audio playback with Visualizations over Ethernet
4. FLAC Audio playback with Visualizations over Ethernet

I know the list may be short but these are the features I use and therefore need to be in good working order
before I take the dive-in and buy a board. I suspect the 3 and 4 are not issues seeing as both can be done
in software on a modest device. The first two I am less sure of…

Thanks and keep up the great work,
Slice

[mijanek]

@slicemaster sofar I was able to test there is pretty much working on it and much better then on Rapsberry Pi (mine is pushed a side from the first day I got my matrix)...

1. don't have the ISO but ripped, without issues now..(no menu testet)
2. same like DVD.
3. Absolutelly no problem, but ceratin visualisations may overload the system bit, the default bars are feally fine (40% CPU of 1 core)
4. don't have that files...

I use it with TVHeadend and testet it with VDR, whereas VDR (xvdr) is much better with about 1 sec. channel change, so I will move to VDR 2.0 later than..
Some recorded streams (drembox HD/SD ts files) does makes some problem with audio syncs, but newer got these on liveTV or lately recorded programs.

If you are thinking of buying such device, I think it would be better to ommit the "lite" versions of that chip, as they have limmited memory bandwidth and slower GC.