Login at Kodi Home

davilla · (This post was last modified: 2013-01-25, 07:04 by davilla.)

sounds like mali did not load.

dmesg
find / -name xbmc.log

noggin · (This post was last modified: 2013-01-25, 15:16 by noggin.)

(2013-01-25, 06:32)davilla Wrote: sounds like mali did not load.

dmesg

Here's my full dmesg log (cut and pasted from a UART session log - is there a way of logging in an SSH client?) : http://pastebin.com/5yvA0uWP

Here's the stuff that looks MALI related :

Code:
[   13.636841] Mali<2>: 

[   13.636860] Mali<2>: Inserting Mali v9 device driver. 

[   13.638683] Mali<2>: Compiled: Jan 24 2013, time: 22:54:46.

[   13.644227] Mali<2>: Svn revision: 

[   13.648360] Mali<2>: MMU memory system initializing

[   13.652938] Mali<2>: Rendercore: subsystem global mutex initialized

[   13.658780] Mali<2>: Core: subsystem_init: Mali-400 PP

[   13.663889] Mali<2>: Core: subsystem_init: Mali-400 GP

[   13.669019] Mali<2>: Mali L2 cache system initializing

[   13.726000] mali_meson_poweron: Interrupt received.

[   13.727061] Mali<2>: MMU registered

[   13.731331] Mali<2>: MMU registered

[   13.732354] Mali<2>: Mali OS memory allocator created with max allocation size of 0x8000000 bytes, cpu_usage_adjust 0x00000000

[   13.750987] Mali<2>: Memory Validator 'Framebuffer' installed for Mali physical address base==0x84000000, size==0x06000000, cpu_adjust==0x00000000

[   13.758626] Mali<2>: Creating Mali L2 cache: Mali-400 L2 cache

[   13.767531] Mali<1>: Mali memory allocators will be used in this order of preference (lowest numbered first) :

[   13.774367] Mali<1>:     5: Mali Memory

[   13.778536] Mali<2>: Attach mmu: 0xe62814e0 to core: Mali-400 PP in subsystem: Mali-400 PP

[   13.786094] Mali<2>: Attach mmu: 0xe6281780 to core: Mali-400 GP in subsystem: Mali-400 GP

[   13.794347] Mali<2>: Mali L2 cache system load complete

[   13.799895] Mali: Mali device driver  loaded

[   13.803758] Mali<2>: malidrv_build_info

It does say Mali : Mali device driver loaded ?

This is then followed by lots of :

Code:
[   13.989327] page 39434 in gc has no object: 0 0 0 

[   13.989646] page 39435 in gc has no object: 0 0 0 

[   13.989962] page 39436 in gc has no object: 0 0 0

Quote:find / -name xbmc.log

This is really odd. I think there is something seriously wrong here : http://pastebin.com/WbxEEA5h

Not the output I was expecting.

When I SSH and try and do this or access /tmp/userdata my UART displays :

Code:
[  109.899628] aml nand write failed at 102916 

[  118.510361] aml nand write failed at 103172 

[  130.650359] aml nand write failed at 103428 

[  135.970362] aml nand write failed at 103684 

[  174.360364] aml nand write failed at 103940

Does this suggest something seriously screwy with how my NAND flash is being mounted and used? If so is there a way of stopping NAND Flash being used (and using SD card or NFS?)

Or am I totally wrong?

Looking at my dmesg and comparing with a bootlog from Android booting out of the box there is an error reported in the buildroot-AmLogic version that isn't there in Android :

Code:
[    2.128530] ######aml_nand_get_read_default_value_hynix 779 read default read retry reg value failed and need read from chip write back to nand using SLC

Could the /usr/tmpdata issue be related to S10setup :

Code:
echo "S10setup: mounting /tmp/userdata"

  # find userdata via mtd partition name, mount it and bind to /root

  for i in `ls /dev/mtd* | grep -o "mtd[0-9]$"`; do

if [[ "`cat /sys/devices/virtual/mtd/$i/name`" == "userdata" ]]; then

USERDATA="/dev/"`ls /sys/devices/virtual/mtd/$i | grep -o mtdblock[0-9]`

      break

fi

done

mkdir -p /tmp/userdata

  mount -t yaffs2 $USERDATA /tmp/userdata

  mkdir -p /tmp/userdata/xios/root

  mount -o bind /tmp/userdata/xios/root /root

which seems to result in this in dmesg :

Code:
[    9.558529] run_init_process /sbin/init.

[   10.001175] yaffs: dev is 32505859 name is "mtdblock3" rw

[   10.001210] yaffs: passed flags ""

[   10.004409] yaffs: Attempting MTD mount of 31.3,"mtdblock3"

[   10.004433] yaffs: NAND geometry problems: chunk size 8192, type is yaffs2, inbandTags 1 

[   10.929912] yaffs: dev is 32505863 name is "mtdblock7" rw

[   10.929946] yaffs: passed flags ""

[   10.933145] yaffs: Attempting MTD mount of 31.7,"mtdblock7"

[   10.933224] yaffs2 check point success.

[   10.961719] yaffs: restored from checkpoint

[   10.963131] yaffs_read_super: isCheckpointed 1

[   10.977155] aml nand write failed at 107692 

[   10.977828] **>> yaffs write required 2 attempts

If I comment out this stuff - which I believe is there to mount /tmp/userdata into the 'userdata' partition on the NAND flash in my box (using a YAFFS file system?), will it instead use a /tmp/userdata folder in my NFS-served rootfs? Or can I mount an SD card partition in its place?

I'm right that this is a totally separate issue to the usr.sqsh squashfs file that is mounted as /usr ? (Is this a read-only file system?)

Earlier NAND stuff (including line that differs between Android and buildroot-AmLogic boot)

Code:
[    1.462509] NAND device id: ad d7 94 da 74 c3 

[    1.466709] aml_chip->hynix_new_nand_type =: 2 

[    1.471233] NAND device: Manufacturer ID: 0xad, Chip ID: 0xad (Hynix B revision 26nm NAND 4GiB H27UBG8T2B)

[    2.042769] 1 NAND chips detected

[    2.042808] #####aml_nand_init, with RB pins and chip->chip_delay:20

[    2.046930] bus_cycle=4, bus_timing=5, start_cycle=5, end_cycle=6,system=5.3ns

[    2.054060] oob size is not enough for selected bch mode: NAND_BCH60_1K_MODE force bch to mode: NAND_BCH40_1K_MODE

[    2.064359] multi plane error for selected plane mode: NAND_TWO_PLANE_MODE force plane to : NAND_SINGLE_PLANE_MODE

[    2.074669]  oob layout use nand base oob layout oobsize = 16,oobmul =1,mtd->oobsize =640,aml_chip->oob_size =640

[    2.084953] Enter aml_nand_get_reg_value_hynix

[    2.089317] aml_nand_get_reg_value_hynix, REG(0xa0):     value:0x25, for chip[0]

[    2.096400] aml_nand_get_reg_value_hynix, REG(0xa1):     value:0x38, for chip[0]

[    2.103513] aml_nand_get_reg_value_hynix, REG(0xb0):     value:0x20, for chip[0]

[    2.110613] aml_nand_get_reg_value_hynix, REG(0xb1):     value:0x34, for chip[0]

[    2.117706] aml_nand_get_reg_value_hynix, REG(0xc9):     value:0x47, for chip[0]

[    2.128530] ######aml_nand_get_read_default_value_hynix 779 read default read retry reg value failed and need read from chip write back to nand using SLC

[    2.138507] Enter aml_nand_get_reg_value_hynix

[    2.142923] aml_nand_get_reg_value_hynix, REG(0xa7):     value:0x21, for chip[0]

[    2.150023] aml_nand_get_reg_value_hynix, REG(0xad):     value:0x21, for chip[0]

[    2.157116] aml_nand_get_reg_value_hynix, REG(0xae):     value:0x72, for chip[0]

[    2.164226] aml_nand_get_reg_value_hynix, REG(0xaf):     value:0xbf, for chip[0]

[    2.213334] aml nand env valid addr: ffffffffffe00000 

[    2.213358] CONFIG_ENV_SIZE=0x2000; ENV_SIZE=0x1ffc; bbt=0x1330; default_environment_size=0xccc

[    2.213374] nand env: nand_env_probe. 

[    2.216876] set nand parts for chip 4096MB

[    2.240671] Creating 8 MTD partitions on "B revision 26nm NAND 4GiB H27UBG8T2B":

[    2.242507] 0x000001400000-0x000001800000 : "ubootenv"

[    2.249335] 0x000001800000-0x000002800000 : "aml_logo"

[    2.258119] 0x000002800000-0x000003800000 : "recovery"

[    2.263516] 0x000003800000-0x000004c00000 : "boot"

[    2.270385] 0x000004c00000-0x000024c00000 : "system"

[    2.432261] 0x000024c00000-0x000030c00000 : "cache"

[    2.491157] 0x000030c00000-0x000050c00000 : "userdata"

[    2.652225] 0x000050c00000-0x000100000000 : "NFTL_Part"

[    4.761191] nftl initilize completely dev size: 0xa1e00000 5

[    4.761871]  avnftl8: unknown partition table

[    4.766656] nftl release flush cache data: 0

davilla · 2013-01-25, 16:33

your nand setup looks borked.

noggin · (This post was last modified: 2013-01-25, 17:19 by noggin.)

(2013-01-25, 16:33)davilla Wrote: your nand setup looks borked.

Yep... Can I bypass the NAND? My aim is to run everything (kernel, rootfs and anything else like userdata) from an SD card - and, whilst developing, from NFS. Or is it a case of working out why my NAND set-up is borked?

Can I just edit S10setup - I guess I could try commenting out the mount yaffs2 /tmp/userdata stuff?

I'm still new to Linux but :

Code:
mkdir -p /tmp/userdata

mount -t yaffs2 $USERDATA /tmp/userdata  

mkdir -p /tmp/userdata/xios/root  

mount -o bind /tmp/userdata/xios/root /root

Looks to mean :

Make a /tmp/userdata folder (whether it exists or not)
Mount $USERDATA (which will be the MTD partition called "userdata") as a YAFFS2 file system at /tmp/userdata
Make a /tmp/userdata/xios/root folder (whether it exists or not)
Mount this folder to /root

So this makes me think that /root (home folder for user /root) is mounted to a folder within the YAFFS2 file-system on the NAND partition? If my NAND stuff doesn't work then I don't have a /root home folder... Which is a problem.

If I don't mount the yaffs2 stuff - will my /tmp/userdata and /tmp/userdata/xios/root be created on my NFS rootfs? (Do I need to do somethign to make this RW?)

davilla · (This post was last modified: 2013-01-25, 18:55 by davilla.)

you are like a machine gun with guestions Smile

pivos buildroot-linux is designed to boot and run from nand but it's also designed to co-boot via tftp/nfs so we can can do quick dev.

booting from sdcard is possible, you just tell uboot to do it.

noggin · 2013-01-25, 23:14

(2013-01-25, 18:48)davilla Wrote: you are like a machine gun with guestions

Sorry

Quote:pivos buildroot-linux is designed to boot and run from nand but it's also designed to co-boot via tftp/nfs so we can can do quick dev.

Thanks - I'll try commenting out the yaffs2 stuff in S10setup.

Quote:booting from sdcard is possible, you just tell uboot to do it.

Yep - managed to do this in uBoot with the same issues with xbmc not starting - I guess for the same NAND-y reasons.

Will get back in my box.

Thanks again

davilla · 2013-01-25, 23:42

post a uboot log, nand partitions.

noggin · 2013-01-26, 02:22

(2013-01-25, 23:42)davilla Wrote: post a uboot log, nand partitions.

Most recent uboot log here : http://pastebin.com/K3V98hx1

Think this is the relevant NAND stuff :

Code:
U-Boot 2010.06(m3_mbox@trunk) (Sep 23 2012 - 18:43:57)

aml_i2c_init 8f89369c 8f895734 

DRAM:  1 GiB

unknow ddr type

NAND:  Amlogic nand flash uboot driver for M3, Version 1.1 (c) 2010 Amlogic Inc.

No NAND device found!!!

NAND device id: ad d7 94 da 74 c3 

aml_chip->hynix_new_nand_type =: 2 

NAND device: Manufacturer ID: 0xad, Chip ID: 0xad (Hynix B revision 26nm NAND 4GiB H27UBG8T2B)

1 NAND chips detected

#####aml_nand_init, with RB pins and chip->chip_delay:20

bus_cycle=4, bus_timing=5, start_cycle=5, end_cycle=6,system=5.3ns

oob size is not enough for selected bch mode: NAND_BCH60_1K_MODE force bch to mode: NAND_BCH40_1K_MODE

multi plane error for selected plane mode: NAND_TWO_PLANE_MODE force plane to : NAND_SINGLE_PLANE_MODE

Enter aml_nand_get_reg_value_hynix

aml_nand_get_reg_value_hynix, REG(0xa0):     value:0x25, for chip[0]

aml_nand_get_reg_value_hynix, REG(0xa1):     value:0x38, for chip[0]

aml_nand_get_reg_value_hynix, REG(0xb0):     value:0x20, for chip[0]

aml_nand_get_reg_value_hynix, REG(0xb1):     value:0x34, for chip[0]

aml_nand_get_reg_value_hynix, REG(0xc9):     value:0x47, for chip[0]

aml_nand_get_read_default_value_hynix 926 get default reg value at blk:0, page:7

aml_nand_get_read_default_value_hynix, REG(0xa7):    value:0x21, for chip[0]

aml_nand_get_read_default_value_hynix, REG(0xad):    value:0x21, for chip[0]

aml_nand_get_read_default_value_hynix, REG(0xae):    value:0x72, for chip[0]

aml_nand_get_read_default_value_hynix, REG(0xaf):    value:0xbf, for chip[0]

aml nand env valid addr: 402000 

Creating 7 MTD partitions on "nandmulti":

0x000000c00000-0x000001c00000 : "aml_logo"

0x000001c00000-0x000002c00000 : "recovery"

0x000002c00000-0x000004000000 : "boot"

0x000004000000-0x000024000000 : "system"

0x000024000000-0x000030000000 : "cache"

0x000030000000-0x000050000000 : "userdata"

0x000050000000-0x000100000000 : "NFTL_Part"

nandmulti initialized ok

davilla · 2013-01-26, 03:58

and same from dmesg ?

noggin · 2013-01-26, 04:59

(2013-01-26, 03:58)davilla Wrote: and same from dmesg ?

http://pastebin.com/5yvA0uWP

Think the bits below are relevant?

Code:
[    1.338002] Amlogic nand flash AVOS driver for M3, Version 1.1 (c) 2010 Amlogic Inc.

[    1.345850] bus_cycle=17, bus_timing=10, start_cycle=10, end_cycle=10,system=5.3ns

[    1.353580] No NAND device found.

[    1.356754] NAND device id: ad d7 94 da 74 c3

[    1.360969] aml_chip->hynix_new_nand_type =: 2

[    1.365461] NAND device: Manufacturer ID: 0xad, Chip ID: 0xad (Hynix B revision 26nm NAND 4GiB H27UBG8T2B)

[    1.375202] #####aml_nand_init, with RB pins and chip->chip_delay:20

[    1.381484] bus_cycle=4, bus_timing=5, start_cycle=5, end_cycle=6,system=5.3ns

[    1.388586] oob size is not enough for selected bch mode: NAND_BCH60_1K_MODE force bch to mode: NAND_BCH40_1K_MODE

[    1.398984] Enter aml_nand_get_reg_value_hynix

[    1.403315] aml_nand_get_reg_value_hynix, REG(0xa0):         value:0x25, for chip[0]

[    1.410425] aml_nand_get_reg_value_hynix, REG(0xa1):         value:0x38, for chip[0]

[    1.417516] aml_nand_get_reg_value_hynix, REG(0xb0):         value:0x20, for chip[0]

[    1.424627] aml_nand_get_reg_value_hynix, REG(0xb1):         value:0x34, for chip[0]

[    1.431729] aml_nand_get_reg_value_hynix, REG(0xc9):         value:0x47, for chip[0]

[    1.438825] Creating 1 MTD partitions on "B revision 26nm NAND 4GiB H27UBG8T2B":

[    1.446195] 0x000000000000-0x000000800000 : "bootloader"

[    1.452155] bus_cycle=17, bus_timing=10, start_cycle=10, end_cycle=10,system=5.3ns

[    1.459354] No NAND device found.

[    1.462509] NAND device id: ad d7 94 da 74 c3

[    1.466709] aml_chip->hynix_new_nand_type =: 2

[    1.471233] NAND device: Manufacturer ID: 0xad, Chip ID: 0xad (Hynix B revision 26nm NAND 4GiB H27UBG8T2B)

[    2.042769] 1 NAND chips detected

[    2.042808] #####aml_nand_init, with RB pins and chip->chip_delay:20

[    2.046930] bus_cycle=4, bus_timing=5, start_cycle=5, end_cycle=6,system=5.3ns

[    2.054060] oob size is not enough for selected bch mode: NAND_BCH60_1K_MODE force bch to mode: NAND_BCH40_1K_MODE

[    2.064359] multi plane error for selected plane mode: NAND_TWO_PLANE_MODE force plane to : NAND_SINGLE_PLANE_MODE

[    2.074669]  oob layout use nand base oob layout oobsize = 16,oobmul =1,mtd->oobsize =640,aml_chip->oob_size =640

[    2.084953] Enter aml_nand_get_reg_value_hynix

[    2.089317] aml_nand_get_reg_value_hynix, REG(0xa0):         value:0x25, for chip[0]

[    2.096400] aml_nand_get_reg_value_hynix, REG(0xa1):         value:0x38, for chip[0]

[    2.103513] aml_nand_get_reg_value_hynix, REG(0xb0):         value:0x20, for chip[0]

[    2.110613] aml_nand_get_reg_value_hynix, REG(0xb1):         value:0x34, for chip[0]

[    2.117706] aml_nand_get_reg_value_hynix, REG(0xc9):         value:0x47, for chip[0]

[    2.128530] ######aml_nand_get_read_default_value_hynix 779 read default read retry reg value failed and need read from chip write back to nand using SLC

[    2.138507] Enter aml_nand_get_reg_value_hynix

[    2.142923] aml_nand_get_reg_value_hynix, REG(0xa7):         value:0x21, for chip[0]

[    2.150023] aml_nand_get_reg_value_hynix, REG(0xad):         value:0x21, for chip[0]

[    2.157116] aml_nand_get_reg_value_hynix, REG(0xae):         value:0x72, for chip[0]

[    2.164226] aml_nand_get_reg_value_hynix, REG(0xaf):         value:0xbf, for chip[0]

[    2.213334] aml nand env valid addr: ffffffffffe00000

[    2.213358] CONFIG_ENV_SIZE=0x2000; ENV_SIZE=0x1ffc; bbt=0x1330; default_environment_size=0xccc

[    2.213374] nand env: nand_env_probe.

[    2.216876] set nand parts for chip 4096MB

[    2.240671] Creating 8 MTD partitions on "B revision 26nm NAND 4GiB H27UBG8T2B":

[    2.242507] 0x000001400000-0x000001800000 : "ubootenv"

[    2.249335] 0x000001800000-0x000002800000 : "aml_logo"

[    2.258119] 0x000002800000-0x000003800000 : "recovery"

[    2.263516] 0x000003800000-0x000004c00000 : "boot"

[    2.270385] 0x000004c00000-0x000024c00000 : "system"

[    2.432261] 0x000024c00000-0x000030c00000 : "cache"

[    2.491157] 0x000030c00000-0x000050c00000 : "userdata"

[    2.652225] 0x000050c00000-0x000100000000 : "NFTL_Part"

[    4.761191] nftl initilize completely dev size: 0xa1e00000 5

[    4.761871]  avnftl8: unknown partition table

[    4.766656] nftl release flush cache data: 0

Code:
[    9.554918] Freeing init memory: 192K

[    9.558529] run_init_process /sbin/init.

[   10.001175] yaffs: dev is 32505859 name is "mtdblock3" rw

[   10.001210] yaffs: passed flags ""

[   10.004409] yaffs: Attempting MTD mount of 31.3,"mtdblock3"

[   10.004433] yaffs: NAND geometry problems: chunk size 8192, type is yaffs2, inbandTags 1

[   10.929912] yaffs: dev is 32505863 name is "mtdblock7" rw

[   10.929946] yaffs: passed flags ""

[   10.933145] yaffs: Attempting MTD mount of 31.7,"mtdblock7"

[   10.933224] yaffs2 check point success.

[   10.961719] yaffs: restored from checkpoint

[   10.963131] yaffs_read_super: isCheckpointed 1

[   10.977155] aml nand write failed at 107692

[   10.977828] **>> yaffs write required 2 attempts

[   11.107238] <30>udevd[108]: starting version 173

[   11.979412] page 39429 in gc has no object: 0 0 0

[   11.979736] page 39430 in gc has no object: 0 0 0

[   11.980052] page 39431 in gc has no object: 0 0 0

[   11.980363] page 39432 in gc has no object: 0 0 0

[   11.980678] page 39433 in gc has no object: 0 0 0

[   12.654897] nftl release flush cache data: 0

[   13.419577] set hash low=2,high=810001

Code:
[    9.554918] Freeing init memory: 192K

[    9.558529] run_init_process /sbin/init.

[   10.001175] yaffs: dev is 32505859 name is "mtdblock3" rw

[   10.001210] yaffs: passed flags ""

[   10.004409] yaffs: Attempting MTD mount of 31.3,"mtdblock3"

[   10.004433] yaffs: NAND geometry problems: chunk size 8192, type is yaffs2, inbandTags 1

[   10.929912] yaffs: dev is 32505863 name is "mtdblock7" rw

[   10.929946] yaffs: passed flags ""

[   10.933145] yaffs: Attempting MTD mount of 31.7,"mtdblock7"

[   10.933224] yaffs2 check point success.

[   10.961719] yaffs: restored from checkpoint

[   10.963131] yaffs_read_super: isCheckpointed 1

[   10.977155] aml nand write failed at 107692

[   10.977828] **>> yaffs write required 2 attempts

[   11.107238] <30>udevd[108]: starting version 173

[   11.979412] page 39429 in gc has no object: 0 0 0

[   11.979736] page 39430 in gc has no object: 0 0 0

[   11.980052] page 39431 in gc has no object: 0 0 0

[   11.980363] page 39432 in gc has no object: 0 0 0

[   11.980678] page 39433 in gc has no object: 0 0 0

[   12.654897] nftl release flush cache data: 0

[   13.419577] set hash low=2,high=810001

davilla · (This post was last modified: 2013-01-26, 07:13 by davilla.)

Creating 7 MTD partitions on "nandmulti":
0x000000c00000-0x000001c00000 : "aml_logo"
0x000001c00000-0x000002c00000 : "recovery"
0x000002c00000-0x000004000000 : "boot"
0x000004000000-0x000024000000 : "system"
0x000024000000-0x000030000000 : "cache"
0x000030000000-0x000050000000 : "userdata"
0x000050000000-0x000100000000 : "NFTL_Part"

[ 2.242507] 0x000001400000-0x000001800000 : "ubootenv"
[ 2.249335] 0x000001800000-0x000002800000 : "aml_logo"
[ 2.258119] 0x000002800000-0x000003800000 : "recovery"
[ 2.263516] 0x000003800000-0x000004c00000 : "boot"
[ 2.270385] 0x000004c00000-0x000024c00000 : "system"
[ 2.432261] 0x000024c00000-0x000030c00000 : "cache"
[ 2.491157] 0x000030c00000-0x000050c00000 : "userdata"
[ 2.652225] 0x000050c00000-0x000100000000 : "NFTL_Part"

Something does not look right here. Your kernel partition map should match your uboot partition map. See below for examples.

Pivos M3:
Creating 6 MTD partitions on "nandmulti":
0x000000800000-0x000000c00000 : "logo"
0x000000c00000-0x000001400000 : "boot"
0x000001400000-0x000021400000 : "system"
0x000021400000-0x000040800000 : "cache"
0x000040800000-0x000060800000 : "NFTL_Part"
0x000060800000-0x000080000000 : "userdata"

[ 2.083441] Creating 6 MTD partitions on "B revision NAND 2GiB H27UAG8T2B":
[ 2.087848] 0x000000800000-0x000000c00000 : "logo"
[ 2.093172] 0x000000c00000-0x000001400000 : "boot"
[ 2.097851] 0x000001400000-0x000021400000 : "system"
[ 2.102722] 0x000021400000-0x000040800000 : "cache"
[ 2.107612] 0x000040800000-0x000060800000 : "NFTL_Part"
[ 2.112755] 0x000060800000-0x000080000000 : "userdata"

ATV510B M3:
Creating 7 MTD partitions on "nandmulti":
0x000000800000-0x000000c00000 : "logo"
0x000000c00000-0x000001400000 : "boot"
0x000001400000-0x000021400000 : "system"
0x000021400000-0x000034000000 : "cache"
0x000034000000-0x000074000000 : "NFTL_Part"
0x000074000000-0x000080800000 : "backup"
0x000080800000-0x000100000000 : "userdata"

[ 2.241703] Creating 7 MTD partitions on "C revision NAND 4GiB MT29F32G-C":
[ 2.247376] 0x000000800000-0x000000c00000 : "logo"
[ 2.253147] 0x000000c00000-0x000001400000 : "boot"
[ 2.257703] 0x000001400000-0x000021400000 : "system"
[ 2.267451] 0x000021400000-0x000034000000 : "cache"
[ 2.272355] 0x000034000000-0x000074000000 : "NFTL_Part"
[ 2.277512] 0x000074000000-0x000080800000 : "backup"
[ 2.282504] 0x000080800000-0x000100000000 : "userdata"

~~j1nx~~ · (This post was last modified: 2013-01-26, 08:49 by j1nx.)

Turn off the CONFIG_AML_NAND_ENV parameter.

noggin · (This post was last modified: 2013-01-26, 14:07 by noggin.)

(2013-01-26, 08:46)j1nx Wrote: Turn off the CONFIG_AML_NAND_ENV parameter.

Thanks j1nx and davilla for your continuing help with this box. Above and beyond - and much appreciated. I'm not going to give up yet!

I think I've done this. (I did it by editing my defconfig in my /buildroot-AmLogic/output/build/linux-<lots of hex>/arch/arm/configs folder and then deleting all but the .stamp_downloaded stamps in the /build/linux-<lots of hex> folder) Doing a make with the defconfig and then a make. (This is what I did to add DVB support and then update the DVB frontend to report symbol rates)

This is what I get in uBoot :

Code:
Creating 7 MTD partitions on "nandmulti":

0x000000c00000-0x000001c00000 : "aml_logo"

0x000001c00000-0x000002c00000 : "recovery"

0x000002c00000-0x000004000000 : "boot"

0x000004000000-0x000024000000 : "system"

0x000024000000-0x000030000000 : "cache"

0x000030000000-0x000050000000 : "userdata"

0x000050000000-0x000100000000 : "NFTL_Part"

This is what I get in dmesg:

Code:
[    2.135596] 0x000000c00000-0x000001c00000 : "aml_logo"

[    2.141311] 0x000001c00000-0x000002c00000 : "recovery"

[    2.146238] 0x000002c00000-0x000004000000 : "boot"

[    2.151035] 0x000004000000-0x000024000000 : "system"

[    2.155910] 0x000024000000-0x000030000000 : "cache"

[    2.160828] 0x000030000000-0x000050000000 : "userdata"

[    2.165926] 0x000050000000-0x000100000000 : "NFTL_Part"

That's now matching isn't it? Which is a good sign?

And I get this :

Code:
[    7.371703] run_init_process /sbin/init.

[    7.709642] yaffs: dev is 32505859 name is "mtdblock3" rw

[    7.709676] yaffs: passed flags ""

[    7.712874] yaffs: Attempting MTD mount of 31.3,"mtdblock3"

[    7.712954] yaffs2 check point success.

[    7.768146] yaffs: restored from checkpoint

[    7.768597] yaffs_read_super: isCheckpointed 1

[   12.098036] yaffs: dev is 32505861 name is "mtdblock5" rw

[   12.098071] yaffs: passed flags ""

[   12.101269] yaffs: Attempting MTD mount of 31.5,"mtdblock5"

[   12.101350] yaffs2 check point success.

[   12.134186] yaffs: restored from checkpoint

[   12.135607] yaffs_read_super: isCheckpointed 1

Which also looks like a good sign?

However I also get a LOT of these :

Code:
[   12.149943] page 39736 in gc has no object: 0 0 0 

[   12.150264] page 39737 in gc has no object: 0 0 0 

[   12.150582] page 39738 in gc has no object: 0 0 0 

[   12.150894] page 39739 in gc has no object: 0 0 0 

[   12.151209] page 39740 in gc has no object: 0 0 0 

[   12.155001] page 39744 in gc has no object: 0 0 0 

[   12.155321] page 39748 in gc has no object: 0 0 0 

[   12.155634] page 39749 in gc has no object: 0 0 0 

[   12.155948] page 39750 in gc has no object: 0 0 0 

[   12.156313] page 39751 in gc has no object: 0 0 0

and when I navigate to /tmp/userdata and do an ls I get screens of blue ?s.

Which suggests something is still wrong with my yaffs2/NAND stuff I think? (Googling suggests 'tags being corrupted')

Thoughts?

Full dmesg here : http://pastebin.com/arv3Utb6

Is there anyway of not using NAND Flash for /tmp/userdata? (Or is that a bad idea?)

~~j1nx~~ · 2013-01-26, 14:54

Just pass on the rootfs to be pointing a (second) partition on your SD card, you can even just use ext if you want. (don't forget to define it then as well in cmd_line parameters)

The in the S10 setup, it creates a userdata directory and mounts the actual nand partition into that directory.

Just create the directory in your ext partition and don't create and mount in S10...

noggin · (This post was last modified: 2013-01-26, 17:21 by noggin.)

(2013-01-26, 14:54)j1nx Wrote: Just pass on the rootfs to be pointing a (second) partition on your SD card, you can even just use ext if you want. (don't forget to define it then as well in cmd_line parameters)

Sorry - being a bit slow here. Are you suggesting I untar my rootfs.tar to an ext/ext2/ext3 partition on my SD card and don't use NFS - or that I just create an ext/ext2/ext3 partition for my /tmp stuff and mount it on the SD card?

Quote:The in the S10 setup, it creates a userdata directory and mounts the actual nand partition into that directory.

Yep - this bit :

Code:
mkdir -p /tmp/userdata

mount -t yaffs2 $USERDATA /tmp/userdata  

mkdir -p /tmp/userdata/xios/root  

mount -o bind /tmp/userdata/xios/root /root

If I comment out the mount -t yaffs2 bit it should just the rootfs /tmp/userdata dir instead?

Quote:Just create the directory in your ext partition and don't create and mount in S10...

If you could suggest example bootargs that would be helpful?

Is it something like : root=/dev/cardblksd2 rw rootfstype=ext3 if I'm using the entire rootfs from SD Card?

* EDIT : Well I tried the above with a /tmp/userdata/xios/root directory on the SD card and got segmentation fault when I used the above.

Does the /tmp dir get deleted at boot up? (I'm guessing it does?)

Code:
[   16.787110] yaffs: dev is 32505859 name is "mtdblock3" rw

[   16.787136] yaffs: passed flags ""

[   16.790528] yaffs2 check point success.

chown: unknown user/group default:default

S10setup: init fb0, fb1

2880+0 records in

2880+0 records out

3686400 bytes (3.5MB) copied, 0.013296 seconds, 264.4MB/s

2880+0 records in

2880+0 records out

3686400 bytes (3.5MB) copied, 0.013154 seconds, 267.3MB/s

S10setup: remount / rw

[   17.045446] EXT3-fs (cardblksd2): using internal journal

S10setup: checking keys

generating rsa key... generating dsa key... [   19.089442] nftl creat stucture completely free blk: 11 erased blk: 1385

S10setup: moving /etc

S10setup: moving /media

S10setup: mounting /tmp/userdata

Segmentation fault

S10setup: remount / ro

S10setup: switching to ondemand

S10setup: done

S10setup: init fb0, fb1

2880+0 records in

2880+0 records out

3686400 bytes (3.5MB) copied, 0.019143 seconds, 183.7MB/s

2880+0 records in

2880+0 records out

3686400 bytes (3.5MB) copied, 0.019237 seconds, 182.8MB/s

S10setup: remount / rw

[   29.382407] EXT3-fs (cardblksd2): using internal journal

S10setup: checking keys

S10setup: moving /etc

cp: '/etc' and '/tmp/etc' are the same file

S10setup: moving /media

cp: '/media' and '/tmp/media' are the same file

S10setup: mounting /tmp/userdata

[   29.493999] yaffs: dev is 32505861 name is "mtdblock5" rw

[   29.494040] yaffs: passed flags ""

[   29.497340] yaffs2 check point success.

S10setup: remount / ro

S10setup: switching to ondemand

S10setup: done

Populating  using udev: Segmentation fault

FAIL

This was with an IP= line in my bootargs still. I tried without the IP= line and didn't get as far.

I'm obviously not understanding something obvious here...