[U-Boot] [bug report] sunxi: booting from eMMC

Wed Oct 12 12:52:07 CEST 2016

In order to exclude possible errors on my board, i get MarsBoard A20 Lite, desolder nand, and put on emmc pads my emmc flash.
So emmc is on sdc2 lines. When iam booting from sd on sdc1, linux kernel can init emmc, and i got /dev/mmcblkp1, works well.
But when i eject sd and try to boot from sdc2 emmc, get the same err 17. 
In this case i have used some old uboot from marsboard site, because mainline u-boot have not config for a20 marsboard.

U-Boot SPL 2014.04-10685-g4af825f-dirty (Jun 14 2014 - 17:55:48)
Board: MarsBoard-A20 & MarsPI
DRAM: 1024 MiB
CPU: 960000000Hz, AXI/AHB/APB: 3/2/2
Card did not respond to voltage select!
spl: mmc init failed: err - -17
### ERROR ### Please RESET the board ###

So, the problem with eMMC KLM4G1FE3B-B001 is exists(
Maybe you know some cheap emmc, which is 100% works with mainline u-boot?

> 7 окт. 2016 г., в 14:44, Alexandr Bochkarev <aneox.inbox at gmail.com> написал(а):
> 
> Hi!
> Very appreciate your help! Hope we will fix it, because my board will be garbage without emmc (
> So, my steps:
> 
> git clone git://git.denx.de/u-boot.git <git://git.denx.de/u-boot.git> --depth 1; cd u-boot;
> make CROSS_COMPILE=arm-linux-gnueabihf- A20-OLinuXino-Lime_defconfig
> make CROSS_COMPILE=arm-linux-gnueabihf-
> 
> Also turn on CONFIG_MMC_TRACE
> 
> After run, got:
> 
> Trying to boot from MMC1
> CMD_SEND:0
> ..ARG... 0x00000000
> ..MMC_RSP_NONE
> CMD_SEND:8
> ..ARG... 0x00000000
> ..RET... -110
> CMD_SEND:55
> ..ARG... 0x00000000
> ..RET... -110
> CMD_SEND:0
> ..ARG... 0x00000000
> ..MMC_RSP_NONE
> CMD_SEND:1
> ..ARG... 0x00000000
> ..MMC_RSP_R3,4.. 0x00000000 
> CMD_SEND:1
> ..ARG... 0x00000000
> ..MMC_RSP_R3,4.. 0x00000000 
> CMD_SEND:0
> ..ARG... 0x00000000
> ..MMC_RSP_NONE
> CMD_SEND:1
> ..ARG... 0x00000000
> ..MMC_RSP_R3,4.. 0x00000000 
> CMD_SEND:1
> ..ARG... 0x00000000
> ..MMC_RSP_R3,4.. 0x00000000 
> CMD_SEND:2
> ..ARG... 0x00000000
> ..MMC_RSP_R2.. 0x00000000 
> ..          .. 0x00000000 
> ..          .. 0x00000000 
> ..          .. 0x00000000 
> 
> .....DUMPING DATA
> .....000 - 00 00 00 00 
> .....004 - 00 00 00 00 
> .....008 - 00 00 00 00 
> .....012 - 00 00 00 00 
> CMD_SEND:3
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:9
> ..ARG... 0x00000000
> ..MMC_RSP_R2.. 0x00000000 
> ..          .. 0x00000000 
> ..          .. 0x00000000 
> ..          .. 0x00000000 
> 
> .....DUMPING DATA
> .....000 - 00 00 00 00 
> .....004 - 00 00 00 00 
> .....008 - 00 00 00 00 
> .....012 - 00 00 00 00 
> CMD_SEND:13
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CURR STATE:3
> CMD_SEND:7
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:8
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:8
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:6
> ..ARG... 0x00000000
> ..MMC_RSP_R1b.. 0x00000000 
> CMD_SEND:13
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CURR STATE:4
> CMD_SEND:8
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:6
> ..ARG... 0x00000000
> ..RET... -110
> CMD_SEND:16
> ..ARG... 0x00000000
> ..MMC_RSP_R1,5,6,7 . 0x00000000 
> CMD_SEND:17
> ..ARG... 0x00000000
> ..RET... -110
> 
> hdr read sector 80, count=0
> mmc_load_image_raw_sector: mmc block read error
> SPL: failed to boot from all boot devices
> ### ERROR ### Please RESET the board ###
> 
> So, 
> spl_mmc.c on line 66, blk_dread return 0 count
> blk.h on line 389, block_dev->block_read(block_dev, start, blkcnt, buffer) return 0
> block_dev->block_read is a pointer, seems like it defined in mmc_legacy.c on line 144, bdesc->block_read = mmc_bread;
> mmc.c, line 285 mmc_read_blocks return Failed to read blocks
> mmc.c, line 223, mmc_send_cmd return 0, cmdidx was 17, like last in mmc trace
> 
> 
>> 7 окт. 2016 г., в 5:48, Jaehoon Chung <jh80.chung at samsung.com <mailto:jh80.chung at samsung.com>> написал(а):
>> 
>> On 10/03/2016 02:39 AM, AneoX wrote:
>>> Emmc Card was the new one, so i just burn a dump of working sd. Previously
>>> sd have been wired to board for test.
>>> 
>>> Board is a custom,  like olinuxino lime which with nand and sd slot, 512 mb
>>> ram. So i using it defconfig and in case of sd card, it works.
>>> Emmc lines has external pullups according to the datasheet.
>> 
>> Well, i needs to get more information..If you guess that seems to return 0 from block_read()..
>> then you can find which condition was hit..is it impossible?
>> 
>> If you can share more debug message..then i will help you.
>> 
>> Best Regards,
>> Jaehoon Chung
>> 
>>> 
>>> 2 окт. 2016 г. 23:28 пользователь "Jagan Teki" <jagannadh.teki at gmail.com <mailto:jagannadh.teki at gmail.com>>
>>> написал:
>>> 
>>> On Sun, Oct 2, 2016 at 10:45 PM, AneoX <aneox.inbox at gmail.com <mailto:aneox.inbox at gmail.com>> wrote:
>>>> Like a sd card. I have emmc to sd adapter, so
>>>> 
>>>> dd if=u-boot-sunxi-with-spl.bin of=/dev/sdb bs=1024 seek=8
>>> 
>>> I guess you cleaned[1] the card before doing this?
>>> 
>>>> 
>>>> and then just solder emmc to my board.
>>> 
>>> OK, which a20 board?
>>> 
>>> [1] http://linux-sunxi.org/Bootable_SD_card <http://linux-sunxi.org/Bootable_SD_card>
>>> 
>>> thanks!
>>> --
>>> Jagan Teki
>>> Free Software Engineer | www.openedev.com <http://www.openedev.com/>
>>> U-Boot, Linux | Upstream Maintainer
>>> Hyderabad, India.
>>> _______________________________________________
>>> U-Boot mailing list
>>> U-Boot at lists.denx.de <mailto:U-Boot at lists.denx.de>
>>> http://lists.denx.de/mailman/listinfo/u-boot <http://lists.denx.de/mailman/listinfo/u-boot>