Hello,
I tried all 3 Linux images on the M3. On none of them i could see my SATA disk. Anyone managed to see the disk via the SATA port? USB works though… but i need SATA (not for speed reasons, for other reasons)
Thanks, Vlad.
The M3 has no SATA, just an onboard GL830 USB-to-SATA bridge.
If USB works then a disk should work. If it’s not working you loose exactly nothing by accessing a disk through an USB enclosure. It will be as slow as with ‘onboard’ USB-to-SATA.
Hello,
Thank you for your answer. I need to work with the SATA connector since i used it with an enclosure. I don’t care about the speed, can be USB speed, but i need the SATA connector. Seams is not working, any idea? Anyone tried?
I just connected an EVO 840. The polarity of the SATA power connector is the same as with other Banana and Orange Pis (therefore not compatible to Linksprite und Cubietech):
[ 11.624803] scsi 0:0:0:0: Direct-Access Samsung SSD 840 EVO 120G 0016 PQ: 0 ANSI: 4
[ 11.637009] EXT4-fs (mmcblk0p2): re-mounted. Opts: errors=remount-ro
[ 11.649766] sd 0:0:0:0: Attached scsi generic sg0 type 0
[ 11.650373] sd 0:0:0:0: [sda] 234441644 512-byte logical blocks: (120 GB/111 GiB)
[ 11.651362] sd 0:0:0:0: [sda] Write Protect is off
[ 11.651370] sd 0:0:0:0: [sda] Mode Sense: 03 00 00 00
[ 11.652362] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.652369] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.655232] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.655239] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.669043] Alternate GPT is invalid, using primary GPT.
[ 11.669093] sda: sda1 sda2
[ 11.748603] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.758588] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.769226] sd 0:0:0:0: [sda] Attached SCSI disk
SAT/SMART works also (small exception below):
root@BPiM3:~# smartctl -a /dev/sda
smartctl 6.4 2014-10-07 r4002 [armv7l-linux-3.4.39-BPI-M3-Kernel] (local build)
Copyright (C) 2002-14, Bruce Allen, Christian Franke, www.smartmontools.org
=== START OF INFORMATION SECTION ===
Model Family: Samsung based SSDs
Device Model: Samsung SSD 840 EVO 120GB
...
=== START OF READ SMART DATA SECTION ===
SMART Status not supported: Incomplete response, ATA output registers missing
SMART overall-health self-assessment test result: PASSED
Warning: This result is based on an Attribute check.
Currently I’m testing performance using iozone (will take some time). And no, I’m not using any of the ‘official’ images but created my own. This is necessary since all 3 linux images are corrupted and they started just recently to fix important stuff. You can use the Raspbian image as a skeleton, clone their M3-BSP and compile the stuff and then combine both: http://linux-sunxi.org/User_talk:Tkaiser#First_steps_with_Banana_Pi_M3
I got 13MB/s sequential writes and 23MB/s sequential reads from a fast SSD. That means a ‘slow’ A20 board like the Banana Pi M1 is 2-3 times faster when accessing USB storage and 3.5-10 times when using SATA.
Just tried it with a 3TB Seagate Barracuda to get a clue whether the GL830 chip suffers from a 2TB limitation or not (a question 100 times discussed here in the forums without a single answer from SinoVoip). It seems to work:
[ 11.564937] sd 0:0:0:0: Attached scsi generic sg0 type 0
[ 11.565654] sd 0:0:0:0: [sda] Very big device. Trying to use READ CAPACITY(16).
[ 11.566156] sd 0:0:0:0: [sda] 5860533164 512-byte logical blocks: (3.00 TB/2.72 TiB)
[ 11.566165] sd 0:0:0:0: [sda] 4096-byte physical blocks
[ 11.567180] sd 0:0:0:0: [sda] Write Protect is off
[ 11.567192] sd 0:0:0:0: [sda] Mode Sense: 03 00 00 00
[ 11.568151] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.576608] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.586772] sd 0:0:0:0: [sda] Very big device. Trying to use READ CAPACITY(16).
[ 11.588893] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.597359] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.629587] Alternate GPT is invalid, using primary GPT.
[ 11.629652] sda: sda1 sda2 sda3 sda4
[ 11.631650] sd 0:0:0:0: [sda] Very big device. Trying to use READ CAPACITY(16).
[ 11.633769] sd 0:0:0:0: [sda] No Caching mode page present
[ 11.642356] sd 0:0:0:0: [sda] Assuming drive cache: write through
[ 11.651583] sd 0:0:0:0: [sda] Attached SCSI disk
root@BPiM3:~# parted -s /dev/sda mklabel gpt
root@BPiM3:~# SectorCount=$(parted -s /dev/sda unit s print | awk -F": " '/^Disk \/dev\/sda/ {print $2}' | tr -d 's')
root@BPiM3:~# LastUsableSector=$(( ${SectorCount} / 2048 ))
root@BPiM3:~# parted -s /dev/sda --align optimal unit s mkpart primary ext4 2048 $(( ${LastUsableSector} * 2048 ))
root@BPiM3:~# mkfs.ext4 /dev/sda1
mke2fs 1.42.12 (29-Aug-2014)
Creating filesystem with 732566272 4k blocks and 183148544 inodes
Filesystem UUID: a111af39-98c5-40ac-8b28-1539d1da6760
Superblock backups stored on blocks:
32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
4096000, 7962624, 11239424, 20480000, 23887872, 71663616, 78675968,
102400000, 214990848, 512000000, 550731776, 644972544
Allocating group tables: done
Writing inode tables: done
Creating journal (32768 blocks): done
Writing superblocks and filesystem accounting information: done
root@BPiM3:~# cat /proc/partitions
major minor #blocks name
179 0 7761920 mmcblk0
179 1 51200 mmcblk0p1
179 2 5242880 mmcblk0p2
179 16 7634944 mmcblk1
179 17 5025792 mmcblk1p1
179 18 16384 mmcblk1p2
179 19 1 mmcblk1p3
179 21 16384 mmcblk1p5
179 22 16384 mmcblk1p6
179 23 1572864 mmcblk1p7
179 24 16384 mmcblk1p8
179 25 32768 mmcblk1p9
179 26 786432 mmcblk1p10
179 27 16384 mmcblk1p11
179 28 16384 mmcblk1p12
179 29 81920 mmcblk1p13
179 48 4096 mmcblk1boot1
179 32 4096 mmcblk1boot0
8 0 2930266582 sda
8 1 2930265088 sda1
root@BPiM3:~# mount /dev/sda1 /mnt/sda1
root@BPiM3:~# df -h
Filesystem Size Used Avail Use% Mounted on
/dev/root 5.0G 3.6G 1.2G 76% /
devtmpfs 751M 0 751M 0% /dev
tmpfs 1007M 0 1007M 0% /dev/shm
tmpfs 1007M 18M 989M 2% /run
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 1007M 0 1007M 0% /sys/fs/cgroup
tmpfs 1007M 4.0K 1007M 1% /tmp
/dev/mmcblk0p1 50M 0 50M 0% /media/boot
tmpfs 202M 0 202M 0% /run/user/108
tmpfs 202M 0 202M 0% /run/user/1000
/dev/sda1 2.7T 201M 2.6T 1% /mnt/sda1
But then I started a simple iozone run to test performance and the whole system stalls. Further investigation needed. The same set of tests with a small 120GB SSD ran flawlessly.
Thank you for the insightful answer! Would it be possible to try your image to see it i can see any SATA disk with my board? I am using a 2TB disk just to play it safe, but it still does not see any device /dev/sda unless i use a USB connected disk and then works very well.
Also as a side note i always use separate power for any HDD (SATA or USB) because will reduce the thermal stress of the board.
Anyhow now i want to know is the image is broken or my SATA bridge is not working.
Vlad
Easy: Your SATA bridge is broken 
This is good news since this is the worst connection method possible (did some more measurements today)
I would like to try the same image you tried before i conclude the HW is defect. How big it is?
I created an image on my own since a few days ago SinoVoip only provided broken OS images (their most recent are also ‘broken’ somewhat since they enabled irqbalanced – you’ll read in a few weeks about this when the forum will be full of stability issues 
)
My image needs an 8 GB SD card since I had to expand the rootfs to be able to test the speed of the SDIO implementation:
I realise this is an old thread. If anyone is still trying to get a stable OS with Sata support working I’ve been using a Raspbian image with a 3TB drive connected without any issues for over 2 years.
I downloaded the Raspbian image from here. install-raspbian-on-banana-pi-m3