SATA and kernel 3.4.75

[oldpenguin]

Hello all

I am testing a SATA hard disk with Olimex A10-LIME board.
It works correctly, but if I disable the sunxi_nand driver in the kernel configuration, SATA stops working with error:

[ 66.033059] ata1: SATA max UDMA/133 mmio [mem 0x01c18000-0x01c18fff] port 0x100 irq 56
[ 66.393493] ata1: SATA link down (SStatus 0 SControl 300)

What is the dependency on sunxi_nand ?
What if I want to use mtd + SATA ?

Best regards

[oliv3r]

mtd requires mainline kernel, so that's a non-issue
As for nand +sata requirements, that's strange, maybe something else got disabled aswel?

[oldpenguin]

Recompiled with SUNXI_NAND as a module.
Tried loading sw_ahci_platform first and nand after This is what I get:

root@debian:# ll /dev/s*
shm/ snd/ stderr stdin stdout sunxi-lcd
root@debian:# modprobe sw_ahci_platform
<6>sw_ahci sw_ahci.0: controller can't do PMP, turning off CAP_PMP
[ 84.752700] sw_ahci sw_ahci.0: controller can't do PMP, turning off CAP_PMP
<4>sw_ahci sw_ahci.0: forcing PORTS_IMPL to 0x1
[ 84.778096] sw_ahci sw_ahci.0: forcing PORTS_IMPL to 0x1
<6>sw_ahci sw_ahci.0: AHCI 0001.0100 32 slots 1 ports 3 Gbps 0x1 impl platform mode
[ 84.791777] sw_ahci sw_ahci.0: AHCI 0001.0100 32 slots 1 ports 3 Gbps 0x1 impl platform mode
<6>sw_ahci sw_ahci.0: flags: ncq sntf pm led clo only pio slum part ccc
[ 84.807280] sw_ahci sw_ahci.0: flags: ncq sntf pm led clo only pio slum part ccc
<6>scsi0 : sw_ahci_platform
[ 84.825220] scsi0 : sw_ahci_platform
<6>ata1: SATA max UDMA/133 mmio [mem 0x01c18000-0x01c18fff] port 0x100 irq 56
[ 84.843344] ata1: SATA max UDMA/133 mmio [mem 0x01c18000-0x01c18fff] port 0x100 irq 56
<6>ata1: SATA link down (SStatus 0 SControl 300)
[ 85.201778] ata1: SATA link down (SStatus 0 SControl 300)
root@debian:#
root@debian:# modprobe nand
[NAND] nand driver version: 0x2 0x9
[ 99.327347] [NAND] nand driver version: 0x2 0x9
<3>ata1: exception Emask 0x10 SAct 0x0 SErr 0x4050002 action 0xe frozen
[ 99.578311] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4050002 action 0xe frozen
<3>ata1: irq_stat 0x00000040, connection status changed
[ 99.591372] ata1: irq_stat 0x00000040, connection status changed
<3>ata1: SError: { RecovComm PHYRdyChg CommWake DevExch }
[ 99.603105] ata1: SError: { RecovComm PHYRdyChg CommWake DevExch }
<6>ata1: hard resetting link
[ 99.612700] ata1: hard resetting link
<6>ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 101.393218] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
<6>ata1.00: ATA-9: ADATA SP600, 2.4, max UDMA/100
[ 101.424164] ata1.00: ATA-9: ADATA SP600, 2.4, max UDMA/100
<6>ata1.00: 62533296 sectors, multi 16: LBA48 NCQ (depth 31/32)
[ 101.453733] ata1.00: 62533296 sectors, multi 16: LBA48 NCQ (depth 31/32)
<6>ata1.00: configured for UDMA/100
[ 101.471261] ata1.00: configured for UDMA/100
<6>ata1: EH complete
[ 101.477672] ata1: EH complete
<5>scsi 0:0:0:0: Direct-Access ATA ADATA SP600 2.4 PQ: 0 ANSI: 5
[ 101.494090] scsi 0:0:0:0: Direct-Access ATA ADATA SP600 2.4 PQ: 0 ANSI: 5
[ 101.533464] sd 0:0:0:0: [sda] 62533296 512-byte logical blocks: (32.0 GB/29.8 GiB)
<5>sd 0:0:0:0: [sda] Write Protect is off
[ 101.584005] sd 0:0:0:0: [sda] Write Protect is off
<7>sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 101.612368] sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
<5>sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 101.635045] sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
[ 101.673434] sda: sda1
<5>sd 0:0:0:0: [sda] Attached SCSI disk
[ 101.697178] sd 0:0:0:0: [sda] Attached SCSI disk
root@debian:# ll /dev/sda*
brw-rw---T 1 root disk 8, 0 Feb 11 18:29 �[0m�[40;33;01m/dev/sda�[0m
brw-rw---T 1 root disk 8, 1 Feb 11 18:29 �[40;33;01m/dev/sda1�[0m
root@debian:#

After loading nand, SATA link is restarted and hard disk is properly recognized.
Any ideas ?

[amery]

can you paste somewhere (sprunge.us?) the deserialized fex of the script.bin you are using?

[oldpenguin]

http://sprunge.us/QXPa

Thanks for looking into this.

[oliv3r]

I do have to ask you to upgrade to the latest 3.4 series as 3.4.75 is pretty ancient!

[oldpenguin]

Updated to 3.4.79. Same happens.
I believe other people on Olimex forum have also verified this behavior.
My config for 3.479 is here http://sprunge.us/BhXh

[amery]

the sata_para section of your .fex doesn't include any of the needed pinmixes

[oldpenguin]

My script.fex is based on Olimex one.
Documentation on http://linux-sunxi.org/Fex_Guide#.5Bsata_para.5D only mentions:

[sata_para]
sata_used = 1
sata_power_en = port:PB08< 1 >< default >< default >< 0 >

Can you tell me what else is missing ?