Package: linux-image-4.8.0-1-marvell
Version: 4.8.7-1
Severity: normal
Tags: upstream
I found that on recent d-i images, the kernel does not detect the hard
drive on my LinkStation. I tried some older images and found that kernel
4.3.3-5 worked and kernel 4.6.2-2 did not.
I suspect the problem is with the device tree file, therefore CCing
Roger Shimizu who contributed that file.
The old kurobox_pro-setup.c contains this code:
static struct mv_sata_platform_data kurobox_pro_sata_data = {
.n_ports = 2,
};
whereas the new orion5x-linkstation.dtsi contains this code:
&sata {
status = "okay";
nr-ports = <1>;
};
The dmesg output seems to reinforce this conclusion.
dmesg from 4.3.0-1:
[ 611.914912] sata_mv sata_mv.0: version 1.28
[ 611.915107] sata_mv sata_mv.0: cannot get optional clkdev
[ 611.916086] sata_mv sata_mv.0: slots 32 ports 2
[ 611.971107] scsi host0: sata_mv
[ 611.984111] scsi host1: sata_mv
[ 611.984881] ata1: SATA max UDMA/133 irq 30
[ 611.984916] ata2: SATA max UDMA/133 irq 30
[ 612.302763] ata1: SATA link down (SStatus 0 SControl 300)
[ 612.778697] ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[ 612.787074] ata2.00: ATA-7: SAMSUNG SP2504C, VT100-41, max UDMA7
[ 612.787118] ata2.00: 488397168 sectors, multi 0: LBA48 NCQ (depth 31/32)
[ 612.819082] ata2.00: configured for UDMA/133
[ 612.820339] scsi 1:0:0:0: Direct-Access ATA SAMSUNG SP2504C 0-41
PQ: 0 ANSI: 5
[ 612.899766] sd 1:0:0:0: [sda] 488397168 512-byte logical blocks: (250 GB/232
GiB)
[ 612.900971] sd 1:0:0:0: [sda] Write Protect is off
[ 612.901031] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
[ 612.901533] sd 1:0:0:0: [sda] Write cache: enabled, read cache: enabled,
doesn't support DPO or FUA
[ 612.924886] sd 1:0:0:0: [sda] Attached SCSI disk
dmesg from 4.8.7-1:
[ 11.140091] sata_mv sata_mv.0: version 1.28
[ 11.140252] sata_mv sata_mv.0: cannot get optional clkdev
[ 11.170717] sata_mv sata_mv.0: slots 32 ports 1
[ 11.226581] scsi host0: sata_mv
[ 11.236985] ata1: SATA max UDMA/133 irq 28
[ 11.554565] ata1: SATA link down (SStatus 0 SControl 300)
In this extract, ata2 is apparently never scanned.
I expect changing nr-ports to 2 will fix this. I have not tested the
change yet, but will try to do so this weekend.
thanks
Ryan
