SSD connecting at SATA 3Gbps instead of 6Gbps when hot-swapped

[jwilliams]

I'm having a problem that I am not sure how to even begin to troubleshoot. I want to determine which component is most at fault and/or easiest to fix.

I have a couple Plextor SSDs, model M3P and M5P, and they both exhibit the same issue. When I hot-swap them on my linux box, they are somehow limited to SATA 3Gbps speed. I determined this by doing a sequential read using fio, and only getting 250MB/s.

If I boot with either of the Plextor SSDs attached (instead of hot-swapping), then they do work at SATA 6Gbps speeds -- sequential reads of over 400MB/s. But if I then hot-swap the SSD out and back in, it reverts to SATA 3Gbps speeds.

I have a Windows 7 machine, Intel DH67GD motherboard with two Intel SATA 6Gbps ports, that can hot-swap the same Plextor SSDs and they work at SATA 6Gbps speeds.

Even more perplexing is that I have two other SSDs, a Samsung 830 and a Samsung 840 Pro, and both of those are able to hot-swap and work at SATA 6Gbps speeds on my linux box.

So I am not sure whether to point the finger at the linux SATA driver, at the Supermicro X9SCM-iiF motherboard (tried both BIOS 2.0 and 2.0a) that is in my linux box, or at the Plextor SSDs themselves. If I blame the Plextor SSDs, then it is tough to explain why they hot-swap fine on my Windows box, but if I blame the linux driver or SM motherboard, then it is tough to explain why the Samsung SSDs hot-swap fine on the linux box.

By the way, I am connecting the Plextor SSDs to one of the two SATA 6Gbps ports on the X9SCM-iiF motherboard. So it is not a matter of me using a 3Gbps SATA port (and besides, it works at 6Gbps on the same port if I boot with the SSD attached). Also, in the BIOS configuration, all of the SATA ports have hot-swap enabled.

Any suggestions on how to troubleshoot this? What linux commands might I try to investigate further?

[jwilliams]

Here is an example of the kernel log when I hotswap various SSDs. Note that the Plextor M5P connects incorrectly at 3.0 Gbps after the hard reset, but the Samsung 840 Pro connects at 6.0 Gbps after the hard reset.

Started with Samsung 830 already connected:
1) removed 830
2) attached M5P
3) removed M5P
4) attached 840 Pro

Nov 18 09:25:13 localhost kernel: [  411.426008] ACPI: Invalid Power Resource to register!<5>[  411.426139] sd 1:0:0:0: [sda] Write Protect is off
Nov 18 09:25:13 localhost kernel: [  411.426145] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
Nov 18 09:25:13 localhost kernel: [  411.426241] sd 1:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Nov 18 09:25:13 localhost kernel: [  412.547220]  sda: sda1
Nov 18 09:25:13 localhost kernel: [  412.548005] sd 1:0:0:0: [sda] Attached SCSI disk
Nov 18 09:25:13 localhost kernel: [  483.938646] fuse init (API version 7.20)
Nov 18 09:25:13 localhost kernel: [  484.159201] EXT4-fs (sdc3): re-mounted. Opts: commit=0
Nov 18 09:25:13 localhost kernel: [  484.163008] EXT4-fs (sdc4): re-mounted. Opts: commit=0
Nov 18 09:25:13 localhost kernel: [  484.163829] EXT4-fs (sdb1): re-mounted. Opts: discard,commit=0
Nov 18 09:25:13 localhost kernel: [58282.411803] Thunar[3514]: segfault at 0 ip 0000000000432132 sp 00007fff121de000 error 4 in thunar[400000+9b000]
Nov 18 09:32:52 localhost kernel: [122030.057934] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4090000 action 0xe frozen
Nov 18 09:32:52 localhost kernel: [122030.058377] ata1: irq_stat 0x00400040, connection status changed
Nov 18 09:32:52 localhost kernel: [122030.058731] ata1: SError: { PHYRdyChg 10B8B DevExch }
Nov 18 09:32:52 localhost kernel: [122030.059030] ata1: hard resetting link
Nov 18 09:32:52 localhost kernel: [122030.775371] ata1: SATA link down (SStatus 0 SControl 300)
Nov 18 09:32:57 localhost kernel: [122035.765628] ata1: hard resetting link
Nov 18 09:32:58 localhost kernel: [122036.091664] ata1: SATA link down (SStatus 0 SControl 300)
Nov 18 09:32:58 localhost kernel: [122036.091675] ata1: limiting SATA link speed to 1.5 Gbps
Nov 18 09:33:03 localhost kernel: [122041.081918] ata1: hard resetting link
Nov 18 09:33:03 localhost kernel: [122041.407959] ata1: SATA link down (SStatus 0 SControl 310)
Nov 18 09:33:03 localhost kernel: [122041.407968] ata1.00: disabled
Nov 18 09:33:03 localhost kernel: [122041.414579] ata1: EH complete
Nov 18 09:33:03 localhost kernel: [122041.414690] ata1.00: detaching (SCSI 1:0:0:0)
Nov 18 09:33:03 localhost kernel: [122041.415225] sd 1:0:0:0: [sda] Synchronizing SCSI cache
Nov 18 09:33:03 localhost kernel: [122041.415263] sd 1:0:0:0: [sda]  
Nov 18 09:33:03 localhost kernel: [122041.415266] Result: hostbyte=0x04 driverbyte=0x00
Nov 18 09:33:03 localhost kernel: [122041.415268] sd 1:0:0:0: [sda] Stopping disk
Nov 18 09:33:03 localhost kernel: [122041.415277] sd 1:0:0:0: [sda] START_STOP FAILED
Nov 18 09:33:03 localhost kernel: [122041.415279] sd 1:0:0:0: [sda]  
Nov 18 09:33:03 localhost kernel: [122041.415280] Result: hostbyte=0x04 driverbyte=0x00
Nov 18 09:33:06 localhost kernel: [122044.939330] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
Nov 18 09:33:06 localhost kernel: [122044.939799] ata1: irq_stat 0x00000040, connection status changed
Nov 18 09:33:06 localhost kernel: [122044.940159] ata1: SError: { CommWake DevExch }
Nov 18 09:33:06 localhost kernel: [122044.940428] ata1: hard resetting link
Nov 18 09:33:07 localhost kernel: [122045.656337] ata1: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
Nov 18 09:33:07 localhost kernel: [122045.656857] ata1.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Nov 18 09:33:07 localhost kernel: [122045.656862] ata1.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Nov 18 09:33:07 localhost kernel: [122045.656865] ata1.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out
Nov 18 09:33:07 localhost kernel: [122045.657068] ata1.00: ATA-8: PLEXTOR PX-128M5P, 1.01, max UDMA/133
Nov 18 09:33:07 localhost kernel: [122045.657072] ata1.00: 250069680 sectors, multi 1: LBA48 NCQ (depth 31/32), AA
Nov 18 09:33:07 localhost kernel: [122045.657590] ata1.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Nov 18 09:33:07 localhost kernel: [122045.657595] ata1.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Nov 18 09:33:07 localhost kernel: [122045.657598] ata1.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out
Nov 18 09:33:07 localhost kernel: [122045.657835] ata1.00: configured for UDMA/133
Nov 18 09:33:07 localhost kernel: [122045.657910] ata1: EH complete
Nov 18 09:33:07 localhost kernel: [122045.658029] scsi 1:0:0:0: Direct-Access     ATA      PLEXTOR PX-128M5 1.01 PQ: 0 ANSI: 5
Nov 18 09:33:07 localhost kernel: [122045.658385] sd 1:0:0:0: [sda] 250069680 512-byte logical blocks: (128 GB/119 GiB)
Nov 18 09:33:07 localhost kernel: [122045.658390] sd 1:0:0:0: [sda] 4096-byte physical blocks
Nov 18 09:33:07 localhost kernel: [122045.658755] sd 1:0:0:0: [sda] Write Protect is off
Nov 18 09:33:07 localhost kernel: [122045.658760] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
Nov 18 09:33:07 localhost kernel: [122045.658878] sd 1:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Nov 18 09:33:07 localhost kernel: [122045.659549]  sda: sda1
Nov 18 09:33:07 localhost kernel: [122045.660302] sd 1:0:0:0: [sda] Attached SCSI disk
Nov 18 09:36:05 localhost kernel: [122223.077933] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4090000 action 0xe frozen
Nov 18 09:36:05 localhost kernel: [122223.078371] ata1: irq_stat 0x00400040, connection status changed
Nov 18 09:36:05 localhost kernel: [122223.078724] ata1: SError: { PHYRdyChg 10B8B DevExch }
Nov 18 09:36:05 localhost kernel: [122223.079022] ata1: hard resetting link
Nov 18 09:36:06 localhost kernel: [122223.795454] ata1: SATA link down (SStatus 0 SControl 300)
Nov 18 09:36:11 localhost kernel: [122228.785713] ata1: hard resetting link
Nov 18 09:36:11 localhost kernel: [122229.111748] ata1: SATA link down (SStatus 0 SControl 300)
Nov 18 09:36:11 localhost kernel: [122229.111759] ata1: limiting SATA link speed to 1.5 Gbps
Nov 18 09:36:16 localhost kernel: [122234.102004] ata1: hard resetting link
Nov 18 09:36:16 localhost kernel: [122234.428043] ata1: SATA link down (SStatus 0 SControl 310)
Nov 18 09:36:16 localhost kernel: [122234.428053] ata1.00: disabled
Nov 18 09:36:16 localhost kernel: [122234.434662] ata1: EH complete
Nov 18 09:36:16 localhost kernel: [122045.658513] ACPI: Invalid Power Resource to register!
Nov 18 09:36:16 localhost kernel: [122234.434775] ata1.00: detaching (SCSI 1:0:0:0)
Nov 18 09:36:16 localhost kernel: [122234.435250] sd 1:0:0:0: [sda] Synchronizing SCSI cache
Nov 18 09:36:16 localhost kernel: [122234.435292] sd 1:0:0:0: [sda]  
Nov 18 09:36:16 localhost kernel: [122234.435295] Result: hostbyte=0x04 driverbyte=0x00
Nov 18 09:36:16 localhost kernel: [122234.435297] sd 1:0:0:0: [sda] Stopping disk
Nov 18 09:36:16 localhost kernel: [122234.435307] sd 1:0:0:0: [sda] START_STOP FAILED
Nov 18 09:36:16 localhost kernel: [122234.435309] sd 1:0:0:0: [sda]  
Nov 18 09:36:16 localhost kernel: [122234.435311] Result: hostbyte=0x04 driverbyte=0x00
Nov 18 09:36:28 localhost kernel: [122246.219908] ata1: exception Emask 0x10 SAct 0x0 SErr 0x4040000 action 0xe frozen
Nov 18 09:36:28 localhost kernel: [122246.220346] ata1: irq_stat 0x00000040, connection status changed
Nov 18 09:36:28 localhost kernel: [122246.220699] ata1: SError: { CommWake DevExch }
Nov 18 09:36:28 localhost kernel: [122246.220961] ata1: hard resetting link
Nov 18 09:36:34 localhost kernel: [122251.674400] ata1: SATA link up 6.0 Gbps (SStatus 133 SControl 300)
Nov 18 09:36:34 localhost kernel: [122251.675341] ata1.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Nov 18 09:36:34 localhost kernel: [122251.675346] ata1.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Nov 18 09:36:34 localhost kernel: [122251.675349] ata1.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out
Nov 18 09:36:34 localhost kernel: [122251.675591] ata1.00: ATA-9: Samsung SSD 840 PRO Series, DXM03B0Q, max UDMA/133
Nov 18 09:36:34 localhost kernel: [122251.675596] ata1.00: 250069680 sectors, multi 16: LBA48 NCQ (depth 31/32), AA
Nov 18 09:36:34 localhost kernel: [122251.676154] ata1.00: ACPI cmd ef/10:06:00:00:00:00 (SET FEATURES) succeeded
Nov 18 09:36:34 localhost kernel: [122251.676159] ata1.00: ACPI cmd f5/00:00:00:00:00:00 (SECURITY FREEZE LOCK) filtered out
Nov 18 09:36:34 localhost kernel: [122251.676162] ata1.00: ACPI cmd b1/c1:00:00:00:00:00 (DEVICE CONFIGURATION OVERLAY) filtered out
Nov 18 09:36:34 localhost kernel: [122251.676400] ata1.00: configured for UDMA/133
Nov 18 09:36:34 localhost kernel: [122251.676457] ata1: EH complete
Nov 18 09:36:34 localhost kernel: [122251.676571] scsi 1:0:0:0: Direct-Access     ATA      Samsung SSD 840  DXM0 PQ: 0 ANSI: 5
Nov 18 09:36:34 localhost kernel: [122251.676899] sd 1:0:0:0: [sda] 250069680 512-byte logical blocks: (128 GB/119 GiB)
Nov 18 09:36:34 localhost kernel: [122251.677224] sd 1:0:0:0: [sda] Write Protect is off
Nov 18 09:36:34 localhost kernel: [122251.677229] sd 1:0:0:0: [sda] Mode Sense: 00 3a 00 00
Nov 18 09:36:34 localhost kernel: [122251.677337] sd 1:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA
Nov 18 09:36:34 localhost kernel: [122251.678096]  sda: unknown partition table
Nov 18 09:36:34 localhost kernel: [122251.678694] sd 1:0:0:0: [sda] Attached SCSI disk