READ/WRITE DMA Errors with SSD. HDDs are working fine

5

1

I have assembled a custom PC, and installed Linux on it. I have tried using several different SSDs of the same kind, and I keep getting the following types of random Read/Write errors:

[  310.697924] ata1: lost interrupt (Status 0x50)
[  310.697949] ata1.00: exception Emask 0x50 SAct 0x0 SErr 0x40d0802 action 0xe frozen
[  310.698036] ata1.00: SError: { RecovComm HostInt PHYRdyChg CommWake 10B8B DevExch }
[  310.698119] ata1.00: failed command: WRITE DMA
[  310.698185] ata1.00: cmd ca/00:20:18:08:8a/00:00:00:00:00/e0 tag 0 dma 16384 out
[  310.698187]          res 40/00:00:00:00:00/00:00:00:00:00/40 Emask 0x54 (ATA bus error)
[  310.698348] ata1.00: status: { DRDY }
[  310.698415] ata1.00: hard resetting link
[  311.421105] ata1.01: hard resetting link
[  311.896619] ata1.00: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  311.896634] ata1.01: SATA link down (SStatus 0 SControl 300)
[  311.936849] ata1.00: configured for UDMA/133
[  311.936858] ata1.00: device reported invalid CHS sector 0
[  311.936866] ata1: EH complete

The same SSDs work fine on another computer. And I tried using some HDDs with the computer in question, and they work fine too. So what could be wrong? Is it the sata cables? Does sata cable quality really matter?

Also, one of my SSDs (which was brand new) was not even being detected. I tried using a different cable, and the system detects it now, but still throws these errors randomly. How can I troubleshoot my computer?

Update:

This motherboard: http://www.supermicro.com/xeon_3400/Motherboard/X8SIL.cfm.

This SSD: http://www.amazon.com/Octane-OCT1-25SAT2-64G-Solid-State-Drive/dp/tech-data/B006C1IIB4

Richard Whitman

Posted 2012-07-25T11:19:39.533

Reputation: 116

Facing same issue and errors on Linux 3.11.0-24-generic, DeLock mSATA<->mSATA converter, Transcend TS64GMSA630 SSD and MSI X58M SLI (MS-7593).

Only difference - lines defining port: [710919.266900] ata1.00: cmd ca/00:08:00:9d:89/00:00:00:00:00/e4 tag 28 dma 4096 out [710919.266900] res 40/00:01:00:00:00/00:00:00:00:00/00 Emask 0x14 (ATA bus error)

– Kamiccolo – 2014-07-29T16:23:07.647

Quick'n'dirty solution for Windows: use USB <-> SATA bridge. USB <-> SATA <-> ePCIm <-> SSD to be exact. Even though it does not work for Linux: same errors thrown. And performance degrades tremendously. – Kamiccolo – 2014-07-31T19:33:23.797

@Kamiccolo can you please post a link to a copy of your full dmesg output? Full,please,no snippets. – MariusMatutiae – 2014-08-03T06:26:52.923

1I would suspect the cables, definitely. – David Schwartz – 2012-07-25T12:53:23.400

What kind of cable would you suggest? – Richard Whitman – 2012-07-25T13:03:08.727

One designed for the highest speed your motherboard and SSD support. – David Schwartz – 2012-07-25T13:04:07.427

The ones I am using are. They are Sata 3 6GBPS. I have used the OkGear gold plated ones as well, with same results. – Richard Whitman – 2012-07-25T13:08:16.930

I pulled out cables from the computer where these SSDs work, and tried using them on the new computer. Still the same errors. Doesn't make any sense to me. – Richard Whitman – 2012-07-25T13:09:29.907

1The common link seems to be the same motherboard in combination with high performance on the SATA bus. Could you post which motherboard (and which chipset) you are using? Maybe there are known problems with it. – Hennes – 2012-07-25T13:32:52.757

updated my question above. – Richard Whitman – 2012-07-25T21:00:21.400

Answers

5

Your motherboard chipset is an Intel 3420, which despite being circa 2009, supports AHCI. I'm not very familiar with Linux, but the logs you posted seem to indicate you are using the legacy IDE or ATA mode.

In your BIOS configuration, change the drive controller mode to AHCI. You might have to reinstall your operating system after you do this.

Jason

Posted 2012-07-25T11:19:39.533

Reputation: 5 925

I'd love to know why this was marked down. It's a SATA drive on a SATA controller and the log shows "configured for UDMA/133". – Jason – 2014-07-31T22:20:23.550

I agree with you. +1 from me. In any case a downvote ought to be explained. – MariusMatutiae – 2014-08-01T06:37:04.047

I am facing this problem currently on a server with a SuperMicro board (hello China, are you listening?) circa 2009. My SATA Mode in the BIOS has been set to AHCI Mode the whole time, so it does not appear that this is a valid solution – Ode – 2019-09-26T18:35:19.273

@Ode You're saying you're using SATA drive and the logs are saying "configured for UDMA/133"? UDMA isn't an option in AHCI mode, so I can't explain why your OS would be trying to use UDMA. Keep in mind, some motherboards (not the X8SIL) have a second drive controller with a separate BIOS and settings. – Jason – 2019-09-26T19:02:57.617