failed command: WRITE FPDMA QUEUED / ATA bus error

Bug #1576634 reported by Davide Aguiari
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
elementary OS
Invalid
Undecided
Unassigned

Bug Description

System: HP Omen 15-5003nl
OS: Linux Elementary Freya 0.3.2 64bit (and Windows 10 64bit)
SSD: SanDisk A110 SD6PP4M-256GB-1006
SSD Firmware: A200906
Kernel: 3.19.0-58-generic and 4.4.6-040406-generic and others (see after)

I recently upgraded my SSD firmware and it fixed a serious ahcpi freeze bug on Windows partition (http://h30434.www3.hp.com/t5/Notebook-Boot-and-Lockup/HP-Omen-system-hangs/td-p/4990788). Sometimes the system freezed some seconds with Linux OS too, so I hoped to resolve all problems with this new firmware.
On the contrary, now my dmesg is full of:

5/4/2016 00:40:26 Dago kernel [ 322.113563] ata5.00: exception Emask 0x10 SAct 0x20000000 SErr 0x50000 action 0xe frozen
5/4/2016 00:40:26 Dago kernel [ 322.113567] ata5.00: irq_stat 0x00400000, PHY RDY changed
5/4/2016 00:40:26 Dago kernel [ 322.113569] ata5: SError: { PHYRdyChg CommWake }
5/4/2016 00:40:26 Dago kernel [ 322.113572] ata5.00: failed command: WRITE FPDMA QUEUED
5/4/2016 00:40:26 Dago kernel [ 322.113575] ata5.00: cmd 61/48:e8:00:3e:7e/01:00:16:00:00/40 tag 29 ncq 167936 out
5/4/2016 00:40:26 Dago kernel [ 322.113575] res 40/00:e4:00:f7:32/00:00:19:00:00/40 Emask 0x10 (ATA bus error)
5/4/2016 00:40:26 Dago kernel [ 322.113577] ata5.00: status: { DRDY }
5/4/2016 00:40:26 Dago kernel [ 322.113580] ata5: hard resetting link
5/4/2016 00:40:27 Dago kernel [ 322.838234] ata5: SATA link up 1.5 Gbps (SStatus 113 SControl 310)
5/4/2016 00:40:27 Dago kernel [ 322.839183] ata5.00: configured for UDMA/33
5/4/2016 00:40:27 Dago kernel [ 322.839247] ata5: EH complete

209 errors in 20minutes.
I did a system ssd check and no errors are found. I put my smartctl on attachment.

Speed test:
sudo hdparm -Tt /dev/sda

/dev/sda:
 Timing cached reads: 19958 MB in 2.00 seconds = 9990.47 MB/sec
 Timing buffered disk reads: 1782 MB in 3.00 seconds = 593.55 MB/sec

Dmesg errors show only with AC power plug plugged in :/
I though It was a kernel bug related, but I tried Ubuntu 16.04 and the dmesg is clean. No errors.
I installed the SAME kernel on Elementary (4.4.0-21) but the errors were still there.

I tried several kernels (4.0.5, 4.0.9, 4.1.0, 4.4, 4.5) without success.

I tried to boot with libata.force=noncq, without success. In this case, the error changed to
[ ] ata5.00: exception Emask 0x10 SAct 0x0 SErr 0x50000 action 0xe frozen [ ... ]
[ ] ata5.00: irq_stat 0x00400000, PHY RDY changed
[ ] ata5.00: failed command: WRITE DMA EXT
[ ] ata5.00: cmd 35/00:90:f0:ec:b2/00:00:17:00:00/e0 tag 26 dma 73728 out
[ ] res 50/00:00:00:00:00/00:00:00:00:00/a0 Emask 0x10 (ATA bus error)

Revision history for this message
Davide Aguiari (gorghino) wrote :
Revision history for this message
Davide Aguiari (gorghino) wrote :

I installed Loki beta today in a different partition.
The errors are missing. I'm starting to think this bug is correlated to a sector corrupted.

Changed in elementaryos:
status: New → Invalid
Revision history for this message
Davide Aguiari (gorghino) wrote :

SOLUTION: Apparently the new SSD firmware doesn't work well with TLP (http://linrunner.de/en/tlp/docs/tlp-linux-advanced-power-management.html) package.

Open /etc/default/tlp and commented

#SATA_LINKPWR_ON_AC=max_performance
#SATA_LINKPWR_ON_BAT=min_power

solved the problem.

Maybe you (Elementary devs) should warn people when they install TLP package.

Changed in elementaryos:
status: Invalid → Confirmed
Revision history for this message
Zisu Andrei (matzipan) wrote :

Hello,

I am glad to hear you found a solution.

I'm afraid there isn't much we can do about either the bug - since we don't manage or maintain tlp - or warning people about this.

Consider opening a bug request upstream: https://github.com/linrunner/TLP/issues

Changed in elementaryos:
status: Confirmed → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Bug attachments

Remote bug watches

Bug watches keep track of this bug in other bug trackers.