| View previous topic :: View next topic |
| Author |
Message |
BlueFusion
Guru
Joined: 08 Mar 2006
Posts: 371
|
 Posted: Tue Aug 26, 2014 3:58 pm Post subject: Marvel SAS issues |
 |
|
I had to start using my ASUS P6T Deluxe's onboard Marvel SAS controller because I had added a few more drives to my PC. I am recieving a bunch of errors when the drives are accessed (see below), and drive performance has suffered. Other instabilities exist - primarily, the PC will no longer wake up from suspend-to-RAM. A Google search doesn't really steer me towards any useful information on how to fix this. Does anybody have any ideas?
I should also mention that the 2 drives connected are both SATA II drives.
The Marvel controller used:
| Quote: | | 05:00.0 RAID bus controller: Marvell Technology Group Ltd. 88SE6440 SAS/SATA PCIe controller (rev 02) |
The dmesg errors received:
| Quote: | [ 45.954522] sd 0:0:0:0: [sda] command ffff8801b527b800 timed out
[ 45.963845] sas: Enter sas_scsi_recover_host busy: 1 failed: 1
[ 45.963850] sas: ata1: end_device-0:0: cmd error handler
[ 45.963925] sas: ata1: end_device-0:0: dev error handler
[ 45.963944] sas: ata2: end_device-0:1: dev error handler
[ 45.963956] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 45.964681] sd 0:0:0:0: [sda] command ffff8801b650bb00 timed out
[ 45.973889] sas: Enter sas_scsi_recover_host busy: 1 failed: 1
[ 45.973893] sas: ata1: end_device-0:0: cmd error handler
[ 45.973935] sas: ata1: end_device-0:0: dev error handler
[ 45.973953] sas: ata2: end_device-0:1: dev error handler
[ 45.973965] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.120876] sd 0:0:0:0: [sda] command ffff8801b527b800 timed out
[ 46.130757] sas: Enter sas_scsi_recover_host busy: 1 failed: 1
[ 46.130762] sas: ata1: end_device-0:0: cmd error handler
[ 46.130835] sas: ata1: end_device-0:0: dev error handler
[ 46.130856] sas: ata2: end_device-0:1: dev error handler
[ 46.130905] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.131114] sas: Enter sas_scsi_recover_host busy: 0 failed: 0
[ 46.131200] sas: ata1: end_device-0:0: dev error handler
[ 46.131225] sas: ata2: end_device-0:1: dev error handler
[ 46.228963] ata1.00: configured for UDMA/133
[ 46.228969] ata1: EH complete
[ 46.229012] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.297327] sd 0:0:1:0: [sdb] command ffff8800b9359980 timed out
[ 46.335357] sas: Enter sas_scsi_recover_host busy: 1 failed: 1
[ 46.335361] sas: ata2: end_device-0:1: cmd error handler
[ 46.335412] sas: ata1: end_device-0:0: dev error handler
[ 46.335413] sas: ata2: end_device-0:1: dev error handler
[ 46.335431] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.336725] sd 0:0:1:0: [sdb] command ffff8801b650ea80 timed out
[ 46.347505] sas: Enter sas_scsi_recover_host busy: 1 failed: 1
[ 46.347510] sas: ata2: end_device-0:1: cmd error handler
[ 46.347520] sas: ata1: end_device-0:0: dev error handler
[ 46.347526] sas: ata2: end_device-0:1: dev error handler
[ 46.347546] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.348568] sas: Enter sas_scsi_recover_host busy: 0 failed: 0
[ 46.348605] sas: ata1: end_device-0:0: dev error handler
[ 46.348607] sas: ata2: end_device-0:1: dev error handler
[ 46.351951] ata2.00: configured for UDMA/133
[ 46.351954] ata2: EH complete
[ 46.351962] sas: --- Exit sas_scsi_recover_host: busy: 0 failed: 0 tries: 1
[ 46.384466] ata4.00: configured for UDMA/133
[ 46.384470] ata4: EH complete
[ 46.408911] ata5.00: supports DRM functions and may not be fully accessible
[ 46.408946] ata5.00: failed to get NCQ Send/Recv Log Emask 0x1
[ 46.409085] ata5.00: supports DRM functions and may not be fully accessible
[ 46.409115] ata5.00: failed to get NCQ Send/Recv Log Emask 0x1 |
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1) |
|
| Back to top |
|
 |
BobWya
Apprentice
Joined: 12 Aug 2012
Posts: 228
Location: Cambridge,UK
|
| Posted: Wed Sep 10, 2014 9:14 pm Post subject: Re: Marvel SAS issues |
|
|
| BlueFusion wrote: | I had to start using my ASUS P6T Deluxe's onboard Marvel SAS controller because I had added a few more drives to my PC. I am recieving a bunch of errors when the drives are accessed (see below), and drive performance has suffered. Other instabilities exist - primarily, the PC will no longer wake up from suspend-to-RAM. A Google search doesn't really steer me towards any useful information on how to fix this. Does anybody have any ideas?
I should also mention that the 2 drives connected are both SATA II drives.
The Marvel controller used:
| Quote: | | 05:00.0 RAID bus controller: Marvell Technology Group Ltd. 88SE6440 SAS/SATA PCIe controller (rev 02) |
The dmesg errors received:
... |
Hi BlueFusion,
I'm just bumping this thread... I've similar issues with the ASUS P6X58D-E - with a very similar hardware setup. I've got an Addonics (Marvel based) 8x PCIe SAS controller added to my board. I've around 6 drives hanging off this controller - all SATA-2 drives. The system has 24Gb of RAM (well heck I need it for Gentoo upgrades  ).
I recently stuck in an Intel(R) Xeon(R) CPU X5650 (6 core, 32nm refresh of the LGA1366 series). Btw a really sweet upgrade for any LGA1366 board (since I only paid 50 GBP for a refurb) - has a lower power envelope than the i7 920 it replaced - great for recompiling your system. 
So anyway (stop rambling!!) I thought lets revisit suspend to RAM (which previously I didn't get very far with) - since electricity is expensive and the system takes about 2-3 minutes to boot
So anyway it's the usual sucky story... From my multi-boot setup, various Linux distros (OpenSUSE, Arch & Gentoo I have tested so far) and MS Windows 7 Professional... Only Windows 7 Professional can suspend to RAM and resume without a kernel panic/lockup.
Please note I've only just started testing this feature this week. On my board/system I can strip out all the add in cards and disable the built-in, add-on Marvel SATA-6G controller. Don't expect an immediate response to testing this - as I've only just rebuilt the system with the new CPU!
Also I hadn't fully ruled out an Nvidia driver bug... But perhaps if your system could previously suspend - this is not an issue?
Do you use OpenRC or systemd as init? I'd like to duplicate your setup - since you say yours previously worked (without the Marvel builtin card being enabled). With a clean install of Gentoo on a single drive in a stripped down system.
For troubleshooting I've yet to boot straight to console mode to test resume functionality... I've tried a suspend-resume cycle in an X/KDE session TTY and I get some random errors. I'm not shit hot on kernel debugging - by any stretch of the imagination ... So any advice would be gratefully received!
I'll stick my system specs in my signature.
All the best,
Robert
_________________
system: G751JT (ASUS-NotebookSKU); processor: Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz; memory: 32GiB System Memory; display: GM204M [GeForce GTX 970M]; disk: 2048GB Samsung SSD 850;BD-CMB UJ172 S;1024GB Samsung SSD 850 |
|
| Back to top |
|
|
BobWya
Apprentice
Joined: 12 Aug 2012
Posts: 228
Location: Cambridge,UK
|
| Posted: Tue Sep 16, 2014 11:58 pm Post subject: Re: Marvel SAS issues |
|
|
| BlueFusion wrote: | I had to start using my ASUS P6T Deluxe's onboard Marvel SAS controller because I had added a few more drives to my PC. I am recieving a bunch of errors when the drives are accessed (see below), and drive performance has suffered. Other instabilities exist - primarily, the PC will no longer wake up from suspend-to-RAM. A Google search doesn't really steer me towards any useful information on how to fix this. Does anybody have any ideas?
...
|
@BlueFusion,
So a quick update (1). I've ripped out my Marvell SAS Controller and put my 6 drives back on the Intel Host Controller (onboard). So funny thing - not only is my boot time decimated (surprise, surprise) - but I have full suspend/resume support back. I'll test with the controller back in (and no drives), etc., etc. Starting to look like it is a kernel/driver bug. The issue definitely isn't a Gentoo specific one (as it also affects Magi ea, OpenSUSE, Ubuntu, Arch) and included kernels 3.12-3.16.
Anyway I've got to thank you for punting me down this direction! Never would have thought of it...
Heck I was even contemplating going back to Windows 7 Professional full-time on the Desktop if I couldn't get the issue resolved...  It's a big deal to have 8800 GTX running all the time (heck it only has one power/clock profile - and that one is big!!) - it just sucks so much juice. 
Sorry I haven't anything more constructive to add (yet!!)
Thanks,
Robert
_________________
system: G751JT (ASUS-NotebookSKU); processor: Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz; memory: 32GiB System Memory; display: GM204M [GeForce GTX 970M]; disk: 2048GB Samsung SSD 850;BD-CMB UJ172 S;1024GB Samsung SSD 850 |
|
| Back to top |
|
|
BlueFusion
Guru
Joined: 08 Mar 2006
Posts: 371
|
| Posted: Mon Oct 20, 2014 4:14 pm Post subject: |
|
|
Hey Robert, sorry for the huuuuge delay - been busy with work.
The problem has been bandaid-ed for now. I bought a 4TB HD to replace a few of the drives in the system. I can now operate with all of my Linux harddrives on the ICH10 controller. Only my Windows Vista drive is connected to the onboard Marvel SAS controller.
I removed the mvsas module from my kernel so it doesn't have any impact on Linux environment. The system has been working fine ever since.
Looks like mvsas driver is just not stable. I've scoured Google and all I see is one complaint after another. Oh well.... Another excuse to build a new computer!
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1) |
|
| Back to top |
|
|
BobWya
Apprentice
Joined: 12 Aug 2012
Posts: 228
Location: Cambridge,UK
|
| Posted: Mon Oct 20, 2014 6:42 pm Post subject: |
|
|
@BlueFusion,
Well heck if there's nothing to update - there's nothing to update
Sorry I've just invested in a Oneplus One (very nice phone - but a little large) - so that's hit my financial liquidity somewhat...  I'm still planning to get the LSI HBA - but it's definitely pushed back a month or two now... It's not really a big deal for me - I doubt the Samsung 830 drives I'm using will be too bottlenecked by the Intel Controller. Certainly the machine feels like the fastest machine I've ever used!
All the best,
Robert
_________________
system: G751JT (ASUS-NotebookSKU); processor: Intel(R) Core(TM) i7-4710HQ CPU @ 2.50GHz; memory: 32GiB System Memory; display: GM204M [GeForce GTX 970M]; disk: 2048GB Samsung SSD 850;BD-CMB UJ172 S;1024GB Samsung SSD 850 |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Kernel & Hardware
