Various problems with MCP55 since 2.6.37

Various problems with MCP55 since 2.6.37

[David Krider]

I'm getting lots of different problems with the MCP55 controller since
kernel 2.6.37. I have an Asus-built Nvidia 780i-based motherboard. I
have both Ubuntu and Windows 7 installed. Windows 7 doesn't seem to be
giving me any trouble at all, so I don't think it's my hardware. I have
spent many hours working on this to no avail. If I stay on Ubuntu 10.10,
which has a 2.6.35-based kernel, I don't have these problems, but that
version of Ubuntu will be rolling off support soon (and I want Gnome 3).

First, the computer won't reboot correctly. If I let it just reboot,
then, when either Windows or Linux is starting, it will just
spontaneously reboot again. This doesn't happen when rebooting from
Windows. If I simply shut down from Linux, then manually power the
computer back on, the BIOS screen will just barely pop up, then the BIOS
itself will restart, and come back up to an error screen saying that the
BIOS "crashed."

I've gotten scrambled file systems twice because of this.

I was hoping the problem was my fakeraid setup, so I bought an SSD and
replaced it. That has brought on a new set of problems. I get lots of
"interesting" messages in dmesg, but I can't figure out if they're just
informational, or if there's a serious, data-losing situation occuring.

[  922.848302] ata3.00: failed command: WRITE FPDMA QUEUED
[  922.848305] ata3.00: cmd 61/10:f0:a0:a3:f0/00:00:10:00:00/40 tag 30
ncq 8192 out
[  922.848306]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask
0x4 (timeout)
[  922.848308] ata3.00: status: { DRDY }
[  922.848311] ata3: hard resetting link
[  922.848312] ata3: nv: skipping hardreset on occupied port
[  928.360008] ata3: link is slow to respond, please be patient (ready=0)
[  932.896009] ata3: SRST failed (errno=-16)
[  932.896014] ata3: hard resetting link
[  932.896016] ata3: nv: skipping hardreset on occupied port
[  933.364026] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  933.380198] ata3.00: configured for UDMA/133
[  933.380203] ata3.00: device reported invalid CHS sector 0

Also, I occasionally get "pauses." Every once in awhile, the drive light
will come on solid, and the whole machine will simply pause for about 60
seconds, and then I get messages like this:

[ 1040.361005] sd 2:0:0:0: [sda]  Result: hostbyte=DID_OK
driverbyte=DRIVER_SENSE
[ 1040.361007] sd 2:0:0:0: [sda]  Sense Key : Aborted Command [current]
[descriptor]
[ 1040.361009] Descriptor sense data with sense descriptors (in hex):
[ 1040.361011]         72 0b 47 00 00 00 00 0c 00 0a 80 00 00 00 00 00
[ 1040.361016]         12 30 d4 48
[ 1040.361019] sd 2:0:0:0: [sda]  Add. Sense: Scsi parity error
[ 1040.361021] sd 2:0:0:0: [sda] CDB: Read(10): 28 00 12 30 d4 f8 00 00
08 00
[ 1040.361026] end_request: I/O error, dev sda, sector 305190136

I've created an Ubuntu bug on the reboot problem that is confirmed, but
I can't get any more information on there.
(https://bugs.launchpad.net/ubuntu/+source/linux/+bug/829413)

I also made a bug report on the kernel bugzilla, but -- obviously --
it's down.

I've tried half a dozen kernel boot flags.

I've tried a dozen different kernels.

I tried to switch to btrfs, but Ubuntu's kernel package has a bug that
prevents it from being installed on there.

In other words, I'm trying to say that I'm doing the work that an
end-user can be expected to do. I don't know what else to try. I'm
really desperate to get this fixed; it really bothers me. At this point,
I'm considering buying a PCIe SATA RAID card, or going back to Gentoo,
where I can really play with kernel versions and patches more
effectively, and I'm not really looking forward to that...

ANY help would be GREATLY appreciated,
dk

Fwd: Various problems with MCP55 since 2.6.37

[David Krider]

I apologize. I see that I'm conflating issues. I reinstalled Ubuntu
10.10, again, which has a 2.6.35 kernel. This fixes the rebooting
problem, but I still have the ata errors.

I've noticed a couple things.

First, I've never gotten scrambled filesystems since getting rid of
fakeraid. I'm going to blame that scrambling on that, and forget about it.

Second, I finally notice, in the ata error below, that ata3 can't be my
SSD. From a message to this list that I just found, I see that it would
be strongly suspected that I have power problems. I have an 800W power
supply, so I don't think it's not giving ENOUGH power, but the quality
might be bad? (I did a lot of damage to an old dual Athlon computer by
having an underspec'ed PS; I didn't make that mistake on this one.) I've
installed lm-sensors and will try to track this.

So I just want to dial into the reboot problem, as I reboot into Windows
(to play games) a lot. That one's pretty clear. If I upgrade to a
post-2.6.35 kernel, I get the problem, whether I'm using an SSD or a
regular HDD. Is there anything that can be done?

-------- Original Message --------
Subject: Various problems with MCP55 since 2.6.37
Date: Sat, 03 Dec 2011 07:29:50 -0500
From: David Krider <david@davidkrider.com>
To: linux-ide@vger.kernel.org

I'm getting lots of different problems with the MCP55 controller since
kernel 2.6.37. I have an Asus-built Nvidia 780i-based motherboard. I
have both Ubuntu and Windows 7 installed. Windows 7 doesn't seem to be
giving me any trouble at all, so I don't think it's my hardware. I have
spent many hours working on this to no avail. If I stay on Ubuntu 10.10,
which has a 2.6.35-based kernel, I don't have these problems, but that
version of Ubuntu will be rolling off support soon (and I want Gnome 3).

First, the computer won't reboot correctly. If I let it just reboot,
then, when either Windows or Linux is starting, it will just
spontaneously reboot again. This doesn't happen when rebooting from
Windows. If I simply shut down from Linux, then manually power the
computer back on, the BIOS screen will just barely pop up, then the BIOS
itself will restart, and come back up to an error screen saying that the
BIOS "crashed."

I've gotten scrambled file systems twice because of this.

I was hoping the problem was my fakeraid setup, so I bought an SSD and
replaced it. That has brought on a new set of problems. I get lots of
"interesting" messages in dmesg, but I can't figure out if they're just
informational, or if there's a serious, data-losing situation occuring.

[  922.848302] ata3.00: failed command: WRITE FPDMA QUEUED
[  922.848305] ata3.00: cmd 61/10:f0:a0:a3:f0/00:00:10:00:00/40 tag 30
ncq 8192 out
[  922.848306]          res 40/00:00:00:00:00/00:00:00:00:00/00 Emask
0x4 (timeout)
[  922.848308] ata3.00: status: { DRDY }
[  922.848311] ata3: hard resetting link
[  922.848312] ata3: nv: skipping hardreset on occupied port
[  928.360008] ata3: link is slow to respond, please be patient (ready=0)
[  932.896009] ata3: SRST failed (errno=-16)
[  932.896014] ata3: hard resetting link
[  932.896016] ata3: nv: skipping hardreset on occupied port
[  933.364026] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[  933.380198] ata3.00: configured for UDMA/133
[  933.380203] ata3.00: device reported invalid CHS sector 0

Also, I occasionally get "pauses." Every once in awhile, the drive light
will come on solid, and the whole machine will simply pause for about 60
seconds, and then I get messages like this:

[ 1040.361005] sd 2:0:0:0: [sda]  Result: hostbyte=DID_OK
driverbyte=DRIVER_SENSE
[ 1040.361007] sd 2:0:0:0: [sda]  Sense Key : Aborted Command [current]
[descriptor]
[ 1040.361009] Descriptor sense data with sense descriptors (in hex):
[ 1040.361011]         72 0b 47 00 00 00 00 0c 00 0a 80 00 00 00 00 00
[ 1040.361016]         12 30 d4 48
[ 1040.361019] sd 2:0:0:0: [sda]  Add. Sense: Scsi parity error
[ 1040.361021] sd 2:0:0:0: [sda] CDB: Read(10): 28 00 12 30 d4 f8 00 00
08 00
[ 1040.361026] end_request: I/O error, dev sda, sector 305190136

I've created an Ubuntu bug on the reboot problem that is confirmed, but
I can't get any more information on there.
(https://bugs.launchpad.net/ubuntu/+source/linux/+bug/829413)

I also made a bug report on the kernel bugzilla, but -- obviously --
it's down.

I've tried half a dozen kernel boot flags.

I've tried a dozen different kernels.

I tried to switch to btrfs, but Ubuntu's kernel package has a bug that
prevents it from being installed on there.

In other words, I'm trying to say that I'm doing the work that an
end-user can be expected to do. I don't know what else to try. I'm
really desperate to get this fixed; it really bothers me. At this point,
I'm considering buying a PCIe SATA RAID card, or going back to Gentoo,
where I can really play with kernel versions and patches more
effectively, and I'm not really looking forward to that...

ANY help would be GREATLY appreciated,
dk

Re: Fwd: Various problems with MCP55 since 2.6.37

[Robert Hancock]

On 12/04/2011 05:56 AM, David Krider wrote:
> I apologize. I see that I'm conflating issues. I reinstalled Ubuntu
> 10.10, again, which has a 2.6.35 kernel. This fixes the rebooting
> problem, but I still have the ata errors.
>
> I've noticed a couple things.
>
> First, I've never gotten scrambled filesystems since getting rid of
> fakeraid. I'm going to blame that scrambling on that, and forget about it.
>
> Second, I finally notice, in the ata error below, that ata3 can't be my
> SSD. From a message to this list that I just found, I see that it would
> be strongly suspected that I have power problems. I have an 800W power
> supply, so I don't think it's not giving ENOUGH power, but the quality
> might be bad? (I did a lot of damage to an old dual Athlon computer by
> having an underspec'ed PS; I didn't make that mistake on this one.) I've
> installed lm-sensors and will try to track this.
>
> So I just want to dial into the reboot problem, as I reboot into Windows
> (to play games) a lot. That one's pretty clear. If I upgrade to a
> post-2.6.35 kernel, I get the problem, whether I'm using an SSD or a
> regular HDD. Is there anything that can be done?

You can try the kernel parameter "reboot=(reboot mode)" to try and make 
the system reboot in different ways:

    warm   Don't set the cold reboot flag
    cold   Set the cold reboot flag
    bios   Reboot by jumping through the BIOS (only for X86_32)
    smp    Reboot by executing reset on BSP or other CPU (only for X86_32)
    triple Force a triple fault (init)
    kbd    Use the keyboard controller. cold reset (default)
    acpi   Use the RESET_REG in the FADT
    efi    Use efi reset_system runtime service
    pci    Use the so-called "PCI reset register", CF9
    force  Avoid anything that could hang.

If a non-default mode works better, then you should send a report to 
linux-kernel about it. Contrary to the list above, I think acpi is the 
default at this point.