Here's the situation. I have a MSI KN8Neo-f motherboard with a Seagate
Barracuda 250 GB SATA drive. I have replaced this drive three times in
the last two weeks due to it failing. Now the only thing in common is
the use of a 2.6.22.9 kernel I built from scratch, before that I was
using 2.6.19 kernel but working on doing some upgrades for the CLFS
project and tried a 2.6.22.9. I will explain the situation and
including the links I used for fix the issues.
It boots up and instantly has a problem detecting so I have to add
irqpoll to my bootup line. {Reference
http://my.opera.com/snowburn/blog/index.dml/tag/failed%20to%20set%20xfermode}
Then I get the adma issue, so I add sata_nv.adma=0 to my bootup
line.{Reference
http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1424.html}
Now after a few times of booting the drive completely fails. So is the
nv_sata drive the cause?? Based on what I have done yes, and the reason
I can say that is because I took another one of those drives and the
same computer loaded Windows(forgive me for using this word!!!) on it
and it worked perfectly.
One other issue that showed up was the drive does not recognize the
drive when it's in the 3gb transfer mode, I have to jumper it to 1.5gb mode.
Here's a link to the drive model:
http://www.seagate.com/ww/v/index.jsp?vgnextoid=a62099f4fa74c010VgnVCM100000dd04090aRCRD&locale=en-US
Here's a link to my motherboard:
http://www.msicomputer.com/product/p_spec.asp?model=K8N_Neo4-F&class=mb
It has 2GB of Memory and a Nvidia PCI-E Graphics Adapter
I will work with any member of the lkml team to help isolate the issue,
the only issue is that when it locks up everything has to be written
down or photographed.
Jim Gifford - CLFS Developer
http://www.cross-lfs.org
Jim Gifford wrote:
> Here's the situation. I have a MSI KN8Neo-f motherboard with a Seagate
> Barracuda 250 GB SATA drive. I have replaced this drive three times in
> the last two weeks due to it failing. Now the only thing in common is
> the use of a 2.6.22.9 kernel I built from scratch, before that I was
> using 2.6.19 kernel but working on doing some upgrades for the CLFS
> project and tried a 2.6.22.9. I will explain the situation and
> including the links I used for fix the issues.
>
> It boots up and instantly has a problem detecting so I have to add
> irqpoll to my bootup line. {Reference
> http://my.opera.com/snowburn/blog/index.dml/tag/failed%20to%20set%20xfermode}
> Then I get the adma issue, so I add sata_nv.adma=0 to my bootup
> line.{Reference
> http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1424.html}
>
> Now after a few times of booting the drive completely fails. So is the
> nv_sata drive the cause?? Based on what I have done yes, and the reason
> I can say that is because I took another one of those drives and the
> same computer loaded Windows(forgive me for using this word!!!) on it
> and it worked perfectly.
You'll have to be more specific about "completely fails".. Also, it
would be useful to get the dmesg output from dmesg directly and not from
syslog, as the entries posted under the second link you gave were
missing some critical information (it seems that syslog can do this in
some cases..) In particular the line showing what command failed would
be useful.
Using "irqpoll" is not a very good workaround, it really is just a
band-aid to get things working. Often this indicates motherboard IRQ
routing issues. You may want to make sure you have the latest BIOS
installed for the board.
>
> One other issue that showed up was the drive does not recognize the
> drive when it's in the 3gb transfer mode, I have to jumper it to 1.5gb
> mode.
>
> Here's a link to the drive model:
> http://www.seagate.com/ww/v/index.jsp?vgnextoid=a62099f4fa74c010VgnVCM100000dd04090aRCRD&locale=en-US
>
> Here's a link to my motherboard:
> http://www.msicomputer.com/product/p_spec.asp?model=K8N_Neo4-F&class=mb
> It has 2GB of Memory and a Nvidia PCI-E Graphics Adapter
>
> I will work with any member of the lkml team to help isolate the issue,
> the only issue is that when it locks up everything has to be written
> down or photographed.
--
Robert Hancock Saskatoon, SK, Canada
To email, remove "nospam" from [email protected]
Home Page: http://www.roberthancock.com/
Robert Hancock skrev:
> Using "irqpoll" is not a very good workaround, it really is just a
> band-aid to get things working. Often this indicates motherboard IRQ
> routing issues. You may want to make sure you have the latest BIOS
> installed for the board.
MSI seems to have problems with the quality of the engineering of there
motherboards.
I've been fighting all year with a MSI K9 Neo and MSI K9 Diamond
motherboard, buy the same chip set from a different vendor and
everything just works.
I'm unfortunately not experienced enough with this and searching for
someone with a clue has been unsuccessful. It seems that disabling
MSI/MSI-X makes the MSI K9 Diamond stable, the Neo one I just gave up on
and moved to a different brand, but same chip set.
Robert Hancock wrote:
> Jim Gifford wrote:
>> Here's the situation. I have a MSI KN8Neo-f motherboard with a
>> Seagate Barracuda 250 GB SATA drive. I have replaced this drive three
>> times in the last two weeks due to it failing. Now the only thing in
>> common is the use of a 2.6.22.9 kernel I built from scratch, before
>> that I was using 2.6.19 kernel but working on doing some upgrades for
>> the CLFS project and tried a 2.6.22.9. I will explain the situation
>> and including the links I used for fix the issues.
>>
>> It boots up and instantly has a problem detecting so I have to add
>> irqpoll to my bootup line. {Reference
>> http://my.opera.com/snowburn/blog/index.dml/tag/failed%20to%20set%20xfermode}
>> Then I get the adma issue, so I add sata_nv.adma=0 to my bootup
>> line.{Reference
>> http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1424.html}
>>
>> Now after a few times of booting the drive completely fails. So is
>> the nv_sata drive the cause?? Based on what I have done yes, and the
>> reason I can say that is because I took another one of those drives
>> and the same computer loaded Windows(forgive me for using this
>> word!!!) on it and it worked perfectly.
>
> You'll have to be more specific about "completely fails".. Also, it
> would be useful to get the dmesg output from dmesg directly and not
> from syslog, as the entries posted under the second link you gave were
> missing some critical information (it seems that syslog can do this in
> some cases..) In particular the line showing what command failed would
> be useful.
It never gets into the system at all, just locks up. The hard drive
starts clicking and eventually doesn't get recognized by BIOS.
>
> Using "irqpoll" is not a very good workaround, it really is just a
> band-aid to get things working. Often this indicates motherboard IRQ
> routing issues. You may want to make sure you have the latest BIOS
> installed for the board.
I'll check for a new BIOS, but believe I have the most current on it.
>
>>
>> One other issue that showed up was the drive does not recognize the
>> drive when it's in the 3gb transfer mode, I have to jumper it to
>> 1.5gb mode.
>>
>> Here's a link to the drive model:
>> http://www.seagate.com/ww/v/index.jsp?vgnextoid=a62099f4fa74c010VgnVCM100000dd04090aRCRD&locale=en-US
>>
>> Here's a link to my motherboard:
>> http://www.msicomputer.com/product/p_spec.asp?model=K8N_Neo4-F&class=mb
>> It has 2GB of Memory and a Nvidia PCI-E Graphics Adapter
>>
>> I will work with any member of the lkml team to help isolate the
>> issue, the only issue is that when it locks up everything has to be
>> written down or photographed.
>
Jim Gifford wrote:
> Robert Hancock wrote:
>> Jim Gifford wrote:
>>> Here's the situation. I have a MSI KN8Neo-f motherboard with a
>>> Seagate Barracuda 250 GB SATA drive. I have replaced this drive three
>>> times in the last two weeks due to it failing. Now the only thing in
>>> common is the use of a 2.6.22.9 kernel I built from scratch, before
>>> that I was using 2.6.19 kernel but working on doing some upgrades for
>>> the CLFS project and tried a 2.6.22.9. I will explain the situation
>>> and including the links I used for fix the issues.
>>>
>>> It boots up and instantly has a problem detecting so I have to add
>>> irqpoll to my bootup line. {Reference
>>> http://my.opera.com/snowburn/blog/index.dml/tag/failed%20to%20set%20xfermode}
>>> Then I get the adma issue, so I add sata_nv.adma=0 to my bootup
>>> line.{Reference
>>> http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1424.html}
>>>
>>> Now after a few times of booting the drive completely fails. So is
>>> the nv_sata drive the cause?? Based on what I have done yes, and the
>>> reason I can say that is because I took another one of those drives
>>> and the same computer loaded Windows(forgive me for using this
>>> word!!!) on it and it worked perfectly.
>>
>> You'll have to be more specific about "completely fails".. Also, it
>> would be useful to get the dmesg output from dmesg directly and not
>> from syslog, as the entries posted under the second link you gave were
>> missing some critical information (it seems that syslog can do this in
>> some cases..) In particular the line showing what command failed would
>> be useful.
> It never gets into the system at all, just locks up. The hard drive
> starts clicking and eventually doesn't get recognized by BIOS.
That doesn't sound like a kernel problem if it's not even recognized by
the BIOS, there's nothing the kernel should be able to do that could
cause that. Sounds more like a hardware issue. If the drive has been
replaced before then it could be a power supply or temperature problem?
Robert Hancock wrote:
> Jim Gifford wrote:
>> Robert Hancock wrote:
>>> Jim Gifford wrote:
>>>> Here's the situation. I have a MSI KN8Neo-f motherboard with a
>>>> Seagate Barracuda 250 GB SATA drive. I have replaced this drive
>>>> three times in the last two weeks due to it failing. Now the only
>>>> thing in common is the use of a 2.6.22.9 kernel I built from
>>>> scratch, before that I was using 2.6.19 kernel but working on doing
>>>> some upgrades for the CLFS project and tried a 2.6.22.9. I will
>>>> explain the situation and including the links I used for fix the
>>>> issues.
>>>>
>>>> It boots up and instantly has a problem detecting so I have to add
>>>> irqpoll to my bootup line. {Reference
>>>> http://my.opera.com/snowburn/blog/index.dml/tag/failed%20to%20set%20xfermode}
>>>> Then I get the adma issue, so I add sata_nv.adma=0 to my bootup
>>>> line.{Reference
>>>> http://www.ussg.iu.edu/hypermail/linux/kernel/0706.1/1424.html}
>>>>
>>>> Now after a few times of booting the drive completely fails. So is
>>>> the nv_sata drive the cause?? Based on what I have done yes, and
>>>> the reason I can say that is because I took another one of those
>>>> drives and the same computer loaded Windows(forgive me for using
>>>> this word!!!) on it and it worked perfectly.
>>>
>>> You'll have to be more specific about "completely fails".. Also, it
>>> would be useful to get the dmesg output from dmesg directly and not
>>> from syslog, as the entries posted under the second link you gave
>>> were missing some critical information (it seems that syslog can do
>>> this in some cases..) In particular the line showing what command
>>> failed would be useful.
>> It never gets into the system at all, just locks up. The hard drive
>> starts clicking and eventually doesn't get recognized by BIOS.
>
> That doesn't sound like a kernel problem if it's not even recognized
> by the BIOS, there's nothing the kernel should be able to do that
> could cause that. Sounds more like a hardware issue. If the drive has
> been replaced before then it could be a power supply or temperature
> problem?
I thought that also, just tried a new power supply, same problem. I can
recreate this problem within minutes with a new drive. Got 2 more drives
left in this batch. Temperature is in the low range of the specs.