Hello all,
We have 64-bit production machines that use the kernel shipped with
a RedHat distribution, linux-2.6.11-1.1369_FC4smp. There have been
no problems until we received recent hardware. With the latest hardware,
which the vendor claims hasn't changed, the machines panic when either
a USB mouse or USB keyboard are plugged in.
Machine:
Supermicro H8DME-2
Two quad-Core AMD Opteron CPUs
64GB DDR2 400 SDRAM
6 SATA2 3.0Gb/s HD Support
2 64-bit 133/100 MHz PCI-X
2 64=bit 100MHz PCI-X
ATI ES1000 Graphics
Chipset nVida MCP55 Pro
NEC uPD720400
USB FUCI, uses ohci_hcd driver.
When mouse or keyboard is inserted, I get a panic with:
"map_single bounce buffer is not DMA'ble."
This comes from: linux-2.6.11-prep/arch/ia64/lib/swiotlb.c,
line 440. I added 64-bit long print hex to the panic statement
so I could display the address it doesn't like.
The address it doesn't like is: 0x000000022a0aa000
Does anybody have a clue what might be the matter and how to fix it?
Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.28 BogoMips).
My book : http://www.AbominableFirebug.com/
_
****************************************************************
The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [email protected] - and destroy all copies of this information, including any attachments, without reading or disclosing them.
Thank you.
On Mon, 3 Mar 2008, linux-os (Dick Johnson) wrote:
> We have 64-bit production machines that use the kernel shipped with
> a RedHat distribution, linux-2.6.11-1.1369_FC4smp.
Try asking on the Fedora lists or perhaps it's even in their bugzilla...
> There have been
> no problems until we received recent hardware. With the latest hardware,
> which the vendor claims hasn't changed,
I'm not following: you changed the hardware, but the vendor said, the hard
did not change?
> the machines panic when either
> a USB mouse or USB keyboard are plugged in.
If you're stuck to this hardware (changed or not): did you try a more
recent kernel? 2.6.11 is pretty old for lkml....
C.
> The information transmitted in this message is confidential and may
> be privileged. Any review, retransmission, dissemination, or other
damn, I retransmitted it :)
--
BOFH excuse #154:
You can tune a file system, but you can't tune a fish (from most tunefs man pages)
On Mon, 3 Mar 2008 19:45:52 +0100 (CET)
Christian Kujau <[email protected]> wrote:
> On Mon, 3 Mar 2008, linux-os (Dick Johnson) wrote:
> > We have 64-bit production machines that use the kernel shipped with
> > a RedHat distribution, linux-2.6.11-1.1369_FC4smp.
>
> Try asking on the Fedora lists or perhaps it's even in their bugzilla...
It's a little late to ask support for Fedora 4. The current released versions
are Fedora 7 and 8; Fedora 9 is the current development version.
--
All Rights Reversed
On Mon, 3 Mar 2008, Rik van Riel wrote:
> On Mon, 3 Mar 2008 19:45:52 +0100 (CET)
> Christian Kujau <[email protected]> wrote:
>
>> On Mon, 3 Mar 2008, linux-os (Dick Johnson) wrote:
>>> We have 64-bit production machines that use the kernel shipped with
>>> a RedHat distribution, linux-2.6.11-1.1369_FC4smp.
>>
>> Try asking on the Fedora lists or perhaps it's even in their bugzilla...
>
> It's a little late to ask support for Fedora 4. The current released versions
> are Fedora 7 and 8; Fedora 9 is the current development version.
>
> --
> All Rights Reversed
>
That's the problem trying to support devices in production.
The last kernel I was able to locate for the x86_64 project
is linux-2.6.23. Its dates are from late last year so maybe
it will work!
It is now being compiled one one of the target machines.
If this works, I'll have to find the differences between
the old and the newer and patch the linux-2.6.11 because
it is a "certified" kernel, i.e., selected by the FAA and
determined to be politically correct!
I was hoping that somebody who works on these projects
would just reply with "just change the xyz mask to ...."
and it would work!
Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.28 BogoMips).
My book : http://www.AbominableFirebug.com/
_
****************************************************************
The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [email protected] - and destroy all copies of this information, including any attachments, without reading or disclosing them.
Thank you.
On Mon, 3 Mar 2008 17:10:53 -0500
"linux-os (Dick Johnson)" <[email protected]> wrote:
> It is now being compiled one one of the target machines.
> If this works, I'll have to find the differences between
> the old and the newer and patch the linux-2.6.11 because
> it is a "certified" kernel, i.e., selected by the FAA and
> determined to be politically correct!
Selecting a kernel that everybody knows will not get
long term support does not strike me as the safest
choice for an FAA project.
--
All Rights Reversed
On 03/03/2008, linux-os (Dick Johnson) <[email protected]> wrote:
>
<snip>
>
> That's the problem trying to support devices in production.
> The last kernel I was able to locate for the x86_64 project
> is linux-2.6.23. Its dates are from late last year so maybe
> it will work!
>
> It is now being compiled one one of the target machines.
> If this works, I'll have to find the differences between
> the old and the newer and patch the linux-2.6.11 because
> it is a "certified" kernel, i.e., selected by the FAA and
> determined to be politically correct!
So, you take a "certified" kernel, then you patch it, which leaves you
with what? Certainly not a certified kernel. You might as well go with
the latest 2.6.24.3 sources then...
>
<snip>
>
--
Jesper Juhl <[email protected]>
Don't top-post http://www.catb.org/~esr/jargon/html/T/top-post.html
Plain text mails only, please http://www.expita.com/nomime.html
On Mon, 3 Mar 2008, Jesper Juhl wrote:
> On 03/03/2008, linux-os (Dick Johnson) <[email protected]> wrote:
>>
> <snip>
>>
>> That's the problem trying to support devices in production.
>> The last kernel I was able to locate for the x86_64 project
>> is linux-2.6.23. Its dates are from late last year so maybe
>> it will work!
>>
>> It is now being compiled one one of the target machines.
>> If this works, I'll have to find the differences between
>> the old and the newer and patch the linux-2.6.11 because
>> it is a "certified" kernel, i.e., selected by the FAA and
>> determined to be politically correct!
>
> So, you take a "certified" kernel, then you patch it, which leaves you
> with what? Certainly not a certified kernel. You might as well go with
> the latest 2.6.24.3 sources then...
>
>>
> <snip>
>>
You are preaching to the choir. When doing government things
we can't resort to logic, only "rules!" Here's an example
of why; We decide to use the latest kernel which, BTW, we are
not allowed to decide that ourselves, we would have to obtain
a consensus from lots of "interested" people, including the
people who move equipment into place (will a later kernel
weigh more? ..They need to know). The process of selecting
a new kernel could take a year. Then, it needs to be
re-certified, which could cause a few million dollars (I am
not kidding...there are people who, working out of their
YMCA rooms, charge the government $600 / hour for such
code-reviews, etc).
That's why we almost always try for a "permissive" change.
In this case, after verifying that the USB works on a
later kernel (it does), I need to find out what changes
were made to make it work and I need to incorporate those
changes into the older driver.
Adding insult to injury, I find that the driver was not
changed much, but the memory allocation scheme for bounce
buffers. To review this stuff takes a lot of nights and
weekends because nobody will give me a hint of the changes
made to make USB work. It's not in any "Changes"
documentation, but somebody knows --and they aren't
telling!
Cheers,
Dick Johnson
Penguin : Linux version 2.6.22.1 on an i686 machine (5588.28 BogoMips).
My book : http://www.AbominableFirebug.com/
_
****************************************************************
The information transmitted in this message is confidential and may be privileged. Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited. If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to [email protected] - and destroy all copies of this information, including any attachments, without reading or disclosing them.
Thank you.
linux-os (Dick Johnson) wrote:
> Hello all,
>
> We have 64-bit production machines that use the kernel shipped with
> a RedHat distribution, linux-2.6.11-1.1369_FC4smp. There have been
> no problems until we received recent hardware. With the latest hardware,
> which the vendor claims hasn't changed, the machines panic when either
> a USB mouse or USB keyboard are plugged in.
>
> Machine:
>
> Supermicro H8DME-2
> Two quad-Core AMD Opteron CPUs
> 64GB DDR2 400 SDRAM
> 6 SATA2 3.0Gb/s HD Support
> 2 64-bit 133/100 MHz PCI-X
> 2 64=bit 100MHz PCI-X
> ATI ES1000 Graphics
>
> Chipset nVida MCP55 Pro
> NEC uPD720400
> USB FUCI, uses ohci_hcd driver.
>
> When mouse or keyboard is inserted, I get a panic with:
> "map_single bounce buffer is not DMA'ble."
>
> This comes from: linux-2.6.11-prep/arch/ia64/lib/swiotlb.c,
> line 440. I added 64-bit long print hex to the panic statement
> so I could display the address it doesn't like.
You're using an Itanium kernel on Opteron?
> The address it doesn't like is: 0x000000022a0aa000
>
> Does anybody have a clue what might be the matter and how to fix it?
If the hardware "didn't change" then it could be a BIOS bug, or just a
different BIOS behavior that your old kernel doesn't handle well. Try
the same version you had on the old hardware. I once had a desktop box
that developed a similar behavior when I twiddled the memory hole
remapping settings in the BIOS.
-- Chris