Blacklisted system / blacklisting bad

Blacklisted system / blacklisting bad

[Dieter Jurzitza]

Dear listmembers,
I recently got a HP VISUALIZE NT workstation, using 2x 1GHz PIII processors 
(no racing car, but that's not the point). After installation ACPI was always 
disabled, the on board USB interface did not work (kernel 2.6.11.4 (SuSE 
linux 9.3)

I tested using acpi=force on the commandline and that solved two problems:

1.) shutdown -h now switches the machine off (should be expected :-):-) )
2.) much more important: the onboard USB interface works now.

After patching dmi_scan.c and removing the blacklist-entry I get the following 
message at boot time:

<4>PCI: Probing PCI hardware (bus 00)
<7>ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
<4>    ACPI-0234: *** Warning: Wrong _BBN value, please reboot and using 
option 'pci=noacpi'
<6>ACPI: PCI _CRS 2 overrides _BBN 0
<6>ACPI: PCI Root Bridge [PCI1] (00:02)
<4>PCI: Probing PCI hardware (bus 02)

If I du so (pci=noacpi at boottime) the on board USB interface gets stuck 
again and refuses to work.

So, blacklisting seems to be malicious for this machine and this kernel (I 
know that for many of you 2.6.11 might be outdated, but that's not the 
point). Even pci=noacpi is bad.

If either pci=noacpi or acpi=ht is given, I get the following message 
in /var/log/messages:

Sep  2 13:04:06 linux kernel: ohci_hcd: 2004 Nov 08 USB 1.1 'Open' Host 
Controller (OHCI) Driver (PCI)
Sep  2 13:04:06 linux kernel: PCI->APIC IRQ transform: 0000:00:0f.2[A] -> IRQ 
0
Sep  2 13:04:06 linux kernel: ohci_hcd 0000:00:0f.2: Found HC with no IRQ.  
Check BIOS/PCI 0000:00:0f.2 setup!

I assume this is why the onboard USB-controller is disfunctional as soon as I 
do what is recommended, i. e. keep the blacklisting / boot with pci=noirq.

Therefore I suggest patching dmi_scan.c and remove the entry for HP VISUALIZE. 
With a freshly built kernel without the blacklisting the machine has happily 
been building kernels for hours now without a single problem.

If there is anything I should test on top of this to verify whatever should be 
verified, please let me know! Any hints are highly appreciated.

To whom this might concern,
take care


Dieter Jurzitza
-- 
-----------------------------------------------------------

                               |
                                \
                 /\_/\           |
                | ~x~ |/-----\   /
                 \   /-       \_/
  ^^__   _        /  _  ____   /
 <°°__ \- \_/     |  |/    |  |
  ||  ||         _| _|    _| _|

if you really want to see the pictures above - use some font
with constant spacing like courier! :-)
-----------------------------------------------------------

-- 
VGER BF report: U 0.508633
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Blacklisted system / blacklisting bad

[Yu Luming]

On Sunday 03 September 2006 17:34, Dieter Jurzitza wrote:
> Dear listmembers,
>  I recently got a HP VISUALIZE NT workstation, using 2x 1GHz PIII
> processors (no racing car, but that's not the point). After installation
> ACPI was always disabled, the on board USB interface did not work (kernel
> 2.6.11.4 (SuSE linux 9.3)
>
>  I tested using acpi=force on the commandline and that solved two problems:
>
I guess this is a pretty old box, and acpi just happen to work.

>  1.) shutdown -h now switches the machine off (should be expected :-):-) )
>  2.) much more important: the onboard USB interface works now.
>
>  After patching dmi_scan.c and removing the blacklist-entry I get the
> following message at boot time:
>
>  <4>PCI: Probing PCI hardware (bus 00)
>  <7>ACPI: PCI Interrupt Routing Table [\_SB_.PCI0._PRT]
>  <4>    ACPI-0234: *** Warning: Wrong _BBN value, please reboot and using
>  option 'pci=noacpi'
>  <6>ACPI: PCI _CRS 2 overrides _BBN 0

You are hitting this issue: 
http://bugzilla.kernel.org/show_bug.cgi?id=1662#c42

If everything works fine, please don't  ignore the warning message.

Thanks,
Luming
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Blacklisted system / blacklisting bad

[Len Brown]

On Sunday 03 September 2006 05:34, Dieter Jurzitza wrote:

> I recently got a HP VISUALIZE NT workstation, using 2x 1GHz PIII processors 
>. After installation ACPI was always  disabled, the on board USB interface did not work
>(kernel 2.6.11.4 (SuSE  linux 9.3)
> 
> I tested using acpi=force on the commandline and that solved two problems:
> 
> 1.) shutdown -h now switches the machine off (should be expected :-):-) )
> 2.) much more important: the onboard USB interface works now.
> 
> After patching dmi_scan.c and removing the blacklist-entry I get the following 
> message at boot time:
....

ACPI is necessary to get the interrupts right for USB to work on this box.
pci=noacpi and acpi=ht will cause ACPI to be disabled for interrupt routing
and thus USB will fail with those.

As Luming pointed out, Shaohua's _CRS workaround for the _BBN BIOS bug
is what your system needed, and that is why "acpi=force" is working.
It seems that perhaps the DMI blacklist entry pre-dated that workaround.

Please verify that 2.6.18-rc7 works with "acpi=force" on this box,
and if it does, please send me a patch to remove the entry in that kernel.

thanks,
-Len

Re: Blacklisted system / blacklisting bad

[Dieter Jurzitza]

Dear Len,
using 2.6.18-rc7 on this box works like charm - using acpi=force. The bad 
thing about it stems from the additional fact, that acpi has been disabled for 
elder bioses by a switch checking for the bios age. I did not know about this 
because 2.6.11 doesn't use the age checking.
However things may be, I tested according to your suggestion and things simply 
work as expected.
You can argue that the compile time option for the age check is optional - but 
i. e. SuSE is using it, at last.
So you may remove the blacklisting entry, but things will not change for the 
user as the age checking forces me to use acpi=force anyway.
IMHO a "normal" user would have failed with this, both kinds of blacklisting 
are bad (sure, for this specific system).
Is there any option to white-list systems in order to avoid the acpi=force 
commandline switch? IMHO it will be hard for a "normal" user to cope with 
this and to understand the relation between an "acpi=force" - switch on the 
boot-commandline and a disfunctional USB interface. Just my two cent here ...


Anyway, patch for 2.6.18 is attached, thanks for your efforts,
take care



Dieter Jurzitza

**************************************************************************************************************
--- linux/arch/i386/kernel/acpi/boot.c  2006-09-13 03:41:36.000000000 +0200
+++ linux/arch/i386/kernel/acpi/boot.c  2006-09-22 21:42:46.000000000 +0200
@@ -951,14 +951,6 @@
         },
        {
         .callback = force_acpi_ht,
-        .ident = "HP VISUALIZE NT Workstation",
-        .matches = {
-                    DMI_MATCH(DMI_BOARD_VENDOR, "Hewlett-Packard"),
-                    DMI_MATCH(DMI_PRODUCT_NAME, "HP VISUALIZE NT 
Workstation"),
-                    },
-        },
-       {
-        .callback = force_acpi_ht,
         .ident = "Compaq Workstation W8000",
         .matches = {
                     DMI_MATCH(DMI_SYS_VENDOR, "Compaq"),
**************************************************************************************************************

Am Mittwoch, 20. September 2006 06:55 schrieb Len Brown:
******
> ACPI is necessary to get the interrupts right for USB to work on this box.
> pci=noacpi and acpi=ht will cause ACPI to be disabled for interrupt routing
> and thus USB will fail with those.
******
-- 
-----------------------------------------------------------

                               |
                                \
                 /\_/\           |
                | ~x~ |/-----\   /
                 \   /-       \_/
  ^^__   _        /  _  ____   /
 <°°__ \- \_/     |  |/    |  |
  ||  ||         _| _|    _| _|

if you really want to see the pictures above - use some font
with constant spacing like courier! :-)
-----------------------------------------------------------
-
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html