Re: [coreboot] r1009 - in coreboot-v3: mainboard/kontron/986lcd-m northbridge/intel/i945 southbridge/intel/i82801gx
svn@coreboot.org wrote:
Fewer errors. The weird part: I had to move all the i82801gx south files to be compiled to the mainboard.
Why? Because the board doesn't use ide support. So you can't compile that in, it's not in the dts.
Oh, the dts is not capable of marking a device unused? This looks like a major shortcoming. I thought we had a discussion about a disabled flag in the past that would allow disabling a device.
But: The board uses IDE support.
Only a quick glance at the DTS stuff, but it's confusing me:
Modified: coreboot-v3/mainboard/kontron/986lcd-m/dts
--- coreboot-v3/mainboard/kontron/986lcd-m/dts 2008-11-12 22:43:50 UTC (rev 1008) +++ coreboot-v3/mainboard/kontron/986lcd-m/dts 2008-11-12 23:09:42 UTC (rev 1009) @@ -181,6 +181,13 @@ pci@1f,0{/* which ich? */ /config/("southbridge/intel/i82801gx/ich7m_dh_lpc.dts"); };
pci@1f,2{/config/("southbridge/intel/i82801gx/sata.dts");};pci@1f,3{/config/("southbridge/intel/i82801gx/smbus.dts");};- }; ioport@2e { /config/("superio/winbond/w83627thg/dts");
So, the generic ICH7 device structure is described in the mainboard dts? We need to mention in the mainboard dts that there's 1f,0 and 1f,2 and which config it uses? That's a rather generic device description
Modified: coreboot-v3/southbridge/intel/i82801gx/sata.dts
--- coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 22:43:50 UTC (rev 1008) +++ coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 23:09:42 UTC (rev 1009) @@ -20,4 +20,9 @@
{ device_operations = "i82801gx_sata_normal_driver";
- ide_legacy_combined = "0";
- ide_enable_primary = "0";
- ide_enable_secondary = "0";
- sata_ahci = "0";
};
--- and --- the mainboard specific variables are in the generic code?
Now, this is definitely very unexpected. Is this a shortcoming in our DTS/DTC concept? Something's really fishy here.
On 13.11.2008 03:20, Stefan Reinauer wrote:
svn@coreboot.org wrote:
Fewer errors. The weird part: I had to move all the i82801gx south files to be compiled to the mainboard.
Why? Because the board doesn't use ide support. So you can't compile that in, it's not in the dts.
Oh, the dts is not capable of marking a device unused? This looks like a major shortcoming. I thought we had a discussion about a disabled flag in the past that would allow disabling a device.
But: The board uses IDE support.
Only a quick glance at the DTS stuff, but it's confusing me:
Modified: coreboot-v3/mainboard/kontron/986lcd-m/dts
--- coreboot-v3/mainboard/kontron/986lcd-m/dts 2008-11-12 22:43:50 UTC (rev 1008) +++ coreboot-v3/mainboard/kontron/986lcd-m/dts 2008-11-12 23:09:42 UTC (rev 1009) @@ -181,6 +181,13 @@ pci@1f,0{/* which ich? */ /config/("southbridge/intel/i82801gx/ich7m_dh_lpc.dts"); };
pci@1f,2{/config/("southbridge/intel/i82801gx/sata.dts");};pci@1f,3{/config/("southbridge/intel/i82801gx/smbus.dts");};- }; ioport@2e { /config/("superio/winbond/w83627thg/dts");
So, the generic ICH7 device structure is described in the mainboard dts? We need to mention in the mainboard dts that there's 1f,0 and 1f,2 and which config it uses? That's a rather generic device description
I agree. That stuff really belongs in the chipset dts. However, that means we need multilevel dts. Is that good or bad?
Modified: coreboot-v3/southbridge/intel/i82801gx/sata.dts
--- coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 22:43:50 UTC (rev 1008) +++ coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 23:09:42 UTC (rev 1009) @@ -20,4 +20,9 @@
{ device_operations = "i82801gx_sata_normal_driver";
- ide_legacy_combined = "0";
- ide_enable_primary = "0";
- ide_enable_secondary = "0";
- sata_ahci = "0";
};
--- and --- the mainboard specific variables are in the generic code?
Now, this is definitely very unexpected. Is this a shortcoming in our DTS/DTC concept? Something's really fishy here.
I need to check if there is an overlay capability for the device tree. That way, only devices which differ from the default need to be specified in the mainboard dts.
Regards, Carl-Daniel
Carl-Daniel Hailfinger wrote:
I need to check if there is an overlay capability for the device tree. That way, only devices which differ from the default need to be specified in the mainboard dts.
I'm thinking maybe have everything disabled by default but I can't quite decide.
//Peter
Try this.
This started out because in the kontron board ide is commented out.
On Wed, Nov 12, 2008 at 6:20 PM, Stefan Reinauer stepan@coresystems.de wrote:
svn@coreboot.org wrote:
Fewer errors. The weird part: I had to move all the i82801gx south files to be compiled to the mainboard.
Why? Because the board doesn't use ide support. So you can't compile that in, it's not in the dts.
Oh, the dts is not capable of marking a device unused? This looks like a major shortcoming. I thought we had a discussion about a disabled flag in the past that would allow disabling a device.
But: The board uses IDE support.
But: it's commented out. #device pci 1f.1 off end # IDE
So, the generic ICH7 device structure is described in the mainboard dts? We need to mention in the mainboard dts that there's 1f,0 and 1f,2 and which config it uses? That's a rather generic device description
It's right ouf of Config.lb. It's how we've written all dts to date.
Modified: coreboot-v3/southbridge/intel/i82801gx/sata.dts
--- coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 22:43:50 UTC (rev 1008) +++ coreboot-v3/southbridge/intel/i82801gx/sata.dts 2008-11-12 23:09:42 UTC (rev 1009) @@ -20,4 +20,9 @@
{ device_operations = "i82801gx_sata_normal_driver";
ide_legacy_combined = "0";ide_enable_primary = "0";ide_enable_secondary = "0";sata_ahci = "0";};
--- and --- the mainboard specific variables are in the generic code?
not sure what you mean. Are these mainboard specific or sata specific? It is hard to tell from the v2 stuff. I am only guessing.
ron
-
Carl-Daniel Hailfinger -
Peter Stuge -
ron minnich -
Stefan Reinauer