On Tue, Nov 25, 2008 at 10:06 AM, ron minnich rminnich@gmail.com wrote:
On Tue, Nov 25, 2008 at 9:04 AM, Myles Watson mylesgw@gmail.com wrote:
I'm comfortable with the idea that this is the way we want to go, but it doesn't work for me yet. We could commit to make patching easier. I
guess
it's up to you how functional you want it before committing.
We'll commit when
- it works for you
- I can build a working dbe62 target (which I still can)
You have the token for now.
I'm not sure that's wise. I'm breaking more than fixing right now.
Here's the latest snippet: DTC mainboard/amd/serengeti/dts (dts->lbh) name_node: pci@0,0(usb11) name_node: pci@0,1(usb12) name_node: pci@0,2(usb2) name_node: pci@1,0(nic) name_node: pci@0,0(multilevel_root) name_node: pci@1,0(lpc) name_node: pci@1,1(ide) name_node: pci@1,2(smbus) name_node: pci@1,3(acpi) name_node: pci@1,5(ac97audio) name_node: pci@1,6(ac97modem) name_node: pci@1,7(dev7) name_node: NULL(nic) name_node: NULL(ide) name_node: ioport@2e(NULL) name_node: NULL(lpc) fixup_properties: before fixups mainboard tree nic() NULL(nic) fixup_properties: before fixups chipnodes tree multilevel_root(pci@0,0) pci@0,0(multilevel_root) pci@0,0(usb11) pci@0,1(usb12) pci@0,2(usb2) pci@1,0(nic) fixup_properties: checking nic() and multilevel_root(pci@0,0) fixup_properties: checking nic() and lpc(pci@1,0) fixup_properties: checking nic() and ide(pci@1,1) fixup_properties: checking nic() and smbus(pci@1,2) fixup_properties: checking nic() and acpi(pci@1,3) fixup_properties: checking nic() and ac97audio(pci@1,5) fixup_properties: checking nic() and ac97modem(pci@1,6) fixup_properties: checking nic() and dev7(pci@1,7) fixup_properties: checking nic() and nic() fixup_properties: last_child NULL() fixup_properties: matched and deleting nic((null)) del_node: matched and deleting nic((null)) del_node: expect seg fault nic((null)) /bin/sh: line 1: 3615 Segmentation fault /home/myles/buildrom/try/buildrom/buildrom-devel/work/coreboot-v3/svn/build/util/dtc/dtc -O lbh mainboard/amd/serengeti/dts >/tmp/statictree.h.$ make: *** [/home/myles/buildrom/try/buildrom/buildrom-devel/work/coreboot-v3/svn/build/statictree.h] Error 139
The "expect seg fault" line means that del_node got passed a node with no parent. That means I don't understand enough what's going on here. I was expecting the parent to be set to pci@0,0 (from the mainboard dts), but it hasn't been made yet (no name_node for it yet)
Help me understand. Should we just take out the sibling links all together from del? Have they not been added yet?
Thanks, Myles
ron