On Tue, Mar 4, 2008 at 5:35 PM, Carl-Daniel Hailfinger c-d.hailfinger.devel.2006@gmx.net wrote:
It complains about left over devices, but the only leftover device is not a PCI device as far as the logs seem to tell. Could be a bug, could be that the db800 and norwich have different sets of PCI devices showing up...
It is a factor of the pci-centric nature of the old v2 device tree.
Needs cleanup.
Hm. Need to find out why dev_find_device on db800 is called 3 times for 1022:2097 and 2 times for 1022:2096, all of them in phase 6. On Norwich, dev_find_device is called 2 times for 1022:2097 and once for 1022:2096.
db800.txt:dev_find_device: find PCI: 1022:2090 db800.txt:dev_find_device: find PCI: 1022:2090 db800.txt:dev_find_device: find PCI: 1022:2095 db800.txt:dev_find_device: find PCI: 1022:2097 db800.txt:dev_find_device: find PCI: 1022:2096 db800.txt:dev_find_device: find PCI: 1022:2097 db800.txt:dev_find_device: find PCI: 1022:2096 db800.txt:dev_find_device: find PCI: 1022:2097 norwich.txt:dev_find_device: find PCI: 1022:2090 norwich.txt:dev_find_device: find PCI: 1022:2090 norwich.txt:dev_find_device: find PCI: 1022:2095 norwich.txt:dev_find_device: find PCI: 1022:2097 norwich.txt:dev_find_device: find PCI: 1022:2096 norwich.txt:dev_find_device: find PCI: 1022:2097
More analysis later.
it's an nxm algorithm. I am pretty sure those multilple prints of same device are the devices in the tree it is trying to match against, not the device it is called with.
ron