Comment 12 for bug 415353

I am a bit confused by the comments: The last comments about the long delays (20-30 min!) was made with 2.6.32-20, a pre-release kernel. There was a fix about lost MSI-X interrupts in 2.6.32-23.37 and two more in the current proposed kernel (2.6.32-26.46), but those seem to address a netpoll crash and a hang.
Timo's comment is that it takes 30s(!) on every boot, which would be the good case that Mattias mentions (5s per interface with 8 interfaces would be 40s which is still a while but likely related to sequential probing with timeouts. If I got the right define the wait_ramrod message tells that it waited for 5s for the state to be FP_HALTED).
So, anyway, it somehow looks to me like Timo is saying Lucid is at least as fixed as Maverick now. Am I misreading things?