Announce: dumpfs v0.01 - common RAS output API
Quoted 5 lines · Andrew
btw, if we simply insert a five-second-pause, what problems does that leave? Network Rx, which is OK. Disk writes will have completed (?). What remains?
I'd disagree: as soon as the IOMMU is reconfigured/reused Network Rx could become a problem: it could point to new io areas of the kexec'ed kernel.
Btw, what's the preferred approach to clear the pci master bit: forcedeth writes to freed buffers after ifdown right now. I'll add a reset into the _close function, but disabling the master bit is probably better.
--
Manfred
Quoted 4 lines · Manfred Spraul
Btw, what's the preferred approach to clear the pci master bit:
forcedeth writes to freed buffers after ifdown right now. I'll add a reset into the_closefunction, but disabling the master bit is probably better.
Can you stick the forcedeth chip into S3 state on close - that might be more direct providing you can get it back again