Messages in this thread | | | Date | Thu, 13 Mar 2008 17:49:19 +0200 | From | Muli Ben-Yehuda <> | Subject | Re: [RFC] [Patch] calgary iommu: Use the first kernel's tce tables in kdump |
| |
On Wed, Mar 12, 2008 at 02:08:06PM -0400, Vivek Goyal wrote:
> Agree. Ignoring all the DMA errors in kdump kernel does not sound > very good. > > On a side note, typically how much time does it take to for DMA > operations to finish. Can we wait for a random amount of time in > second kernel, hoping all the DMA operations are complete, and then > setup a new table? > > This can be atleast a stop gap solution till we come up with a good > solution?
The problem as I understand it is that when we boot into the kdump kernel we have left-over devices from the old kernel which are active and initiating DMA's. This is fine as long as Calgayr is also active (with the old kernel's mappings). Once we initialize Calgary (in the kdump) kernel with the new mappings, however, it starts trapping all those DMA's. We can't wait for them to stop, because that will only happen when the device is re-initialized (or otherwise quiesced).
Cheers, Muli
| |