* Alexandra N. Kossovsky ([email protected]) wrote:
> It is not clear from the documentation if iommu_map()/iommu_unmap()
> functions may be called from atomic context.
> In case of Intel IOMMU, it works; in case of AMD IOMMU, it does not.
> 
> Was it done by purpose?  May I propose a patch for AMD IOMMU replacing
> GFP_KERNEL by GFP_ATOMIC to make things better?

Looks like omap calls iopte_alloc which does GFP_KERNEL allocation too.

> We use IOMMU API in OpenOnload project http://www.openonload.org/,
> and we get better latency with Intel IOMMU because we are not
> forced to use threaded IRQ.

Have you looked at VFIO (and possibly libusnic is still up to date)?

thanks,
-chris
_______________________________________________
iommu mailing list
[email protected]
https://lists.linuxfoundation.org/mailman/listinfo/iommu

Reply via email to