Re: [Xen-devel] [PATCH] iommu: leave IOMMU enabled by default during kexec crash transition

[Igor Druzhinin <igor.druzhinin@xxxxxxxxxx>]

On 20/02/2019 08:48, Jan Beulich wrote:
> 
> Some entity needs to decide whether to add the respective command
> line option to the crash kernel's command line. It should be this same
> entity to tell Xen whether to keep the IOMMU enabled while invoking
> the crash kernel. I am merely guessing that this entity is the kexec
> tool.
> 

I was just double checking and it seems (assuming the device reset
correctly in the crash kernel) everything seem to work even without
enabling intel_iommu in the command line - newer kernels handle this
case by explicitly disabling translation in any case: they expect it to
be enabled by the previous kernel.

I intended Xen command line option be more of a fallback mechanism for
those still using ancient kdump kernels. For all the others the
transition should be transparent. With the above said, having an
orchestration tool on top of that seems a bit of overkill.

>> I'm not following your last question. Could you elaborate?
> 
> I'm simply asking what the bare metal equivalent behavior is here,
> i.e. how command line option addition and IOMMU state are being
> kept in sync in that case. Just for reference, in the hope that what
> they do is sane, and hence we could follow that model.
> 

As mentioned above, Linux (as the main kernel) currently only supports
secondary kernels that expect IOMMU to be enabled and have basic DMAR
understanding which means crash kernel is supposed to be up to date with
the main one. On the other hand, Linux (as a crash kernel) expects the
main kernel (which might be Xen) to do either of things: keep IOMMU
enabled or disabled - both cases are handled.

Igor


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel