El 26/2/16 a les 9:13, Gustau Pérez ha escrit:
> 
> 
> El 25/02/16 a les 20:30, Roger Pau Monné ha escrit:
>> Thanks, I've also realized about this error and fixed it in new version.
>> I've just send the libxl side of this work for review:
>>
>> http://lists.xenproject.org/archives/html/xen-devel/2016-02/msg03669.html
>>
>> I've also realized that at least one change is needed in the hypervisor
>> for this to work properly, I will prepare the series so that you can
>> apply them to the xen-kernel and xen-tools ports, together with the
>> FreeBSD blkback changes that are needed for this to work. I will
>> hopefully be able to send them tomorrow morning.
>>
>    Hi Roger,
> 
>    thank you for your efforts. I'll wait for your patches.

Hello,

I've rebased everything on top of Xen 4.5, so it can be applied to the
xen-tools and xen-kernel packages without issues, I've uploaded all the
patches to:

https://people.freebsd.org/~royger/hotplug/

As you can see, each folder contains the patches that should be applied
to each component. With this I've been able to use the simple block
hotplug script that's included in the series.

There's a limitation with block hotplug scripts, that is not specific to
FreeBSD (the same applies to Linux for example). You cannot use block
hotplug scripts with HVM guests. This is because HVM guests require a
QEMU instance, and in order to provide a virtual disk to QEMU it must be
a file or block device present on the system. Then when using block
hotplug scripts the block device can only be made available using what
we call a local attach (attaching the virtual disk to Dom0), and libxl
doesn't do that yet. Of course patches are very welcome to remove this
limitation ;). The code to perform local attaching to Dom0 is already in
libxl, it's just not used for this case.

Roger.
_______________________________________________
freebsd-xen@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-xen
To unsubscribe, send any mail to "freebsd-xen-unsubscr...@freebsd.org"

Reply via email to