| |
|
 |
|
 |
|
|
|
| |
|
|
xen-devel
Re: [Xen-devel] Porting linux network driver to Xen
Thank you Kaushik for your help. Very much appreciated.
Thanks
-RK
On Wed, Nov 4, 2009 at 3:07 PM, Kaushik Kumar Ram <kaushik@xxxxxxxx> wrote:
I am confused now, probably I should ask what changes are needed such that I can access NIC from guest domains(domUs) for Tx and Rx? i.e if I compile native linux driver and run on dom0, how can I access it from domU's
are the changes just the configuration such that path is setup from domU/netfront -> dom0/netback->Bridge->NIC?
Yes. Just make sure the NIC corresponding to this driver is attached to the bridge. If this NIC is not your default interface you may need to tweak the configuration.
-Kaushik
On Wed, Nov 4, 2009 at 1:47 PM, Kaushik Kumar Ram <kaushik@xxxxxxxx> wrote:
Again, you don't have to make any changes to the driver. Netback/netfront do not interact directly with the
NIC's driver.
-Kaushik On Nov 4, 2009, at 3:18 PM, ravi kerur wrote: Thanks kaushik, How does netfront/netback interact with this driver? I believe I need netback functionality running in dom0 and netfront running in domU's for the existing driver. One of the documents has following explanation for Tx/Rx path.
"For transmit (TX)
requests the driver domain(dom0) uses a hypercall to map
the guest page(domU) into its
address space before sending the request through the
bridge. When the physical device driver frees the
page a callback function is automatically
invoked to return a response to netfront
which then revokes the grant.
For RX requests netfront posts I/O buffer page grants
to the RX I/O channel. When netback receives a packet from the
bridge it retrieves a posted grant from the I/O channel
and issues a grant copy hypercall to copy the
packet to the guest page(domU). Finally, netback sends a response
to the guest
via the RX channel indicating a packet is available."
I am not quite sure what changes or api's I need to call such that existing Tx/Rx buffers(rings) used for DMA in the native drivers starts talking via netfront/netback.
Inputs appreciated.
Thanks
-RK
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
|
|
| |
|
Home