ice(4): Add host SR-IOV support #1573
Conversation
Would you be willing to update the manual page for this change? |
|
If I was sufficiently motivated, I could maybe take a crack at it. @kgalazka have you tried this on E830 devices? The last time I tried it didn't work; I don't know if it's worth calling out somewhere that it doesn't work on those since E830 hasn't been released yet, but I imagine once they are people will expect this to work on those and will be surprised that it mysteriously does not. |
| @@ -369,6 +369,14 @@ ice_iov_add_vf(struct ice_softc *sc, uint16_t vfnum, const nvlist_t *params) | |||
| goto release_imap; | |||
| } | |||
|
|
|||
| /* Add the broadcast address */ | |||
| error = ice_add_vsi_mac_filter(vsi, broadcastaddr); | |||
There was a problem hiding this comment.
Doing this here might mean that this doesn't get re-added after a reset, right? Though, I want to swear that broadcast promiscuous mode should be getting enabled on this VF VSI already.
There was a problem hiding this comment.
Doing this here might mean that this doesn't get re-added after a reset, right?
Yeah, it's not the best place and I intend to move it.
Though, I want to swear that broadcast promiscuous mode should be getting enabled on this VF VSI already.
Not in ICE as far as I know. We do that in IXL though.
kgalazka
force-pushed
the
ice_iov_support
branch
from
82fdeba to
e0e855e
Compare
Not there yet. I'm trying to make it ready for full blown testing on E810 first. |
kgalazka
force-pushed
the
ice_iov_support
branch
from
e0e855e to
a47886e
Compare
Enable basic SR-IOV support for E800 adapters. Authored-by: Eric Joyner <erj@FreeBSD.org> Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
Filter for unicast MAC address is configured with a virtual channel message, but filter for a broadcast traffic was missing. It caused issues with e.g. ARP. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
VF driver may request to configure MSI-X vectors for less queues than assigned by PF. Don't try to configure unassigned vectors to avoid panic. While at that make the loop process whole array of vectors received in a VIRTCHNL_OP_CFG_IRQ_MAP message from a VF. It's not guarantied that vector '0', which is used for other interrupt causes and is not mapped to a queue, will be always on the last position. Condition inside the loop already handles that vector correctly. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
PF driver needs to tear down and setup VF configuration after a reset event, e.g. due to reloading a VF driver. Re-use ice_reset_vf function for that by adding new parameter, which decides if new reset has to be triggered. This most likely does not cover all necessary steps and will be extended in future commits. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
On receiving a virtual channel request from VF driver tried to configure and enable Tx and Rx queues without making sure that they were disabled. It caused issue with reloading a VF driver without a reset e.g. in case it crashed. Fix that by always disabling all Rx and Tx queues. While at that make sure that only queues requested by VF driver are enabled. VF driver may use less queues than assigned to the function when it was created. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
Remove existing RSS HENA configuration to make sure that only config requested by VF is applied and allow VFs to disable RSS completely. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
Let VF driver know that HW supports WB_ON_ITR feature. It allows polling mode driver such as DPDK to configure an immediate write-back of completed descriptors. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
Adds two new parameters to iovctl config for VLAN filter limits and MAC filter limits and gives them defaults of 64 and 16, respectively. These are intended to limit the number of resources that a VF can consume so that any one VF cannot starve an other VFs or the PF of filters. Signed-off-by: Eric Joyner <eric.joyner@intel.com> Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
MAC address assigned to a Virtual Function using iovctl configuration was not passed to the VF, but its presence caused ICE to reject random address generated by the VF. This lead to VF not being able to send or recieve traffic. Send the administratively assigned MAC address to a VF in a response to the 'Get VF Resources' virtual channel command. While at that replace bcopy with memcpy for consistency, and shorten an over 80 characters line. Signed-off-by: Krzysztof Galazka <krzysztof.galazka@intel.com>
kgalazka
force-pushed
the
ice_iov_support
branch
from
7baca85 to
f6bdd84
Compare
Allows creating Virtual Functions on E800 adapters with up to 16 queues.
Changing MAC address and using promiscuous mode for each VF
may be enabled in configuration.
Please be advised that this is still work in progress and may be unstable.
Authored-by: Eric Joyner erj@FreeBSD.org
Signed-off-by: Krzysztof Galazka krzysztof.galazka@intel.com