viona: reset device more completely #1033
Merged
+32
−25
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
when I reboot a Debian 13 guest the NIC works again, but judging by viona.d output I think it's by.. accident? The reboot `SET_PAIRS 0xb` still returns 0x10 (EBUSY), and there's no `SET_USEPAIRS 0xb`. instead, there's a `SET_USEPAIRS 0x1`? so I think this fails to actually get all the viona rings back to VRS_RESET and I don't think the guest actually configures multiqueue on the reboot.
iximeow
added
networking
iximeow
force-pushed
the
not-really-a-fix-for-1030
branch
from
f55650b to
53e0874
Compare
iximeow
changed the title
dancrossnyc
approved these changes
Feb 6, 2026
Contributor
dancrossnyc
left a comment
dancrossnyc
left a comment
There was a problem hiding this comment.
I think this is right. Thanks!
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
when I reboot a Debian 13 guest the NIC works again, but judging by viona.d output I think it's by.. accident? The rebootI think what I was seeing is that the guest negotiates multi-queue, configures the queues, then sets features again with the same bits (including MQ). We then try to SET_PAIRS to viona again but the queues are already out of reset because the guest just set them up. So our SET_PAIRS to the same value we'd set before fails, we don't SET_USEPAIRS to the same value again, and things continue "happily". This happens on first boot too, I just didn't see it above all theSET_PAIRS 0xbstill returns 0x10 (EBUSY), and there's noSET_USEPAIRS 0xb. instead, there's aSET_USEPAIRS 0x1?RING_SET_MSIand such.@dancrossnyc suggested that it would be better to have a method on
VirtQueuesto go reset all the viona rings regardless of whatVirtQueues.len()happens to be, which works out well.I noticed along the way that we don't SET_PAIRS back down to 1 on reset. I think in the case that a guest would boot with a legacy driver, reconfigure the NIC for multiqueue, then reboot to a legacy driver (I'm thinking iPXE -> Linux -> reboot as a hypothesis), the driver would be pretty confused? So I'm SET_PAIRS and SET_USEPAIRS back to 1 on reset to get closer to the initial device state.