Skip to content

set-features should be idempotent? #1053

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Draft
iximeow wants to merge 1 commit into
base: master
Choose a base branch
from
Draft

set-features should be idempotent? #1053

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
26 changes: 19 additions & 7 deletions lib/propolis/src/hw/virtio/viona.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@ use std::io::{self, Error, ErrorKind};
use std::num::NonZeroU16;
use std::os::unix::io::{AsRawFd, RawFd};
use std::sync::{Arc, Condvar, Mutex, Weak};
use std::sync::atomic::{Ordering, AtomicBool};
Copy link
Member

@hawkw hawkw Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

clippy is kind of obnoxiously mad about alphabetization here :|


use crate::common::{RWOp, ReadOp};
use crate::hw::pci;
Expand Down Expand Up @@ -308,6 +309,7 @@ pub struct PciVirtioViona {
mac_addr: [u8; ETHERADDRL],
mtu: Option<u16>,
hdl: VionaHdl,
mq_active: AtomicBool,
Copy link
Member

@hawkw hawkw Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

would like a comment on this explaining what it means and why it's tracked here?

Copy link
Member

@hawkw hawkw Feb 9, 2026

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

i found the description of the suspected issue in #1048 and #1052 quite useful and felt like something like that (though not necessarily from the angle of "it was broken") would be nice to capture closer to the code.

inner: Mutex<Inner>,
}

Expand Down Expand Up @@ -396,6 +398,7 @@ impl PciVirtioViona {
dev_features,
mac_addr: [0; ETHERADDRL],
mtu: info.mtu,
mq_active: AtomicBool::new(false),
hdl,
inner: Mutex::new(Inner::new(nqueues)),
};
Expand Down Expand Up @@ -735,13 +738,21 @@ impl VirtioDevice for PciVirtioViona {

fn set_features(&self, feat: u64) -> Result<(), ()> {
self.hdl.set_features(feat).map_err(|_| ())?;
if (feat & VIRTIO_NET_F_MQ) != 0 {
self.hdl.set_pairs(PROPOLIS_MAX_MQ_PAIRS).map_err(|_| ())?;
probes::virtio_viona_mq_set_use_pairs!(|| (
MqSetPairsCause::MqEnabled as u8,
PROPOLIS_MAX_MQ_PAIRS
));
self.set_use_pairs(PROPOLIS_MAX_MQ_PAIRS)?;
let want_mq = feat & VIRTIO_NET_F_MQ != 0;

if want_mq {
// This might be the first we're hearing from the guest about
// wanting multi-queue. If it is, we'll have a bit of work to do.
let mq_active = self.mq_active.swap(true, Ordering::Relaxed);

if !mq_active {
self.hdl.set_pairs(PROPOLIS_MAX_MQ_PAIRS).map_err(|_| ())?;
probes::virtio_viona_mq_set_use_pairs!(|| (
MqSetPairsCause::MqEnabled as u8,
PROPOLIS_MAX_MQ_PAIRS
));
self.set_use_pairs(PROPOLIS_MAX_MQ_PAIRS)?;
}
}
Ok(())
}
Expand Down Expand Up @@ -838,6 +849,7 @@ impl Lifecycle for PciVirtioViona {
));
self.set_use_pairs(1).expect("can set viona back to one queue pair");
self.hdl.set_pairs(1).expect("can set viona back to one queue pair");
self.mq_active.store(false, Ordering::Relaxed);
self.virtio_state.queues.reset_peak();
}
fn start(&self) -> anyhow::Result<()> {
Expand Down
Loading