Complete WSL implementation in Podman 5

[n1hility]

Complete WSL implementation, refactor a few areas

Also addresses a number of issues:
- StopHostNetworking isn't plumbed, win-sshproxy leaks on hyperv
- Wait api and print output doesn't work properly on Windows
- API forwarding doesn't work on WSL
- Terminal corruption with after start/stop on Windows
- Gvproxy is forcefully killed vs gracefully quit
- Switching rootful/rootless does not update /var/run/docker.sock on the guest
- File already closed error on init
- HyperV backend is publishing Unix sockets when it should be named pipes
- User-mode networking doesn't always work
- Stop state outside of lock boundaries
- WSL blocks parallel machined (should be supported)

none

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: n1hility

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [n1hility]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[n1hility]

/hold

[packit-as-a-service]

Ephemeral COPR build failed. @containers/packit-build please check.

[baude]

the stop failure is legit problem ...

[baude]

@n1hility this patch resolves the test problem. My only pause is I am not dead sure if I am treating a symptom instead of fixing the real problem. But on that front, I do not see something like a missed .Close() ... maybe take a quick peek before applying my patch. The patch also explains whats going on where the cause/effect? is ... maybe someone else will see it too.

diff --git a/pkg/machine/qemu/machine.go b/pkg/machine/qemu/machine.go
index 1bcb46ce46..65040ed457 100644
--- a/pkg/machine/qemu/machine.go
+++ b/pkg/machine/qemu/machine.go
@@ -6,6 +6,7 @@ import (
        "encoding/json"
        "errors"
        "fmt"
+       "io"
        "io/fs"
        "os"
        "os/exec"
@@ -266,6 +267,21 @@ func (q *QEMUStubber) State(mc *vmconfigs.MachineConfig, bypass bool) (define.St
                return "", err
        }
        if err := monitor.Connect(); err != nil {
+               // There is a case where if we stop the same vm (from running) two
+               // consecutive times we can get an econnreset when trying to get the
+               // state
+               if errors.Is(err, syscall.ECONNRESET) {
+                       // try again
+                       logrus.Debug("received ECCONNRESET from QEMU monitor; trying again")
+                       secondTry := monitor.Connect()
+                       if errors.Is(secondTry, io.EOF) {
+                               return define.Stopped, nil
+                       }
+                       if secondTry != nil {
+                               logrus.Debugf("second attempt to connect to QEMU monitor failed")
+                               return "", secondTry
+                       }
+               }
                return "", err
        }
        defer func() {

[n1hility]

@baude thanks! I applied your patch and rebased again. i mentioned in chat but just in case you miss it that i noticed some underlying raciness in qemu start/stop that we can fixup later (e.g. the logic has the netdev sock setup overwriting the qmp monitor sock which confuse client connection code)

[baude]

LGTM

[arixmkii]

@n1hility regarding this comment #21597 (comment) Why don't just eliminate the root cause and don't reuse QMP for VLAN in QEMU. It is what I proposed in #21594 with a bigger change. But this could be also achieved w/o removing this FD. Just use dedicated socket for the purpose.

[ashley-cui]

LGTM other than the prev reviews.

[n1hility]

@n1hility regarding this comment #21597 (comment) Why don't just eliminate the root cause and don't reuse QMP for VLAN in QEMU. It is what I proposed in #21594 with a bigger change. But this could be also achieved w/o removing this FD. Just use dedicated socket for the purpose.

I totally agree we need to fix this, and soon, but I also don't want this PR to grow beyond what it is (already pretty huge) and its holding up WSL working in 5. That and other stuff can come in other commits.

[n1hility]

I added the small comment review charge in a separate commit to address the review, yet still show the record of CI passing on the earlier commits.

[baude]

i think there is consensus to get this PR in after @n1hility addressed the first round of comments ...

/hold
/lgtm

[Luap99]

LGTM but for the size it is hard to properly review so it is hard to catch things.
Looks like windows CI is still disabled, does that mean we can turn it on after this merges?

[baude]

/hold cancel