qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Damien Hedde <damien.hedde@greensocs.com>
To: qemu-devel@nongnu.org
Cc: mark.burton@greensocs.com, edgari@xilinx.com,
	"Damien Hedde" <damien.hedde@greensocs.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>,
	"Daniel P. Berrangé" <berrange@redhat.com>,
	"Eduardo Habkost" <eduardo@habkost.net>
Subject: [PATCH v5 2/6] machine&vl: introduce phase_until() to handle phase transitions
Date: Thu, 19 May 2022 17:33:58 +0200	[thread overview]
Message-ID: <20220519153402.41540-3-damien.hedde@greensocs.com> (raw)
In-Reply-To: <20220519153402.41540-1-damien.hedde@greensocs.com>

phase_until() is implemented in vl.c and is meant to be used
to make startup progress up to a specified phase being reached().
At this point, no behavior change is introduced: phase_until()
only supports a single double transition corresponding
to the functionality of qmp_exit_preconfig():
+ accel-created -> machine-initialized -> machine-ready

As a result qmp_exit_preconfig() now uses phase_until().

This commit is a preparation to support cold plugging a device
using qapi command (which will be introduced in a following commit).
For this we need fine grain control of the phase.

Signed-off-by: Damien Hedde <damien.hedde@greensocs.com>
---

v5:
  + refactor to avoid indentation change
---
 include/hw/qdev-core.h | 14 +++++++++++++
 softmmu/vl.c           | 46 ++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 60 insertions(+)

diff --git a/include/hw/qdev-core.h b/include/hw/qdev-core.h
index e29c705b74..5f73d06408 100644
--- a/include/hw/qdev-core.h
+++ b/include/hw/qdev-core.h
@@ -909,4 +909,18 @@ extern bool phase_check(MachineInitPhase phase);
  */
 extern void phase_advance(MachineInitPhase phase);
 
+/**
+ * @phase_until:
+ * @phase: the target phase
+ * @errp: error report
+ *
+ * Make the machine init progress until the target phase is reached.
+ *
+ * Its is a no-op is the target phase is the current or an earlier
+ * phase.
+ *
+ * Returns true in case of success.
+ */
+extern bool phase_until(MachineInitPhase phase, Error **errp);
+
 #endif
diff --git a/softmmu/vl.c b/softmmu/vl.c
index 84a31eba76..7f8d15b5b8 100644
--- a/softmmu/vl.c
+++ b/softmmu/vl.c
@@ -2702,11 +2702,17 @@ void qmp_x_exit_preconfig(Error **errp)
         error_setg(errp, "The command is permitted only before machine initialization");
         return;
     }
+    phase_until(PHASE_MACHINE_READY, errp);
+}
 
+static void qemu_phase_ready(Error **errp)
+{
     qemu_init_board();
+    /* phase is now PHASE_MACHINE_INITIALIZED. */
     qemu_create_cli_devices();
     cxl_fixed_memory_window_link_targets(errp);
     qemu_machine_creation_done();
+    /* Phase is now PHASE_MACHINE_READY. */
 
     if (loadvm) {
         load_snapshot(loadvm, NULL, false, NULL, &error_fatal);
@@ -2729,6 +2735,46 @@ void qmp_x_exit_preconfig(Error **errp)
     }
 }
 
+bool phase_until(MachineInitPhase phase, Error **errp)
+{
+    ERRP_GUARD();
+    if (!phase_check(PHASE_ACCEL_CREATED)) {
+        error_setg(errp, "Phase transition is not supported until accelerator"
+                   " is created");
+        return false;
+    }
+
+    while (!phase_check(phase)) {
+        MachineInitPhase cur_phase = phase_get();
+
+        switch (cur_phase) {
+        case PHASE_ACCEL_CREATED:
+            qemu_phase_ready(errp);
+            break;
+
+        default:
+            /*
+             * If we end up here, it is because we miss a case above.
+             */
+            error_setg(&error_abort, "Requested phase transition is not"
+                       " implemented");
+            return false;
+        }
+
+        if (*errp) {
+            return false;
+        }
+
+        /*
+         * Ensure we made some progress.
+         * With the default case above, it should be enough to prevent
+         * any infinite loop.
+         */
+        assert(cur_phase < phase_get());
+    }
+    return true;
+}
+
 void qemu_init(int argc, char **argv, char **envp)
 {
     QemuOpts *opts;
-- 
2.36.1



  parent reply	other threads:[~2022-05-19 15:44 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-05-19 15:33 [PATCH v5 0/6] QAPI support for device cold-plug Damien Hedde
2022-05-19 15:33 ` [PATCH v5 1/6] machine: add phase_get() and document phase_check()/advance() Damien Hedde
2022-05-24 19:56   ` Jim Shu
2022-05-19 15:33 ` Damien Hedde [this message]
2022-05-24 19:56   ` [PATCH v5 2/6] machine&vl: introduce phase_until() to handle phase transitions Jim Shu
2022-05-19 15:33 ` [PATCH v5 3/6] vl: support machine-initialized target in phase_until() Damien Hedde
2022-05-24 19:56   ` Jim Shu
2022-05-19 15:34 ` [PATCH v5 4/6] qapi/device_add: compute is_hotplug flag Damien Hedde
2022-05-24 19:57   ` Jim Shu
2022-05-19 15:34 ` [PATCH v5 5/6] RFC qapi/device_add: handle the rom_order_override when cold-plugging Damien Hedde
2022-05-19 15:34 ` [PATCH v5 6/6] qapi/device_add: Allow execution in machine initialized phase Damien Hedde
2022-05-24 19:58   ` Jim Shu

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20220519153402.41540-3-damien.hedde@greensocs.com \
    --to=damien.hedde@greensocs.com \
    --cc=berrange@redhat.com \
    --cc=edgari@xilinx.com \
    --cc=eduardo@habkost.net \
    --cc=mark.burton@greensocs.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).