qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Peter Xu <peterx@redhat.com>
To: qemu-devel@nongnu.org
Cc: "Daniel P . Berrangé" <berrange@redhat.com>,
	"Eduardo Habkost" <ehabkost@redhat.com>,
	"Michael S . Tsirkin" <mst@redhat.com>,
	"Jason Wang" <jasowang@redhat.com>,
	"Markus Armbruster" <armbru@redhat.com>,
	peterx@redhat.com,
	"Dr . David Alan Gilbert" <dgilbert@redhat.com>,
	"Eric Auger" <eric.auger@redhat.com>,
	"Alex Williamson" <alex.williamson@redhat.com>,
	"Paolo Bonzini" <pbonzini@redhat.com>
Subject: [PATCH 4/4] vl: Prioritize realizations of devices
Date: Wed, 18 Aug 2021 15:43:18 -0400	[thread overview]
Message-ID: <20210818194318.110993-1-peterx@redhat.com> (raw)
In-Reply-To: <20210818194217.110451-1-peterx@redhat.com>

QEMU creates -device objects in order as specified by the user's cmdline.
However that ordering may not be the ideal order.  For example, some platform
devices (vIOMMUs) may want to be created earlier than most of the rest
devices (e.g., vfio-pci, virtio).

This patch orders the QemuOptsList of '-device's so they'll be sorted first
before kicking off the device realizations.  This will allow the device
realization code to be able to use APIs like pci_device_iommu_address_space()
correctly, because those functions rely on the platfrom devices being realized.

Now we rely on vmsd->priority which is defined as MigrationPriority to provide
the ordering, as either VM init and migration completes will need such an
ordering.  In the future we can move that priority information out of vmsd.

Signed-off-by: Peter Xu <peterx@redhat.com>
---
 softmmu/vl.c | 35 +++++++++++++++++++++++++++++++++++
 1 file changed, 35 insertions(+)

diff --git a/softmmu/vl.c b/softmmu/vl.c
index 5ca11e7469..3a30dfe27d 100644
--- a/softmmu/vl.c
+++ b/softmmu/vl.c
@@ -126,6 +126,7 @@
 #include "qapi/qmp/qerror.h"
 #include "sysemu/iothread.h"
 #include "qemu/guest-random.h"
+#include "migration/vmstate.h"
 
 #include "config-host.h"
 
@@ -2627,6 +2628,35 @@ static void qemu_init_board(void)
     }
 }
 
+/* Return the priority of the device option; zero is the default priority */
+static int qemu_opts_device_priority(QemuOpts *opts)
+{
+    const char *driver;
+    DeviceClass *dc;
+
+    driver = qemu_opt_get(opts, "driver");
+    if (!driver) {
+        return 0;
+    }
+
+    dc = qdev_get_device_class(&driver, NULL);
+    if (!dc) {
+        return 0;
+    }
+
+    if (!dc->vmsd) {
+        return 0;
+    }
+
+    /*
+     * Currently we rely on vmsd priority because that's solving the same
+     * problem for device realization ordering but just for migration.  In the
+     * future, we can move it out of vmsd, but that's not urgently needed.
+     * Return the negative of it so it'll be sorted with descendant order.
+     */
+    return -dc->vmsd->priority;
+}
+
 static void qemu_create_cli_devices(void)
 {
     soundhw_init();
@@ -2642,6 +2672,11 @@ static void qemu_create_cli_devices(void)
 
     /* init generic devices */
     rom_set_order_override(FW_CFG_ORDER_OVERRIDE_DEVICE);
+    /*
+     * Sort all the -device parameters; e.g., platform devices like vIOMMU
+     * should be initialized earlier.
+     */
+    qemu_sort_opts("device", qemu_opts_device_priority);
     qemu_opts_foreach(qemu_find_opts("device"),
                       device_init_func, NULL, &error_fatal);
     rom_reset_order_override();
-- 
2.31.1



  parent reply	other threads:[~2021-08-18 20:26 UTC|newest]

Thread overview: 52+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-18 19:42 [PATCH 0/4] vl: Prioritize device realizations Peter Xu
2021-08-18 19:42 ` [PATCH 1/4] qdev-monitor: Trace qdev creation Peter Xu
2021-08-18 19:43 ` [PATCH 2/4] qemu-config: Allow in-place sorting of QemuOptsList Peter Xu
2021-08-18 19:43 ` [PATCH 3/4] qdev: Export qdev_get_device_class() Peter Xu
2021-08-18 19:43 ` Peter Xu [this message]
2021-08-23 18:49   ` [PATCH 4/4] vl: Prioritize realizations of devices Eduardo Habkost
2021-08-23 19:18     ` Peter Xu
2021-08-23 21:07       ` Eduardo Habkost
2021-08-23 21:31         ` Peter Xu
2021-08-23 21:54           ` Michael S. Tsirkin
2021-08-23 22:51             ` Peter Xu
2021-08-23 21:56           ` Eduardo Habkost
2021-08-23 23:05             ` Peter Xu
2021-08-25  9:39               ` Markus Armbruster
2021-08-25 12:28                 ` Markus Armbruster
2021-08-25 21:50                   ` Peter Xu
2021-08-26  3:50                     ` Peter Xu
2021-08-26  8:01                       ` Markus Armbruster
2021-08-26 11:36                         ` Igor Mammedov
2021-08-26 13:43                           ` Peter Xu
2021-08-30 19:02                             ` Peter Xu
2021-08-31 11:35                               ` Markus Armbruster
2021-09-02  8:26                               ` Igor Mammedov
2021-09-02 13:45                                 ` Peter Xu
2021-09-02 13:53                                   ` Daniel P. Berrangé
2021-09-02 14:21                                     ` Peter Xu
2021-09-02 14:57                                       ` Markus Armbruster
2021-09-03 15:48                                         ` Peter Xu
2021-09-02 15:06                                       ` Daniel P. Berrangé
2021-09-02 15:26                                   ` Markus Armbruster
2021-09-03 13:00                                   ` Igor Mammedov
2021-09-03 16:03                                     ` Peter Xu
2021-09-06  8:49                                       ` Igor Mammedov
2021-09-02  7:46                             ` Igor Mammedov
2021-08-26  4:57                     ` Markus Armbruster
2021-08-23 22:05       ` Michael S. Tsirkin
2021-08-23 22:36         ` Peter Xu
2021-08-24  2:52           ` Jason Wang
2021-08-24 15:50             ` Peter Xu
2021-08-25  4:23               ` Jason Wang
2021-09-06  9:22                 ` Eric Auger
2021-08-24 16:24         ` David Hildenbrand
2021-08-24 19:52           ` Peter Xu
2021-08-25  8:08             ` David Hildenbrand
2021-08-24  2:51       ` Jason Wang
2021-10-20 13:44 ` [PATCH 0/4] vl: Prioritize device realizations David Hildenbrand
2021-10-20 13:48   ` Daniel P. Berrangé
2021-10-20 13:58     ` David Hildenbrand
2021-10-21  4:20   ` Peter Xu
2021-10-21  7:17     ` David Hildenbrand
2021-10-21  8:00       ` Peter Xu
2021-10-21 16:54         ` David Hildenbrand

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=20210818194318.110993-1-peterx@redhat.com \
    --to=peterx@redhat.com \
    --cc=alex.williamson@redhat.com \
    --cc=armbru@redhat.com \
    --cc=berrange@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=ehabkost@redhat.com \
    --cc=eric.auger@redhat.com \
    --cc=jasowang@redhat.com \
    --cc=mst@redhat.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).