From: Paolo Bonzini <pbonzini@redhat.com>
To: qemu-devel@nongnu.org
Cc: aliguori@us.ibm.com, mst@redhat.com
Subject: [Qemu-devel] [PATCH 12/15] qdev: document reset semantics
Date: Mon, 17 Dec 2012 17:24:47 +0100 [thread overview]
Message-ID: <1355761490-10073-13-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1355761490-10073-1-git-send-email-pbonzini@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
hw/qdev-core.h | 51 ++++++++++++++++++++++++++++++++++++++++++++++++++-
1 file changed, 50 insertions(+), 1 deletion(-)
diff --git a/hw/qdev-core.h b/hw/qdev-core.h
index 28f12a4..1274994 100644
--- a/hw/qdev-core.h
+++ b/hw/qdev-core.h
@@ -36,7 +36,24 @@ typedef struct DeviceClass {
Property *props;
int no_user;
- /* callbacks */
+ /* reset:
+ * @dev: Device being reset.
+ *
+ * Performs the device-specific part of a device-level ("soft") reset
+ * of @dev. This is called by QEMU internally as part of initialization
+ * or resetting a bus, but it is often accessible using a device register
+ * as well. Devices can access it using qdev_reset_all.
+ *
+ * In general, a soft reset will not reset any state that depends on the
+ * type of bus that the device resides on. For example, PCI devices do not
+ * reset their base address registers or configuration space in the reset
+ * callback. Resetting these registers is handled by the bus, not by the
+ * device.
+ *
+ * A device-level reset also includes a hard reset of all the buses exposed
+ * by @dev (and all devices below those, recursively). However, this is
+ * handled by qdev_reset_all and this callback need not care about it.
+ */
void (*reset)(DeviceState *dev);
/* device state */
@@ -86,6 +103,23 @@ struct BusClass {
* bindings can be found at http://playground.sun.com/1275/bindings/.
*/
char *(*get_fw_dev_path)(DeviceState *dev);
+
+ /* reset:
+ * @bus: Bus being reset.
+ *
+ * This callback performs a reset of @bus. It is usually done when
+ * qdev_reset_all is called on the parent of @bus.
+ *
+ * Resetting a bus includes a bus-level ("hard") reset of all devices on
+ * the bus itself. Compared to a device-level ("soft") reset, this will
+ * also remove all the configuration for devices. In the case of PCI, for
+ * example, this means the base address registers, configuration space,
+ * etc.
+ *
+ * The callback may take care of calling qdev_reset_all on all devices
+ * to perform the soft reset, or not. In the first case, it should return
+ * 1; in the second case, it should return 0 (see also qbus_walkerfn).
+ */
int (*reset)(BusState *bus);
};
@@ -186,6 +220,21 @@ int qdev_walk_children(DeviceState *dev,
qdev_walkerfn *post_devfn, qbus_walkerfn *post_busfn,
void *opaque);
+/**
+ * @qdev_reset_all:
+ * @dev: Device to be reset.
+ *
+ * Perform a device-level ("soft") reset of dev, including all buses
+ * and all devices connected to those buses. A soft reset means that
+ * qdev_reset_all will not reset any state that depends on the type of
+ * bus that the device resides on. For example, PCI devices will not
+ * reset their base address registers or configuration space.
+ *
+ * However, this is not true for device connected to @dev's buses; these are
+ * reset completely. For example, if @dev is a PCI-to-PCI bridge, the base
+ * address registers will be reset for devices on the bridge, but not for @dev
+ * itself.
+ */
void qdev_reset_all(DeviceState *dev);
/**
--
1.8.0.2
next prev parent reply other threads:[~2012-12-17 16:25 UTC|newest]
Thread overview: 60+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-12-17 16:24 [Qemu-devel] [PATCH 00/15] qdev: make reset semantics more clear and consistent, reset qbuses under virtio devices Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 01/15] qdev: do not reset a device until the parent has been initialized Paolo Bonzini
2012-12-17 16:52 ` Michael S. Tsirkin
2012-12-17 16:53 ` Michael S. Tsirkin
2012-12-17 17:06 ` Paolo Bonzini
2012-12-17 21:57 ` Andreas Färber
2012-12-17 16:24 ` [Qemu-devel] [PATCH 02/15] intel-hda: do not reset codecs from intel_hda_reset Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 03/15] pci: clean up resetting of IRQs Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 04/15] virtio-pci: reset device before PCI layer Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 05/15] virtio-s390: add a reset function to virtio-s390 devices Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 06/15] qdev: add qbus_reset_all Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 07/15] pci: do not export pci_bus_reset Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 08/15] lsi: use qbus_reset_all to reset SCSI bus Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 09/15] qdev: allow both pre- and post-order vists in qdev walking functions Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 10/15] qdev: switch reset to post-order Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 11/15] qdev: remove device_reset Paolo Bonzini
2012-12-17 16:24 ` Paolo Bonzini [this message]
2012-12-17 16:24 ` [Qemu-devel] [PATCH 13/15] virtio-pci: reset all qbuses too when writing to the status field Paolo Bonzini
2012-12-17 16:48 ` Michael S. Tsirkin
2012-12-17 16:54 ` Paolo Bonzini
2012-12-17 17:08 ` Michael S. Tsirkin
2012-12-17 17:09 ` Paolo Bonzini
2012-12-17 16:24 ` [Qemu-devel] [PATCH 14/15] virtio-s390: " Paolo Bonzini
2012-12-17 16:29 ` Alexander Graf
2012-12-17 16:24 ` [Qemu-devel] [PATCH 15/15] virtio-serial: do not perform bus reset by hand Paolo Bonzini
2012-12-17 21:43 ` [Qemu-devel] [PATCH 00/15] qdev: make reset semantics more clear and consistent, reset qbuses under virtio devices Michael S. Tsirkin
2012-12-18 7:27 ` Paolo Bonzini
2012-12-18 8:35 ` Paolo Bonzini
2012-12-18 9:49 ` Michael S. Tsirkin
2012-12-18 11:40 ` Paolo Bonzini
2013-01-07 17:46 ` Michael S. Tsirkin
2013-01-07 19:10 ` Anthony Liguori
2013-01-07 19:57 ` Peter Maydell
2013-01-07 20:20 ` Anthony Liguori
2013-01-07 20:28 ` Peter Maydell
2013-01-07 20:51 ` Anthony Liguori
2013-01-09 9:33 ` Paolo Bonzini
2013-01-09 10:22 ` Michael S. Tsirkin
2013-01-09 10:53 ` Paolo Bonzini
2013-01-09 11:09 ` Michael S. Tsirkin
2013-01-09 11:12 ` Paolo Bonzini
2013-01-09 12:10 ` Michael S. Tsirkin
2013-01-09 17:46 ` Paolo Bonzini
2013-01-09 20:40 ` Anthony Liguori
2013-01-09 21:22 ` Paolo Bonzini
2013-01-09 21:40 ` Michael S. Tsirkin
2013-01-10 8:31 ` Paolo Bonzini
2013-01-10 11:32 ` Michael S. Tsirkin
2013-01-10 11:46 ` Peter Maydell
2013-01-10 11:47 ` Paolo Bonzini
2013-01-10 11:59 ` Peter Maydell
2013-01-10 12:12 ` Paolo Bonzini
2013-01-10 12:31 ` Peter Maydell
2013-01-10 12:45 ` Paolo Bonzini
2013-01-10 13:01 ` Peter Maydell
2013-01-10 13:32 ` Paolo Bonzini
2013-01-10 14:14 ` Anthony Liguori
2013-01-10 14:38 ` Paolo Bonzini
2013-01-10 15:01 ` Michael S. Tsirkin
2013-01-08 13:58 ` Michael S. Tsirkin
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=1355761490-10073-13-git-send-email-pbonzini@redhat.com \
--to=pbonzini@redhat.com \
--cc=aliguori@us.ibm.com \
--cc=mst@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).