From: Stefan Hajnoczi <stefanha@gmail.com>
To: Maxim Levitsky <mlevitsk@redhat.com>
Cc: "Fam Zheng" <fam@euphon.net>,
"Daniel P. Berrangé" <berrange@redhat.com>,
"Eduardo Habkost" <ehabkost@redhat.com>,
"Michael S. Tsirkin" <mst@redhat.com>,
qemu-devel@nongnu.org, "Paolo Bonzini" <pbonzini@redhat.com>
Subject: Re: [PATCH v2 4/7] device-core: use atomic_set on .realized property
Date: Wed, 27 May 2020 16:00:53 +0100 [thread overview]
Message-ID: <20200527150053.GJ29137@stefanha-x1.localdomain> (raw)
In-Reply-To: <20200511160951.8733-5-mlevitsk@redhat.com>
[-- Attachment #1: Type: text/plain, Size: 2093 bytes --]
On Mon, May 11, 2020 at 07:09:48PM +0300, Maxim Levitsky wrote:
> Some code might race with placement of new devices on a bus.
> We currently first place a (unrealized) device on the bus
> and then realize it.
>
> As a workaround, users that scan the child device list, can
> check the realized property to see if it is safe to access such a device.
> Use an atomic write here too to aid with this.
>
> A separate discussion is what to do with devices that are unrealized:
> It looks like for this case we only call the hotplug handler's unplug
> callback and its up to it to unrealize the device.
> An atomic operation doesn't cause harm for this code path though.
>
> Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
> ---
> hw/core/qdev.c | 15 ++++++++++++++-
> 1 file changed, 14 insertions(+), 1 deletion(-)
Please add a comment to struct DeviceState saying the realized field
must be accessed with atomic_load_acquire() when used outside the QEMU
global mutex.
> diff --git a/hw/core/qdev.c b/hw/core/qdev.c
> index 732789e2b7..d530c5922f 100644
> --- a/hw/core/qdev.c
> +++ b/hw/core/qdev.c
> @@ -964,7 +964,20 @@ static void device_set_realized(Object *obj, bool value, Error **errp)
> }
> }
>
> + atomic_store_release(&dev->realized, value);
> +
> } else if (!value && dev->realized) {
> +
> + /*
> + * Change the value so that any concurrent users are aware
> + * that the device is going to be unrealized
> + *
> + * TODO: change .realized property to enum that states
> + * each phase of the device realization/unrealization
> + */
> +
> + atomic_store_release(&dev->realized, value);
I'm not sure if atomic_store_release() is strong enough in the true ->
false case:
Operations coming after ``atomic_store_release()`` can still be
reordered before it.
A reader may already seen changes made to unrealize the DeviceState even
though realized still appears to be true. A full write memory barrier
seems safer here.
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]
next prev parent reply other threads:[~2020-05-27 15:02 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-11 16:09 [PATCH v2 0/7] RFC/WIP: Fix scsi devices plug/unplug races w.r.t virtio-scsi iothread Maxim Levitsky
2020-05-11 16:09 ` [PATCH v2 1/7] scsi/scsi_bus: switch search direction in scsi_device_find Maxim Levitsky
2020-05-27 12:53 ` Stefan Hajnoczi
2020-05-11 16:09 ` [PATCH v2 2/7] Implement drain_call_rcu and use it in hmp_device_del Maxim Levitsky
2020-05-27 13:11 ` Stefan Hajnoczi
2020-07-09 9:34 ` Maxim Levitsky
2020-07-09 11:42 ` Markus Armbruster
2020-07-09 11:56 ` Maxim Levitsky
2020-07-09 12:02 ` Paolo Bonzini
2020-05-11 16:09 ` [PATCH v2 3/7] device-core: use RCU for list of childs of a bus Maxim Levitsky
2020-05-27 14:45 ` Stefan Hajnoczi
2020-07-09 9:40 ` Maxim Levitsky
2020-05-11 16:09 ` [PATCH v2 4/7] device-core: use atomic_set on .realized property Maxim Levitsky
2020-05-27 15:00 ` Stefan Hajnoczi [this message]
2020-07-09 10:24 ` Maxim Levitsky
2020-05-11 16:09 ` [PATCH v2 5/7] virtio-scsi: don't touch scsi devices that are not yet realized or about to be un-realized Maxim Levitsky
2020-05-27 15:08 ` Stefan Hajnoczi
2020-05-11 16:09 ` [PATCH v2 6/7] scsi: Add scsi_device_get Maxim Levitsky
2020-05-27 15:27 ` Stefan Hajnoczi
2020-07-09 10:35 ` Maxim Levitsky
2020-05-11 16:09 ` [PATCH v2 7/7] virtio-scsi: use scsi_device_get Maxim Levitsky
2020-05-27 15:50 ` Stefan Hajnoczi
2020-07-09 10:43 ` Maxim Levitsky
2020-05-27 15:50 ` Stefan Hajnoczi
2020-05-11 18:03 ` [PATCH v2 0/7] RFC/WIP: Fix scsi devices plug/unplug races w.r.t virtio-scsi iothread no-reply
2020-05-11 18:03 ` no-reply
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=20200527150053.GJ29137@stefanha-x1.localdomain \
--to=stefanha@gmail.com \
--cc=berrange@redhat.com \
--cc=ehabkost@redhat.com \
--cc=fam@euphon.net \
--cc=mlevitsk@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).