From: Paolo Bonzini <pbonzini@redhat.com>
To: alexander_barabash@mentor.com
Cc: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH] qom: In function object_set_link_property(), first call object_ref(), then object_unref().
Date: Wed, 22 Feb 2012 18:17:18 +0100 [thread overview]
Message-ID: <4F45231E.2030600@redhat.com> (raw)
In-Reply-To: <1329930831-26837-1-git-send-email-alexander_barabash@mentor.com>
On 02/22/2012 06:13 PM, alexander_barabash@mentor.com wrote:
> From: Alexander Barabash <alexander_barabash@mentor.com>
>
> In the old implementation, if the new value of the property links
> to the same object, as the old value, that object is first unref-ed,
> and then ref-ed. This leads to unintended deinitialization of that object.
>
> In the new implementation, this is fixed.
>
> Signed-off-by: Alexander Barabash <alexander_barabash@mentor.com>
> ---
> qom/object.c | 18 +++++++++++++-----
> 1 files changed, 13 insertions(+), 5 deletions(-)
>
> diff --git a/qom/object.c b/qom/object.c
> index 941c291..d1b3ac7 100644
> --- a/qom/object.c
> +++ b/qom/object.c
> @@ -892,19 +892,19 @@ static void object_set_link_property(Object *obj, Visitor *v, void *opaque,
> const char *name, Error **errp)
> {
> Object **child = opaque;
> + Object *old_target;
> bool ambiguous = false;
> const char *type;
> char *path;
> gchar *target_type;
> + bool clear_old_target = true;
>
> type = object_property_get_type(obj, name, NULL);
>
> visit_type_str(v, &path, name, errp);
>
> - if (*child) {
> - object_unref(*child);
> - *child = NULL;
> - }
> + old_target = *child;
> + *child = NULL;
You can just remove the unref here...
> if (strcmp(path, "") != 0) {
> Object *target;
> @@ -916,7 +916,11 @@ static void object_set_link_property(Object *obj, Visitor *v, void *opaque,
> if (ambiguous) {
> error_set(errp, QERR_AMBIGUOUS_PATH, path);
> } else if (target) {
> - object_ref(target);
> + if (target != old_target) {
> + object_ref(target);
... leave the unconditional ref to target here...
> + } else {
> + clear_old_target = false;
> + }
> *child = target;
> } else {
> target = object_resolve_path(path, &ambiguous);
> @@ -930,6 +934,10 @@ static void object_set_link_property(Object *obj, Visitor *v, void *opaque,
> }
>
> g_free(path);
> +
> + if (clear_old_target && (old_target != NULL)) {
> + object_unref(old_target);
... and leave this unref on old_target, without the need for
clear_old_target.
> + }
> }
>
> void object_property_add_link(Object *obj, const char *name,
Paolo
next prev parent reply other threads:[~2012-02-22 17:17 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-02-22 17:13 [Qemu-devel] [PATCH] qom: In function object_set_link_property(), first call object_ref(), then object_unref() alexander_barabash
2012-02-22 17:17 ` Paolo Bonzini [this message]
2012-02-22 17:19 ` Alexander Barabash
-- strict thread matches above, loose matches on Subject: below --
2012-02-22 17:22 alexander_barabash
2012-02-22 17:24 ` Paolo Bonzini
2012-02-22 17:25 ` Alexander Barabash
2012-02-22 17:52 ` Eric Blake
2012-02-24 15:32 ` Anthony Liguori
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=4F45231E.2030600@redhat.com \
--to=pbonzini@redhat.com \
--cc=alexander_barabash@mentor.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).