qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Anthony Liguori <anthony@codemonkey.ws>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH v2 04/27] qom: avoid useless conversions from string to type
Date: Mon, 06 Feb 2012 08:14:45 -0600	[thread overview]
Message-ID: <4F2FE055.30307@codemonkey.ws> (raw)
In-Reply-To: <1328342577-25732-5-git-send-email-pbonzini@redhat.com>

On 02/04/2012 02:02 AM, Paolo Bonzini wrote:
> Signed-off-by: Paolo Bonzini<pbonzini@redhat.com>
> ---
>   qom/object.c |   16 +++++++++-------
>   1 files changed, 9 insertions(+), 7 deletions(-)
>
> diff --git a/qom/object.c b/qom/object.c
> index 4d21f0a..a89e9e3 100644
> --- a/qom/object.c
> +++ b/qom/object.c
> @@ -368,9 +368,10 @@ void object_delete(Object *obj)
>       g_free(obj);
>   }
>
> -static bool object_is_type(Object *obj, const char *typename)
> +static Type type_interface;
> +

Statics belong at the top of the file.  But I'd rather not introduce a global 
unless we have some sort of compelling data that shows that it's a performance 
problem.

The type lookup is done through a hash table, at most it's going to be 2-3 
integer comparisons followed by a strcmp() of a small word.  It shouldn't be 
that slow.

Regards,

Anthony Liguori

> +static bool object_is_type(Object *obj, TypeImpl *target_type)
>   {
> -    TypeImpl *target_type = type_get_by_name(typename);
>       TypeImpl *type = obj->class->type;
>
>       /* Check if typename is a direct ancestor of type */
> @@ -387,10 +388,11 @@ static bool object_is_type(Object *obj, const char *typename)
>
>   Object *object_dynamic_cast(Object *obj, const char *typename)
>   {
> +    TypeImpl *target_type = type_get_by_name(typename);
>       GSList *i;
>
>       /* Check if typename is a direct ancestor */
> -    if (object_is_type(obj, typename)) {
> +    if (object_is_type(obj, target_type)) {
>           return obj;
>       }
>
> @@ -404,10 +406,10 @@ Object *object_dynamic_cast(Object *obj, const char *typename)
>        * (run-time) type safety and to speed up tests like this one.  If we ever
>        * do that we can revisit the order here.
>        */
> -    if (object_is_type(obj, TYPE_INTERFACE)) {
> +    if (object_is_type(obj, type_interface)) {
>           assert(!obj->interfaces);
>           obj = INTERFACE(obj)->obj;
> -        if (object_is_type(obj, typename)) {
> +        if (object_is_type(obj, target_type)) {
>               return obj;
>           }
>       }
> @@ -416,7 +418,7 @@ Object *object_dynamic_cast(Object *obj, const char *typename)
>       for (i = obj->interfaces; i; i = i->next) {
>           Interface *iface = i->data;
>
> -        if (object_is_type(OBJECT(iface), typename)) {
> +        if (object_is_type(OBJECT(iface), target_type)) {
>               return OBJECT(iface);
>           }
>       }
> @@ -433,7 +435,7 @@ static void register_interface(void)
>           .abstract = true,
>       };
>
> -    type_register_static(&interface_info);
> +    type_interface = type_register_static(&interface_info);
>   }
>
>   device_init(register_interface);

  reply	other threads:[~2012-02-06 14:14 UTC|newest]

Thread overview: 62+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-02-04  8:02 [Qemu-devel] [PATCH v2 00/27] next steps for qdev & QOM Paolo Bonzini
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 01/27] qom: clean up cast macros Paolo Bonzini
2012-02-06 14:03   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 02/27] qom: more documentation on subclassing Paolo Bonzini
2012-02-06 14:04   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 03/27] qom: clean up/optimize object_dynamic_cast Paolo Bonzini
2012-02-06 14:10   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 04/27] qom: avoid useless conversions from string to type Paolo Bonzini
2012-02-06 14:14   ` Anthony Liguori [this message]
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 05/27] qom: do not include qdev header file Paolo Bonzini
2012-02-06 14:15   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 06/27] qom: add QObject-based property get/set wrappers Paolo Bonzini
2012-02-06 14:16   ` Anthony Liguori
2012-02-07  9:12     ` Paolo Bonzini
2012-02-08 12:29       ` Luiz Capitulino
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 07/27] qom: add property get/set wrappers for C types Paolo Bonzini
2012-02-06 14:17   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 08/27] qom: fix off-by-one Paolo Bonzini
2012-02-06 14:19   ` Anthony Liguori
2012-02-07  9:13     ` Paolo Bonzini
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 09/27] qom: add object_resolve_path_type Paolo Bonzini
2012-02-06 14:21   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 10/27] qom: use object_resolve_path_type for links Paolo Bonzini
2012-02-06 14:24   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 11/27] qom: fix canonical paths vs. interfaces Paolo Bonzini
2012-02-06 14:24   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 12/27] qom: add property get/set wrappers for links Paolo Bonzini
2012-02-06 14:27   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 13/27] qdev: remove direct calls to print/parse Paolo Bonzini
2012-02-06 14:30   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 14/27] qdev: allow reusing get/set for legacy property Paolo Bonzini
2012-02-06 14:31   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 15/27] qdev: remove parse method for string properties Paolo Bonzini
2012-02-06 14:31   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 16/27] qdev: remove print/parse methods from LostTickPolicy properties Paolo Bonzini
2012-02-06 14:32   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 17/27] qdev: remove parse/print methods for mac properties Paolo Bonzini
2012-02-06 14:32   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 18/27] qdev: make the non-legacy pci address property accept an integer Paolo Bonzini
2012-02-06 14:33   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 19/27] qdev: remove parse/print methods for pointer properties Paolo Bonzini
2012-02-06 14:34   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 20/27] qdev: let QOM free properties Paolo Bonzini
2012-02-06 14:35   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 21/27] qdev: fix off-by-one Paolo Bonzini
2012-02-06 14:35   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 22/27] qdev: access properties via QOM Paolo Bonzini
2012-02-06 14:36   ` Anthony Liguori
2012-02-11 15:46   ` Blue Swirl
2012-02-13  7:53     ` Paolo Bonzini
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 23/27] qdev: inline qdev_prop_set into qdev_prop_set_ptr Paolo Bonzini
2012-02-06 14:37   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 24/27] qdev: initialize properties via QOM Paolo Bonzini
2012-02-06 14:38   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 25/27] qdev: remove unused fields from PropertyInfo Paolo Bonzini
2012-02-06 14:39   ` Anthony Liguori
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 26/27] omap_clk: convert to QOM Paolo Bonzini
2012-02-06 14:42   ` Anthony Liguori
2012-02-06 16:32   ` Peter Maydell
2012-02-04  8:02 ` [Qemu-devel] [PATCH v2 27/27] omap: remove PROP_PTR properties Paolo Bonzini
2012-02-06 14:43   ` Anthony Liguori
2012-02-07 13:20 ` [Qemu-devel] [PATCH v2 00/27] next steps for qdev & QOM Paolo Bonzini

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=4F2FE055.30307@codemonkey.ws \
    --to=anthony@codemonkey.ws \
    --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).