Re: [Qemu-devel] [PATCH] Add interface to traverse the qmp command list by QmpCommand

[Michael Roth <mdroth@linux.vnet.ibm.com>]

Quoting Eric Blake (2013-09-25 21:57:39)
> On 09/25/2013 07:56 PM, Mark Wu wrote:
> > In the original code, qmp_get_command_list is used to construct
> > a list of all commands' name. To get the information of all qga
> > commands, it traverses the name list and search the command info
> > with its name.  So it can cause O(n^2) in the number of commands.
> > 
> > This patch adds an interface to traverse the qmp command list by
> > QmpCommand to replace qmp_get_command_list. It can decrease the
> > complexity from O(n) to O(n^2)
> 
> from O(n^2) to O(n)
> 
> > 
> > Signed-off-by: Mark Wu <wudxw@linux.vnet.ibm.com>
> > ---
> >  include/qapi/qmp/dispatch.h |  3 +-
> >  qapi/qmp-registry.c         | 28 ++-----------------
> >  qga/commands.c              | 39 ++++++++++----------------
> >  qga/main.c                  | 68 +++++++++++++++++----------------------------
> >  4 files changed, 45 insertions(+), 93 deletions(-)
> > 
> 
> >  
> > -bool qmp_command_is_enabled(const char *name)
> 
> I think one of Michael's suggestions was that you may still want
> qmp_command_is_enabled, but with a new signature:
> 
> bool qmp_command_is_enabled(const QmpCommand *cmd)
> {
>     return cmd->enabled;
> }

That would still be my preference, as I can see the underlying
fields changing somewhat in a future where QMP starts using the same
infrastructure, and new users come along like qemu-ga guest->host
commands we did as part of GSoC.

Wouldn't hold things up in the meantime, but if we're re-spinning
anyway let's maintain the precedence of using getter/setter funcs
for QmpCommand.

> 
> > -struct GuestAgentInfo *qmp_guest_info(Error **err)
> > +static void qmp_command_info(QmpCommand *cmd, void *opaque)
> >   {
> > -    GuestAgentInfo *info = g_malloc0(sizeof(GuestAgentInfo));
> > +    GuestAgentInfo *info = (GuestAgentInfo *)opaque; 
> 
> 
> > -        cmd_info->enabled = qmp_command_is_enabled(cmd_info->name);
> >  
> > -        cmd_info_list = g_malloc0(sizeof(GuestAgentCommandInfoList));
> > -        cmd_info_list->value = cmd_info;
> > -        cmd_info_list->next = info->supported_commands;
> > -        info->supported_commands = cmd_info_list;
> > +    cmd_info = g_malloc0(sizeof(GuestAgentCommandInfo));
> > +    cmd_info->name = g_strdup(cmd->name);
> > +    cmd_info->enabled = cmd->enabled;
> 
> I guess it all depends on whether we want QmpCommand to be an opaque
> type outside of a single file.  But I don't have a strong argument for
> making it opaque, and your approach works if we don't mind exposing the
> details of QmpCommand across multiple files.  So I can live with your
> patch as-is.
> 
> 
> > +static void ga_enable_non_blacklisted(QmpCommand *cmd, void *opaque)
> >  {
> 
> > +    GList *blacklist = (GList *)opaque;
> 
> Again, the cast is not necessary.
> 
> Overall, I like the patch.  Just a few tweaks suggested, but I can live
> with you adding:
> 
> Reviewed-by: Eric Blake <eblake@redhat.com>
> 
> -- 
> Eric Blake   eblake redhat com    +1-919-301-3266
> Libvirt virtualization library http://libvirt.org