Re: [Qemu-devel] [PATCH v2] Man page: Add -global description

[Anthony Liguori]

On 03/14/2012 01:21 PM, Peter Maydell wrote:

On 14 March 2012 08:53, Miroslav Rezanina<mreza...@redhat.com>  wrote:

--- a/qemu-options.hx
+++ b/qemu-options.hx
@@ -292,9 +292,13 @@ DEF("global", HAS_ARG, QEMU_OPTION_global,
     "                set a global default for a driver property\n",
     QEMU_ARCH_ALL)
  STEXI
-@item -global
+@item -global @var{driver}.@var{property}=@var{value}

We seem to use @var{prop}, not @var{property}, elsewhere in the docs.

  @findex -global
-TODO
+Set default value of @var{driver}'s @var{property} to @var{value}, e.g.:
+
+@example
+qemu -global ide-drive.physical_block_size=4096 -drive file=file,if=ide,index=0
+@end example
  ETEXI

This is missing any motivation for why you would want to actually
use this option. How about:

"In particular, you can use this to set driver properties for devices which
are created automatically by the machine model. (To create a device which is
not created automatically and set properties on it, use -device.)"

?

That's still not great, but I think it helps a little.

(ideally if -device/-global are the new standard interface we should have a
section explaining the general concepts and syntax and then documentation of
how to do specific things like networking via -device, and relegate all the
'legacy' options to a section clearly marked as 'legacy' with pointers back
to the new ways of doing the same thing. That would be a much bigger job,
though.)

Just while we're here, I'll be posting the following shortly:

commit 82aff428155d469ab705294486cc26cb34947999
Author: Anthony Liguori <aligu...@us.ibm.com>
Date:   Fri Dec 23 11:30:45 2011 -0600

qdev: don't allow globals to be set by bus name

This is technically a compatibility breaker.  However:

1) libvirt does not rely on this (it always uses the driver name)

2) This behavior isn't actually documented anywhere (the docs just say driver).

3) I suspect there are less than three people on earth that even know this is
   possible (minus the people reading this message).

So I think we can safely break it :-)

Signed-off-by: Anthony Liguori <aligu...@us.ibm.com>

Regards,

Anthony Liguori

-- PMM