On 05/10/2013 11:28 PM, Eric Blake wrote:
On 05/07/2013 04:39 PM, John Baboval wrote:
From: "John V. Baboval" <john.babo...@virtualcomputer.com>
This parameter will cause writes to tty backed chardevs to return
-EAGAIN if the backing tty has buffered more than the specified
number of characters. When data is sent, the TIOCOUTQ ioctl is invoked
to determine the current TTY output buffer depth.
+++ b/qapi-schema.json
@@ -3182,11 +3182,14 @@
#
# @device: The name of the special file for the device,
# i.e. /dev/ttyS0 on Unix or COM1: on Windows
-# @type: What kind of device this is.
+#
+# @maxqdepth: #optional The maximum depth of the underlying tty
+# output queue (Unix) (Since 1.6)
#
# Since: 1.4
##
-{ 'type': 'ChardevHostdev', 'data': { 'device' : 'str' } }
+{ 'type': 'ChardevHostdev', 'data': { 'device' : 'str',
+ '*maxqdepth' : 'int' } }
Thanks; this interface change looks better (I'd still like to see
someone working on introspection, but it doesn't have to be you). I'll
still leave the implementation details to others more qualified for that
part of the review. In particular, since you are claiming this optional
attribute is Linux-only, that means we'd need introspection to know
whether a given qemu build supports the field (compiled on Linux) or not
(compiled on mingw), not just whether the qemu is new enough (1.6) or
older (1.4).
I believe (though I haven't tested on Windows) that the option at least
does no harm there. You can still set it - It's just a no-op.
