|
From: | Anthony Liguori |
Subject: | [Qemu-devel] Re: [PATCH v7 4/6] char: Add qemu_chr_write_nb() for nonblocking writes |
Date: | Wed, 05 May 2010 14:33:15 -0500 |
User-agent: | Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.5) Gecko/20091209 Fedora/3.0-4.fc12 Lightning/1.0pre Thunderbird/3.0 |
On 05/05/2010 02:16 PM, Gerd Hoffmann wrote:
On 05/05/10 20:49, Anthony Liguori wrote:On 05/05/2010 01:43 PM, Gerd Hoffmann wrote:On 05/05/10 15:53, Anthony Liguori wrote:This patch adds optional control flow in a pretty crufty way to *some* backends but not all. This just adds a bunch of complexity that will certainly introduce bugs.Amit plans to add support to the others as well. Beside that there is a clearly defined backup plan: In case the non-blocking interface isn't supported by $chardev it will fallback to use the blocking mode.If we have a second interface, we'll have two interfaces forever. I'd rather see us aggressive remove the blocking interface instead of introducing a second interface.I'm all for killing the blocking interface. Problem is that converting over all users isn't exactly trivial and we have plenty of them. IMHO it isn't realistic to do the switch with a single patch series.
If we're agreed we ought to kill the blocking interface, let's define a new proper interface, rename all previous users to something ugly and deprecated, and approach it that way.
Let's make it clear what interfaces are supported and which interfaces are scheduled to be removed.
Regards, Anthony Liguori
cheers, Gerd
[Prev in Thread] | Current Thread | [Next in Thread] |