On Wed, Dec 23, 2015 at 12:34:25PM +0100, Paolo Bonzini wrote:
>
>
> On 22/12/2015 19:17, Daniel P. Berrange wrote:
> > + if (common->has_logfile) {
> > + int flags = O_WRONLY | O_CREAT;
> > + if (!common->has_logappend ||
> > + !common->logappend) {
> > + flags |= O_TRUNC;
> > + }
>
> Should it use O_APPEND if logappend is absent or true, or perhaps
> always? I can take care of the change myself.
Yes it should use O_APPEND in the other branch of the if,
that's a bug eric pointed out too. Basically same logic
as the recently added 'append' flag in qmp_chardev_open_file
> You are also missing a few qemu_chr_alloc calls outside qemu-char.c,
> which makes me wonder if it's better to add the new logic in
> qemu_chr_alloc itself.
Oh, yes, perhaps I should just pass ChardevCommon straight
into qemu_chr_alloc(). I'll look at that when doing the other
changes Eric suggested wrt to removing use of backend.u.data
Regards,
Daniel
--
|: http://berrange.com -o- http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org -o- http://virt-manager.org :|
|: http://autobuild.org -o- http://search.cpan.org/~danberr/ :|
|: http://entangle-photo.org -o- http://live.gnome.org/gtk-vnc :|
