On Mon, May 28, 2018 at 05:19:08PM +0200, Markus Armbruster wrote:
[...]
> >>
> >> > + * Meanwhile it can also be used even at the end of main. Let's keep
> >> > + * it initialized for the whole lifecycle of QEMU.
> >> > + */
> >>
> >> Awkward question, since our main() is such a tangled
Peter Xu writes:
> On Thu, May 24, 2018 at 11:03:55AM +0200, Markus Armbruster wrote:
>> Peter Xu writes:
>>
>> > Similar to previous patch, but introduce a new global big lock for
>> > mon_fdsets. Take it where needed.
>>
>> The previous patch is
On Thu, May 24, 2018 at 11:03:55AM +0200, Markus Armbruster wrote:
> Peter Xu writes:
>
> > Similar to previous patch, but introduce a new global big lock for
> > mon_fdsets. Take it where needed.
>
> The previous patch is "monitor: more comments on lock-free
>
On Fri, May 25, 2018 at 10:01:57AM +0100, Stefan Hajnoczi wrote:
> On Fri, May 25, 2018 at 11:30:22AM +0800, Peter Xu wrote:
> > On Thu, May 24, 2018 at 10:28:37AM +0100, Stefan Hajnoczi wrote:
> > > On Thu, May 24, 2018 at 12:39:52PM +0800, Peter Xu wrote:
> > > > int
On Fri, May 25, 2018 at 11:30:22AM +0800, Peter Xu wrote:
> On Thu, May 24, 2018 at 10:28:37AM +0100, Stefan Hajnoczi wrote:
> > On Thu, May 24, 2018 at 12:39:52PM +0800, Peter Xu wrote:
> > > int monitor_fdset_get_fd(int64_t fdset_id, int flags)
> > > {
> > > -#ifndef _WIN32
> > > +#ifdef
On Thu, May 24, 2018 at 10:28:37AM +0100, Stefan Hajnoczi wrote:
> On Thu, May 24, 2018 at 12:39:52PM +0800, Peter Xu wrote:
> > int monitor_fdset_get_fd(int64_t fdset_id, int flags)
> > {
> > -#ifndef _WIN32
> > +#ifdef _WIN32
> > +return -ENOENT;
>
> stubs/fdset.c:monitor_fdset_get_fd()
On Thu, May 24, 2018 at 12:39:52PM +0800, Peter Xu wrote:
> int monitor_fdset_get_fd(int64_t fdset_id, int flags)
> {
> -#ifndef _WIN32
> +#ifdef _WIN32
> +return -ENOENT;
stubs/fdset.c:monitor_fdset_get_fd() should return -ENOENT instead of -1
now.
signature.asc
Description: PGP
Peter Xu writes:
> Similar to previous patch, but introduce a new global big lock for
> mon_fdsets. Take it where needed.
The previous patch is "monitor: more comments on lock-free
fleids/funcs". Sure you mean that one?
>
> The monitor_fdset_get_fd() handling is a bit
Similar to previous patch, but introduce a new global big lock for
mon_fdsets. Take it where needed.
The monitor_fdset_get_fd() handling is a bit tricky: now we need to call
qemu_mutex_unlock() which might pollute errno, so we need to make sure
the correct errno be passed up to the callers. To