Re: [Qemu-devel] [PATCH] Asynchronous reliable and configurable cache flush

[Jamie Lokier <jamie@shareable.org>]

Paul Brook wrote:
> > Looks that way to me too.  It might be simplest to open the device
> > with O_DSYNC when !s->write_cache and user actually wants fdatasync,
> > so that async write can be used.  I suspect every platform with useful
> > Posix AIO has O_DSYNC.
> 
> The cache is dynamically enabled/disabled by the target. This means
> you've got to close and repopen the file every time it changes,
> which is likely to get really hairy.

Just open two descriptors :-)

I'm not sure if F_SETFL can be used.

Both descriptors are useful with the cache enabled, if the SATA FUA
(force unit access) bit is set on a write command.  Only those writes
would use the O_DSYNC descriptor.

> > Otherwise, chaining where the completion of aio_write triggers
> > aio_fsync instead of reporting completion to the guest?
> 
> Yes, that should do it.

Yes, though O_DSYNC will save a system call.

-- Jamie