On Thu, Oct 28, 2021 at 07:55:39PM +0200, Laszlo Ersek wrote:
> On 10/28/21 16:24, Richard W.M. Jones wrote:
> > +#ifdef HAVE_STDATOMIC_H
> > +#include <stdatomic.h>
> > +#else
> > +/* Some old platforms lack atomic types, but this is only used for
> > + * debug messages.
> > + */
> > +#define _Atomic /**/
> > +#endif
...
> > +/* Send command to the background thread and wait for completion.
> > + *
> > + * Returns 0 for OK
> > + * On error, calls nbdkit_error and returns -1.
> > + */
> > +int
> > +send_command_and_wait (struct vddk_handle *h, struct command *cmd)
> > +{
> > + static _Atomic uint64_t id = 0;
>
> Ugh.
>
> _Atomic is from C11, and it comes with a huge theoretical baggage
> (section "5.1.2.4 Multi-threaded executions and data races"). Do we
> really need this? For example:
[theoretical stuff snipped]
Sure thing, I'll put it in the handle structure and protect it with
the command queue lock.
I think I'll push this with the above change.
There are quite a few other cases of _Atomic in nbdkit & libnbd if
you're feeling spirited enough to check any of them :-)
Thanks,
Rich.
--
Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones
Read my programming and virtualization blog: http://rwmj.wordpress.com
virt-builder quickly builds VMs from scratch
http://libguestfs.org/virt-builder.1.html
_______________________________________________
Libguestfs mailing list
[email protected]
https://listman.redhat.com/mailman/listinfo/libguestfs
