Am 16.01.20 um 18:23 schrieb Jens Axboe:
> The application currently has no way of knowing if a given opcode is
> supported or not without having to try and issue one and see if we get
> -EINVAL or not. And even this approach is fraught with peril, as maybe
> we're getting -EINVAL due to some fields being missing, or maybe it's
> just not that easy to issue that particular command without doing some
> other leg work in terms of setup first.
> 
> This adds IORING_REGISTER_PROBE, which fills in a structure with info
> on what it supported or not. This will work even with sparse opcode
> fields, which may happen in the future or even today if someone
> backports specific features to older kernels.

That's funny I was just thinking about exactly that topic before
I opened the io-uring mail folder:-)

That's will make it much easier to write a portable
vfs backend for samba that doesn't depend on the kernel
features at build time.

> +	p->last_op = IORING_OP_LAST - 1;
> +	/* stock kernel isn't sparse, so everything is supported */
> +	for (i = 0; i < nr_args; i++) {
> +		p->ops[i].op = i;

Shouldn't there be an if (i <= p->last_op) before we pretent to support
an opcode? Or we need to truncate nr_args

> +		p->ops[i].flags = IO_URING_OP_SUPPORTED;
> +	}

metze