Re: [PATCH V3 3/3] xen-blkback: handle trim request in backend driver

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

> +static void xen_blk_trim(struct xen_blkif *blkif, struct blkif_request *req)

Just call it discard: s/trim/discard/ through all the patches.

Also make sure you do that for the functions. And for the 'tr' - change it to
'discard' maybe?

.. snip..
> +int xen_blkbk_trim(struct xenbus_transaction xbt, struct backend_info *be)

Call it discard.

.. snip ..
> +	err = xenbus_printf(xbt, dev->nodename, "feature-trim",

>From one hand it would be really nice to call this 'feature-discard'
but the Citrix frontends (and perhaps the SuSE ones too?) expect
it as 'feature-trim'. Maybe we should just call it 'feature-discard'
here and add a backwards compatible patch that will call it
'feature-trim'?

Sadly, the 'feature-trim' has been enumareted in the blkif.h so it
kind of is written in stone.

Keir, what is your feeling on this?

Can we change the name to 'feature-discard'?


> +			    "%d", state);
> +	if (err)
> +		xenbus_dev_fatal(dev, err, "writing feature-trim");
> +kfree:
> +	kfree(type);
> +out:
> +	return err;
> +}
> +
>  /*
>   * Entry point to this code when a new device is created.  Allocate the basic
>   * structures, and watch the store waiting for the hotplug scripts to tell us
> @@ -650,6 +708,10 @@ again:
>  	if (err)
>  		goto abort;
>  
> +	err = xen_blkbk_trim(xbt, be);
> +	if (err)
> +		goto abort;
> +

No need really. We don't need to abort b/c we can't establish discard
support.

>  	err = xenbus_printf(xbt, dev->nodename, "sectors", "%llu",
>  			    (unsigned long long)vbd_sz(&be->blkif->vbd));
>  	if (err) {
> -- 
> 1.7.6
> 
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xensource.com
> http://lists.xensource.com/xen-devel