On Tue, Feb 11, 2020 at 06:16:25PM +0200, Nikolay Ivanets wrote:
вт, 11 лют. 2020 о 17:20 Richard W.M. Jones <rjones(a)redhat.com>
пише:
>
> I pushed this with some trailing whitespace fixes, and I dropped the
> change to tmp/.gitignore since the test does clean up after itself. I
> also fixed test-qemu-drive-with-blocksize-libvirt.sh so it doesn't
> actually open /dev/sda etc on the host (don't run tests as root!)
Thanks!
> However ...
>
> We already use blocksize as an optional parameter to mkfs. While they
> don't directly conflict, it is confusing. Is there a reason we
> shouldn't call this new parameter "sectorsize"?
>
> We can change the parameter name any time up til we make the next
> stable release.
But mkfs has 'sectorsize' optional parameter as well. :-)
Here are my thoughts:
1. Our 'blocksize' parameter is related to disks and it's using everywhere
in
context of disks. It hardly can confuse users with mkfs' blocksize which is
related to a file system.
2. Under the hood this parameter is tied to physical/logical block size so it
named accordingly.
OK fair enough. I didn't spot that it had "sectorsize" too.
Let's leave it alone.
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