David Sterba wrote on 2016/05/17 15:14 +0200:
On Wed, Apr 27, 2016 at 09:29:29AM +0800, Qu Wenruo wrote:
@@ -5542,6 +5606,10 @@ long btrfs_ioctl(struct file *file, unsigned int
return btrfs_ioctl_get_fslabel(file, argp);
case BTRFS_IOC_SET_FSLABEL:
return btrfs_ioctl_set_fslabel(file, argp);
Would you mind me to add a new kernel config "Btrfs experimental
features-> dedupe ioctl" for case like dedupe and further experimental
btrfs features?
The BTRFS_DEBUG seems quite odd for me.
Although in-band dedupe is quite good at exposing bugs of
backref/qgroup/delayed_refs, but I still think it's not a debug tool.
So I hope to use "BTRFS_EXPERIMENTAL_DEDUPE_IOCTL" and add corresponding
Kconfig interfaces.
The point was not to add new config options, but if you find _DEBUG odd
we can rethink the "hide it under config option" approach. The alternate
way is to get the code into a good shape and add it to for-next. I don't
want to track yet another set of branches with the too experimental
stuff.
OK, I'll try the alternative method, make the code into good shape for
in-memory backend.
As for the ioctl interface, I'll post some ideas on this to wiki and
hopes we can get a good solution for it.
Like adding new 'force' flag, to make 'enable --force' to be stateless,
and without that flag, go to normal enable/config stateful method.
Just going completely stateful method is still another alternative though.
Thanks,
Qu
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
