On 2015-04-06 21:28, 인정식 wrote:
> Hello BTRFS developers,
> 
> I am requesting your opion.
> 
> I am planning to design and implement DFS version of BTRFS.
> Roughly it will be done by
>     1. Extending current DeviceID to NodeID:DeviceID to support multi-node, 
> and
>     2. Implementing inter-node data and meta-data access over TCP.
> Do you think it would be feasible?
> 
> For development, I would like to port the kernel code to user space.
> I could continue to run in user mode or be ported back to kernel space.
> But I don't know how can I start to porting the kernel code to user space.
> Is there anyone who can help?
> 
> Thanks,
> JeongsikN떑꿩�r툤y鉉싕b쾊Ф푤v�^�)頻{.n�+돴쪐{켷雹�)�鍊w*jgП�텎쉸듶줷/곴�z받뻿�2듷솳鈺�&�)傘첺뛴�췍쳺�h��j:+v돣둾�明
> 
Personally, my suggestion would be to start with a port to FUSE, and
then start implementing your modifications off of that.  That said, much
of the code should already be available in userspace through libbtrfs,
which is packaged with btrfs-progs.

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to