[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: [dm-devel] Latest dm-userspace kernel code
- From: FUJITA Tomonori <fujita tomonori lab ntt co jp>
- To: dm-devel redhat com
- Subject: Re: [dm-devel] Latest dm-userspace kernel code
- Date: Sun, 01 Oct 2006 02:27:48 +0900
From: Dan Smith <danms us ibm com>
Subject: Re: [dm-devel] Latest dm-userspace kernel code
Date: Sat, 30 Sep 2006 08:46:10 -0700
> FT> I thought that dm-userpsace supports a feature to map a single
> FT> request to multiple devices for kinda raid. For that, I guess that
> FT> struct dmu_msg_map_response includes several sets of dst_maj and
> FT> dst_min. Or you might use 'linked' dmu_msg_map_response, which
> FT> enables a user-space daemon to send multiple dmu_msg_map_responses
> FT> for one dmu_msg_map_request (if a dmu_msg_map_response is marked
> FT> 'linked', dm-userspace module clones the bio and does not free
> FT> struct the dmu_request).
>
> Hmm, I had not planned to implement this kind of behavior, but I can
> see the benefit. I think it would be best to focus on performance for
> now and add that capability later.
As I said before, once the code is merged into mainline (hopefully),
you cannot touch dm-usersapce.h. So at least, you need to have a
proper kernel/user interface for possible features.
OK. Let me work on this.
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]