[Top] [All Lists]

Re: Mounting real-time device - can it be done?

To: Annius Groenink <avg@xxxxxxxxxxxx>
Subject: Re: Mounting real-time device - can it be done?
From: Steve Lord <lord@xxxxxxx>
Date: Tue, 13 Feb 2001 17:41:58 -0600
Cc: linux-xfs@xxxxxxxxxxx, Joost Remijn <remijnj@xxxxxxxxxxxx>
In-reply-to: Message from Annius Groenink <avg@xxxxxxxxxxxx> of "Wed, 14 Feb 2001 00:36:47 +0100." <Pine.LNX.4.20.0102140032480.5197-100000@xxxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
> Hi.
> I am new to XFS and this list.  Great initiative.
> I tried the following:
> mkfs.xfs /dev/hdb4 -r size=1g,extsize=65536,rtdev=/dev/hdb3 -f
> i.e. setting up an XFS filesystem on /dev/hdb4 with a real-time section on
> /dev/hdb3.  Mkfs says it's happy about it.
> But when I try to mount, I get the following response (whereas mounting an
> XFS without a real-time section will go just fine):
> iep:/usr/include/linux# mount /dev/hdb4
> mount: wrong fs type, bad option, bad superblock on /dev/hdb4,
>        or too many mounted file systems


mount -t xfs -o rtdev=/dev/hdb3 /dev/hdb4 /xfs

This should let you mount, I suspect you will find problems with actual file
I/O into realtime files, but we can fix those as we find them. The realtime
subvolume has not had any attention payed to it in quite a while.


> Isn't that a pity!  Is this an unsupported feature?  The header
> files seem to do all the necessary ioctl() calls to open a file
> as part of the real-time section...
> How much work would it be to implement the feature, if it is
> indeed missing?  We'd be prepared to put some effort into this.
> --
> dr Annius V. Groenink
> groenink@xxxxxxxxxxxx
> tel +31 20 888 4126 fax 4001
> Kruislaan 400                 Eidetica, Amsterdam
> NL 1098 SM Amsterdam
> http://www.eidetica.com/      Knowledge services - Internet intelligence

<Prev in Thread] Current Thread [Next in Thread>