xfs
[Top] [All Lists]

Re: Linus merges XFS

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: Linus merges XFS
From: Michael Best <mbest@xxxxxxxxxxxxx>
Date: Tue, 17 Sep 2002 07:12:55 -0600
Cc: linux-xfs@xxxxxxxxxxx
References: <20020917000532.A18289@xxxxxxxxxx> <20020917072301.C81FF69123@xxxxxxxxxxxxxxxxxxxxx> <3D8723BE.8050702@xxxxxxxxxxxxx> <20020917135954.A25165@xxxxxxxxxxxxx>
Sender: linux-xfs-bounce@xxxxxxxxxxx
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.1) Gecko/20020831
Christoph Hellwig wrote:
> On Tue, Sep 17, 2002 at 06:44:46AM -0600, Michael Best wrote:
> 
>>My discussions with some of the Gentoo developers there seems to be a
>>number of them that have/had experienced data loss in the past and I
>>think some of them have a problem that it only journals metadata.
> 
> 
> Some of their patches/patch combination almost guarntee data loss.
> I wouldn't trust gentoo for throw-away bk repo :)

Just curious which of their patches/patch combinations were the worst?

They have a separate xfs_sources available, and since keeping XFS out of
their main tree anyways, I think I'll just stick to trying to get things
like newer xfsprogs into their package system and perhaps a couple of
things (like rml preempt) into their xfs kernel sources.

I was also hoping to get their maintainer to build a weekly CVS snapshot
kernel (as there appears to be a weekly XFS-patch snapshot) that would
be masked/hidden, just in case developers wanted to test things.

Other than the gcc-3.2 - O_DIRECT on i586 problem I know about I am
using an XFS CVS kernel from a couple of weeks back.  (using athlon so
I'm not affected by the O_DIRECT problem)

-Mike





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