xfs
[Top] [All Lists]

Re: [PATCH 3/2] delalloc support for xfs fiemap

To: Eric Sandeen <sandeen@xxxxxxxxxxx>, xfs-oss <xfs@xxxxxxxxxxx>
Subject: Re: [PATCH 3/2] delalloc support for xfs fiemap
From: Eric Sandeen <sandeen@xxxxxxxxxxx>
Date: Fri, 24 Oct 2008 08:01:16 -0500
In-reply-to: <20081024052930.GI18495@disturbed>
References: <49014887.6050309@xxxxxxxxxxx> <20081024052930.GI18495@disturbed>
User-agent: Thunderbird 2.0.0.17 (Macintosh/20080914)
Dave Chinner wrote:
> On Thu, Oct 23, 2008 at 11:01:11PM -0500, Eric Sandeen wrote:
>> Ok, got the kinks worked out of the delalloc portion of the
>> fiemap hookup I think.
>>
>> This introduces a new input flag (BMV_IF_DELALLOC) and output
>> flag (BMV_OF_DELALLOC) to request & show the delayed allocation
>> segments.
> ....
>> @@ -5827,6 +5827,7 @@ xfs_getbmap(
>>                                               * preallocated data space */
>>      int                     sh_unwritten;   /* true, if unwritten */
>>                                              /* extents listed separately */
>> +    int                     sh_delalloc;    /* show delayed extents */
>>      int                     iflags;         /* interface flags */
>>      int                     bmapi_flags;    /* flags for xfs_bmapi */
> 
> Do we really need this "sh_delalloc" flag? Checking for
> (iflags & BMV_IF_DELALLOC) is fine by me, same for replacing
> sh_unwritten. I've never really liked the "sh" abbreviation for
> "show"....

Sure, I'd agree with that, though some of the tests for sh_$FOO are a
bit wordy already ... I'll make the change, though.

> Otherwise it looks ok. Perhaps we should also add delalloc extent
> display to xfs_bmap....

Yep, that'd be good.

BTW thanks for the help looking through the bmap flag stuff!

I'll resend the whole series; there are some 80+ char lines etc
to fix up as well.

Thanks,
-Eric

> Cheers,
> 
> Dave.

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