xfs
[Top] [All Lists]

Re: hole punching performance

To: "Bradley C. Kuszmaul" <kuszmaul@xxxxxxxxx>
Subject: Re: hole punching performance
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Sun, 6 Jan 2013 09:54:31 +1100
Cc: xfs@xxxxxxxxxxx
In-reply-to: <CAKSyJXct+kRvDqGg1LtjU6q+P713bbsZDdik9LhLRvH2q5GEhw@xxxxxxxxxxxxxx>
References: <CAKSyJXf66H2U-BF-aYnSr2fF24_6LJw6swOx1RhUc_3Eqayaiw@xxxxxxxxxxxxxx> <20130102232706.GD3120@dastard> <CAKSyJXd8wsH3MDOOaqo24M8DSYQjE2S2hF3QcFnEWjYTOo3s4w@xxxxxxxxxxxxxx> <20130103220717.GI3120@dastard> <CAKSyJXct+kRvDqGg1LtjU6q+P713bbsZDdik9LhLRvH2q5GEhw@xxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Sat, Jan 05, 2013 at 11:31:19AM -0500, Bradley C. Kuszmaul wrote:
> What are the requirements for alignment of punched holes?    Is 512-byte
> alignment good enough?  (I'll have 512-byte alignment anyway since I'll be
> using direct I/O on these files.)

No alignment at all - they have byte range granularity.

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

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