xfs
[Top] [All Lists]

Re: [PATCH] mkfs.xfs: don't detect geometry values <= physblk size

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH] mkfs.xfs: don't detect geometry values <= physblk size
From: Eric Sandeen <sandeen@xxxxxxxxxx>
Date: Wed, 12 Dec 2012 17:24:33 -0600
Cc: xfs-oss <xfs@xxxxxxxxxxx>
In-reply-to: <20121212231642.GZ16353@dastard>
References: <50C9093C.8080400@xxxxxxxxxx> <20121212231642.GZ16353@dastard>
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.8; rv:17.0) Gecko/17.0 Thunderbird/17.0
On 12/12/12 5:16 PM, Dave Chinner wrote:
> On Wed, Dec 12, 2012 at 04:46:20PM -0600, Eric Sandeen wrote:
>> blkid_get_topology() ignores devices which report 512
>> as their minimum & optimal IO size, but we should ignore
>> anything up to the physical block size; otherwise
>> 4k devices will report a "stripe size" of 4k, and warn
>> if anything larger is specified:
> 
> As i read this, I thought no, that's wrong, it should read
> "we should ignore anything up to the sector size". I had to read the
> code to find out that's exactly what the change does.
> 
> Perhaps you should be talking about sector sizes here?
> 
>> -     * If the reported values are just the normal 512 byte block size
>> -     * do not bother to report anything.  It will just causes warnings
>> -     * if people specifier larger stripe units or widths manually.
>> +     * If the reported values are the same as the physical block size
> 
> "... same as the sector size of the device ...."
> 
> Otherwise seems sane...
> 
> Cheers,
> 
> Dave.
> 

Sorry, yeah, s/block/sector/

I blame the meds (and hch's prior comment) ;)  Will resend.

-Eric

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