xfs
[Top] [All Lists]

Re: [RFC PATCH] xfs: pass KM_SLEEP flag to kmem_realloc() in xlog_recove

To: Mitsuo Hayasaka <mitsuo.hayasaka.hu@xxxxxxxxxxx>
Subject: Re: [RFC PATCH] xfs: pass KM_SLEEP flag to kmem_realloc() in xlog_recover_add_to_cnt_trans()
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Fri, 27 Jan 2012 05:52:53 -0500
Cc: Ben Myers <bpm@xxxxxxx>, Alex Elder <aelder@xxxxxxx>, Christoph Hellwig <hch@xxxxxx>, Alex Elder <elder@xxxxxxxxxx>, linux-kernel@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, xfs-masters@xxxxxxxxxxx, yrl.pp-manager.tt@xxxxxxxxxxx
In-reply-to: <20120127063726.16227.681.stgit@xxxxxxxxxxxxxxxxxxxxxxxx>
References: <20120127063726.16227.681.stgit@xxxxxxxxxxxxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Fri, Jan 27, 2012 at 03:37:26PM +0900, Mitsuo Hayasaka wrote:
> The kmem_realloc() in xfs is given KM_* memory allocation flags. And it
> allocates memory using kmalloc() after they are converted to gfp_mask
> flags. In xlog_recover_add_to_cont_trans(), 0u is passed to kmem_realloc(),
> instead of them. I guess it is preferred to use them, and here memory must
> be allocated but don't have to be done with GFP_ATOMIC. So, this patch
> changes it to KM_SLEEP.

Looks good.  I actually just noticed it when changing the kmem_realloc
implementation, but let's put your patch in before that and send it off
to Linus ASAP.

Reviewed-by: Christoph Hellwig <hch@xxxxxx>

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