xfs
[Top] [All Lists]

Re: [PATCH 4/4] xfs: fix the logspace waiting algorithm

To: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Subject: Re: [PATCH 4/4] xfs: fix the logspace waiting algorithm
From: Chandra Seetharaman <sekharan@xxxxxxxxxx>
Date: Fri, 02 Dec 2011 10:02:39 -0600
Cc: xfs@xxxxxxxxxxx
In-reply-to: <20111202111649.GA3072@xxxxxxxxxxxxx>
Organization: IBM
References: <20111128081732.350228200@xxxxxxxxxxxxxxxxxxxxxx> <20111128081925.981681380@xxxxxxxxxxxxxxxxxxxxxx> <1322767713.7115.1548.camel@xxxxxxxxxxxxxxxxxxxxxxxxxxxx> <20111202111649.GA3072@xxxxxxxxxxxxx>
Reply-to: sekharan@xxxxxxxxxx
Then it is all good

Reviewed-by: Chandra Seetharaman <sekharan@xxxxxxxxxx>
  
On Fri, 2011-12-02 at 06:16 -0500, Christoph Hellwig wrote:
> On Thu, Dec 01, 2011 at 01:28:33PM -0600, Chandra Seetharaman wrote:
> > Tested the patch with testcases 234 and 273. They ran for more than 350
> > iterations without getting into the hang situation.
> > 
> > Tested-by: Chandra Seetharaman <sekharan@xxxxxxxxxx>
> > 
> > Few generic comments on the patch
> > 
> > 1. xlog_*_wake could use something to indicate that they are looking for
> > log space in the specific queue. ex: xlog_reserveq_available()
> > 
> > 2. All new functions expect a lock to be held on entry. Can be
> > explicitly specified in a comment.
> > 
> > 3. Change the trace function names to reflect the names of the
> > function ?!
> 
> All this is going to change with the refactoring of these to use a
> common data structure which I have in my queue for 3.3.
> 


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