Re: [PATCH] src/seek_sanity_test: ensure file size is big enough

[Eryu Guan <eguan@redhat.com>]

On Mon, May 08, 2017 at 10:45:45AM +0100, Luis Henriques wrote:
> On Mon, May 08, 2017 at 10:37:23AM +0100, Luis Henriques wrote:
> > On Mon, May 08, 2017 at 04:52:34PM +0800, Eryu Guan wrote:
> > > On Fri, May 05, 2017 at 03:57:23PM +0100, Luis Henriques wrote:
> > > > Tests test07, test08, and test09 preallocate a file and assume the file
> > > > size used is bigger than 10xbufsz (100xbufsz for test09).  This patch
> > > > adjusts the file size so this assumption is always true.
> > > > 
> > > > As an example, here's test06 output for cephfs, where the allocation size
> > >                         ^^^^^^ meant test07?
> > 
> > ups, yeah I did.
> > 
> > > > is set to 4194304, and the output is (4194304 * 10 + 4194304)
> > > > 
> > > >   07. Test file with unwritten extents, only have dirty pages
> > > >   07.01 SEEK_HOLE expected 0 or 4194304, got 46137344.              FAIL
> > > >   07.02 SEEK_HOLE expected 1 or 4194304, got 46137344.              FAIL
> > > > 
> > > > (Note: The test will be skipped if an integer overflow occurs.)
> > > > 
> > > > Signed-off-by: Luis Henriques <lhenriques@suse.com>
> > > > ---
> > > >  src/seek_sanity_test.c | 21 +++++++++++++++++++++
> > > >  1 file changed, 21 insertions(+)
> > > > 
> > > > diff --git a/src/seek_sanity_test.c b/src/seek_sanity_test.c
> > > > index a6dd48cc257b..41da59801212 100644
> > > > --- a/src/seek_sanity_test.c
> > > > +++ b/src/seek_sanity_test.c
> > > > @@ -282,6 +282,13 @@ static int test09(int fd, int testnum)
> > > >  	int bufsz = alloc_size;
> > > >  	int filsz = 8 << 20;
> > > >  
> > > > +	while ((filsz < (bufsz * 100 + bufsz)) && (filsz > 0))
> > > > +		filsz <<= 1;
> > > > +	if (filsz <= 0) {
> > > > +		fprintf(stdout, "Test skipped due to int overflow.\n");
> > > > +		return ret;
> > > > +	}
> > > > +
> > > 
> > > Hmm, why not just set filsz to (bufsz * 100 + bufsz)? This works for me
> > > on XFS/ext4/btrfs and NFS, I guess it should work for cepthfs too?
> > 
> > Right, that should work too.  I've decided to keep it as a power of 2 as I
> > assumed the original author had a reason for using that value initially
> > (4Mb).  But yeah if no one objects, I'll submit v2 with your suggestion.
> > 
> 
> BTW, do you think it's worth keeping that int overflow test with this
> change?

I have no strong preference on this, but seems if filsz overflows int,
there might be something goes wrong, letting it fail is better than
skipping it sliently.

Thanks,
Eryu