Re: [PATCH 7/8] xfstests: Add fallocate zero range operation to fsx

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, 28 Feb 2014, Andreas Dilger wrote:

> Date: Fri, 28 Feb 2014 12:08:47 -0700
> From: Andreas Dilger <adilger@xxxxxxxxx>
> To: Lukas Czerner <lczerner@xxxxxxxxxx>
> Cc: linux-ext4@xxxxxxxxxxxxxxx, xfs@xxxxxxxxxxx, linux-fsdevel@xxxxxxxxxxxxxxx
> Subject: Re: [PATCH 7/8] xfstests: Add fallocate zero range operation to fsx
> 
> On Feb 28, 2014, at 9:11 AM, Lukas Czerner <lczerner@xxxxxxxxxx> wrote:
> > This commit adds fallocate FALLOC_FL_ZERO_RANGE support for fsx.
> 
> It looks like this patch breaks the option handling for these
> fallocate features.
> 
> >  int     fallocate_calls = 1;            /* -F flag disables */
> >  int     punch_hole_calls = 1;           /* -H flag disables */
> > +int     zero_range_calls = 1;           /* -z flag disables */
> 
> So these fallocate tests are on by default...
> 
> > +int
> > +test_fallocate(int mode)
> >  {
> >  #ifdef HAVE_LINUX_FALLOC_H
> > +	int ret = 0;
> > 	if (!lite && fallocate_calls) {
                     ^^^^^^^^^^^^^^^

Right, I forgot about this.


> > +		if (fallocate(fd, mode, 0, 1) && errno == EOPNOTSUPP) {
> > 			if(!quiet)
> > 				warn("main: filesystem does not support fallocate, disabling\n");
> > 		} else {
> > +			ret = 1;
> > 			ftruncate(fd, 0);
> > 		}
> > 	}
> > #endif
> > }
> 
> And this returns 1 or 0 depending if they are supported or not...
> 
> > +	while ((ch = getopt(argc, argv, "b:c:dfl:m:no:p:qr:s:t:w:xyAD:FHzLN:OP:RS:WZ"))
> > 	       != EOF)
> > 		switch (ch) {
> > 		case 'H':
> > 			punch_hole_calls = 0;
> > 			break;
> > +		case 'z':
> > +			zero_range_calls = 0;
> > +			break;
> 
> And the option parsing sets the values to zero if they are disabled,
> so far, so good...
> 
> > +	fallocate_calls = test_fallocate(0);
> > +	punch_hole_calls = test_fallocate(FALLOC_FL_PUNCH_HOLE | FALLOC_FL_KEEP_SIZE);
> > +	zero_range_calls = test_fallocate(FALLOC_FL_ZERO_RANGE);
> 
> But here, the values set by option parsing are clobbered and the
> tests are only enabled or disabled depending on whether the kernel
> supports that feature or not.  I think you need something like:

Right, and the 'fallocate_calls' test needs to disappear from
test_fallocate().

Thanks!
-Lukas

> 
>         if (fallocate_calls)
>                 fallocate_calls = test_fallocate(0);
>         if (punch_hole_calls)
>                 punch_hole_calls = test_fallocate(FALLOC_FL_PUNCH_HOLE |
>                                                   FALLOC_FL_KEEP_SIZE);
>         if (zero_range_calls)
>                 zero_range_calls = test_fallocate(FALLOC_FL_ZERO_RANGE);
> 
> Cheers, Andreas
> 
> 
> 
> 
> 
> 

_______________________________________________
xfs mailing list
xfs@xxxxxxxxxxx
http://oss.sgi.com/mailman/listinfo/xfs




[Index of Archives]     [Linux XFS Devel]     [Linux Filesystem Development]     [Filesystem Testing]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux