On Sat, Dec 13, 2014 at 11:29:17AM -0500, Theodore Ts'o wrote:
> On Fri, Nov 07, 2014 at 01:51:52PM -0800, Darrick J. Wong wrote:
> > + ret = ioctl(data->dev, BLKZEROOUT, range);
> > + if (ret)
> > + goto err;
> > + if (!(data->flags & IO_FLAG_DIRECT_IO)) {
> > +#ifdef POSIX_FADV_DONTNEED
> > + ret = posix_fadvise(data->dev, range[0], range[1],
> > + POSIX_FADV_DONTNEED);
> > + if (ret == 0)
> > + goto err;
> > +#endif
> > +#ifdef BLKFLSBUF
> > + ret = ioctl(data->dev, BLKFLSBUF, 0);
> > +#endif
>
> Why do you need BLKFLSBUF after calling POSIX_FADV_DONTNEED? The
I don't; if the DONTNEED call returns 0, we exit.
(The 'goto err' could be 'return 0' to be more clear, I suppose.)
--D
> problem with BLKFLSBUF is that it has a special meaning for ramdisks,
> dating back to when the ramdisk was stored only in the buffer cache, namely:
>
> http://www.memegen.com/meme/hotw4i
>
> We could add yet another special case for the ramdisk, but I'm
> seriously wondering whether the use of BLKZEROOUT is worth it. Sure,
> it will speed up mke2fs, but with the advent of metadata checksums, we
> can skip clearing the journal and the inode table.
>
> - Ted
--
To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
