On Thu, 9 August 2007 03:39:06 +0100, Christoph Hellwig wrote: > On Thu, Aug 09, 2007 at 04:32:17AM +0200, J??rn Engel wrote: > > On Thu, 9 August 2007 01:19:13 +0100, Christoph Hellwig wrote: > > > On Wed, Aug 08, 2007 at 06:24:47PM +0200, J??rn Engel wrote: > > > > --- linux-2.6.21logfs/include/linux/Kbuild~logfs 2007-08-08 02:57:37.000000000 +0200 > > > > +++ linux-2.6.21logfs/include/linux/Kbuild 2007-08-08 02:57:37.000000000 +0200 > > > > @@ -101,6 +101,7 @@ header-y += jffs2.h > > > > header-y += keyctl.h > > > > header-y += limits.h > > > > header-y += lock_dlm_plock.h > > > > +header-y += logfs.h > > > > > > That file doesn't define any kind of user<->kernel interface and thus should > > > not be exported. > > > > mklogfs needs the header. Does that not qualify? > > No. We been through this example (filesystem layouts) a few times. > It's of course not a kernel interface, and sharing it actually is harmfull, > because it means you can't change the structure layout for your filesystem > definition anymore, which unlike an actual ondisk change happens very > frequently, e.g. when introducing new unions keyed off a feature flag, > or similar things. So please ship a copy with your userspace tools > package. Fair enough. Does the file still make sense in include/linux/ then? Jörn -- All art is but imitation of nature. -- Lucius Annaeus Seneca - To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
