[linux-lvm] Re: devmapper-ioctl 2.4.21 ...

[Herbert Pötzl]

On Wed, Aug 27, 2003 at 03:57:43PM +0100, Joe Thornber wrote:
> On Wed, Aug 27, 2003 at 04:42:40PM +0200, Herbert P?tzl wrote:
> > 
> > Hi!
> > 
> > stumbled over the following code:
> > 
> > +	/* Don't update times for block devices using O_DIRECT */
> > +	if (!(file->f_flags & O_DIRECT) || !S_ISBLK(inode->i_mode)) {
> > +		inode->i_ctime = inode->i_mtime = CURRENT_TIME;
> > +		mark_inode_dirty_sync(inode);
> > +	}
> > 
> > but [A:] S_ISBLK() should be true if the inode is a block device
> > [B:] (file->f_flags & O_DIRECT) should be true if O_DIRECT is requested
> > 
> > !(A || !B)  ==  (!A && B) (De Morgan)
> 
> but it's actually
> (!A ||!B)
> 

*argl*, okay, simply didn't see it ... sorry for the noise ...

thanks for the quick response ...

ahh, by the way, updated the devmapper-ioctl for 2.4.22 final,
are you interrested in the rediff?

best,
Herbert

> - Joe