Re: [PATCH 5/5] mkfs: substitute slashes with spaces in protofiles

["Darrick J. Wong" <djwong@kernel.org>]

On Thu, Feb 23, 2023 at 09:48:39AM +0100, Carlos Maiolino wrote:
> On Wed, Feb 22, 2023 at 08:37:03AM -0800, Darrick J. Wong wrote:
> > On Wed, Feb 22, 2023 at 10:03:03AM +0100, Carlos Maiolino wrote:
> > > On Thu, Feb 16, 2023 at 01:53:15PM -0800, Darrick J. Wong wrote:
> > > > From: Darrick J. Wong <djwong@kernel.org>
> > > >
> > > > A user requested the ability to specify directory entry names in a
> > > > protofile that have spaces in them.  The protofile format itself does
> > > > not allow spaces (yay 1973-era protofiles!) but it does allow slashes.
> > > > Slashes aren't allowed in directory entry names, so we'll permit this
> > > > one gross hack.
> > > >
> > > > /
> > > > 0 0
> > > > d--775 1000 1000
> > > > : Descending path /code/t/fstests
> > > >  get/isk.sh   ---775 1000 1000 /code/t/fstests/getdisk.sh
> > > > $
> > > >
> > > > Will produce "get isk.h" in the root directory.
> > > >
> > >
> > > While I don't really mind this patch, it seems strange to me to simply replace a
> > > slash with a space in lieu of failing the prototype with an 'invalid character'
> > > error message, or something like that.
> > > With this patch, an user could be mistakenly assuming the get/isk.sh path will
> > > be created, and instead, what's gonna be created is a file with a space.
> > > I don't really mind it, but I think we could be misleading users.
> > 
> > Hmm.  I agree that it /does/ look weird.  I guess we could invent
> > suboptions for mkfs -p:
> 
> Sorry, I don't mean to give you more trouble, but silently replacing a character
> by another one does indeed looks weird to me.

Agreed.

> > 
> >    mkfs.xfs -p slashes_are_spaces=1,/tmp/protofile
> 
> 
> > 
> > Which would turn on this odd looking functionalty?  How does that sound?
> 
> this sounds better to me :)

Ok, will do.

> > 
> > Sidebar:
> > 
> > As it is now, the slash gets passed to _addname, with the result that
> > mkfs will error out.
> > 
> > Another thing I (just) remembered is that the proto.c will also pick up
> > nulls and pass them to addname, so:
> > 
> >    get\000isk.sh ---775 1000 1000 /code/t/fstests/getdisk.sh
> > 
> > will also cause mkfs to fail the format.  I suppose that will also
> > require patching...
> 
> Interesting.
> /me never played much with protofiles
> 
> I'll pickup the first 4 patches of this series tomorrow and skip this one for
> now. I plan to release a 6.2 soon, as we already have linux 6.2 release, but I
> want to do a libxfs sync first, hopefully I'll have the bandwidth tomorrow and
> next week release 6.2. will see :)

Ok.  The slashes_are_spaces=1 patches are pretty trivial, so I'll send
them in Tuesday's batch.

--D

> 
> > 
> > >
> > > > Requested-by: Daan De Meyer <daan.j.demeyer@gmail.com>
> > > > Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> > > > ---
> > > >  mkfs/proto.c |   23 ++++++++++++++++++++++-
> > > >  1 file changed, 22 insertions(+), 1 deletion(-)
> > > >
> > > >
> > > > diff --git a/mkfs/proto.c b/mkfs/proto.c
> > > > index 68ecdbf3632..bf8de0189db 100644
> > > > --- a/mkfs/proto.c
> > > > +++ b/mkfs/proto.c
> > > > @@ -171,6 +171,27 @@ getstr(
> > > >  	return NULL;
> > > >  }
> > > >
> > > > +/* Extract directory entry name from a protofile. */
> > > > +static char *
> > > > +getdirentname(
> > > > +	char	**pp)
> > > > +{
> > > > +	char	*p = getstr(pp);
> > > > +	char	*c = p;
> > > > +
> > > > +	if (!p)
> > > > +		return NULL;
> > > > +
> > > > +	/* Replace slash with space because slashes aren't allowed. */
> > > > +	while (*c) {
> > > > +		if (*c == '/')
> > > > +			*c = ' ';
> > > > +		c++;
> > > > +	}
> > > > +
> > > > +	return p;
> > > > +}
> > > > +
> > > >  static void
> > > >  rsvfile(
> > > >  	xfs_mount_t	*mp,
> > > > @@ -580,7 +601,7 @@ parseproto(
> > > >  			rtinit(mp);
> > > >  		tp = NULL;
> > > >  		for (;;) {
> > > > -			name = getstr(pp);
> > > > +			name = getdirentname(pp);
> > > >  			if (!name)
> > > >  				break;
> > > >  			if (strcmp(name, "$") == 0)
> > > >
> > >
> > > --
> > > Carlos Maiolino
> 
> -- 
> Carlos Maiolino