Re: [PATCH] fs/9p/xattr.c: avoid format-overflow warning

[Christian Schoenebeck <linux_oss@crudebyte.com>]

On Tuesday, October 10, 2023 4:23:06 AM CEST asmadeus@codewreck.org wrote:
> Christian Schoenebeck wrote on Mon, Oct 09, 2023 at 08:34:15PM +0200:
> > > +++ b/fs/9p/xattr.c
> > > @@ -139,7 +139,7 @@ int v9fs_fid_xattr_set(struct p9_fid *fid, const char *name,
> > >  
> > >  ssize_t v9fs_listxattr(struct dentry *dentry, char *buffer, size_t buffer_size)
> > >  {
> > > -	return v9fs_xattr_get(dentry, NULL, buffer, buffer_size);
> > > +	return v9fs_xattr_get(dentry, "", buffer, buffer_size);
> > >  }
> > >  
> > >  static int v9fs_xattr_handler_get(const struct xattr_handler *handler,
> > > 
> > 
> > Mmm, that's not the same is it? Have you tested this change?
> > 
> > Currently this function causes a 'Txattrwalk' 9p message to be sent to 9p
> > server with its name[s] field being NULL, and the latter being the magical
> > hint to 9p server to not send an attribute, but rather the list of attributes.
> > 
> > With your change I would assume that it would rather ask server for one
> > attribute called "". I have not tested myself, just worrying that it might
> > break behaviour.
> 
> p9pdu_vwritef should output the same (just a 0 length) for both NULL and
> "" so I think it should be ok, but it definitely needs testing.
> 
> I'll try to find time to check (getfattr -d should be enough) later this
> week and add it to the pile

Yeah, I think you are right Dominique, it should end up the same as both cases
result in a string length of 0.

/Christian