[PATCH v2 nvme-cli 4/4] fabrics: Take the hostnqn parameter from a conf file if not given

[J Freyensee]

On Mon, 2016-08-08 at 15:36 +0200, Christoph Hellwig wrote:
> On Mon, Aug 08, 2016 at 02:58:00PM +0300, Sagi Grimberg wrote:
> > 
> > In order to allow persistent hostnqns, take the hostnqn parameter
> > for /etc/nvme/hostnqn if exists.
> > 
> > Signed-off-by: Sagi Grimberg <sagi at grimberg.me>
> > ---
> >  fabrics.c | 22 +++++++++++++++++++++-
> >  1 file changed, 21 insertions(+), 1 deletion(-)
> > 
> > diff --git a/fabrics.c b/fabrics.c
> > index 54ed8e17b527..90308f07d0fb 100644
> > --- a/fabrics.c
> > +++ b/fabrics.c
> > @@ -57,6 +57,7 @@ struct config {
> >  #define BUF_SIZE		4096
> >  #define PATH_NVME_FABRICS	"/dev/nvme-fabrics"
> >  #define PATH_NVMF_DISC		"/etc/nvme/nvmf_disc"
> > +#define PATH_NVMF_HOSTNQN	"/etc/nvme/hostnqn"
> >  #define MAX_DISC_ARGS		10
> >  
> >  enum {
> > @@ -405,6 +406,25 @@ static void save_discovery_log(struct
> > nvmf_disc_rsp_page_hdr *log, int numrec)
> >  	close(fd);
> >  }
> >  
> > +static int nvmf_hostnqn_file(void)
> > +{
> > +	FILE *f;
> > +	char hostnqn[255];
> > +
> > +	f = fopen(PATH_NVMF_HOSTNQN, "r");
> > +	if (f == NULL)
> > +		return false;
> > +
> > +	if (fgets(hostnqn, sizeof(hostnqn), f) == NULL)
> > +		return false;
> > +
> > +	cfg.hostnqn = strdup(hostnqn);
> > +	if (!cfg.hostnqn)
> > +		return false;
> > +
> > +	return true;
> 
> No fclose?  The leak probably isn't bad for a short running program,
> but sooner or later someone is going to turn it into a library.
> 
> Also shouldn't we just read the 223 bytes of the actual NQN value
> instead
> of the on the wire field length? (which should be 256 anyway IIRC).

That's true, nqn names are 223, unless you have to store the name and
keep it byte-aligned, then having a variable with length 256 is more
appropriate, but then you'll need #define's, 1 for the NQN length, one
to define the array (reminder to null-terminate the string too).

>