From mboxrd@z Thu Jan 1 00:00:00 1970 From: Elise Lennion Subject: Re: [PATCH nft v2] datatype: Display pre-defined inet_service values in host byte order Date: Thu, 8 Dec 2016 15:16:06 -0200 Message-ID: <20161208171606.GA14123@lennorien.com> References: <20161207190331.GA7206@lennorien.com> <20161207192510.GA3314@salvia> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: netfilter-devel@vger.kernel.org To: Pablo Neira Ayuso Return-path: Received: from mail-ua0-f194.google.com ([209.85.217.194]:32940 "EHLO mail-ua0-f194.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753596AbcLHRQN (ORCPT ); Thu, 8 Dec 2016 12:16:13 -0500 Received: by mail-ua0-f194.google.com with SMTP id 20so43262487uak.0 for ; Thu, 08 Dec 2016 09:16:13 -0800 (PST) Content-Disposition: inline In-Reply-To: <20161207192510.GA3314@salvia> Sender: netfilter-devel-owner@vger.kernel.org List-ID: On Wed, Dec 07, 2016 at 08:25:10PM +0100, Pablo Neira Ayuso wrote: > Hi Elise, > > On Wed, Dec 07, 2016 at 05:03:31PM -0200, Elise Lennion wrote: > > nft describe displays, to the user, which values are available for a selector, > > then the values should be in host byte order. > > > > Reported-by: Pablo Neira Ayuso > > Fixes: ccc5da470e76 ("datatype: Replace getnameinfo() by internal lookup table") > > Signed-off-by: Elise Lennion > > --- > > > > v2: Created a function to convert different types and number of bytes > > > > include/datatype.h | 3 ++- > > src/datatype.c | 27 ++++++++++++++++++++++++--- > > src/expression.c | 3 ++- > > 3 files changed, 28 insertions(+), 5 deletions(-) > > > > diff --git a/include/datatype.h b/include/datatype.h > > index d4fe817..a7db1df 100644 > > --- a/include/datatype.h > > +++ b/include/datatype.h > > @@ -191,7 +191,8 @@ extern struct error_record *symbolic_constant_parse(const struct expr *sym, > > extern void symbolic_constant_print(const struct symbol_table *tbl, > > const struct expr *expr, bool quotes); > > extern void symbol_table_print(const struct symbol_table *tbl, > > - const struct datatype *dtype); > > + const struct datatype *dtype, > > + enum byteorder byteorder); > > > > extern struct symbol_table *rt_symbol_table_init(const char *filename); > > extern void rt_symbol_table_free(struct symbol_table *tbl); > > diff --git a/src/datatype.c b/src/datatype.c > > index b5d73bc..c884171 100644 > > --- a/src/datatype.c > > +++ b/src/datatype.c > > @@ -180,15 +180,36 @@ void symbolic_constant_print(const struct symbol_table *tbl, > > printf("%s", s->identifier); > > } > > > > +static void big_to_host_byteorder(void *value, const unsigned int nbytes) > > +{ > > + unsigned char *p = value; > > + unsigned int i; > > + > > + for (i = 0; i < nbytes / 2; i++) { > > + p[i] ^= p[nbytes - 1 - i]; > > + p[nbytes - 1 - i] ^= p[i]; > > + p[i] ^= p[nbytes - 1 - i]; > > + } > > Sorry, probably I should have specified a bit more. > > I suggest you use libgmp for this, that allows any arbitrary word > size. See mpz_init() to initialize the variable, then > mpz_switch_byteorder() to change byteorder. Then, export it via > mpz_export_data() back to fixed standard datatypes. > > libgmp is well documented, please have a look at it. > > Thanks. Hello Pablo, Should I write a function that switches byteorder of different types, for possible general use, or just use libgmp inside symbol_table_print? In symbol_table_print the converted values have a specific type, so, something similar to the following is enough, right? ... if (byteorder == BYTEORDER_BIG_ENDIAN) { mpz_init_set_ui(v, value); mpz_switch_byteorder(v, nbytes); mpz_export_data(&value, v, BYTEORDER_HOST_ENDIAN, nbytes); mpz_clear(v); } ... Thank you, elise