Re: [RFC 10/16] ieee802154: 6lowpan: add dispatch evalualtion helpers

[Alexander Aring <alex.aring@gmail.com>]

On Wed, Aug 12, 2015 at 02:51:10PM +0200, Stefan Schmidt wrote:
> Hello.
> 
> typo in subject. s/evalualtion/evaluation/
> 

ok.

> On 03/08/15 08:23, Alexander Aring wrote:
> >This patch introduce some static inline function for checking the right
> >dispatch value. This also fixes an bug to detect the right fragmentation
> >dispatch value, we currently use "0xe0" as masking the fragmentation
> >dispatch. Correct is "0xf8" as mask.
> >
> >Signed-off-by: Alexander Aring<alex.aring@gmail.com>
> >---
> >  include/net/6lowpan.h              | 18 +++++++++++++-----
> >  net/ieee802154/6lowpan/6lowpan_i.h |  3 +++
> >  net/ieee802154/6lowpan/rx.c        | 30 +++++++++++++++++++++++++-----
> >  3 files changed, 41 insertions(+), 10 deletions(-)
> >
> >diff --git a/include/net/6lowpan.h b/include/net/6lowpan.h
> >index e16763f..073c791 100644
> >--- a/include/net/6lowpan.h
> >+++ b/include/net/6lowpan.h
> >@@ -126,11 +126,19 @@
> >  	 (((a)[6]) == 0xFF) &&	\
> >  	 (((a)[7]) == 0xFF))
> >-#define LOWPAN_DISPATCH_IPV6	0x41 /* 01000001 = 65 */
> >-#define LOWPAN_DISPATCH_HC1	0x42 /* 01000010 = 66 */
> >-#define LOWPAN_DISPATCH_IPHC	0x60 /* 011xxxxx = ... */
> >-#define LOWPAN_DISPATCH_FRAG1	0xc0 /* 11000xxx */
> >-#define LOWPAN_DISPATCH_FRAGN	0xe0 /* 11100xxx */
> >+#define LOWPAN_DISPATCH_IPV6		0x41 /* 01000001 = 65 */
> >+#define LOWPAN_DISPATCH_IPHC		0x60 /* 011xxxxx = ... */
> >+#define LOWPAN_DISPATCH_IPHC_MASK	0xe0
> >+
> >+static inline bool lowpan_is_ipv6(u8 dispatch)
> >+{
> >+	return dispatch == LOWPAN_DISPATCH_IPV6;
> >+}
> >+
> >+static inline bool lowpan_is_iphc(u8 dispatch)
> >+{
> >+	return (dispatch & LOWPAN_DISPATCH_IPHC_MASK) == LOWPAN_DISPATCH_IPHC;
> >+}
> >  #define LOWPAN_DISPATCH_MASK	0xf8 /* 11111000 */
> >diff --git a/net/ieee802154/6lowpan/6lowpan_i.h b/net/ieee802154/6lowpan/6lowpan_i.h
> >index d62046e..e009a4a 100644
> >--- a/net/ieee802154/6lowpan/6lowpan_i.h
> >+++ b/net/ieee802154/6lowpan/6lowpan_i.h
> >@@ -7,6 +7,9 @@
> >  #include <net/inet_frag.h>
> >  #include <net/6lowpan.h>
> >+#define LOWPAN_DISPATCH_FRAG1		0xc0
> >+#define LOWPAN_DISPATCH_FRAGN		0xe0
> >+
> >  struct lowpan_create_arg {
> >  	u16 tag;
> >  	u16 d_size;
> >diff --git a/net/ieee802154/6lowpan/rx.c b/net/ieee802154/6lowpan/rx.c
> >index c7afd4a..b0066b7 100644
> >--- a/net/ieee802154/6lowpan/rx.c
> >+++ b/net/ieee802154/6lowpan/rx.c
> >@@ -21,6 +21,16 @@ typedef unsigned __bitwise__ lowpan_rx_result;
> >  #define RX_DROP			((__force lowpan_rx_result) 2u)
> >  #define RX_QUEUED		((__force lowpan_rx_result) 3u)
> >+#define LOWPAN_DISPATCH_FIRST		0xc0
> >+#define LOWPAN_DISPATCH_FRAG_MASK	0xf8
> >+#define LOWPAN_DISPATCH_IPHC_MASK	0xe0
> >+
> >+#define LOWPAN_DISPATCH_NALP		0x00
> >+#define LOWPAN_DISPATCH_HC1		0x42
> >+#define LOWPAN_DISPATCH_BC0		0x50
> >+#define LOWPAN_DISPATCH_ESC		0x7f
> >+#define LOWPAN_DISPATCH_MESH		0x80
> >+
> 
> Any specific reason you moved the *_HC1 define here (same for the *_FRAG
> defines above)?
> 
> I would think having them all in 6lowpan.h would make most sense. You might
> need them only in this file for now but later one maybe in another file. It
> could be personal taste but having them all in one place should be easier.
> imho.

There exists 6LoWPAN functionality which is shared between BTLE 6LoWPAN
and IEEE 802.15.4 6LoWPAN, these are:

 - IPHC
 - IPV6

they use both the same dispatch values.

The file "include/net/6lowpan.h" is part of the generic 6LoWPAN branch,
so functionality which belongs to "IPHC" and "IPv6" (IPv6 dispatch is
the easiest dispatch, there is only maybe this one inline function).


The other dispatches are not defined for BTLE, e.g. fragmentation is
handled by MAC layer of BTLE so far I know. Also HC1 is not defined for
BTLE 6LoWPAN.

HC1 is also deprecated, I think there will come no other 6LoWPAN
standard which supports this format. (IPHC is the replacement, See [0],
"MAY implement decompression according to Section 10 of [RFC4944] but
 SHOULD NOT send packets compressed according to.." with otherwords we
should support receiving of HC1 but not sending HC1 anymore.)

If there comming new 6LoWPAN standards then we can later move these
dispatch values to generic 6LoWPAN branch (and maybe more, like
fragmentation framework) if they share these dispatches and also
implement if it looks almost the same.

Nevertheless 6LoWPAN things which is not shared between two or more
link-layer 6LoWPAN types should NOT go into the 6LoWPAN generic branch.

btw:
This is currently not true, "net/include/6lowpan.h" contains some
defines which we should put into net/6lowpan/iphc.c (all IPHC defines
which is used inside "iphc.c") and some things which are used in
802.15.4 6LoWPAN only.

This is all currently not clear and we need some another round for this
to cleanup. (Usually I do send patch-series under 25 patches to have
some milestones).

- Alex

[0] https://tools.ietf.org/html/rfc6282#section-2