Skip to contentSkip to conversation

: use explicit enums for CCID 3 states

PATCH Applied Gerrit Renker 2 participants 4 messages Nov 15, 2006 – Nov 16, 2006 Networking

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 : use explicit enums for CCID 3 states

Hi Arnaldo,

thank you for the link @-)

|  > The following is the revised patch with three macros added, I liked the
|  > DCCP_BUG_ON():
|  
|  will look into that barring social life that is taking me so much time
|  that I should be working on DCCP :-P :-) Well, at least its not "Real
|  Work (tm)" :-)

I really hope you can get some quality time - there should be funding for all the good work you are putting into this.

I have made a terrible mess of the previous patch (with the warn semantics, it would not have worked correctly). I have tested it and below is what I had intended, but was too tired to get clear.

Best,
Gerrit

net/dccp/dccp.h +8 -0
@@ -18,6 +18,14 @@
 #include <net/tcp.h>
 #include "ackvec.h"
 
+#define DCCP_CRIT(fmt, a...) LIMIT_NETDEBUG(KERN_CRIT fmt " at %s:%d/%s()\n",  \
+					 ##a, __FILE__, __LINE__, __FUNCTION__)
+#define DCCP_BUG(fmt, a...)  do { DCCP_CRIT(fmt, ##a); dump_stack(); } while (0)
+#define DCCP_BUG_ON(cond)    do { if (unlikely((cond) = 0))                   \
+					DCCP_BUG("BUG: condition \"%s\" fails",\
+						 __stringify((cond)));	       \
+			     } while (0)
+
 #ifdef MODULE
 #define DCCP_PRINTK(enable, fmt, args...)	do { if (enable)	     \
 							printk(fmt, ##args); \
net/dccp/ccids/ccid3.h +17 -2
@@ -73,6 +73,14 @@ struct ccid3_options_received {
 	u32 ccid3or_receive_rate;
 };
 
+/* TFRC sender states */
+enum ccid3_hc_tx_states {
+       	TFRC_SSTATE_NO_SENT = 1,
+	TFRC_SSTATE_NO_FBACK,
+	TFRC_SSTATE_FBACK,
+	TFRC_SSTATE_TERM,
+};
+
 /** struct ccid3_hc_tx_sock - CCID3 sender half connection sock
  *
   * @ccid3hctx_state - Sender state
@@ -103,7 +111,7 @@ struct ccid3_hc_tx_sock {
 #define ccid3hctx_t_rto			ccid3hctx_tfrc.tfrctx_rto
 #define ccid3hctx_t_ipi			ccid3hctx_tfrc.tfrctx_ipi
 	u16				ccid3hctx_s;
-  	u8				ccid3hctx_state;
+  	enum ccid3_hc_tx_states 	ccid3hctx_state:8;
 	u8				ccid3hctx_last_win_count;
 	u8				ccid3hctx_idle;
 	struct timeval			ccid3hctx_t_last_win_count;
@@ -115,6 +123,13 @@ struct ccid3_hc_tx_sock {
 	struct ccid3_options_received	ccid3hctx_options_received;
 };
 
+/* TFRC receiver states */
+enum ccid3_hc_rx_states {
+       	TFRC_RSTATE_NO_DATA = 1,
+	TFRC_RSTATE_DATA,
+	TFRC_RSTATE_TERM    = 127,
+};
+
 struct ccid3_hc_rx_sock {
 	struct tfrc_rx_info	ccid3hcrx_tfrc;
 #define ccid3hcrx_x_recv	ccid3hcrx_tfrc.tfrcrx_x_recv
@@ -122,8 +137,8 @@ struct ccid3_hc_rx_sock {
 #define ccid3hcrx_p		ccid3hcrx_tfrc.tfrcrx_p
   	u64			ccid3hcrx_seqno_nonloss:48,
 				ccid3hcrx_ccval_nonloss:4,
-				ccid3hcrx_state:8,
 				ccid3hcrx_ccval_last_counter:4;
+	enum ccid3_hc_rx_states	ccid3hcrx_state:8;
   	u32			ccid3hcrx_bytes_recv;
   	struct timeval		ccid3hcrx_tstamp_last_feedback;
   	struct timeval		ccid3hcrx_tstamp_last_ack;
net/dccp/ccids/ccid3.c +12 -33

93 lines not on this page · raw message

#1 lore raw
Quoted 4 lines · Gerrit Renker

Hi Arnaldo,

thank you for the link @-)

Quoted 8 lines
|  > The following is the revised patch with three macros added, I liked the
|  > DCCP_BUG_ON():
|
|  will look into that barring social life that is taking me so much time
|  that I should be working on DCCP :-P :-) Well, at least its not "Real
|  Work (tm)" :-)

I really hope you can get some quality time - there should be funding for all the good work you are putting into this.

Thanks, Mandriva allows me to work part time on what I think is important, barring when there is some important customer project, that was the case for at least 6 months this year, when I mostly disappeared from my kernel projects.

Quoted 3 lines

I have made a terrible mess of the previous patch (with the warn semantics, it would not have worked correctly). I have tested it and below is what I had intended, but was too tired to get clear.

:-)

Seems ok now:

[acme@newtoy net-2.6.20]$ codiff --verbose /tmp/ccid3.o.before /tmp/ccid3.o.enum /pub/scm/linux/kernel/git/acme/net-2.6.20/net/dccp/ccids/ccid3.c:
struct ccid3_hc_tx_sock | +0

    ccid3hctx_state;
     from: u8                    /*    30(0)     1(0) */
     to:   enum ccid3_hc_tx_states /*    28(8)     4(8) */
  struct ccid3_hc_rx_sock       |   +0
    ccid3hcrx_state:8;
     from: u64                   /*    16(36)     8(8) */
     to:   enum ccid3_hc_rx_states /*    16(0)     4(8) */
    ccid3hcrx_ccval_last_counter:4;
     from: u64                   /*    16(32)     8(4) */
     to:   u64                   /*    16(40)     8(4) */
 2 structs changed
  ccid3_hc_rx_set_state         |  -28
  ccid3_hc_rx_send_feedback     |  +26
  ccid3_hc_rx_init              |  -11
  ccid3_hc_tx_no_feedback_timer |  +32
  ccid3_hc_tx_packet_recv       |  +32
  ccid3_hc_rx_packet_recv       |  -16
  ccid3_hc_rx_get_info          |   -4
  ccid3_hc_tx_send_packet       |  +27
  ccid3_hc_tx_packet_sent       |  +32

9 functions changed, 149 bytes added, 59 bytes removed

Thanks, applied.

- Arnaldo

#2 lore raw
In reply to Gerrit Renker

Hi Arnaldo,

| Thanks, applied.
many thanks for applying this; it was sooner as expected and I think it would be good to tidy this up and make it consistent throughout the code. Since I initiated this, I feel a bit obliged to improve this.

In particular, I have thought about the following scheme which divides warning messages into two (plus debug) categories

a) GRADE 1: Severe Protocol/Internal Faults
This includes
* violated pre- / postconditions,
* illegal states
* malfunctioning (something is NULL)
I think these messages should not be rate-limited and, catching up on the discussion, should dump the stack after the warning message.

b) GRADE 2: Plain Misbehaviour
* out-of-memory messages
* Request floods
* protocol violations (wrong message/option format)

To make this consistent, I would like input whether the following is ok with you:

1) use DCCP_BUG for all Grade 1 type of messages, i.e. it would replace
* BUG()
* WARN_ON() (of course with the appropriate message)
* all sections using 'printk(KERN_CRIT ... ); dump_stack()'

2) either keep the LIMIT_NETDEBUG(KERN_WARNING ... ) for Grade 2 or put it into a macro which would add the __FUNCTION__ thing to it

The point is that, if it is done consistently, (1) can later be replaced with the more conventional BUG() or WARN_ON() once everyone is sure that the code behaves exactly as it shoulds. For the moment, as Ian has pointed out, bringing the machine to a grinding halt for every little problem is probably too much of a nuisance (in the same vein, I would really like to suggest the return of Unload Hack :0).

If you are ok with this basic scheme, I shall set about and tidy things up, otherwise please let me know what the constraints should be with regard to warning/bug messages.

Best,
Gerrit

#3 lore raw
In reply to Gerrit Renker
Quoted 6 lines · Gerrit Renker

Hi Arnaldo,

| Thanks, applied.
many thanks for applying this; it was sooner as expected and I think it would be good to tidy this up and make it consistent throughout the code. Since I initiated this, I feel a bit obliged to improve this.

:)

Quoted 23 lines

In particular, I have thought about the following scheme which divides warning messages into two (plus debug) categories

a) GRADE 1: Severe Protocol/Internal Faults
This includes
* violated pre- / postconditions,
* illegal states
* malfunctioning (something is NULL)
I think these messages should not be rate-limited and, catching up on the discussion, should dump the stack after the warning message.

b) GRADE 2: Plain Misbehaviour
* out-of-memory messages
* Request floods
* protocol violations (wrong message/option format)

To make this consistent, I would like input whether the following is ok with you:

1) use DCCP_BUG for all Grade 1 type of messages, i.e. it would replace
* BUG()
* WARN_ON() (of course with the appropriate message)
* all sections using 'printk(KERN_CRIT ... ); dump_stack()'

Quoted 5 lines

2) either keep the LIMIT_NETDEBUG(KERN_WARNING ... ) for Grade 2 or put it into a macro which would add the __FUNCTION__ thing to it

The point is that, if it is done consistently, (1) can later be replaced with the more conventional BUG() or WARN_ON() once everyone is sure that the code behaves exactly as it shoulds.

Using a DCCP_ prefix for BUG(), BUG_ON(), WARN_ON() and a new WARN() with the difference being none would panic the machine, but warn, in some cases ratelimited, seems to be the way to go, later, as you suggests, we just remove the prefix.

Quoted 3 lines

For the moment, as Ian has pointed out, bringing the machine to a grinding halt for every little problem is probably too much of a nuisance (in the same vein, I would really like to suggest the return of Unload Hack :0).

I'll look at what to do to avoid allowing the removal while there are TW sockets on the death row :-)

Quoted 6 lines

If you are ok with this basic scheme, I shall set about and tidy things up, otherwise please let me know what the constraints should be with regard to warning/bug messages.

Best,
Gerrit