All of lore.kernel.org
 help / color / mirror / Atom feed
* [Xenomai-core] Extended CAN frame filtering
@ 2007-02-14 16:28 Jan Kiszka
  2007-02-15  8:34 ` [Xenomai-core] " Wolfgang Grandegger
  2007-02-16 12:37 ` Wolfgang Grandegger
  0 siblings, 2 replies; 10+ messages in thread
From: Jan Kiszka @ 2007-02-14 16:28 UTC (permalink / raw)
  To: Wolfgang Grandegger; +Cc: xenomai-core

[-- Attachment #1: Type: text/plain, Size: 2556 bytes --]

Hi Wolfgang,

unless I messed something up, the first patch aligns the implementation of
Socket-CAN filters in Xenomai with their current specification. Right now, if you
set a filter on a standard frame ID, you will also receive extended frames with
the same ID. In contrast, when the extended bit is set, only extended frames are
received, not standard frames with the same ID (that's again spec-conforming).

--- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
+++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
@@ -55,11 +55,11 @@ void rtcan_raw_print_filter(struct rtcan
 static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
 					  can_filter_t *filter)
 {
-   if (filter->can_id & CAN_EFF_FLAG)
-	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
-				 CAN_EFF_FLAG);
+    if (filter->can_id & CAN_EFF_FLAG)
+	recv_filter->can_mask = filter->can_mask & CAN_EFF_MASK;
     else
-	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
+	recv_filter->can_mask = filter->can_mask & CAN_SFF_MASK;
+    recv_filter->can_mask |= CAN_EFF_FLAG;
 
     recv_filter->can_id = filter->can_id & recv_filter->can_mask;
 } 


However, I wonder if this behaviour is useful. You can now either set a filter
for extended frames or standard frame, not for both frame type, just varying on
the ID length. If we consider EFF just as another bit of the CAN ID, we could
take this into account for the mask:

--- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
+++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
@@ -55,11 +55,12 @@ void rtcan_raw_print_filter(struct rtcan
 static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
 					  can_filter_t *filter)
 {
-   if (filter->can_id & CAN_EFF_FLAG)
-	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
-				 CAN_EFF_FLAG);
+    if (filter->can_id & CAN_EFF_FLAG)
+	recv_filter->can_mask = filter->can_mask &
+				(CAN_EFF_MASK | CAN_EFF_FLAG);
     else
-	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
+	recv_filter->can_mask = filter->can_mask &
+				(CAN_SFF_MASK | CAN_EFF_FLAG);
 
     recv_filter->can_id = filter->can_id & recv_filter->can_mask;
 }


Note: this alternative patch would also require a patch to rtdm/rtcan.h.

Actually, the second variant was what I intuitively expected. What is the
behaviour of non-RT Socket-CAN here?

Jan


PS: A few lines above those hunks is still a "#if 0" code block. Either make
this configurable or please clean it up.



[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 250 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-14 16:28 [Xenomai-core] Extended CAN frame filtering Jan Kiszka
@ 2007-02-15  8:34 ` Wolfgang Grandegger
  2007-02-16 12:37 ` Wolfgang Grandegger
  1 sibling, 0 replies; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-15  8:34 UTC (permalink / raw)
  To: Jan Kiszka; +Cc: xenomai-core

Hi Jan,

Jan Kiszka wrote:
> Hi Wolfgang,
> 
> unless I messed something up, the first patch aligns the implementation of
> Socket-CAN filters in Xenomai with their current specification. Right now, if you
> set a filter on a standard frame ID, you will also receive extended frames with
> the same ID. In contrast, when the extended bit is set, only extended frames are
> received, not standard frames with the same ID (that's again spec-conforming).
> 
> --- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
> +++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
> @@ -55,11 +55,11 @@ void rtcan_raw_print_filter(struct rtcan
>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>  					  can_filter_t *filter)
>  {
> -   if (filter->can_id & CAN_EFF_FLAG)
> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
> -				 CAN_EFF_FLAG);
> +    if (filter->can_id & CAN_EFF_FLAG)
> +	recv_filter->can_mask = filter->can_mask & CAN_EFF_MASK;
>      else
> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
> +	recv_filter->can_mask = filter->can_mask & CAN_SFF_MASK;
> +    recv_filter->can_mask |= CAN_EFF_FLAG;
>  
>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>  } 
> 
> 
> However, I wonder if this behaviour is useful. You can now either set a filter
> for extended frames or standard frame, not for both frame type, just varying on
> the ID length. If we consider EFF just as another bit of the CAN ID, we could
> take this into account for the mask:
> 
> --- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
> +++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
> @@ -55,11 +55,12 @@ void rtcan_raw_print_filter(struct rtcan
>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>  					  can_filter_t *filter)
>  {
> -   if (filter->can_id & CAN_EFF_FLAG)
> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
> -				 CAN_EFF_FLAG);
> +    if (filter->can_id & CAN_EFF_FLAG)
> +	recv_filter->can_mask = filter->can_mask &
> +				(CAN_EFF_MASK | CAN_EFF_FLAG);
>      else
> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
> +	recv_filter->can_mask = filter->can_mask &
> +				(CAN_SFF_MASK | CAN_EFF_FLAG);
>  
>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>  }
> 
> 
> Note: this alternative patch would also require a patch to rtdm/rtcan.h.
> 
> Actually, the second variant was what I intuitively expected. What is the
> behaviour of non-RT Socket-CAN here?

Hm, good question. I'm going to ask on the Socket-CAN-ML later today.

> 
> Jan
> 
> 
> PS: A few lines above those hunks is still a "#if 0" code block. Either make
> this configurable or please clean it up.

OK.

Wolfgang.



^ permalink raw reply	[flat|nested] 10+ messages in thread

* [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-14 16:28 [Xenomai-core] Extended CAN frame filtering Jan Kiszka
  2007-02-15  8:34 ` [Xenomai-core] " Wolfgang Grandegger
@ 2007-02-16 12:37 ` Wolfgang Grandegger
  2007-02-17 14:31   ` Wolfgang Grandegger
  1 sibling, 1 reply; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-16 12:37 UTC (permalink / raw)
  To: Jan Kiszka; +Cc: xenomai-core

Jan Kiszka wrote:
> Hi Wolfgang,
> 
> unless I messed something up, the first patch aligns the implementation of
> Socket-CAN filters in Xenomai with their current specification. Right now, if you
> set a filter on a standard frame ID, you will also receive extended frames with
> the same ID. In contrast, when the extended bit is set, only extended frames are
> received, not standard frames with the same ID (that's again spec-conforming).
> 
> --- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
> +++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
> @@ -55,11 +55,11 @@ void rtcan_raw_print_filter(struct rtcan
>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>  					  can_filter_t *filter)
>  {
> -   if (filter->can_id & CAN_EFF_FLAG)
> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
> -				 CAN_EFF_FLAG);
> +    if (filter->can_id & CAN_EFF_FLAG)
> +	recv_filter->can_mask = filter->can_mask & CAN_EFF_MASK;
>      else
> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
> +	recv_filter->can_mask = filter->can_mask & CAN_SFF_MASK;
> +    recv_filter->can_mask |= CAN_EFF_FLAG;
>  
>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>  } 
> 
> 
> However, I wonder if this behaviour is useful. You can now either set a filter
> for extended frames or standard frame, not for both frame type, just varying on
> the ID length. If we consider EFF just as another bit of the CAN ID, we could
> take this into account for the mask:
> 
> --- ksrc/drivers/can/rtcan_raw_filter.c	(Revision 2178)
> +++ ksrc/drivers/can/rtcan_raw_filter.c	(Arbeitskopie)
> @@ -55,11 +55,12 @@ void rtcan_raw_print_filter(struct rtcan
>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>  					  can_filter_t *filter)
>  {
> -   if (filter->can_id & CAN_EFF_FLAG)
> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
> -				 CAN_EFF_FLAG);
> +    if (filter->can_id & CAN_EFF_FLAG)
> +	recv_filter->can_mask = filter->can_mask &
> +				(CAN_EFF_MASK | CAN_EFF_FLAG);
>      else
> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
> +	recv_filter->can_mask = filter->can_mask &
> +				(CAN_SFF_MASK | CAN_EFF_FLAG);
>  
>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>  }
> 
> 
> Note: this alternative patch would also require a patch to rtdm/rtcan.h.
> 
> Actually, the second variant was what I intuitively expected. What is the
> behaviour of non-RT Socket-CAN here?

I checked the filter related code of Socket-CAN. They use a more 
sophisticated filtering scheme:

   if      (filter.can_id & 0x40000000) [CAN_RTR_FLAG]
     accept if ((can_id & filter->can_mask) == filter->can_id)
   else if (filter.can_id & 0x20000000) [CAN_INV_FILTER]
     accept if ((can_id & filter->can_mask) != filter->can_id)
   else if (filter->can_mask == 0)
     accept all messages
   else if filter.can_id & 0x80000000) [CAN_EFF_FLAG]
     accept if (can_id == filter->can_id)
   else
     accept if (can_id == filter->can_id)

In other words, you can define a positive or negative filter using id 
and mask, accept all messages or require an exact id match for standard 
or extended frames. Well, we need something compatible and for this 
reason I triggered a discussion on the Socket-CAN-ML.

Wolfgang.


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-16 12:37 ` Wolfgang Grandegger
@ 2007-02-17 14:31   ` Wolfgang Grandegger
  2007-02-17 17:55     ` Jan Kiszka
  0 siblings, 1 reply; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-17 14:31 UTC (permalink / raw)
  To: Wolfgang Grandegger; +Cc: Jan Kiszka, xenomai-core

[-- Attachment #1: Type: text/plain, Size: 3763 bytes --]

Wolfgang Grandegger wrote:
> Jan Kiszka wrote:
>> Hi Wolfgang,
>>
>> unless I messed something up, the first patch aligns the 
>> implementation of
>> Socket-CAN filters in Xenomai with their current specification. Right 
>> now, if you
>> set a filter on a standard frame ID, you will also receive extended 
>> frames with
>> the same ID. In contrast, when the extended bit is set, only extended 
>> frames are
>> received, not standard frames with the same ID (that's again 
>> spec-conforming).
>>
>> --- ksrc/drivers/can/rtcan_raw_filter.c    (Revision 2178)
>> +++ ksrc/drivers/can/rtcan_raw_filter.c    (Arbeitskopie)
>> @@ -55,11 +55,11 @@ void rtcan_raw_print_filter(struct rtcan
>>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>>                        can_filter_t *filter)
>>  {
>> -   if (filter->can_id & CAN_EFF_FLAG)
>> -    recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
>> -                 CAN_EFF_FLAG);
>> +    if (filter->can_id & CAN_EFF_FLAG)
>> +    recv_filter->can_mask = filter->can_mask & CAN_EFF_MASK;
>>      else
>> -    recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
>> +    recv_filter->can_mask = filter->can_mask & CAN_SFF_MASK;
>> +    recv_filter->can_mask |= CAN_EFF_FLAG;
>>  
>>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>>  }
>>
>> However, I wonder if this behaviour is useful. You can now either set 
>> a filter
>> for extended frames or standard frame, not for both frame type, just 
>> varying on
>> the ID length. If we consider EFF just as another bit of the CAN ID, 
>> we could
>> take this into account for the mask:
>>
>> --- ksrc/drivers/can/rtcan_raw_filter.c    (Revision 2178)
>> +++ ksrc/drivers/can/rtcan_raw_filter.c    (Arbeitskopie)
>> @@ -55,11 +55,12 @@ void rtcan_raw_print_filter(struct rtcan
>>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>>                        can_filter_t *filter)
>>  {
>> -   if (filter->can_id & CAN_EFF_FLAG)
>> -    recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
>> -                 CAN_EFF_FLAG);
>> +    if (filter->can_id & CAN_EFF_FLAG)
>> +    recv_filter->can_mask = filter->can_mask &
>> +                (CAN_EFF_MASK | CAN_EFF_FLAG);
>>      else
>> -    recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
>> +    recv_filter->can_mask = filter->can_mask &
>> +                (CAN_SFF_MASK | CAN_EFF_FLAG);
>>  
>>      recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>>  }
>>
>>
>> Note: this alternative patch would also require a patch to rtdm/rtcan.h.
>>
>> Actually, the second variant was what I intuitively expected. What is the
>> behaviour of non-RT Socket-CAN here?
> 
> I checked the filter related code of Socket-CAN. They use a more 
> sophisticated filtering scheme:
> 
>   if      (filter.can_id & 0x40000000) [CAN_RTR_FLAG]
>     accept if ((can_id & filter->can_mask) == filter->can_id)
>   else if (filter.can_id & 0x20000000) [CAN_INV_FILTER]
>     accept if ((can_id & filter->can_mask) != filter->can_id)
>   else if (filter->can_mask == 0)
>     accept all messages
>   else if filter.can_id & 0x80000000) [CAN_EFF_FLAG]
>     accept if (can_id == filter->can_id)
>   else
>     accept if (can_id == filter->can_id)
> 
> In other words, you can define a positive or negative filter using id 
> and mask, accept all messages or require an exact id match for standard 
> or extended frames. Well, we need something compatible and for this 
> reason I triggered a discussion on the Socket-CAN-ML.

Attached is a patch adding the CAN_INV_FILTER feature. It also updates 
the doc. Now the filter scheme should be compatible with Socket-CAN. If 
it's OK, I will check it in (or you can do it).

Wolfgang.

[-- Attachment #2: xenomai-rtcan-filter.patch --]
[-- Type: text/x-patch, Size: 5563 bytes --]

Index: include/rtdm/rtcan.h
===================================================================
--- include/rtdm/rtcan.h	(revision 2193)
+++ include/rtdm/rtcan.h	(working copy)
@@ -441,28 +441,38 @@ typedef enum CAN_STATE can_state_t;
 
 #define CAN_STATE_OPERATING(state) ((state) < CAN_STATE_BUS_OFF)
 
+/** Invert CAN filter definition */
+#define CAN_INV_FILTER      CAN_ERR_FLAG
+
 /**
  * Filter for reception of CAN messages.
  *
  * This filter works as follows:
  * A received CAN ID is AND'ed bitwise with @c can_mask and then compared to
- * @c can_id. If this comparison is true the message will be received by the
- * socket.
+ * @c can_id. This also includes the @ref CAN_EFF_FLAG and @ref CAN_RTR_FLAG
+ * of @ref CAN_xxx_FLAG. If this comparison is true, the message will be
+ * received by the socket. The logic can be inverted with the @c can_id flag
+ * @ref CAN_INV_FILTER :
+ *
+ * @code
+ * if (can_id & CAN_INV_FILTER) {
+ *    if ((received_can_id & can_mask) != (can_id & ~CAN_INV_FILTER))
+ *       accept-message;
+ * } else {
+ *    if ((received_can_id & can_mask) == can_id)
+ *       accept-message;
+ * }
+ * @endcode
  *
- * Multiple filters can be arranged in a filter list and set with 
- * @ref Sockopts. If one of these filters matches a CAN ID upon reception 
+ * Multiple filters can be arranged in a filter list and set with
+ * @ref Sockopts. If one of these filters matches a CAN ID upon reception
  * of a CAN frame, this frame is accepted.
  *
- * @note Only @ref CAN_EFF_FLAG of @ref CAN_xxx_FLAG "CAN ID flags" is
- * valid for @c can_id and none for @c can_mask. This means that the RTR bit
- * is not taken into account while filtering messages.
- *
- * Extended IDs are received only if @ref CAN_EFF_FLAG is set in
- * @c can_id. If it is cleared only standard IDs are accepted.
  */
 typedef struct can_filter {
 
-    /** CAN ID which must match with incoming IDs after passing the mask */
+    /** CAN ID which must match with incoming IDs after passing the mask.
+     *  The filter logic can be inverted with the flag @ref CAN_INV_FILTER. */
     uint32_t    can_id;
 
     /** Mask which is applied to incoming IDs. See @ref CAN_xxx_MASK
@@ -470,12 +480,11 @@ typedef struct can_filter {
     uint32_t    can_mask;
 } can_filter_t;
 
-
 /**
  * Socket address structure for the CAN address family
  */
 struct sockaddr_can {
-    /** CAN address family, must be @c AF_CAN */    
+    /** CAN address family, must be @c AF_CAN */
     sa_family_t  can_family;
     /** Interface index of CAN controller. See @ref SIOCGIFINDEX. */
     int          can_ifindex;
Index: ksrc/drivers/can/rtcan_module.c
===================================================================
--- ksrc/drivers/can/rtcan_module.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_module.c	(working copy)
@@ -262,11 +262,11 @@ static int rtcan_read_proc_filter(char *
     rtdm_lockctx_t lock_ctx;
     RTCAN_PROC_PRINT_VARS(80);
 
-    /*  fd __CAN_ID__ _CAN_Mask_ MatchCount
-     *   3 0x12345678 0x12345678 1234567890
+    /*  fd __CAN_ID__ _CAN_Mask_ Inv MatchCount
+     *   3 0x12345678 0x12345678  no 1234567890
      */
-    
-    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ MatchCount\n"))
+
+    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ Inv MatchCount\n"))
         goto done;
 
     rtdm_lock_get_irqsave(&rtcan_recv_list_lock, lock_ctx);
@@ -275,10 +275,13 @@ static int rtcan_read_proc_filter(char *
     while (recv_listener != NULL) {
 	context = rtcan_socket_context(recv_listener->sock);
 
-	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %10d\n",
-			      context->fd, 
+	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %s %10d\n",
+			      context->fd,
 			      recv_listener->can_filter.can_id,
-			      recv_listener->can_filter.can_mask,
+			      recv_listener->can_filter.can_mask &
+			      ~CAN_INV_FILTER,
+			      (recv_listener->can_filter.can_mask &
+			       CAN_INV_FILTER) ? "yes" : " no",
 			      recv_listener->match_count))
 	    break;
 	recv_listener = recv_listener->next;
Index: ksrc/drivers/can/rtcan_raw.c
===================================================================
--- ksrc/drivers/can/rtcan_raw.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_raw.c	(working copy)
@@ -67,7 +67,10 @@ static struct rtdm_device rtcan_proto_ra
 
 static inline int rtcan_accept_msg(uint32_t can_id, can_filter_t *filter)
 {
-    return ((can_id & filter->can_mask) == filter->can_id);
+    if ((filter->can_mask & CAN_INV_FILTER))
+	return ((can_id & filter->can_mask) != filter->can_id);
+    else
+	return ((can_id & filter->can_mask) == filter->can_id);
 }
 
 
Index: ksrc/drivers/can/rtcan_raw_filter.c
===================================================================
--- ksrc/drivers/can/rtcan_raw_filter.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_raw_filter.c	(working copy)
@@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
 static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
 					  can_filter_t *filter)
 {
-   if (filter->can_id & CAN_EFF_FLAG)
-	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
-				 CAN_EFF_FLAG);
-    else
-	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
-
-    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
+    if (filter->can_id & CAN_INV_FILTER) {
+	recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
+	recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
+    } else {
+	recv_filter->can_id = filter->can_id;
+	recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
+    }
 }
 
 

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-17 14:31   ` Wolfgang Grandegger
@ 2007-02-17 17:55     ` Jan Kiszka
  2007-02-17 18:20       ` Wolfgang Grandegger
  0 siblings, 1 reply; 10+ messages in thread
From: Jan Kiszka @ 2007-02-17 17:55 UTC (permalink / raw)
  To: Wolfgang Grandegger; +Cc: xenomai-core

[-- Attachment #1: Type: text/plain, Size: 6754 bytes --]

Wolfgang Grandegger wrote:
...
> 
> Attached is a patch adding the CAN_INV_FILTER feature. It also updates
> the doc. Now the filter scheme should be compatible with Socket-CAN. If
> it's OK, I will check it in (or you can do it).
> 
> Wolfgang.
> 
> 
> ------------------------------------------------------------------------
> 
> Index: include/rtdm/rtcan.h
> ===================================================================
> --- include/rtdm/rtcan.h	(revision 2193)
> +++ include/rtdm/rtcan.h	(working copy)

Please increment the profile revision (2 spots: #define and
documentation text).

> @@ -441,28 +441,38 @@ typedef enum CAN_STATE can_state_t;
>  
>  #define CAN_STATE_OPERATING(state) ((state) < CAN_STATE_BUS_OFF)
>  
> +/** Invert CAN filter definition */
> +#define CAN_INV_FILTER      CAN_ERR_FLAG
> +

I would move this close to the other CAN ID flags and explain explicitly
which one is valid in which context. Both are sharing the same bit, so
confusion can arise.

>  /**
>   * Filter for reception of CAN messages.
>   *
>   * This filter works as follows:
>   * A received CAN ID is AND'ed bitwise with @c can_mask and then compared to
> - * @c can_id. If this comparison is true the message will be received by the
> - * socket.
> + * @c can_id. This also includes the @ref CAN_EFF_FLAG and @ref CAN_RTR_FLAG
> + * of @ref CAN_xxx_FLAG. If this comparison is true, the message will be
> + * received by the socket. The logic can be inverted with the @c can_id flag
> + * @ref CAN_INV_FILTER :
> + *
> + * @code
> + * if (can_id & CAN_INV_FILTER) {
> + *    if ((received_can_id & can_mask) != (can_id & ~CAN_INV_FILTER))
> + *       accept-message;
> + * } else {
> + *    if ((received_can_id & can_mask) == can_id)
> + *       accept-message;
> + * }
> + * @endcode

Nice and helpful!

>   *
> - * Multiple filters can be arranged in a filter list and set with 
> - * @ref Sockopts. If one of these filters matches a CAN ID upon reception 
> + * Multiple filters can be arranged in a filter list and set with
> + * @ref Sockopts. If one of these filters matches a CAN ID upon reception
>   * of a CAN frame, this frame is accepted.
>   *
> - * @note Only @ref CAN_EFF_FLAG of @ref CAN_xxx_FLAG "CAN ID flags" is
> - * valid for @c can_id and none for @c can_mask. This means that the RTR bit
> - * is not taken into account while filtering messages.
> - *
> - * Extended IDs are received only if @ref CAN_EFF_FLAG is set in
> - * @c can_id. If it is cleared only standard IDs are accepted.
>   */
>  typedef struct can_filter {
>  
> -    /** CAN ID which must match with incoming IDs after passing the mask */
> +    /** CAN ID which must match with incoming IDs after passing the mask.
> +     *  The filter logic can be inverted with the flag @ref CAN_INV_FILTER. */
>      uint32_t    can_id;
>  
>      /** Mask which is applied to incoming IDs. See @ref CAN_xxx_MASK
> @@ -470,12 +480,11 @@ typedef struct can_filter {
>      uint32_t    can_mask;
>  } can_filter_t;
>  
> -
>  /**
>   * Socket address structure for the CAN address family
>   */
>  struct sockaddr_can {
> -    /** CAN address family, must be @c AF_CAN */    
> +    /** CAN address family, must be @c AF_CAN */
>      sa_family_t  can_family;
>      /** Interface index of CAN controller. See @ref SIOCGIFINDEX. */
>      int          can_ifindex;
> Index: ksrc/drivers/can/rtcan_module.c
> ===================================================================
> --- ksrc/drivers/can/rtcan_module.c	(revision 2193)
> +++ ksrc/drivers/can/rtcan_module.c	(working copy)
> @@ -262,11 +262,11 @@ static int rtcan_read_proc_filter(char *
>      rtdm_lockctx_t lock_ctx;
>      RTCAN_PROC_PRINT_VARS(80);
>  
> -    /*  fd __CAN_ID__ _CAN_Mask_ MatchCount
> -     *   3 0x12345678 0x12345678 1234567890
> +    /*  fd __CAN_ID__ _CAN_Mask_ Inv MatchCount
> +     *   3 0x12345678 0x12345678  no 1234567890
>       */
> -    
> -    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ MatchCount\n"))
> +
> +    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ Inv MatchCount\n"))
>          goto done;
>  
>      rtdm_lock_get_irqsave(&rtcan_recv_list_lock, lock_ctx);
> @@ -275,10 +275,13 @@ static int rtcan_read_proc_filter(char *
>      while (recv_listener != NULL) {
>  	context = rtcan_socket_context(recv_listener->sock);
>  
> -	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %10d\n",
> -			      context->fd, 
> +	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %s %10d\n",
> +			      context->fd,
>  			      recv_listener->can_filter.can_id,
> -			      recv_listener->can_filter.can_mask,
> +			      recv_listener->can_filter.can_mask &
> +			      ~CAN_INV_FILTER,
> +			      (recv_listener->can_filter.can_mask &
> +			       CAN_INV_FILTER) ? "yes" : " no",
>  			      recv_listener->match_count))
>  	    break;
>  	recv_listener = recv_listener->next;
> Index: ksrc/drivers/can/rtcan_raw.c
> ===================================================================
> --- ksrc/drivers/can/rtcan_raw.c	(revision 2193)
> +++ ksrc/drivers/can/rtcan_raw.c	(working copy)
> @@ -67,7 +67,10 @@ static struct rtdm_device rtcan_proto_ra
>  
>  static inline int rtcan_accept_msg(uint32_t can_id, can_filter_t *filter)
>  {
> -    return ((can_id & filter->can_mask) == filter->can_id);
> +    if ((filter->can_mask & CAN_INV_FILTER))
> +	return ((can_id & filter->can_mask) != filter->can_id);
> +    else
> +	return ((can_id & filter->can_mask) == filter->can_id);
>  }
>  
>  
> Index: ksrc/drivers/can/rtcan_raw_filter.c
> ===================================================================
> --- ksrc/drivers/can/rtcan_raw_filter.c	(revision 2193)
> +++ ksrc/drivers/can/rtcan_raw_filter.c	(working copy)
> @@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>  					  can_filter_t *filter)
>  {
> -   if (filter->can_id & CAN_EFF_FLAG)
> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
> -				 CAN_EFF_FLAG);
> -    else
> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
> -
> -    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
> +    if (filter->can_id & CAN_INV_FILTER) {
> +	recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
> +	recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
> +    } else {
> +	recv_filter->can_id = filter->can_id;
> +	recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
> +    }

Why do you push CAN_INV_FILTER internally into the mask instead of
keeping it in the filter's ID - as the pseudo code above states?

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 250 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-17 17:55     ` Jan Kiszka
@ 2007-02-17 18:20       ` Wolfgang Grandegger
  2007-02-17 18:54         ` Jan Kiszka
  0 siblings, 1 reply; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-17 18:20 UTC (permalink / raw)
  To: Jan Kiszka; +Cc: xenomai-core

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
> ...
>> Attached is a patch adding the CAN_INV_FILTER feature. It also updates
>> the doc. Now the filter scheme should be compatible with Socket-CAN. If
>> it's OK, I will check it in (or you can do it).
>>
>> Wolfgang.
>>
>>
>> ------------------------------------------------------------------------
>>
>> Index: include/rtdm/rtcan.h
>> ===================================================================
>> --- include/rtdm/rtcan.h	(revision 2193)
>> +++ include/rtdm/rtcan.h	(working copy)
> 
> Please increment the profile revision (2 spots: #define and
> documentation text).
> 
>> @@ -441,28 +441,38 @@ typedef enum CAN_STATE can_state_t;
>>  
>>  #define CAN_STATE_OPERATING(state) ((state) < CAN_STATE_BUS_OFF)
>>  
>> +/** Invert CAN filter definition */
>> +#define CAN_INV_FILTER      CAN_ERR_FLAG
>> +
> 
> I would move this close to the other CAN ID flags and explain explicitly
> which one is valid in which context. Both are sharing the same bit, so
> confusion can arise.

I already thought about it but was unsure if it's less confusing. OK, 
one more pro, I will change it.

> 
>>  /**
>>   * Filter for reception of CAN messages.
>>   *
>>   * This filter works as follows:
>>   * A received CAN ID is AND'ed bitwise with @c can_mask and then compared to
>> - * @c can_id. If this comparison is true the message will be received by the
>> - * socket.
>> + * @c can_id. This also includes the @ref CAN_EFF_FLAG and @ref CAN_RTR_FLAG
>> + * of @ref CAN_xxx_FLAG. If this comparison is true, the message will be
>> + * received by the socket. The logic can be inverted with the @c can_id flag
>> + * @ref CAN_INV_FILTER :
>> + *
>> + * @code
>> + * if (can_id & CAN_INV_FILTER) {
>> + *    if ((received_can_id & can_mask) != (can_id & ~CAN_INV_FILTER))
>> + *       accept-message;
>> + * } else {
>> + *    if ((received_can_id & can_mask) == can_id)
>> + *       accept-message;
>> + * }
>> + * @endcode
> 
> Nice and helpful!
> 
>>   *
>> - * Multiple filters can be arranged in a filter list and set with 
>> - * @ref Sockopts. If one of these filters matches a CAN ID upon reception 
>> + * Multiple filters can be arranged in a filter list and set with
>> + * @ref Sockopts. If one of these filters matches a CAN ID upon reception
>>   * of a CAN frame, this frame is accepted.
>>   *
>> - * @note Only @ref CAN_EFF_FLAG of @ref CAN_xxx_FLAG "CAN ID flags" is
>> - * valid for @c can_id and none for @c can_mask. This means that the RTR bit
>> - * is not taken into account while filtering messages.
>> - *
>> - * Extended IDs are received only if @ref CAN_EFF_FLAG is set in
>> - * @c can_id. If it is cleared only standard IDs are accepted.
>>   */
>>  typedef struct can_filter {
>>  
>> -    /** CAN ID which must match with incoming IDs after passing the mask */
>> +    /** CAN ID which must match with incoming IDs after passing the mask.
>> +     *  The filter logic can be inverted with the flag @ref CAN_INV_FILTER. */
>>      uint32_t    can_id;
>>  
>>      /** Mask which is applied to incoming IDs. See @ref CAN_xxx_MASK
>> @@ -470,12 +480,11 @@ typedef struct can_filter {
>>      uint32_t    can_mask;
>>  } can_filter_t;
>>  
>> -
>>  /**
>>   * Socket address structure for the CAN address family
>>   */
>>  struct sockaddr_can {
>> -    /** CAN address family, must be @c AF_CAN */    
>> +    /** CAN address family, must be @c AF_CAN */
>>      sa_family_t  can_family;
>>      /** Interface index of CAN controller. See @ref SIOCGIFINDEX. */
>>      int          can_ifindex;
>> Index: ksrc/drivers/can/rtcan_module.c
>> ===================================================================
>> --- ksrc/drivers/can/rtcan_module.c	(revision 2193)
>> +++ ksrc/drivers/can/rtcan_module.c	(working copy)
>> @@ -262,11 +262,11 @@ static int rtcan_read_proc_filter(char *
>>      rtdm_lockctx_t lock_ctx;
>>      RTCAN_PROC_PRINT_VARS(80);
>>  
>> -    /*  fd __CAN_ID__ _CAN_Mask_ MatchCount
>> -     *   3 0x12345678 0x12345678 1234567890
>> +    /*  fd __CAN_ID__ _CAN_Mask_ Inv MatchCount
>> +     *   3 0x12345678 0x12345678  no 1234567890
>>       */
>> -    
>> -    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ MatchCount\n"))
>> +
>> +    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ Inv MatchCount\n"))
>>          goto done;
>>  
>>      rtdm_lock_get_irqsave(&rtcan_recv_list_lock, lock_ctx);
>> @@ -275,10 +275,13 @@ static int rtcan_read_proc_filter(char *
>>      while (recv_listener != NULL) {
>>  	context = rtcan_socket_context(recv_listener->sock);
>>  
>> -	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %10d\n",
>> -			      context->fd, 
>> +	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %s %10d\n",
>> +			      context->fd,
>>  			      recv_listener->can_filter.can_id,
>> -			      recv_listener->can_filter.can_mask,
>> +			      recv_listener->can_filter.can_mask &
>> +			      ~CAN_INV_FILTER,
>> +			      (recv_listener->can_filter.can_mask &
>> +			       CAN_INV_FILTER) ? "yes" : " no",
>>  			      recv_listener->match_count))
>>  	    break;
>>  	recv_listener = recv_listener->next;
>> Index: ksrc/drivers/can/rtcan_raw.c
>> ===================================================================
>> --- ksrc/drivers/can/rtcan_raw.c	(revision 2193)
>> +++ ksrc/drivers/can/rtcan_raw.c	(working copy)
>> @@ -67,7 +67,10 @@ static struct rtdm_device rtcan_proto_ra
>>  
>>  static inline int rtcan_accept_msg(uint32_t can_id, can_filter_t *filter)
>>  {
>> -    return ((can_id & filter->can_mask) == filter->can_id);
>> +    if ((filter->can_mask & CAN_INV_FILTER))
>> +	return ((can_id & filter->can_mask) != filter->can_id);
>> +    else
>> +	return ((can_id & filter->can_mask) == filter->can_id);
>>  }
>>  
>>  
>> Index: ksrc/drivers/can/rtcan_raw_filter.c
>> ===================================================================
>> --- ksrc/drivers/can/rtcan_raw_filter.c	(revision 2193)
>> +++ ksrc/drivers/can/rtcan_raw_filter.c	(working copy)
>> @@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
>>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>>  					  can_filter_t *filter)
>>  {
>> -   if (filter->can_id & CAN_EFF_FLAG)
>> -	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
>> -				 CAN_EFF_FLAG);
>> -    else
>> -	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
>> -
>> -    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>> +    if (filter->can_id & CAN_INV_FILTER) {
>> +	recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
>> +	recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
>> +    } else {
>> +	recv_filter->can_id = filter->can_id;
>> +	recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
>> +    }
> 
> Why do you push CAN_INV_FILTER internally into the mask instead of
> keeping it in the filter's ID - as the pseudo code above states?

To simplify the filter calculation. It actually avoids the expression 
(filter->can_id & ~CAN_INV_FILTER). As this is in a very frequently 
called function, I think it's worth the trick.

Wolfgang.



^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-17 18:20       ` Wolfgang Grandegger
@ 2007-02-17 18:54         ` Jan Kiszka
  2007-02-17 20:49           ` Wolfgang Grandegger
  0 siblings, 1 reply; 10+ messages in thread
From: Jan Kiszka @ 2007-02-17 18:54 UTC (permalink / raw)
  To: Wolfgang Grandegger; +Cc: xenomai-core

[-- Attachment #1: Type: text/plain, Size: 1521 bytes --]

Wolfgang Grandegger wrote:
>>> Index: ksrc/drivers/can/rtcan_raw_filter.c
>>> ===================================================================
>>> --- ksrc/drivers/can/rtcan_raw_filter.c    (revision 2193)
>>> +++ ksrc/drivers/can/rtcan_raw_filter.c    (working copy)
>>> @@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
>>>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>>>                        can_filter_t *filter)
>>>  {
>>> -   if (filter->can_id & CAN_EFF_FLAG)
>>> -    recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
>>> -                 CAN_EFF_FLAG);
>>> -    else
>>> -    recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
>>> -
>>> -    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>>> +    if (filter->can_id & CAN_INV_FILTER) {
>>> +    recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
>>> +    recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
>>> +    } else {
>>> +    recv_filter->can_id = filter->can_id;
>>> +    recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
>>> +    }
>>
>> Why do you push CAN_INV_FILTER internally into the mask instead of
>> keeping it in the filter's ID - as the pseudo code above states?
> 
> To simplify the filter calculation. It actually avoids the expression
> (filter->can_id & ~CAN_INV_FILTER). As this is in a very frequently
> called function, I think it's worth the trick.
> 

Ack. I missed that point on first run.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 250 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-17 18:54         ` Jan Kiszka
@ 2007-02-17 20:49           ` Wolfgang Grandegger
  2007-02-18 18:18             ` Jan Kiszka
  0 siblings, 1 reply; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-17 20:49 UTC (permalink / raw)
  To: Jan Kiszka; +Cc: xenomai-core

[-- Attachment #1: Type: text/plain, Size: 1714 bytes --]

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
>>>> Index: ksrc/drivers/can/rtcan_raw_filter.c
>>>> ===================================================================
>>>> --- ksrc/drivers/can/rtcan_raw_filter.c    (revision 2193)
>>>> +++ ksrc/drivers/can/rtcan_raw_filter.c    (working copy)
>>>> @@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
>>>>  static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
>>>>                        can_filter_t *filter)
>>>>  {
>>>> -   if (filter->can_id & CAN_EFF_FLAG)
>>>> -    recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
>>>> -                 CAN_EFF_FLAG);
>>>> -    else
>>>> -    recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
>>>> -
>>>> -    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
>>>> +    if (filter->can_id & CAN_INV_FILTER) {
>>>> +    recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
>>>> +    recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
>>>> +    } else {
>>>> +    recv_filter->can_id = filter->can_id;
>>>> +    recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
>>>> +    }
>>> Why do you push CAN_INV_FILTER internally into the mask instead of
>>> keeping it in the filter's ID - as the pseudo code above states?
>> To simplify the filter calculation. It actually avoids the expression
>> (filter->can_id & ~CAN_INV_FILTER). As this is in a very frequently
>> called function, I think it's worth the trick.
>>
> 
> Ack. I missed that point on first run.

Attached is a revised patch including change log entry. As this patch 
just extends the existing filter capabilities, it could be applied to 
trunk and the 2.3.x branch as well.

Wolfgang.

[-- Attachment #2: xenomai-rtcan-filter-2.patch --]
[-- Type: text/x-patch, Size: 6808 bytes --]

Index: include/rtdm/rtcan.h
===================================================================
--- include/rtdm/rtcan.h	(revision 2193)
+++ include/rtdm/rtcan.h	(working copy)
@@ -289,9 +289,15 @@ typedef can_id_t can_err_mask_t;
  * @anchor CAN_xxx_FLAG @name CAN ID flags
  * Flags within a CAN ID indicating special CAN frame attributes
  * @{ */
-#define CAN_EFF_FLAG  0x80000000 /**< extended frame           */
-#define CAN_RTR_FLAG  0x40000000 /**< remote transmission flag */
-#define CAN_ERR_FLAG  0x20000000 /**< error frame (see @ref Errors) */
+/** Extended frame */
+#define CAN_EFF_FLAG  0x80000000
+/** Remote transmission frame */
+#define CAN_RTR_FLAG  0x40000000
+/** Error frame (see @ref Errors) */
+#define CAN_ERR_FLAG  0x20000000
+/** Invert CAN filter definition (used for struct can_filter) */
+#define CAN_INV_FILTER CAN_ERR_FLAG
+
 /** @} */
 
 
@@ -446,23 +452,30 @@ typedef enum CAN_STATE can_state_t;
  *
  * This filter works as follows:
  * A received CAN ID is AND'ed bitwise with @c can_mask and then compared to
- * @c can_id. If this comparison is true the message will be received by the
- * socket.
+ * @c can_id. This also includes the @ref CAN_EFF_FLAG and @ref CAN_RTR_FLAG
+ * of @ref CAN_xxx_FLAG. If this comparison is true, the message will be
+ * received by the socket. The logic can be inverted with the @c can_id flag
+ * @ref CAN_INV_FILTER :
+ *
+ * @code
+ * if (can_id & CAN_INV_FILTER) {
+ *    if ((received_can_id & can_mask) != (can_id & ~CAN_INV_FILTER))
+ *       accept-message;
+ * } else {
+ *    if ((received_can_id & can_mask) == can_id)
+ *       accept-message;
+ * }
+ * @endcode
  *
- * Multiple filters can be arranged in a filter list and set with 
- * @ref Sockopts. If one of these filters matches a CAN ID upon reception 
+ * Multiple filters can be arranged in a filter list and set with
+ * @ref Sockopts. If one of these filters matches a CAN ID upon reception
  * of a CAN frame, this frame is accepted.
  *
- * @note Only @ref CAN_EFF_FLAG of @ref CAN_xxx_FLAG "CAN ID flags" is
- * valid for @c can_id and none for @c can_mask. This means that the RTR bit
- * is not taken into account while filtering messages.
- *
- * Extended IDs are received only if @ref CAN_EFF_FLAG is set in
- * @c can_id. If it is cleared only standard IDs are accepted.
  */
 typedef struct can_filter {
 
-    /** CAN ID which must match with incoming IDs after passing the mask */
+    /** CAN ID which must match with incoming IDs after passing the mask.
+     *  The filter logic can be inverted with the flag @ref CAN_INV_FILTER. */
     uint32_t    can_id;
 
     /** Mask which is applied to incoming IDs. See @ref CAN_xxx_MASK
@@ -470,12 +483,11 @@ typedef struct can_filter {
     uint32_t    can_mask;
 } can_filter_t;
 
-
 /**
  * Socket address structure for the CAN address family
  */
 struct sockaddr_can {
-    /** CAN address family, must be @c AF_CAN */    
+    /** CAN address family, must be @c AF_CAN */
     sa_family_t  can_family;
     /** Interface index of CAN controller. See @ref SIOCGIFINDEX. */
     int          can_ifindex;
Index: ChangeLog
===================================================================
--- ChangeLog	(revision 2193)
+++ ChangeLog	(working copy)
@@ -1,3 +1,9 @@
+2007-02-17  Wolfgang Grandegger  <wg@domain.hid>
+
+	* ksrc/drivers/can/rtcan_raw.c, ksrc/drivers/can/rtcan_raw_filter.c,
+	ksrc/drivers/can/rtcan_module.c, include/rtdm/rtcan.h: The CAN filter
+	definition can now be inverted with the can_id flag CAN_INV_FILTER.
+
 2007-02-16  Philippe Gerum  <rpm@xenomai.org>
 
 	* ksrc/skins/native/queue.c (rt_queue_delete): 
@@ -57,7 +63,6 @@
 
 	* ksrc/nucleus/pipe.c (xnpipe_write): Use regular copy_from_user()
 	and check return value.
-
 2007-02-07  Philippe Gerum  <rpm@xenomai.org>
 
 	* ksrc/arch/i386/patches: Upgrade to 2.6.19-1.7-01.
Index: ksrc/drivers/can/rtcan_module.c
===================================================================
--- ksrc/drivers/can/rtcan_module.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_module.c	(working copy)
@@ -262,11 +262,11 @@ static int rtcan_read_proc_filter(char *
     rtdm_lockctx_t lock_ctx;
     RTCAN_PROC_PRINT_VARS(80);
 
-    /*  fd __CAN_ID__ _CAN_Mask_ MatchCount
-     *   3 0x12345678 0x12345678 1234567890
+    /*  fd __CAN_ID__ _CAN_Mask_ Inv MatchCount
+     *   3 0x12345678 0x12345678  no 1234567890
      */
-    
-    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ MatchCount\n"))
+
+    if (!RTCAN_PROC_PRINT("fd __CAN_ID__ _CAN_Mask_ Inv MatchCount\n"))
         goto done;
 
     rtdm_lock_get_irqsave(&rtcan_recv_list_lock, lock_ctx);
@@ -275,10 +275,13 @@ static int rtcan_read_proc_filter(char *
     while (recv_listener != NULL) {
 	context = rtcan_socket_context(recv_listener->sock);
 
-	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %10d\n",
-			      context->fd, 
+	if (!RTCAN_PROC_PRINT("%2d 0x%08x 0x%08x %s %10d\n",
+			      context->fd,
 			      recv_listener->can_filter.can_id,
-			      recv_listener->can_filter.can_mask,
+			      recv_listener->can_filter.can_mask &
+			      ~CAN_INV_FILTER,
+			      (recv_listener->can_filter.can_mask &
+			       CAN_INV_FILTER) ? "yes" : " no",
 			      recv_listener->match_count))
 	    break;
 	recv_listener = recv_listener->next;
Index: ksrc/drivers/can/rtcan_raw.c
===================================================================
--- ksrc/drivers/can/rtcan_raw.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_raw.c	(working copy)
@@ -67,7 +67,10 @@ static struct rtdm_device rtcan_proto_ra
 
 static inline int rtcan_accept_msg(uint32_t can_id, can_filter_t *filter)
 {
-    return ((can_id & filter->can_mask) == filter->can_id);
+    if ((filter->can_mask & CAN_INV_FILTER))
+	return ((can_id & filter->can_mask) != filter->can_id);
+    else
+	return ((can_id & filter->can_mask) == filter->can_id);
 }
 
 
Index: ksrc/drivers/can/rtcan_raw_filter.c
===================================================================
--- ksrc/drivers/can/rtcan_raw_filter.c	(revision 2193)
+++ ksrc/drivers/can/rtcan_raw_filter.c	(working copy)
@@ -55,13 +55,13 @@ void rtcan_raw_print_filter(struct rtcan
 static inline void rtcan_raw_mount_filter(can_filter_t *recv_filter,
 					  can_filter_t *filter)
 {
-   if (filter->can_id & CAN_EFF_FLAG)
-	recv_filter->can_mask = ((filter->can_mask & CAN_EFF_MASK) |
-				 CAN_EFF_FLAG);
-    else
-	recv_filter->can_mask = (filter->can_mask & CAN_SFF_MASK);
-
-    recv_filter->can_id = filter->can_id & recv_filter->can_mask;
+    if (filter->can_id & CAN_INV_FILTER) {
+	recv_filter->can_id = filter->can_id & ~CAN_INV_FILTER;
+	recv_filter->can_mask = filter->can_mask | CAN_INV_FILTER;
+    } else {
+	recv_filter->can_id = filter->can_id;
+	recv_filter->can_mask = filter->can_mask & ~CAN_INV_FILTER;
+    }
 }
 
 

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-17 20:49           ` Wolfgang Grandegger
@ 2007-02-18 18:18             ` Jan Kiszka
  2007-02-18 19:49               ` Wolfgang Grandegger
  0 siblings, 1 reply; 10+ messages in thread
From: Jan Kiszka @ 2007-02-18 18:18 UTC (permalink / raw)
  To: Wolfgang Grandegger; +Cc: xenomai-core

[-- Attachment #1: Type: text/plain, Size: 563 bytes --]

Wolfgang Grandegger wrote:
> ...
> Attached is a revised patch including change log entry. As this patch
> just extends the existing filter capabilities, it could be applied to
> trunk and the 2.3.x branch as well.

You are just forgetting the other changes to the filter mechanism you
implemented.

Thanks for the patch, I just merged your patch into trunk. For v2.3.x, I
applied my first patch I sent on Wednesday to align it to what was
specified on release of 2.3. No breakage inside the stable series unless
something really forces us.

Jan


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 249 bytes --]

^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: [Xenomai-core] Re: Extended CAN frame filtering
  2007-02-18 18:18             ` Jan Kiszka
@ 2007-02-18 19:49               ` Wolfgang Grandegger
  0 siblings, 0 replies; 10+ messages in thread
From: Wolfgang Grandegger @ 2007-02-18 19:49 UTC (permalink / raw)
  To: Jan Kiszka; +Cc: xenomai-core

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
>> ...
>> Attached is a revised patch including change log entry. As this patch
>> just extends the existing filter capabilities, it could be applied to
>> trunk and the 2.3.x branch as well.
> 
> You are just forgetting the other changes to the filter mechanism you
> implemented.
> 
> Thanks for the patch, I just merged your patch into trunk. For v2.3.x, I
> applied my first patch I sent on Wednesday to align it to what was
> specified on release of 2.3. No breakage inside the stable series unless
> something really forces us.

OK, thanks.

Wolfgang.


^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2007-02-18 19:49 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2007-02-14 16:28 [Xenomai-core] Extended CAN frame filtering Jan Kiszka
2007-02-15  8:34 ` [Xenomai-core] " Wolfgang Grandegger
2007-02-16 12:37 ` Wolfgang Grandegger
2007-02-17 14:31   ` Wolfgang Grandegger
2007-02-17 17:55     ` Jan Kiszka
2007-02-17 18:20       ` Wolfgang Grandegger
2007-02-17 18:54         ` Jan Kiszka
2007-02-17 20:49           ` Wolfgang Grandegger
2007-02-18 18:18             ` Jan Kiszka
2007-02-18 19:49               ` Wolfgang Grandegger

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.