[Xenomai-help] rtcan_mscan compilation.

[Xenomai-help] rtcan_mscan compilation.

[Syed Amer Gilani]

I wanted to test rtcan but the mscan Driver fails to compile:

  CC      drivers/xenomai/can/mscan/rtcan_mscan.o
drivers/xenomai/can/mscan/rtcan_mscan.c:85: error: syntax error before
string constant
drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: type defaults to
'int' in declaration of 'MODULE_PARM'
drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: function
declaration isn't a prototype
drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: data definition
has no type or storage class
drivers/xenomai/can/mscan/rtcan_mscan.c:89: error: syntax error before
string constant
drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: type defaults to
'int' in declaration of 'MODULE_PARM'
drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: function
declaration isn't a prototype
drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: data definition
has no type or storage class
make[4]: *** [drivers/xenomai/can/mscan/rtcan_mscan.o] Fehler 1

When i understand the Sources correct it has something to do with the
value of the config variables. But i left them all on the default
values:

CONFIG_XENO_DRIVERS_RTCAN_MSCAN=y
CONFIG_XENO_DRIVERS_RTCAN_MSCAN_1=y
CONFIG_XENO_DRIVERS_RTCAN_MSCAN_2=y
CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK=66000000
CONFIG_XENO_DRIVERS_RTCAN_MSCAN_ALT=y
# CONFIG_XENO_DRIVERS_RTCAN_MSCAN_PSC2 is not set

Regards,
Syed Amer Gilani

Re: [Xenomai-help] rtcan_mscan compilation.

[Wolfgang Grandegger]

Syed Amer Gilani wrote:
> I wanted to test rtcan but the mscan Driver fails to compile:
> 
>  CC      drivers/xenomai/can/mscan/rtcan_mscan.o
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: error: syntax error before
> string constant
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: type defaults to
> 'int' in declaration of 'MODULE_PARM'
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: function
> declaration isn't a prototype
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: data definition
> has no type or storage class
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: error: syntax error before
> string constant
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: type defaults to
> 'int' in declaration of 'MODULE_PARM'
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: function
> declaration isn't a prototype
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: data definition
> has no type or storage class
> make[4]: *** [drivers/xenomai/can/mscan/rtcan_mscan.o] Fehler 1
> 
> When i understand the Sources correct it has something to do with the
> value of the config variables. But i left them all on the default
> values:
> 
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_1=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_2=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK=66000000
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_ALT=y
> # CONFIG_XENO_DRIVERS_RTCAN_MSCAN_PSC2 is not set

Yes, I know since yesterday and it proofs again, untested software does 
  not work :-(. I'm working on a fix. Should be available today or tomorrow.

Wolfgang.

Re: [Xenomai-help] rtcan_mscan compilation.

[Wolfgang Grandegger]

[-- Attachment #1: Type: text/plain, Size: 1898 bytes --]

Syed Amer Gilani wrote:
> I wanted to test rtcan but the mscan Driver fails to compile:
> 
>  CC      drivers/xenomai/can/mscan/rtcan_mscan.o
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: error: syntax error before
> string constant
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: type defaults to
> 'int' in declaration of 'MODULE_PARM'
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: function
> declaration isn't a prototype
> drivers/xenomai/can/mscan/rtcan_mscan.c:85: warning: data definition
> has no type or storage class
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: error: syntax error before
> string constant
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: type defaults to
> 'int' in declaration of 'MODULE_PARM'
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: function
> declaration isn't a prototype
> drivers/xenomai/can/mscan/rtcan_mscan.c:89: warning: data definition
> has no type or storage class
> make[4]: *** [drivers/xenomai/can/mscan/rtcan_mscan.o] Fehler 1
> 
> When i understand the Sources correct it has something to do with the
> value of the config variables. But i left them all on the default
> values:
> 
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_1=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_2=y
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK=66000000
> CONFIG_XENO_DRIVERS_RTCAN_MSCAN_ALT=y
> # CONFIG_XENO_DRIVERS_RTCAN_MSCAN_PSC2 is not set

It did at least compile under Linux 2.6.14, but MODULE_PARM has been 
removed in 2.6.17 :-(. Attached is a patch to get the MSCAN driver 
working under Linux 2.6.17. Philippe or Jan, could you please apply it, 
thanks.

Also be aware that on the MPC5200B the MSCAN clock source used is the 
oscillator clock (SYS_XTAL_IN) while on the MPC5200 it is the IP bus 
clock (IP_CLK) due to hardware bugs (see chip errata). Likely 
SYS_XTAL_IN runs at 33 MHz on your system.

Wolfgang.

[-- Attachment #2: xenomai-rtcan-mscan-2.6.patch --]
[-- Type: text/x-patch, Size: 5444 bytes --]

Index: ChangeLog
===================================================================
--- ChangeLog	(revision 1695)
+++ ChangeLog	(working copy)
@@ -1,3 +1,18 @@
+2006-10-06  Wolfgang Grandegger  <wg@domain.hid>
+
+	* ksrc/drivers/can/rtcan_module.c (rtcan_dev_get_state_name): Bug
+	fix. Function did return invalid state name.
+
+	* ksrc/drivers/can/rtcan_internal.h: MODULE_PARM has been removed
+	in Linux 2.6.17. Replaced with compat_module_*_param definitions.
+	
+	* ksrc/drivers/can/mscan/rtcan_mscan.c: Use compat_module_*
+	functions in the MSCAN driver as well.
+
+	* ksrc/drivers/can/mscan/rtcan_mscan_regs.h: Fix invalid address
+	for MPC5200 GPIO registers. Now the MSCAN driver works for Linux
+	2.6.17.
+	
 2006-10-06  Jan Kiszka  <jan.kiszka@domain.hid>
 
 	* src/testsuite/cyclic/cyclictest.c: Align to latest v0.11,
Index: ksrc/drivers/can/rtcan_internal.h
===================================================================
--- ksrc/drivers/can/rtcan_internal.h	(revision 1695)
+++ ksrc/drivers/can/rtcan_internal.h	(working copy)
@@ -111,6 +111,10 @@
 
 #ifndef compat_module_int_param_array
 #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
+# define compat_module_int_param(name) \
+    MODULE_PARM(name, "i")
+# define compat_module_charp_param(name) \
+    MODULE_PARM(name, "s")
 # define compat_module_byte_param_array(name, count) \
     MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
 # define compat_module_short_param_array(name, count) \
@@ -118,6 +122,10 @@
 # define compat_module_int_param_array(name, count) \
     MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
 #else
+# define compat_module_int_param(name) \
+    module_param(name, int, 0444)
+# define compat_module_charp_param(name) \
+    module_param(name, charp, 0444)
 # define compat_module_byte_param_array(name, count) \
     module_param_array(name, byte, NULL, 0444)
 # define compat_module_short_param_array(name, count) \
Index: ksrc/drivers/can/rtcan_module.c
===================================================================
--- ksrc/drivers/can/rtcan_module.c	(revision 1695)
+++ ksrc/drivers/can/rtcan_module.c	(working copy)
@@ -54,15 +54,15 @@
 }
 
 static char *rtcan_state_names[] = {
-    "active", "stopped", "sleeping", "warning", 
-    "passive" , "bus-off", "scanning"
+    "active", "warning", "passive" , "bus-off", 
+    "scanning", "stopped", "sleeping"
 };
 
 static void rtcan_dev_get_state_name(can_state_t state, 
 				     char* name, int max_len)
 {
     if (state >= CAN_STATE_ACTIVE && 
-	state <= CAN_STATE_SCANNING_BAUDRATE)
+	state <= CAN_STATE_SLEEPING)
 	strncpy(name, rtcan_state_names[state], max_len);
     else
 	strncpy(name, "unknown", max_len);
Index: ksrc/drivers/can/mscan/rtcan_mscan.c
===================================================================
--- ksrc/drivers/can/mscan/rtcan_mscan.c	(revision 1695)
+++ ksrc/drivers/can/mscan/rtcan_mscan.c	(working copy)
@@ -37,6 +37,7 @@
 #include <rtdm/rtcan.h>
 #include "rtcan_dev.h"
 #include "rtcan_raw.h"
+#include "rtcan_internal.h"
 #include "rtcan_mscan_regs.h"
 
 extern int rtcan_mscan_create_proc(struct rtcan_device* dev);
@@ -56,7 +57,7 @@
 
 /** Module parameter for the CAN controllers' */
 
-static int port[RTCAN_MSCAN_DEVS] = {
+int port[RTCAN_MSCAN_DEVS] = {
 #ifdef CONFIG_XENO_DRIVERS_RTCAN_MSCAN_1
 #ifdef CONFIG_XENO_DRIVERS_RTCAN_MSCAN_2
 	1, 2  /* Enable CAN 1 and 2 */
@@ -71,22 +72,20 @@
 #endif
 #endif
 };
-
-static int port_count = RTCAN_MSCAN_DEVS;
-
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,10)
-module_param_array(port, int, &port_count, 0400);
-#else
-MODULE_PARM(port, "2" __MODULE_STRING(RTCAN_MSCAN_DEVS) "i");
-#endif
+compat_module_int_param_array(port, RTCAN_MSCAN_DEVS);
 MODULE_PARM_DESC(port, "Port numbers of enabled controllers, e.g. 1,2");
 
-unsigned long mscan_clock = CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK;
-MODULE_PARM(mscan_clock,"i");
+/* 
+ * Note: on the MPC5200 the MSCAN clock source is the IP bus 
+ * clock (IP_CLK) while on the MPC5200B it is the oscillator 
+ * clock (SYS_XTAL_IN).
+ */
+unsigned int mscan_clock = CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK;
+compat_module_int_param(mscan_clock);
 MODULE_PARM_DESC(mscan_clock, "Clock frequency in Hz");
 
 char *mscan_pins = NULL;
-MODULE_PARM(mscan_pins,"s");
+compat_module_charp_param(mscan_pins);
 MODULE_PARM_DESC(mscan_pins, "Routing to GPIO pins (PSC2 or I2C1/TMR01)");
 
  
@@ -739,7 +738,6 @@
     }
 }
 
-/** Init module */
 int __init rtcan_mscan_init_one(int idx)
 {
     int ret, irq;
@@ -836,7 +834,6 @@
 
 }
 
-/** Init module */
 static int __init rtcan_mscan_init(void)
 {
     int i, ret;
@@ -854,14 +851,12 @@
     return 0;
 }
 
-
-/** Cleanup module */
 static void __exit rtcan_mscan_exit(void)
 {
     int i;
     struct rtcan_device *dev;
 
-    for (i = 0; i < port_count; i++) {
+    for (i = 0; i < RTCAN_MSCAN_DEVS; i++) {
 	if ((dev = rtcan_mscan_devs[i]) == NULL)
 	    continue;
 	
Index: ksrc/drivers/can/mscan/rtcan_mscan_regs.h
===================================================================
--- ksrc/drivers/can/mscan/rtcan_mscan_regs.h	(revision 1695)
+++ ksrc/drivers/can/mscan/rtcan_mscan_regs.h	(working copy)
@@ -35,7 +35,7 @@
 #define MSCAN_MBAR	MPC52xx_MBAR
 #define MSCAN_CAN1_IRQ	MPC52xx_MSCAN1_IRQ
 #define MSCAN_CAN2_IRQ	MPC52xx_MSCAN2_IRQ
-#define MPC5xxx_GPIO	MPC52xx_GPIO_OFFSET 
+#define MPC5xxx_GPIO	MPC52xx_VA(MPC52xx_GPIO_OFFSET) 
 #define mpc5xxx_gpio	mpc52xx_gpio
 #endif
 

Re: [Xenomai-help] rtcan_mscan compilation.

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 1400 bytes --]

Wolfgang Grandegger wrote:
> ...
> Index: ksrc/drivers/can/rtcan_internal.h
> ===================================================================
> --- ksrc/drivers/can/rtcan_internal.h	(revision 1695)
> +++ ksrc/drivers/can/rtcan_internal.h	(working copy)
> @@ -111,6 +111,10 @@
>  
>  #ifndef compat_module_int_param_array
>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
> +# define compat_module_int_param(name) \
> +    MODULE_PARM(name, "i")
> +# define compat_module_charp_param(name) \
> +    MODULE_PARM(name, "s")
>  # define compat_module_byte_param_array(name, count) \
>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>  # define compat_module_short_param_array(name, count) \
> @@ -118,6 +122,10 @@
>  # define compat_module_int_param_array(name, count) \
>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>  #else
> +# define compat_module_int_param(name) \
> +    module_param(name, int, 0444)
> +# define compat_module_charp_param(name) \
> +    module_param(name, charp, 0444)
>  # define compat_module_byte_param_array(name, count) \
>      module_param_array(name, byte, NULL, 0444)
>  # define compat_module_short_param_array(name, count) \

No need, module_param comes even with 2.4. We only need wrapping for
parameter arrays (due to different semantics of the macro arguments).

So, this patch can be shorter... :)

Jan


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

Re: [Xenomai-help] rtcan_mscan compilation.

[Wolfgang Grandegger]

[-- Attachment #1: Type: text/plain, Size: 2571 bytes --]

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
>> ...
>> Index: ksrc/drivers/can/rtcan_internal.h
>> ===================================================================
>> --- ksrc/drivers/can/rtcan_internal.h	(revision 1695)
>> +++ ksrc/drivers/can/rtcan_internal.h	(working copy)
>> @@ -111,6 +111,10 @@
>>  
>>  #ifndef compat_module_int_param_array
>>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
>> +# define compat_module_int_param(name) \
>> +    MODULE_PARM(name, "i")
>> +# define compat_module_charp_param(name) \
>> +    MODULE_PARM(name, "s")
>>  # define compat_module_byte_param_array(name, count) \
>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>>  # define compat_module_short_param_array(name, count) \
>> @@ -118,6 +122,10 @@
>>  # define compat_module_int_param_array(name, count) \
>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>>  #else
>> +# define compat_module_int_param(name) \
>> +    module_param(name, int, 0444)
>> +# define compat_module_charp_param(name) \
>> +    module_param(name, charp, 0444)
>>  # define compat_module_byte_param_array(name, count) \
>>      module_param_array(name, byte, NULL, 0444)
>>  # define compat_module_short_param_array(name, count) \
> 
> No need, module_param comes even with 2.4. We only need wrapping for
> parameter arrays (due to different semantics of the macro arguments).

But with limitations, unfortunately:

   /* type is byte, short, ushort, int, uint, long, ulong, bool. (2.6
      has more, but they are not supported).  perm is permissions when
      it appears in sysfs: 0 means doens't appear, 0444 means read-only
      by everyone, 0644 means changable dynamically by root, etc.  name
      must be in scope (unlike MODULE_PARM).
   */
   #define module_param(name, type, perm) \
         static inline void *__check_existence_##name(void) { return
              &name; } \
         MODULE_PARM(name, _MODULE_PARM_STRING_ ## type)

   #define _MODULE_PARM_STRING_byte "b"
   #define _MODULE_PARM_STRING_short "h"
   #define _MODULE_PARM_STRING_ushort "h"
   #define _MODULE_PARM_STRING_int "i"
   #define _MODULE_PARM_STRING_uint "i"
   #define _MODULE_PARM_STRING_long "l"
   #define _MODULE_PARM_STRING_ulong "l"
   #define _MODULE_PARM_STRING_bool "i"

Especially "s" is missing. But it could be wrapped with:

   #define _MODULE_PARM_STRING_charp "s"

Any idea why this is not already done?

I also want to replace the compat_module_*_param_array functions with 
the attached code snippet.

Wolfgang.


> So, this patch can be shorter... :)
> 
> Jan
> 


[-- Attachment #2: compat_module_param_array.c --]
[-- Type: text/x-csrc, Size: 465 bytes --]

#ifndef compat_module_param_array
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
# define _MODULE_PARM_STRING_charp "s"
# define compat_module_param_array(name, type, count, perm) \
    static inline void *__check_existence_##name(void) { return &name; } \
    MODULE_PARM(name, "1-" __MODULE_STRING(count) _MODULE_PARM_STRING_##type)
#else
# define compat_module_param_array(name, type, count, perm) \
    module_param_array(name, type, NULL, perm)
#endif
#endif

Re: [Xenomai-help] rtcan_mscan compilation.

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 3174 bytes --]

Wolfgang Grandegger wrote:
> Jan Kiszka wrote:
>> Wolfgang Grandegger wrote:
>>> ...
>>> Index: ksrc/drivers/can/rtcan_internal.h
>>> ===================================================================
>>> --- ksrc/drivers/can/rtcan_internal.h    (revision 1695)
>>> +++ ksrc/drivers/can/rtcan_internal.h    (working copy)
>>> @@ -111,6 +111,10 @@
>>>  
>>>  #ifndef compat_module_int_param_array
>>>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
>>> +# define compat_module_int_param(name) \
>>> +    MODULE_PARM(name, "i")
>>> +# define compat_module_charp_param(name) \
>>> +    MODULE_PARM(name, "s")
>>>  # define compat_module_byte_param_array(name, count) \
>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>>>  # define compat_module_short_param_array(name, count) \
>>> @@ -118,6 +122,10 @@
>>>  # define compat_module_int_param_array(name, count) \
>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>>>  #else
>>> +# define compat_module_int_param(name) \
>>> +    module_param(name, int, 0444)
>>> +# define compat_module_charp_param(name) \
>>> +    module_param(name, charp, 0444)
>>>  # define compat_module_byte_param_array(name, count) \
>>>      module_param_array(name, byte, NULL, 0444)
>>>  # define compat_module_short_param_array(name, count) \
>>
>> No need, module_param comes even with 2.4. We only need wrapping for
>> parameter arrays (due to different semantics of the macro arguments).
> 
> But with limitations, unfortunately:
> 
>   /* type is byte, short, ushort, int, uint, long, ulong, bool. (2.6
>      has more, but they are not supported).  perm is permissions when
>      it appears in sysfs: 0 means doens't appear, 0444 means read-only
>      by everyone, 0644 means changable dynamically by root, etc.  name
>      must be in scope (unlike MODULE_PARM).
>   */
>   #define module_param(name, type, perm) \
>         static inline void *__check_existence_##name(void) { return
>              &name; } \
>         MODULE_PARM(name, _MODULE_PARM_STRING_ ## type)
> 
>   #define _MODULE_PARM_STRING_byte "b"
>   #define _MODULE_PARM_STRING_short "h"
>   #define _MODULE_PARM_STRING_ushort "h"
>   #define _MODULE_PARM_STRING_int "i"
>   #define _MODULE_PARM_STRING_uint "i"
>   #define _MODULE_PARM_STRING_long "l"
>   #define _MODULE_PARM_STRING_ulong "l"
>   #define _MODULE_PARM_STRING_bool "i"
> 
> Especially "s" is missing. But it could be wrapped with:
> 
>   #define _MODULE_PARM_STRING_charp "s"
> 
> Any idea why this is not already done?

Likely because you mostly don't need it? I think mscan is an exception
here, at least from my experience.

When _MODULE_PARM_STRING_charp helps to enhance module_param on 2.4,
lets put it in Xenomai's wrapping header. I think we should also move
the parameter array over (directing it officially through the RTDM layer
can still be done later, at the latest when providing user-space
support). I would support a patch doing this as well.

> 
> I also want to replace the compat_module_*_param_array functions with
> the attached code snippet.

Ok, I'm awaiting a revised patch.

Jan


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

Re: [Xenomai-help] rtcan_mscan compilation.

[Wolfgang Grandegger]

[-- Attachment #1: Type: text/plain, Size: 3728 bytes --]

Ji Jan,

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
>> Jan Kiszka wrote:
>>> Wolfgang Grandegger wrote:
>>>> ...
>>>> Index: ksrc/drivers/can/rtcan_internal.h
>>>> ===================================================================
>>>> --- ksrc/drivers/can/rtcan_internal.h    (revision 1695)
>>>> +++ ksrc/drivers/can/rtcan_internal.h    (working copy)
>>>> @@ -111,6 +111,10 @@
>>>>  
>>>>  #ifndef compat_module_int_param_array
>>>>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
>>>> +# define compat_module_int_param(name) \
>>>> +    MODULE_PARM(name, "i")
>>>> +# define compat_module_charp_param(name) \
>>>> +    MODULE_PARM(name, "s")
>>>>  # define compat_module_byte_param_array(name, count) \
>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>>>>  # define compat_module_short_param_array(name, count) \
>>>> @@ -118,6 +122,10 @@
>>>>  # define compat_module_int_param_array(name, count) \
>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>>>>  #else
>>>> +# define compat_module_int_param(name) \
>>>> +    module_param(name, int, 0444)
>>>> +# define compat_module_charp_param(name) \
>>>> +    module_param(name, charp, 0444)
>>>>  # define compat_module_byte_param_array(name, count) \
>>>>      module_param_array(name, byte, NULL, 0444)
>>>>  # define compat_module_short_param_array(name, count) \
>>> No need, module_param comes even with 2.4. We only need wrapping for
>>> parameter arrays (due to different semantics of the macro arguments).
>> But with limitations, unfortunately:
>>
>>   /* type is byte, short, ushort, int, uint, long, ulong, bool. (2.6
>>      has more, but they are not supported).  perm is permissions when
>>      it appears in sysfs: 0 means doens't appear, 0444 means read-only
>>      by everyone, 0644 means changable dynamically by root, etc.  name
>>      must be in scope (unlike MODULE_PARM).
>>   */
>>   #define module_param(name, type, perm) \
>>         static inline void *__check_existence_##name(void) { return
>>              &name; } \
>>         MODULE_PARM(name, _MODULE_PARM_STRING_ ## type)
>>
>>   #define _MODULE_PARM_STRING_byte "b"
>>   #define _MODULE_PARM_STRING_short "h"
>>   #define _MODULE_PARM_STRING_ushort "h"
>>   #define _MODULE_PARM_STRING_int "i"
>>   #define _MODULE_PARM_STRING_uint "i"
>>   #define _MODULE_PARM_STRING_long "l"
>>   #define _MODULE_PARM_STRING_ulong "l"
>>   #define _MODULE_PARM_STRING_bool "i"
>>
>> Especially "s" is missing. But it could be wrapped with:
>>
>>   #define _MODULE_PARM_STRING_charp "s"
>>
>> Any idea why this is not already done?
> 
> Likely because you mostly don't need it? I think mscan is an exception
> here, at least from my experience.

Well, a "find . -name '*.c'|xargs grep charp|grep module" reveals plenty 
of references. And in 2.6 there is "module_param_string* as well, which 
will copy the string directly into the preallocated array.

> When _MODULE_PARM_STRING_charp helps to enhance module_param on 2.4,
> lets put it in Xenomai's wrapping header. I think we should also move
> the parameter array over (directing it officially through the RTDM layer
> can still be done later, at the latest when providing user-space
> support). I would support a patch doing this as well.

With the proposed modifications almost all module parameter use cases 
can be handled in a portable way, including strings and arrays of 
strings. I have moved it to the Xenomai's wrapping header file, 
including compat_module_param_array ...

>> I also want to replace the compat_module_*_param_array functions with
>> the attached code snippet.
> 
> Ok, I'm awaiting a revised patch.

... and updated the ISA and MEM driver as well in the attached revised 
patch.

Wolfgang.



[-- Attachment #2: xenomai-rtcan-mscan-2.6-2.patch --]
[-- Type: text/x-patch, Size: 10491 bytes --]

Index: include/asm-generic/wrappers.h
===================================================================
--- include/asm-generic/wrappers.h	(revision 1695)
+++ include/asm-generic/wrappers.h	(working copy)
@@ -46,6 +46,10 @@
 #endif
 
 #define module_param_named(name,var,type,mode)  module_param(var,type,mode)
+#define _MODULE_PARM_STRING_charp "s"
+#define compat_module_param_array(name, type, count, perm) \
+	static inline void *__check_existence_##name(void) { return &name; } \
+	MODULE_PARM(name, "1-" __MODULE_STRING(count) _MODULE_PARM_STRING_##type)
 
 #define container_of(ptr, type, member) ({			\
 	const typeof( ((type *)0)->member ) *__mptr = (ptr);	\
@@ -176,6 +180,9 @@
 
 #else /* LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,0) */
 
+#define compat_module_param_array(name, type, count, perm) \
+	module_param_array(name, type, NULL, perm)
+
 /* VM */
 
 #ifdef CONFIG_MMU
Index: ChangeLog
===================================================================
--- ChangeLog	(revision 1695)
+++ ChangeLog	(working copy)
@@ -1,3 +1,24 @@
+2006-10-08  Wolfgang Grandegger  <wg@domain.hid>
+
+	* include/asm-generic/wrappers.h: MODULE_PARM has been removed
+	in Linux 2.6.17. Module parameter compatibility functions 
+	reimplemented and moved to Xenomai's wrapping header. 
+	
+	* ksrc/drivers/can/mscan/rtcan_mscan.c, 
+	ksrc/drivers/can/sja1000/rtcan_isa.c,
+	ksrc/drivers/can/sja1000/rtcan_mem.c: Use new module parameter 
+	compatibility functions. 	
+	
+	* ksrc/drivers/can/mscan/rtcan_mscan.c (rtcan_mscan_init): Error 
+	handling corrected.
+
+	* ksrc/drivers/can/mscan/rtcan_mscan_regs.h: Fix invalid address
+	for MPC5200 GPIO registers. Now the MSCAN driver works for Linux
+	2.6.17.
+
+	* ksrc/drivers/can/rtcan_module.c (rtcan_dev_get_state_name): Bug
+	fix. Function did return invalid state name.
+
 2006-10-06  Jan Kiszka  <jan.kiszka@domain.hid>
 
 	* src/testsuite/cyclic/cyclictest.c: Align to latest v0.11,
Index: ksrc/drivers/can/rtcan_internal.h
===================================================================
--- ksrc/drivers/can/rtcan_internal.h	(revision 1695)
+++ ksrc/drivers/can/rtcan_internal.h	(working copy)
@@ -108,25 +108,6 @@
 
 #endif /* CONFIG_PROC_FS */
 
-
-#ifndef compat_module_int_param_array
-#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
-# define compat_module_byte_param_array(name, count) \
-    MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
-# define compat_module_short_param_array(name, count) \
-    MODULE_PARM(name, "1-" __MODULE_STRING(count) "h")
-# define compat_module_int_param_array(name, count) \
-    MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
-#else
-# define compat_module_byte_param_array(name, count) \
-    module_param_array(name, byte, NULL, 0444)
-# define compat_module_short_param_array(name, count) \
-    module_param_array(name, short, NULL, 0444)
-# define compat_module_int_param_array(name, count) \
-    module_param_array(name, int, NULL, 0444)
-#endif
-#endif
-
 #ifdef CONFIG_XENO_DRIVERS_RTCAN_DEBUG
 # define RTCAN_DBG(fmt,args...) do { printk(fmt ,##args); } while (0)
 # define RTCAN_RTDM_DBG(fmt,args...) do { rtdm_printk(fmt ,##args); } while (0)
Index: ksrc/drivers/can/rtcan_module.c
===================================================================
--- ksrc/drivers/can/rtcan_module.c	(revision 1695)
+++ ksrc/drivers/can/rtcan_module.c	(working copy)
@@ -54,15 +54,15 @@
 }
 
 static char *rtcan_state_names[] = {
-    "active", "stopped", "sleeping", "warning", 
-    "passive" , "bus-off", "scanning"
+    "active", "warning", "passive" , "bus-off", 
+    "scanning", "stopped", "sleeping"
 };
 
 static void rtcan_dev_get_state_name(can_state_t state, 
 				     char* name, int max_len)
 {
     if (state >= CAN_STATE_ACTIVE && 
-	state <= CAN_STATE_SCANNING_BAUDRATE)
+	state <= CAN_STATE_SLEEPING)
 	strncpy(name, rtcan_state_names[state], max_len);
     else
 	strncpy(name, "unknown", max_len);
Index: ksrc/drivers/can/mscan/rtcan_mscan.c
===================================================================
--- ksrc/drivers/can/mscan/rtcan_mscan.c	(revision 1695)
+++ ksrc/drivers/can/mscan/rtcan_mscan.c	(working copy)
@@ -37,6 +37,7 @@
 #include <rtdm/rtcan.h>
 #include "rtcan_dev.h"
 #include "rtcan_raw.h"
+#include "rtcan_internal.h"
 #include "rtcan_mscan_regs.h"
 
 extern int rtcan_mscan_create_proc(struct rtcan_device* dev);
@@ -56,7 +57,7 @@
 
 /** Module parameter for the CAN controllers' */
 
-static int port[RTCAN_MSCAN_DEVS] = {
+int port[RTCAN_MSCAN_DEVS] = {
 #ifdef CONFIG_XENO_DRIVERS_RTCAN_MSCAN_1
 #ifdef CONFIG_XENO_DRIVERS_RTCAN_MSCAN_2
 	1, 2  /* Enable CAN 1 and 2 */
@@ -71,24 +72,21 @@
 #endif
 #endif
 };
-
-static int port_count = RTCAN_MSCAN_DEVS;
-
-#if LINUX_VERSION_CODE >= KERNEL_VERSION(2,6,10)
-module_param_array(port, int, &port_count, 0400);
-#else
-MODULE_PARM(port, "2" __MODULE_STRING(RTCAN_MSCAN_DEVS) "i");
-#endif
+compat_module_param_array(port, int, RTCAN_MSCAN_DEVS, 0444);
 MODULE_PARM_DESC(port, "Port numbers of enabled controllers, e.g. 1,2");
 
-unsigned long mscan_clock = CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK;
-MODULE_PARM(mscan_clock,"i");
+/* 
+ * Note: on the MPC5200 the MSCAN clock source is the IP bus 
+ * clock (IP_CLK) while on the MPC5200B it is the oscillator 
+ * clock (SYS_XTAL_IN).
+ */
+unsigned int mscan_clock = CONFIG_XENO_DRIVERS_RTCAN_MSCAN_CLOCK;
+module_param(mscan_clock, int, 0444);
 MODULE_PARM_DESC(mscan_clock, "Clock frequency in Hz");
 
 char *mscan_pins = NULL;
-MODULE_PARM(mscan_pins,"s");
+module_param(mscan_pins, charp, 0444);
 MODULE_PARM_DESC(mscan_pins, "Routing to GPIO pins (PSC2 or I2C1/TMR01)");
-
  
 struct rtcan_device *rtcan_mscan_devs[RTCAN_MSCAN_DEVS];
 
@@ -739,7 +737,6 @@
     }
 }
 
-/** Init module */
 int __init rtcan_mscan_init_one(int idx)
 {
     int ret, irq;
@@ -836,32 +833,13 @@
 
 }
 
-/** Init module */
-static int __init rtcan_mscan_init(void)
-{
-    int i, ret;
-
-    mscan_gpio_config();
-
-    for (i = 0; i < RTCAN_MSCAN_DEVS; i++) {
-	if (port[i] < 1 || port[i] > RTCAN_MSCAN_DEVS)
-	    continue;
-
-	if ((ret = rtcan_mscan_init_one(i) != 0))
-	    return ret;
-    }
-
-    return 0;
-}
-
-
-/** Cleanup module */
 static void __exit rtcan_mscan_exit(void)
 {
     int i;
     struct rtcan_device *dev;
 
-    for (i = 0; i < port_count; i++) {
+    for (i = 0; i < RTCAN_MSCAN_DEVS; i++) {
+
 	if ((dev = rtcan_mscan_devs[i]) == NULL)
 	    continue;
 	
@@ -876,5 +854,24 @@
 
 }
 
+static int __init rtcan_mscan_init(void)
+{
+    int i, err;
+
+    mscan_gpio_config();
+
+    for (i = 0; i < RTCAN_MSCAN_DEVS; i++) {
+	if (port[i] < 1 || port[i] > RTCAN_MSCAN_DEVS)
+	    continue;
+
+	if ((err = rtcan_mscan_init_one(i) != 0)) {
+	    rtcan_mscan_exit();
+	    return err;
+	}
+    }
+
+    return 0;
+}
+
 module_init(rtcan_mscan_init);
 module_exit(rtcan_mscan_exit);
Index: ksrc/drivers/can/mscan/rtcan_mscan_regs.h
===================================================================
--- ksrc/drivers/can/mscan/rtcan_mscan_regs.h	(revision 1695)
+++ ksrc/drivers/can/mscan/rtcan_mscan_regs.h	(working copy)
@@ -35,7 +35,7 @@
 #define MSCAN_MBAR	MPC52xx_MBAR
 #define MSCAN_CAN1_IRQ	MPC52xx_MSCAN1_IRQ
 #define MSCAN_CAN2_IRQ	MPC52xx_MSCAN2_IRQ
-#define MPC5xxx_GPIO	MPC52xx_GPIO_OFFSET 
+#define MPC5xxx_GPIO	MPC52xx_VA(MPC52xx_GPIO_OFFSET) 
 #define mpc5xxx_gpio	mpc52xx_gpio
 #endif
 
Index: ksrc/drivers/can/sja1000/rtcan_peak_dng.c
===================================================================
--- ksrc/drivers/can/sja1000/rtcan_peak_dng.c	(revision 1695)
+++ ksrc/drivers/can/sja1000/rtcan_peak_dng.c	(working copy)
@@ -50,15 +50,9 @@
 static ushort io[RTCAN_PEAK_DNG_MAX_DEV];
 static char   irq[RTCAN_PEAK_DNG_MAX_DEV];
 
-#if LINUX_VERSION_CODE > KERNEL_VERSION(2,6,10)
-module_param_array(type, charp,  NULL, 0444);
-module_param_array(io,   ushort, NULL, 0444);
-module_param_array(irq,  byte,   NULL, 0444);
-#else
-MODULE_PARM(type, "0-" __MODULE_STRING(RTCAN_PEAK_DNG_MAX_DEV) "s");
-MODULE_PARM(io,   "0-" __MODULE_STRING(RTCAN_PEAK_DNG_MAX_DEV) "h");
-MODULE_PARM(irq,  "0-" __MODULE_STRING(RTCAN_PEAK_DNG_MAX_DEV) "b");
-#endif
+compat_module_param_array(type, charp,  RTCAN_PEAK_DNG_MAX_DEV, 0444);
+compat_module_param_array(io,   ushort, RTCAN_PEAK_DNG_MAX_DEV, 0444);
+compat_module_param_array(irq,  byte,   RTCAN_PEAK_DNG_MAX_DEV, 0444);
 
 MODULE_PARM_DESC(type, "The type of interface (sp, epp)");
 MODULE_PARM_DESC(io,   "The io-port address");
Index: ksrc/drivers/can/sja1000/rtcan_isa.c
===================================================================
--- ksrc/drivers/can/sja1000/rtcan_isa.c	(revision 1695)
+++ ksrc/drivers/can/sja1000/rtcan_isa.c	(working copy)
@@ -51,11 +51,11 @@
 static u8 ocr[RTCAN_ISA_MAX_DEV];
 static u8 cdr[RTCAN_ISA_MAX_DEV];
 
-compat_module_short_param_array(io, RTCAN_ISA_MAX_DEV);
-compat_module_int_param_array(irq, RTCAN_ISA_MAX_DEV);
-compat_module_int_param_array(clock, RTCAN_ISA_MAX_DEV);
-compat_module_byte_param_array(ocr, RTCAN_ISA_MAX_DEV);
-compat_module_byte_param_array(cdr, RTCAN_ISA_MAX_DEV);
+compat_module_param_array(io, ushort, RTCAN_ISA_MAX_DEV, 0444);
+compat_module_param_array(irq, int, RTCAN_ISA_MAX_DEV, 0444);
+compat_module_param_array(clock, uint, RTCAN_ISA_MAX_DEV, 0444);
+compat_module_param_array(ocr, byte, RTCAN_ISA_MAX_DEV, 0444);
+compat_module_param_array(cdr, byte, RTCAN_ISA_MAX_DEV, 0444);
 
 MODULE_PARM_DESC(io, "The io-port address");
 MODULE_PARM_DESC(irq, "The interrupt number");
Index: ksrc/drivers/can/sja1000/rtcan_mem.c
===================================================================
--- ksrc/drivers/can/sja1000/rtcan_mem.c	(revision 1695)
+++ ksrc/drivers/can/sja1000/rtcan_mem.c	(working copy)
@@ -59,11 +59,11 @@
 static u8 ocr[RTCAN_MEM_MAX_DEV];
 static u8 cdr[RTCAN_MEM_MAX_DEV];
 
-compat_module_int_param_array(mem, RTCAN_MEM_MAX_DEV);
-compat_module_int_param_array(irq, RTCAN_MEM_MAX_DEV);
-compat_module_int_param_array(clock, RTCAN_MEM_MAX_DEV);
-compat_module_byte_param_array(ocr, RTCAN_MEM_MAX_DEV);
-compat_module_byte_param_array(cdr, RTCAN_MEM_MAX_DEV);
+compat_module_param_array(mem, uint, RTCAN_MEM_MAX_DEV, 0444);
+compat_module_param_array(irq, int, RTCAN_MEM_MAX_DEV, 0444);
+compat_module_param_array(clock, uint, RTCAN_MEM_MAX_DEV, 0444);
+compat_module_param_array(ocr, byte, RTCAN_MEM_MAX_DEV, 0444);
+compat_module_param_array(cdr, byte, RTCAN_MEM_MAX_DEV, 0444);
 
 MODULE_PARM_DESC(mem, "The io-memory address");
 MODULE_PARM_DESC(irq, "The interrupt number");

Re: [Xenomai-help] rtcan_mscan compilation.

[Jan Kiszka]

[-- Attachment #1: Type: text/plain, Size: 4142 bytes --]

Wolfgang Grandegger wrote:
> Ji Jan,
> 
> Jan Kiszka wrote:
>> Wolfgang Grandegger wrote:
>>> Jan Kiszka wrote:
>>>> Wolfgang Grandegger wrote:
>>>>> ...
>>>>> Index: ksrc/drivers/can/rtcan_internal.h
>>>>> ===================================================================
>>>>> --- ksrc/drivers/can/rtcan_internal.h    (revision 1695)
>>>>> +++ ksrc/drivers/can/rtcan_internal.h    (working copy)
>>>>> @@ -111,6 +111,10 @@
>>>>>  
>>>>>  #ifndef compat_module_int_param_array
>>>>>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
>>>>> +# define compat_module_int_param(name) \
>>>>> +    MODULE_PARM(name, "i")
>>>>> +# define compat_module_charp_param(name) \
>>>>> +    MODULE_PARM(name, "s")
>>>>>  # define compat_module_byte_param_array(name, count) \
>>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>>>>>  # define compat_module_short_param_array(name, count) \
>>>>> @@ -118,6 +122,10 @@
>>>>>  # define compat_module_int_param_array(name, count) \
>>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>>>>>  #else
>>>>> +# define compat_module_int_param(name) \
>>>>> +    module_param(name, int, 0444)
>>>>> +# define compat_module_charp_param(name) \
>>>>> +    module_param(name, charp, 0444)
>>>>>  # define compat_module_byte_param_array(name, count) \
>>>>>      module_param_array(name, byte, NULL, 0444)
>>>>>  # define compat_module_short_param_array(name, count) \
>>>> No need, module_param comes even with 2.4. We only need wrapping for
>>>> parameter arrays (due to different semantics of the macro arguments).
>>> But with limitations, unfortunately:
>>>
>>>   /* type is byte, short, ushort, int, uint, long, ulong, bool. (2.6
>>>      has more, but they are not supported).  perm is permissions when
>>>      it appears in sysfs: 0 means doens't appear, 0444 means read-only
>>>      by everyone, 0644 means changable dynamically by root, etc.  name
>>>      must be in scope (unlike MODULE_PARM).
>>>   */
>>>   #define module_param(name, type, perm) \
>>>         static inline void *__check_existence_##name(void) { return
>>>              &name; } \
>>>         MODULE_PARM(name, _MODULE_PARM_STRING_ ## type)
>>>
>>>   #define _MODULE_PARM_STRING_byte "b"
>>>   #define _MODULE_PARM_STRING_short "h"
>>>   #define _MODULE_PARM_STRING_ushort "h"
>>>   #define _MODULE_PARM_STRING_int "i"
>>>   #define _MODULE_PARM_STRING_uint "i"
>>>   #define _MODULE_PARM_STRING_long "l"
>>>   #define _MODULE_PARM_STRING_ulong "l"
>>>   #define _MODULE_PARM_STRING_bool "i"
>>>
>>> Especially "s" is missing. But it could be wrapped with:
>>>
>>>   #define _MODULE_PARM_STRING_charp "s"
>>>
>>> Any idea why this is not already done?
>>
>> Likely because you mostly don't need it? I think mscan is an exception
>> here, at least from my experience.
> 
> Well, a "find . -name '*.c'|xargs grep charp|grep module" reveals plenty
> of references. And in 2.6 there is "module_param_string* as well, which
> will copy the string directly into the preallocated array.
> 
>> When _MODULE_PARM_STRING_charp helps to enhance module_param on 2.4,
>> lets put it in Xenomai's wrapping header. I think we should also move
>> the parameter array over (directing it officially through the RTDM layer
>> can still be done later, at the latest when providing user-space
>> support). I would support a patch doing this as well.
> 
> With the proposed modifications almost all module parameter use cases
> can be handled in a portable way, including strings and arrays of
> strings. I have moved it to the Xenomai's wrapping header file,
> including compat_module_param_array ...
> 
>>> I also want to replace the compat_module_*_param_array functions with
>>> the attached code snippet.
>>
>> Ok, I'm awaiting a revised patch.
> 
> ... and updated the ISA and MEM driver as well in the attached revised
> patch.

Thanks, applied.

Jan


PS: I fixed a few (not all) trailing whitespaces. Maybe you can
configure your editor to highlight them. Once we have Lindent'ed the CAN
stack, this will become relevant anyway.


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

Re: [Xenomai-help] rtcan_mscan compilation.

[Wolfgang Grandegger]

Jan Kiszka wrote:
> Wolfgang Grandegger wrote:
>> Ji Jan,
>>
>> Jan Kiszka wrote:
>>> Wolfgang Grandegger wrote:
>>>> Jan Kiszka wrote:
>>>>> Wolfgang Grandegger wrote:
>>>>>> ...
>>>>>> Index: ksrc/drivers/can/rtcan_internal.h
>>>>>> ===================================================================
>>>>>> --- ksrc/drivers/can/rtcan_internal.h    (revision 1695)
>>>>>> +++ ksrc/drivers/can/rtcan_internal.h    (working copy)
>>>>>> @@ -111,6 +111,10 @@
>>>>>>  
>>>>>>  #ifndef compat_module_int_param_array
>>>>>>  #if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,10)
>>>>>> +# define compat_module_int_param(name) \
>>>>>> +    MODULE_PARM(name, "i")
>>>>>> +# define compat_module_charp_param(name) \
>>>>>> +    MODULE_PARM(name, "s")
>>>>>>  # define compat_module_byte_param_array(name, count) \
>>>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "b")
>>>>>>  # define compat_module_short_param_array(name, count) \
>>>>>> @@ -118,6 +122,10 @@
>>>>>>  # define compat_module_int_param_array(name, count) \
>>>>>>      MODULE_PARM(name, "1-" __MODULE_STRING(count) "i")
>>>>>>  #else
>>>>>> +# define compat_module_int_param(name) \
>>>>>> +    module_param(name, int, 0444)
>>>>>> +# define compat_module_charp_param(name) \
>>>>>> +    module_param(name, charp, 0444)
>>>>>>  # define compat_module_byte_param_array(name, count) \
>>>>>>      module_param_array(name, byte, NULL, 0444)
>>>>>>  # define compat_module_short_param_array(name, count) \
>>>>> No need, module_param comes even with 2.4. We only need wrapping for
>>>>> parameter arrays (due to different semantics of the macro arguments).
>>>> But with limitations, unfortunately:
>>>>
>>>>   /* type is byte, short, ushort, int, uint, long, ulong, bool. (2.6
>>>>      has more, but they are not supported).  perm is permissions when
>>>>      it appears in sysfs: 0 means doens't appear, 0444 means read-only
>>>>      by everyone, 0644 means changable dynamically by root, etc.  name
>>>>      must be in scope (unlike MODULE_PARM).
>>>>   */
>>>>   #define module_param(name, type, perm) \
>>>>         static inline void *__check_existence_##name(void) { return
>>>>              &name; } \
>>>>         MODULE_PARM(name, _MODULE_PARM_STRING_ ## type)
>>>>
>>>>   #define _MODULE_PARM_STRING_byte "b"
>>>>   #define _MODULE_PARM_STRING_short "h"
>>>>   #define _MODULE_PARM_STRING_ushort "h"
>>>>   #define _MODULE_PARM_STRING_int "i"
>>>>   #define _MODULE_PARM_STRING_uint "i"
>>>>   #define _MODULE_PARM_STRING_long "l"
>>>>   #define _MODULE_PARM_STRING_ulong "l"
>>>>   #define _MODULE_PARM_STRING_bool "i"
>>>>
>>>> Especially "s" is missing. But it could be wrapped with:
>>>>
>>>>   #define _MODULE_PARM_STRING_charp "s"
>>>>
>>>> Any idea why this is not already done?
>>> Likely because you mostly don't need it? I think mscan is an exception
>>> here, at least from my experience.
>> Well, a "find . -name '*.c'|xargs grep charp|grep module" reveals plenty
>> of references. And in 2.6 there is "module_param_string* as well, which
>> will copy the string directly into the preallocated array.
>>
>>> When _MODULE_PARM_STRING_charp helps to enhance module_param on 2.4,
>>> lets put it in Xenomai's wrapping header. I think we should also move
>>> the parameter array over (directing it officially through the RTDM layer
>>> can still be done later, at the latest when providing user-space
>>> support). I would support a patch doing this as well.
>> With the proposed modifications almost all module parameter use cases
>> can be handled in a portable way, including strings and arrays of
>> strings. I have moved it to the Xenomai's wrapping header file,
>> including compat_module_param_array ...
>>
>>>> I also want to replace the compat_module_*_param_array functions with
>>>> the attached code snippet.
>>> Ok, I'm awaiting a revised patch.
>> ... and updated the ISA and MEM driver as well in the attached revised
>> patch.
> 
> Thanks, applied.
> 
> Jan
> 
> 
> PS: I fixed a few (not all) trailing whitespaces. Maybe you can
> configure your editor to highlight them. Once we have Lindent'ed the CAN
> stack, this will become relevant anyway.

Ah, oh, sorry, thanks. My editor now shows trailing white space in _red_.

Wolfgang.