Subject:
|
Re: LegOS 0.2.5 and LNP
|
Newsgroups:
|
lugnet.robotics.rcx.legos
|
Date:
|
Sat, 20 Jan 2001 17:14:41 GMT
|
Viewed:
|
2000 times
|
| |
|
|
Bernardo Dal Seno <piu1759@cdc8g5.cdc.polimi.it> writes:
> 1. Now lnp_logical_write() is no more thread-safe. Furthermore calling
> lnp_logical_write() screw up the semaphore (tx_sem) and so the other two
> lnp_XXX_write() become thread-unsafe.
>
> 2. On the host computer LNP is still thread unsafe.
I cleaned up my code (found one bug!), the patch follows. As you may see
now there are two semaphore, because two are the resources to give access
to: the LNP buffer and the IR port. I modified the calling of the handlers
to avoid race condition in a multithead envinronment on a PC. My only
worry is the symbol the semaphore-related code in lnp.c should depend on.
I used the conditional
#if defined CONF_TM || defined CONF_HOST
Perhaps that breaks something in the utilities and we should introduce a
new symbol (CONF_HOST_TM?), I don't know for sure. I leave this point to
more savvy people.
bye
Bernardo
Patch:
diff -Naur legOS-0.2.5/include/lnp/lnp.h legOS-0.2.5+/include/lnp/lnp.h
--- legOS-0.2.5/include/lnp/lnp.h Mon Aug 14 11:03:42 2000
+++ legOS-0.2.5+/include/lnp/lnp.h Wed Jan 17 23:00:16 2001
@@ -64,10 +64,10 @@
//! there are no ports for integrity layer packets, so there's just
// one handler.
-extern lnp_integrity_handler_t lnp_integrity_handler;
+extern volatile lnp_integrity_handler_t lnp_integrity_handler;
//! addressing layer packets may be directed to a variety of ports.
-extern lnp_addressing_handler_t lnp_addressing_handler[];
+extern volatile lnp_addressing_handler_t lnp_addressing_handler[];
///////////////////////////////////////////////////////////////////////
//
diff -Naur legOS-0.2.5/kernel/lnp-logical.c legOS-0.2.5+/kernel/lnp-logical.c
--- legOS-0.2.5/kernel/lnp-logical.c Mon Aug 14 22:43:26 2000
+++ legOS-0.2.5+/kernel/lnp-logical.c Sat Jan 20 17:57:08 2001
@@ -66,7 +66,7 @@
static time_t allow_tx; //!< time to allow new transmission
#ifdef CONF_TM
-sem_t tx_sem; //!< transmitter access semaphore
+static sem_t tx_sem; //!< transmitter access semaphore
#endif
///////////////////////////////////////////////////////////////////////////////
@@ -259,10 +259,10 @@
*/
int lnp_logical_write(const void* buf,size_t len) {
unsigned char tmp;
-
-//#ifdef CONF_TM
-// sem_wait(&tx_sem);
-//#endif
+
+#ifdef CONF_TM
+ sem_wait(&tx_sem);
+#endif
#ifdef CONF_AUTOSHUTOFF
shutoff_restart();
diff -Naur legOS-0.2.5/kernel/lnp.c legOS-0.2.5+/kernel/lnp.c
--- legOS-0.2.5/kernel/lnp.c Mon Aug 14 22:43:26 2000
+++ legOS-0.2.5+/kernel/lnp.c Fri Jan 19 19:18:02 2001
@@ -54,16 +54,15 @@
//
///////////////////////////////////////////////////////////////////////////////
-#ifndef CONF_HOST
-#ifdef CONF_TM
+
+#if defined CONF_TM || defined CONF_HOST
#include <semaphore.h>
-//!< transmitter access semaphore
-extern sem_t tx_sem;
+//!< tansmit buffer access semaphore
+static sem_t buf_sem;
#endif
-#endif
//! the timeout counter in ms
volatile unsigned short lnp_timeout_counter;
@@ -77,13 +76,13 @@
//! there are no ports for integrity layer packets, so just one handler.
/*! FIXME: uninitialized
*/
-lnp_integrity_handler_t lnp_integrity_handler;
+volatile lnp_integrity_handler_t lnp_integrity_handler;
//! addressing layer packets may be directed to a variety of ports.
/*! FIXME: uninitialized
FIXME: inefficient if LNP_PORT_MASK doesn't adhere to 0...01...1 scheme.
*/
-lnp_addressing_handler_t lnp_addressing_handler[LNP_PORTMASK+1];
+volatile lnp_addressing_handler_t lnp_addressing_handler[LNP_PORTMASK+1];
//! the LNP transmit buffer
static unsigned char lnp_buffer[256+3];
@@ -141,10 +140,10 @@
*/
int lnp_integrity_write(const unsigned char *data,unsigned char length) {
-#ifndef CONF_HOST
-#ifdef CONF_TM
- sem_wait(&tx_sem);
-#endif
+ int r;
+
+#if defined CONF_TM || defined CONF_HOST
+ sem_wait(&buf_sem);
#endif
lnp_buffer[0]=0xf0;
@@ -152,7 +151,12 @@
memcpy(lnp_buffer+2,data,length);
lnp_buffer[length+2]=(unsigned char) lnp_checksum(lnp_buffer,length+2);
- return lnp_logical_write(lnp_buffer,length+3);
+ r = lnp_logical_write(lnp_buffer,length+3);
+
+#if defined CONF_TM || defined CONF_HOST
+ sem_post(&buf_sem);
+#endif
+ return r;
}
//! send a LNP addressing layer packet of given length
@@ -161,10 +165,10 @@
int lnp_addressing_write(const unsigned char *data,unsigned char length,
unsigned char dest,unsigned char srcport) {
-#ifndef CONF_HOST
-#ifdef CONF_TM
- sem_wait(&tx_sem);
-#endif
+ int r;
+
+#if defined CONF_TM || defined CONF_HOST
+ sem_wait(&buf_sem);
#endif
lnp_buffer[0]=0xf1;
@@ -174,23 +178,31 @@
memcpy(lnp_buffer+4,data,length);
lnp_buffer[length+4]=(unsigned char) lnp_checksum(lnp_buffer,length+4);
- return lnp_logical_write(lnp_buffer,length+5);
+ r = lnp_logical_write(lnp_buffer,length+5);
+
+#if defined CONF_TM || defined CONF_HOST
+ sem_post(&buf_sem);
+#endif
+
+ return r;
}
//! handle LNP packet from the integrity layer
void lnp_receive_packet(const unsigned char *data) {
unsigned char header=*(data++);
unsigned char length=*(data++);
+ lnp_integrity_handler_t intgh;
+ lnp_addressing_handler_t addrh;
// only handle non-degenerate packets in boot protocol 0xf0
//
switch(header) {
case 0xf0: // raw integrity layer packet, no addressing.
- if(lnp_integrity_handler) {
+ if((intgh = lnp_integrity_handler)) {
#ifdef CONF_AUTOSHUTOFF
shutoff_restart();
#endif
- lnp_integrity_handler(data,length);
+ intgh(data,length);
}
break;
@@ -201,12 +213,12 @@
if(LNP_HOSTADDR == (dest & LNP_HOSTMASK)) {
unsigned char port=dest & LNP_PORTMASK;
- if(lnp_addressing_handler[port]) {
+ if((addrh = lnp_addressing_handler[port])) {
unsigned char src=*(data++);
#ifdef CONF_AUTOSHUTOFF
shutoff_restart();
#endif
- lnp_addressing_handler[port](data,length-2,src);
+ addrh(data,length-2,src);
}
}
}
@@ -320,6 +332,10 @@
for(k=1;k<=LNP_PORTMASK;k++)
lnp_addressing_handler[k]=LNP_DUMMY_ADDRESSING;
lnp_integrity_handler=LNP_DUMMY_INTEGRITY;
+
+#if defined CONF_TM || defined CONF_HOST
+ sem_init(&buf_sem,0,1);
+#endif
}
#endif // CONF_LNP
|
|
Message has 1 Reply:
 | | Re: LegOS 0.2.5 and LNP
|
| (...) you may see (...) give access (...) the handlers (...) My only (...) depend on. (...) introduce a (...) this point to (...) I think that we can use safely (CONF_TM || defined CONF_HOST) condition. I'm checking the whole patch and I think that (...) (24 years ago, 21-Jan-01, to lugnet.robotics.rcx.legos)
|
Message is in Reply To:
| | LegOS 0.2.5 and LNP
|
| Yesterday I downloaded legOS 0.2.5 and peeked at the code. Btw. README is not up to date. For last two month I was working with a friend of mine on a transport protocol implemented on top of LNP for the transmission of packets of unlimited size and (...) (24 years ago, 17-Jan-01, to lugnet.robotics.rcx.legos)
|
9 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
