This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Problems with eCos 2.0 tests for viper


On Wed, 2002-05-29 at 20:49, Paul Randall wrote:
> Hi all,
> 
> Yesterday, I did a CVS download of the eCos Version 2.0 alpha source and
> I was able to build the library without errors (although a lot of
> warnings were generated - I'm not sure if this is normal).  However, I
> got some errors while building the tests.
> 
> The first error occurred while building the eCos kernel tests.  I got
> the message "No rule to make target 'tests/thread2tests/kclock0.d'".  I
> traced this to the makefile in the build/kernel/current directory where
> I deleted the /thread2tests part of TESTS := line.  When I tried to
> build the tests again, this test built ok.  However, when I reloaded the
> configuration, the makefile was regenerated and /thread2tests
> reappeared.  I'm not sure where this /thread2tests is coming from -
> there was a /thread2tests on line 9201 of the *.ecc file but it was in a
> section of comments.  I deleted it, but the /thread2tests was still
> generated in the makefile.
> 

Make sure you get the most up-to-date version of "ecosconfig" since the
CDL has changed in ways that older tools can't handle.

> The next error I got was in the tcp_echo.c test where I got the message
> "lib/libkern/libkern.h : no such file or directory".  I took this test
> out of the net.cdl file so as to skip this test.  I got a similar error
> while building the bridge.c test where I got the message
> "net/if_bridge.h : no such file or directory".  I also removed the
> bridge test from the net.cdl file.

I assume that you used the "new_net" (the FreeBSD TCP/IP stack).  The
bridge test should not be built as bridging isn't currently supported
with that stack.

As for the other error, this patch will fix it:
Index: net/common/current/tests/bridge.c
===================================================================
RCS file: /home/cvs/ecc/ecc/net/common/current/tests/bridge.c,v
retrieving revision 1.2
diff -u -5 -p -r1.2 bridge.c
--- net/common/current/tests/bridge.c	18 May 2002 18:38:32 -0000	1.2
+++ net/common/current/tests/bridge.c	30 May 2002 08:49:53 -0000
@@ -65,10 +65,11 @@
 
 #include <stdio.h>
 #include <network.h>
 #include <sys/types.h>
 #include <stdlib.h>
+#ifdef CYGPKG_NET_BRIDGE
 #include <net/if.h>
 #include <net/if_dl.h>
 #include <netinet/in.h>
 #include <netinet/if_ether.h>
 #include <net/if_bridge.h>
@@ -981,6 +982,13 @@ cyg_user_start(void) {
                     &main_handle, &main_thread);
   cyg_thread_resume(main_handle);
   printf("cyg_user_start done\n");
 }
 
+#else // CYGPKG_NET_BRIDGE
+
+void
+cyg_user_start(void) {
+    printf("No bridge support available\n");
+}
+#endif
 
Index: net/common/current/tests/tcp_echo.c
===================================================================
RCS file: /home/cvs/ecc/ecc/net/common/current/tests/tcp_echo.c,v
retrieving revision 1.2
diff -u -5 -p -r1.2 tcp_echo.c
--- net/common/current/tests/tcp_echo.c	18 May 2002 18:38:32 -0000	1.2
+++ net/common/current/tests/tcp_echo.c	30 May 2002 08:37:31 -0000
@@ -63,13 +63,17 @@
 #endif
 

 // Network throughput test code
 
-#include <lib/libkern/libkern.h>
-
 #include <network.h>
+
+static __inline__ int
+max(int m, int n)
+{
+    return m > n ? m : n;
+}
 
 #define SOURCE_PORT 9990
 #define SINK_PORT   9991
 
 #define MAX_BUF 8192
> 
> I tried to run the ftp_test (which built successfully) using Insight,
> but once the test was downloaded to the viper, nothing happened.  There
> was no diagnostic output on Insight's console, and execution did not
> stop at any breakpoints I set.  I also tried the ping_test which did
> nothing as well.  These tests worked with the pre-2.0 version of eCos
> (although ftp_test took a long time to complete - about 90 seconds,
> which is why I'm trying to get Version 2.0 working).  I did notice a
> difference in the way ftp_test was downloaded to the viper.  When I was
> running with the pre-2.0 version of eCos, the Start address was 0x20000
> and the load size was 497836.  With eCos 2.0, the load address is now
> 0x40000 and the load size is 289500.  Could this be preventing the
> program from executing?
> 

I'll give this a shot and see what's up.

> I am running eCos on a Windows 98 system and debugging over the viper's
> serial port.
> 
> Paul Randall
> Delta Information Systems
> 
> 
> -- 
> Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
> and search the list archive: http://sources.redhat.com/ml/ecos-discuss



-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]