Re: [LTP] mtest06/mmap1.c test really broken !?

[Subrata M. <sub...@in...>]

This is also through.

Regards & Thanks--
Subrata Modak,



                                                                           
             Carmelo AMOROSO                                               
             <carmelo.amoroso@                                             
             st.com>                                                    To 
             Sent by:                  Carmelo AMOROSO                     
             ltp-list-bounces@         <car...@st...>            
             lists.sourceforge                                          cc 
             .net                      ltp...@li...,     
                                       Filippo ARCIDIACONO                 
                                       <fil...@st...>        
             05/14/07 05:25 PM                                     Subject 
                                       Re: [LTP] mtest06/mmap1.c test      
                                       really broken !?                    
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




Carmelo AMOROSO wrote:
> Carmelo AMOROSO wrote:
>> Subrata Modak wrote:
>>> Hi Carmelo,
>>>
>>> Applied your Patch Manually as it created some problems with the
>>> APRIL version. The attached is the diff-ed patch that i generated
>>> against APRIL release. Please verify it. I also found some page
>>> faults in my machines when i ran the modified mmap1. Please see the
>>> attached tar containing outputs and logs of the test results done on
>>> "2.6.9-42.0.3.EL" and "2.6.18-1.2961.el5" kernels. But logs say that
>>> tests passed.
>>>
>>> Regards--
>>> Subrata
>> Hi Subrata,
>> for what I can see, your modified patch is rejecting all my changes,
>> so it seems that you run the test using the original code.
>> About the original log, the test says PASS just because it perfomors
>> the check againt the thread exit value in the wrong way.
>> To have evidence of what it is happening you need to run the test
>> with the -v (verbose) option ON, but without our patch
>> it doesnt' work.
>> So I split the patch into three parts and attached the log for each run.
>> I executed the test with the following command line: ./mmap1 -l 10 -v
>> , where -l 10 it used to set the iterations number to 10
>> (just to reduce the execution time).
>>
> ..................... snippet out ............................
>
> A last minute bug: the usage of OPT_MISSING is wrong. It should be
> called into the if(..............) statements and not into the
> else.
> A forthcoming patch will be posted.
>
> Regards,
> Carmelo
>
>
Hi All, please find attached a proposal to fix the wrong usage of
OPT_MISSING macro.
Plase note that the patch has been created against 20070228 code base,
with all previous posted patches already applied.

Regards,
Carmelo



--- mmap1.c.orig         2007-05-14 13:23:27.195996000 +0200
+++ mmap1.c        2007-05-14 13:47:41.702766000 +0200
@@ -116,6 +116,7 @@
 #define OPT_MISSING(prog, opt) do {\
                         fprintf(stderr, "%s: option -%c ", prog, opt); \
                         fprintf(stderr, "requires and argument\n");\
+
usage(prog); \
                             } while(0)


@@ -528,7 +529,7 @@
     num_iter = 1000;
     exec_time = 24;

-    while ((c =  getopt(argc, argv, "h:l:s:vx:")) != -1)
+    while ((c =  getopt(argc, argv, "hvl:s:x:")) != -1)
     {
         switch(c)
             {
@@ -537,25 +538,29 @@
                         break;
                 case 'l':
                         if ((num_iter = atoi(optarg)) == (int)NULL)
-                            num_iter = 1000;
-                else
-                            OPT_MISSING(argv[0], optopt);
+                                    OPT_MISSING(argv[0], optopt);
+        else
+                    if (num_iter < 0)
+                                                fprintf(stdout,"WARNING:
bad argument. Using default %d\n", (num_iter = 1000));
                         break;
                 case 's':
                         if ((file_size = atoi(optarg)) == 0)
-                            file_size = 1024;
-                else
-                            OPT_MISSING(argv[0], optopt);
+                                        OPT_MISSING(argv[0], optopt);
+        else
+                    if (file_size < 0)
+                                                fprintf(stdout,"WARNING:
bad argument. Using default %d\n", (file_size = 1024));
                         break;
                 case 'v':
                         verbose_print = TRUE;
                         break;
                 case 'x':
                         if ((exec_time = atoi(optarg)) == (int)NULL)
-                            exec_time = 24;
-                else
-                            OPT_MISSING(argv[0], optopt);
-                    run_once = FALSE;
+                                    OPT_MISSING(argv[0], optopt);
+                        else
+                    if (exec_time < 0) {
+                                                fprintf(stdout,"WARNING:
bad argument. Using default %d\n", (exec_time = 24));
+                                    } else

+                                                run_once = FALSE;
                         break;
                 default :
                         usage(argv[0]);
-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
Ltp-list mailing list
Ltp...@li...
https://lists.sourceforge.net/lists/listinfo/ltp-list