[PATCH] New fstest to check rename POSIX conformance

[Andre Marques]

On 02/21/14 13:43, Sebastian Huber wrote:
> On 2014-02-21 14:38, Andre Marques wrote:
>>>> I tried to create a ramdisk as in
>>>>
>>>> http://www.rtems.org/onlinedocs/doxygen/cpukit/html/group__rtems__ramdisk.html#gac6c99eed9f3b92bb4cf5184b25972e65 
>>>>
>>>>
>>>>
>>>>
>>>> but the program exits during the rtems_blkdev_create() call, if I use
>>>> "/dev/rda" as the device path.
>>>
>>> How does the code look like?
>>
>> With
>>
>>    rtems_status_code sc;
>>    ramdisk *rd;
>>
>>    const char device [] = "/dev/rda";
>>    uint32_t media_block_size = 512;
>>    rtems_blkdev_bnum media_block_count = 4000;
>
> So this are 2048000 bytes.  Has your target this amount of RAM?
>

I'm running qemu with -m 256M, so that should be plenty?

>>
>>     rd = ramdisk_allocate(NULL, media_block_size, media_block_count, 
>> false);
>>     if (rd != NULL) {
>>       sc = rtems_blkdev_create(
>>         device,
>>         media_block_size,
>>         media_block_count,
>>         ramdisk_ioctl,
>>         rd);
>>
>>       perror("rtems_blkdev_create -> ");
>>     }
>>
>>   the program just exits without giving any reason before the 
>> perror() call. I
>> assumed it was because the /dev directory wasn't there, so I created 
>> the /dev
>> directory before the rtems_blkdev_create() call and then it gives
>>
>> rtems_blkdev_create -> : Not supported
>
> You get a status code as the return value.  Use something like this
>
> assert(sc == RTEMS_SUCCESSFUL)
>

I moved on from this to test_initialize_filesystem () and 
test_shutdown_filesystem () instead, which seem to be working as I 
wanted. I've made some questions at the very end of my previous e-mail 
regarding that, don't know if you saw them.