Questions about Configuration for shell prompt & file system
Gedare Bloom
gedare at rtems.org
Wed Jan 10 16:59:52 UTC 2018
On Thu, Dec 28, 2017 at 3:34 AM, jaeho jo <hot486two at gmail.com> wrote:
> Hi, I have 2 Questions about RTEMS!
> 1. What is the relationship between shell prompt and unlimited_object?
> When I using RTEMS 4.11.2, I need 'CONFIGURE_UNLIMITED_OBJECTS' to use
> shell prompt.
> I already read 'RTEMS C User Manual - 24.7. Unlimited Objects ' page,
> but I don't understand exactly why I need that configuration.
You shouldn't need unlimited objects for the shell. Unlimited objects
is provided to simplify porting and to allow users to ignore resource
requirements during development. For deployed/production systems,
fine-tuning of the needed kernel objects is recommended to reduce size
overheads.
Why did you think that you need this unlimited objects to use the
shell? See for example libtests/capture01 that uses the shell without
unlimited.
> What is the relationship between shell prompt and unlimited_object?
> Also, Can you briefly describe the structural differences associated
> with this? between RTEMS 4.10.2 and RTEMS 4.11.2.
>
CONFIGURE_UNLIMITED_OBJECTS does not exist in 4.10.
> 2. I wonder if it is designed to ignore file system options when mounting a
> file system in RTEMS 4.10.2
> I mounted the file system using OSAL source code.
>
>
> https://github.com/nasa/osal/blob/master/src/os/rtems/osfilesys.c#L529
> The 4th parameter of this function determines whether a write
> operation is possible to the file system
>
> Option for file system
> typedef enum {
> RTEMS_FILESYSTEM_READ_ONLY,
> RTEMS_FILESYSTEM_READ_WRITE,
> RTEMS_FILESYSTEM_BAD_OPTIONS
> } rtems_filesystem_options_t; // libio.h
>
> With RTEMS 4.10.2, Even if RTEMS_FILESYSTEM_READ_ONLY(0) is used as an
> option of the file system, the file writing to the file system is working
> With RTEMS 4.11.2, if RTEMS_FILESYSTEM_READ_ONLY(0) is used as an option
> of the file system, the file writing to the file system is not working.
> if RTEMS_FILESYSTEM_READ_WRITE(1) is
> used as an option of the file system, the file writing to the file system is
> working.
There were a number of changes made in the filesystem/block layer
during RTEMS 4.11 including heavy changes to how file paths, modes,
and permissions are evaluated. It is not surprising to see different
behavior between the two versions.
> As far as I can see, RTEMS 4.11.2 seems to be working properly.
> I wonder if it is designed to ignore file system options when mounting a
> file system in RTEMS 4.10.2 or is it a bug?
>
This sounds like a bug. It could just be that no one cared to validate
that write-protection worked before.
>
> Thank you for reading the long questions.
>
> jhjo
>
> _______________________________________________
> users mailing list
> users at rtems.org
> http://lists.rtems.org/mailman/listinfo/users
More information about the users
mailing list