[Bug 1517] mount does not use the provided filesystem handler table functions

bugzilla-daemon at rtems.org bugzilla-daemon at rtems.org
Tue Jun 8 02:41:07 UTC 2010


https://www.rtems.org/bugzilla/show_bug.cgi?id=1517





--- Comment #11 from Chris Johns <chrisj at rtems.org>  2010-06-07 21:41:05 ---
I am not sure how to answer the file system unregister not being thread safe.
There are so many holes in this area of the code picking on this function is
difficult to respond to. If I fix the ability to control the list extraction,
which I assume is the reason you mentioned this (and correctly so), then what
happens if the file system is in use ? This whole area of RTEMS needs to be
fixed and my hope is the cleaning up of this will provide the infrastructure to
support these file system calls being thread safe.

I consider the registration of file systems as a system level operations and
encourage users to manage this at the application level until such time as
RTEMS has better file system support for threading. Currently this interface is
provided for development of new file systems. I suspect almost all users will
use the static confdefs.h approach to managing which file systems they have in
a system. Having said this, dynamic loading of modules will expose this
argument as weak and we will have address the issue at some point in the
future.

I do not agree to the need to change the way the iterator works. It is modeled
on the chains type iterator and further I have worked on an open source package
recently that used this a lot and I considered it confusing and obtuse. Having
to design code to work through a specific handler interface complicates users
of the iterator plus designs tend to end up having iterators of this type
nesting to do anything useful. I discourage the use of this pattern.

-- 
Configure bugmail: https://www.rtems.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.



More information about the bugs mailing list