On 11/26/2014 01:32 AM, Sebastian Huber wrote:
Hello Joel,
On 25/11/14 20:30, Joel Sherrill wrote:
Coverity Id 1255348 in fs-rtems.c can't be ignored on the grounds
the JFFS2 project will see it.
this is one-to-one copy from the eCos code. I don't know how and why it
works.
Based on history of usage, I was fairly comfortable
trusting the core JFFS2 code. Then I found a Coverity
annotation on one case and decided the core code had
been through some review. So I trust all the non-target
specific code has been reviewed by someone knowledgeable.
To be perfectly blunt, this code does not have that provenance
and I have no confidence anyone has every looked at it. The logic
in this method is indeed twisted.
If you could make an attempt analysing it, it would be greatly
appreciated. If Coverity is right, we have an insidious bug in which
memory is accessed after it is freed which is not good in a
multitasking system and even worse on an SMP one.
--joel
_______________________________________________
devel mailing list
devel@rtems.org
http://lists.rtems.org/mailman/listinfo/devel
