[RTEMS SMP ISVV] : Timer Server
Luis Alves
lmbalves at gmail.com
Fri Sep 23 11:19:55 UTC 2022
Dear RTEMS community,
I’m part of the team performing the Independent Software Verification and
Validation (ISVV) of the RTEMS SMP Qualification Data Pack (which is being
provided through https://rtems-qual.io.esa.int/).
While trying to setup tests to assert the overall “Timer Server” feature,
one has realized that:
- Although rtems_timer_initiate_server() is mentioned within section
4.10 of the SW Configuration File (SCF) as a pre-qualified item (“*which
may be directly used by applications*”), one gets build issues while
trying to use it, as it depends on rtems_timer_create().
- The same SCF (within section 4.9.3) also says: “*The
rtems_task_create() directive is pre-qualified, however, it is only
available if a custom stack allocator is configured. It is recommended to
use rtems_task_construct() instead*”:
- Isn’t this contradictory, as:
- Both rtems_timer_initiate_server() and tems_task_create() are
listed within SCF section 4.10 and
- at the start of section 4.10, it is said that listed items
are “*pre-qualified
items .. which may be directly used by applications*”?
- Given this, shouldn’t rtems_timer_initiate_server() natively use of
rtems_task_construct() rather than rtems_timer_create() as advised?
- This makes other two functions listed within SCF section 4.10 as not
directly usable, namely rtems_timer_server_fire_after() and
rtems_timer_server_fire_when() as those are to be used in conjunction with
rtems_timer_initiate_server().
- In a nutshell, this compromises the overall “Timer Server” feature in
the qual-only setup of the QDP.
Can the community help to clarify these points, and tell us what and how
can one improve the RTEMS SMP QDP w.r.t. this (at the documentation and
qual-only implementation level)?
Thanks in advance for the clarifications,
Kind Regards
Luís Alves
Critical Software
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20220923/5202edcf/attachment-0001.htm>
More information about the devel
mailing list