beforeQuery and afterConfig are less-than-perfectly documented (after
looking at the three relevant documents). The excuse for this is that
the idea of batches of properties is normally used only when OpenCPI is
used underneath an SCA core framework, where the higher level SCA APIs
indeed have "batch" reading and writing of property values.
In any case, here is are paragraphs that should be in the docs somewhere:
An implementation (in its OWD) can set the "readsync" or "writesync"
attributes of some of its configuration properties. This means that
those properties are intended to be read and written atomically with
all other properties with that same sync attribute. For reading,
atomically means that the implementation will implement and respond
to the "beforeQuery" operation by taking a "snapshot" of these
properties so that when they are individually read, the values are
all consistent as they where at the time of the "beforeQuery"
operation. Similarly, for "writesync" properties, the
implementation will essentially buffer up the individually written
properties until an "afterConfigure" operation, and then apply them
all atomically.
This implies that a control application that sets and gets property
values needs to know when beforeQuery and afterConfigure are
needed. Note that for implementations that don't bother
implementing beforeQuery or afterConfigure, they are no-ops and
ignored in any case.
Thus beforeQuery can always be done before reading a set of
properties and afterConfigure can be done after writing a set of
properties. In the control API, properties have a readSync() and
writeSync() member functions to test for this attribute (soon to be
released - these were previously internal-only, not API).
Let us know if this is not clear enough.
On 3/8/12 1:10 AM, Ziersch, Troy (Contractor) wrote:
*UNCLASSIFIED*
Hi,
Thanks for the earlier info Jim.
I'm not too clear on the relationship between ReadSync, BeforeQuery
and Read Cycles in HDL workers..
Say I've got a HDL worker with three ULong properties with
ReadSync=True because I want to be able to read three
counters atomically.
Because of the existence of the ReadSync property I'm assuming that
that OCPI will automatically give the HDL worker a BeforeQuery command
when I read these parameters.
There are other assumptions I can make about the relationship between
BeforeQuery commands and parameter reads but I'd just like to confirm
before coding an assumption.
The question is, is there an order I need to read the parameters in to
get one BeforeQuery command followed by three reads?
Thanks,
Troy
*IMPORTANT*: This email remains the property of the Department of
Defence and is subject to the jurisdiction of section 70 of the Crimes
Act 1914. If you have received this email in error, you are requested
to contact the sender and delete the email.
_______________________________________________
opencpi_dev mailing list
[email protected]
http://lists.opencpi.org/listinfo.cgi/opencpi_dev-opencpi.org
_______________________________________________
opencpi_dev mailing list
[email protected]
http://lists.opencpi.org/listinfo.cgi/opencpi_dev-opencpi.org
