This is the mail archive of the ecos-discuss@sources.redhat.com mailing list for the eCos project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: posix time functions


David Brennan <eCos@brennanhome.com> writes:

> We discovered a (perhaps) un-documented feature in the eCos posix
> clock_settime and clock_gettime functions. Setting the time, changes
> the system clock (number of ticks). The system clock should only be
> changeable by the interrupt routine which is incrementing it. All
> other code should really only have read access.

Hmm, we have also had complaints that clock_settime() doesn't change
enough and want it to alter the wallclock time as well. I guess you
cannot please everyone.

> Is there a chance this
> could be changed? That is, if we fix it in the kernel, would that type
> of patch be accepted.

A patch that maintains an offet for CLOCK_REALTIME from the main
timer and used that when calculating timer timeouts would be
fine. Depending on how much gets changed, we might need an assignment
for it, though.

> I would think that setting the clock and
> possibly losing alarms would be something that other people have
> noticed before.

To be honest, I don't think that many people call clock_settime().
It's mainly there for completeness.

-- 
Nick Garnett                    eCos Kernel Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]