On Wed, Nov 07, 2018 at 04:40:50PM +0100, Dimitrios Katsaros wrote:
> I was instead checking for the clock_gettime syscall. Would that be an
> acceptable solution?
We should check for clock_adjtime() as well.
Thanks,
Richard
___
Linuxptp-devel
Ah and one more clarification,
Because my device is reprisented using the POSIX interface, i do not
support the
phc_get_caps option. if I were to use the -p option to pass any clock, I
would have to
change this to something more basic, like I was doing in the posix_open
function, where
I was
Thanks for the quick reply,
I was more afraid that changing the phc to not be expected to be in the
pattern /dev/ptp%
and to remove the phc_get_caps check would find more resistance. I can make
a new patch
with your suggestion and get back to you.
Thanks!
Dimitrios
On Wed, Nov 7, 2018 at 4:07
On Wed, Nov 07, 2018 at 03:16:22PM +0100, Dimitrios Katsaros wrote:
> However, there is one issue for this setup to work correctly: fd clock
> support is limited. The kernel does provide the ability to implement your
> own POSIX clocks, but the daemons providing ptp clock support do not have
> the
Hello,
I was not very descriptive with the patches, sorry about that.
I am working on a set of devices where I wish to expose custom system
clocks. Because we will be using different technologies for different
hardware setups, the goal is to develop custom clock drivers using the
POSIX interface
On Mon, Nov 05, 2018 at 12:00:40PM +0100, Dimitrios Katsaros wrote:
> This patch adds dedicated logic for opening and closing posix clocks.
> We will need this for adding logic to ptp4l for using clocks that
> may not be ptp clocks.
Tell us about the use case, please.
Thanks,
Richard
This patch adds dedicated logic for opening and closing posix clocks.
We will need this for adding logic to ptp4l for using clocks that
may not be ptp clocks.
Signed-off-by: Dimitrios Katsaros
---
makefile | 2 +-
missing.h| 10 +-
posixclock.c | 51