Ian Lepore
2017-07-04 23:17:12 UTC
FYI, I've put up two related phab reviews for changes to realtime clock
support.
https://reviews.freebsd.org/D11483
https://reviews.freebsd.org/D11484
D11483 changes some amd64 EFI code to use a hardware-specific mutex
instead of the registered-clock mutex when accessing the atrtc hardware
clock; this just clears the way for the main change.
The main change in D11484 adds support for multiple concurrent realtime
clocks, and eliminates restrictions on locking and sleeping in the
clock driver gettime and settime methods. The sleeping restriction has
been especially problematic for systems using an i2c realtime clock
chip, as i2c bus access often requires sleeping.
-- Ian
support.
https://reviews.freebsd.org/D11483
https://reviews.freebsd.org/D11484
D11483 changes some amd64 EFI code to use a hardware-specific mutex
instead of the registered-clock mutex when accessing the atrtc hardware
clock; this just clears the way for the main change.
The main change in D11484 adds support for multiple concurrent realtime
clocks, and eliminates restrictions on locking and sleeping in the
clock driver gettime and settime methods. The sleeping restriction has
been especially problematic for systems using an i2c realtime clock
chip, as i2c bus access often requires sleeping.
-- Ian