Since a rework of CAMSS power domain management for newer platforms done in
commit 2f6f8af67203 ("media: camss: Refactor VFE power domain toggling"),
all operations over CAMSS imply enablement of all power domains described
in the correspondent device tree node. Apparently it's too excessive and it's
quite a complicated scheme to allow simple addition of newer platforms with
even more power domains.
I would appreciate, if somebody can test the changes on db820 for any probable
regressions.
The change is based on changes in clock framework [1] and on a recent fix-up [2]
in camss.
[1] https://lore.kernel.org/linux-clk/20220519214133.1728979-1-vladimir.zapolskiy@xxxxxxxxxx/
[2] https://lore.kernel.org/linux-media/20220704220814.629130-1-vladimir.zapolskiy@xxxxxxxxxx/
Vladimir Zapolskiy (2):
media: camss: Collect information about a number of lite VFEs
media: camss: Split power domain management
.../media/platform/qcom/camss/camss-vfe-170.c | 20 +++++++-
.../media/platform/qcom/camss/camss-vfe-480.c | 20 +++++++-
drivers/media/platform/qcom/camss/camss.c | 50 ++++++++++---------
drivers/media/platform/qcom/camss/camss.h | 1 +
4 files changed, 66 insertions(+), 25 deletions(-)
--
2.33.0
