Hi Stephen,
On 8/26/21 8:47 PM, Stephen Boyd wrote:
Quoting Bjorn Andersson (2021-08-25 13:45:17)
Using parent_data and parent_hws, instead of parent_names, does protect
against some cases of incompletely defined clock trees. While it turns
out that the bug being chased this time was totally unrelated, this
patch converts the SDM660 GCC driver to avoid such issues.
The "xo" fixed_factor clock is unused within the gcc driver, but
referenced from the DSI PHY. So it's left in place until the DSI driver
is updated.
Tested-by: Marijn Suijten <marijn.suijten@xxxxxxxxxxxxxx>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@xxxxxxxxxxxxxx>
Reviewed-by: Marijn Suijten <marijn.suijten@xxxxxxxxxxxxxx>
Signed-off-by: Bjorn Andersson <bjorn.andersson@xxxxxxxxxx>
---
Applied to clk-next after moving back the arrays so the diff is smaller.
Unfortunately the arrays cannot be moved back up since they (for the
most part) reference gpll clocks in `.hw = &gpll*` fields - these have
to be defined first or forward-declared otherwise, which is why the
arrays were moved down in the first place.
The general structure nowadays seems to place these parent maps/data
right above the clock that uses them, making it easier to read.
- Marijn
