diff options
| author | Andreas Kemnade <andreas@kemnade.info> | 2026-09-09 14:16:36 +0200 |
|---|---|---|
| committer | Brian Masney <bmasney@redhat.com> | 2026-09-11 19:08:36 -0400 |
| commit | 59e0cc31cc5c43586c6eaada62c2c863bfbcbfb5 (patch) | |
| tree | c4ba2998347f2ff972304d4c4ca0b101d357a416 /lib/zstd | |
| parent | df2908090cda368b01ff43709f51890076c56157 (diff) | |
| download | linux-stable-59e0cc31cc5c43586c6eaada62c2c863bfbcbfb5.tar.gz linux-stable-59e0cc31cc5c43586c6eaada62c2c863bfbcbfb5.zip | |
clk: ti: composite: resolve parent clocks by name again
Trying to resolve them by DT index causes massive havoc on OMAP3, Clocks
around the timers used as a system clocksource are not properly resolved
causing ealy boot failures. The problem seems to be that parents must be
resolved by looking into the clocks property of the component node marked
with "ti,composite-mux-clock", with the switch to dt index, that node
was not used anymore. It was used by the of_clk_parent_fill() call.
To a lesser extent also OMAP4/5 boards are affected.
Since this patch was introduced just because of a cleanup request
and not to solve the actual problem
(https://lore.kernel.org/linux-omap/alkZmw-XmnCOZfOD@redhat.com/)
just revert it. Cleanup needs really more thought here.
The similar change to the TI mux clock, which solves a problem on the AM3
platform, seems to be harmless.
So just revert
commit fe3dd92ac54a ("clk: ti: composite: resolve parent clocks by DT index, not by name")
for now.
Fixes: fe3dd92ac54a ("clk: ti: composite: resolve parent clocks by DT index, not by name")
Reviewed-by: Mathieu Dubois-Briand <mathieu.dubois-briand@bootlin.com>
Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
Signed-off-by: Brian Masney <bmasney@redhat.com>
Diffstat (limited to 'lib/zstd')
0 files changed, 0 insertions, 0 deletions
