Skip to contentSkip to conversation

clk: thead: Updates for v6.17, part 2

GIT PULL Drew Fustini 2 participants 3 messages Jul 24, 2025

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi Stephen,

I'm sending this followup PR in case there is still time to include it for your 6.17 PR. Yao Zi refactored mux clk registration which fixes an orphan mux clk issue which would occur when booting with mainline uboot.

Thanks,
Drew

The following changes since commit 19272b37aa4f83ca52bdf9c16d5d81bdd1354494:

Linux 6.16-rc1 (2025-06-08 13:44:43 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/fustini/linux.git tags/thead-clk-for-v6.17-p2

for you to fetch changes up to 54edba916e2913b0893b0f6404b73155d48374ea:

clk: thead: th1520-ap: Describe mux clocks with clk_mux (2025-07-22 15:40:54 -0700)

----------------------------------------------------------------
T-HEAD clock changes for v6.17, part 2

One more improvement for the T-HEAD TH1520 clock controller this cycle:

Yao Zi has fixed an issue where the c910 mux clk could end up as an orphan in CCF when the bootloader reparents it to the c910-i0 mux clk. The solution is to refactor the handling of mux clocks by embedding a clk_mux structure directly in ccu_mux. This allows the mux clocks to be registered with devm_clk_hw_register() without allocating any new clk_hw pointer which solves the orphan issue.

This change has been tested in linux-next. The LPi4a still boots okay without clk_ignore_unused and peripherals like serial, emmc and ethernet are functional. The file /sys/kernel/debug/clk/c910/clk_possible_parents now correctly outputs: "c910-i0 cpu-pll1"

----------------------------------------------------------------
Michal Wilczynski (1):
clk: thead: Mark essential bus clocks as CLK_IGNORE_UNUSED

Yao Zi (2):
clk: thead: th1520-ap: Correctly refer the parent of osc_12m
clk: thead: th1520-ap: Describe mux clocks with clk_mux

 drivers/clk/thead/clk-th1520-ap.c | 109 +++++++++++++++++---------------------
 1 file changed, 48 insertions(+), 61 deletions(-)
#1 lore raw
Quoted 5 lines · Drew Fustini

Hi Stephen,

I'm sending this followup PR in case there is still time to include it for your 6.17 PR. Yao Zi refactored mux clk registration which fixes an orphan mux clk issue which would occur when booting with mainline uboot.

[snip]

Quoted 9 lines

Michal Wilczynski (1):
clk: thead: Mark essential bus clocks as CLK_IGNORE_UNUSED

Yao Zi (2):
clk: thead: th1520-ap: Correctly refer the parent of osc_12m
clk: thead: th1520-ap: Describe mux clocks with clk_mux

 drivers/clk/thead/clk-th1520-ap.c | 109 +++++++++++++++++---------------------
 1 file changed, 48 insertions(+), 61 deletions(-)

My apologies, I failed to just send the delta between thead-clk-for-v6.17 and thead-clk-for-v6.17-p2.

The following is the correct text of the pull request. Hopefull this is acceptable?

Thanks,
Drew

--
The following changes since commit d274c77ffa202b70ad01d579f33b73b4de123375:

clk: thead: th1520-ap: Correctly refer the parent of osc_12m (2025-07-13 12:46:58 -0700)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/fustini/linux.git tags/thead-clk-for-v6.17-p2

for you to fetch changes up to 54edba916e2913b0893b0f6404b73155d48374ea:

clk: thead: th1520-ap: Describe mux clocks with clk_mux (2025-07-22 15:40:54 -0700)

----------------------------------------------------------------
T-HEAD clock changes for v6.17, part 2

One more improvement for the T-HEAD TH1520 clock controller this cycle:

Yao Zi has fixed an issue where the c910 mux clk could end up as an orphan in CCF when the bootloader reparents it to the c910-i0 mux clk. The solution is to refactor the handling of mux clocks by embedding a clk_mux structure directly in ccu_mux. This allows the mux clocks to be registered with devm_clk_hw_register() without allocating any new clk_hw pointer which solves the orphan issue.

This change has been tested in linux-next. The LPi4a still boots okay without clk_ignore_unused and peripherals like serial, emmc and ethernet are functional. The file /sys/kernel/debug/clk/c910/clk_possible_parents now correctly outputs: "c910-i0 cpu-pll1"

----------------------------------------------------------------
Yao Zi (1):
clk: thead: th1520-ap: Describe mux clocks with clk_mux

 drivers/clk/thead/clk-th1520-ap.c | 95 +++++++++++++++------------------------
 1 file changed, 37 insertions(+), 58 deletions(-)
maintainer
#2 lore raw
In reply to Drew Fustini

Quoting Drew Fustini (2025-07-24 13:48:23)

Quoted 15 lines

Hi Stephen,

I'm sending this followup PR in case there is still time to include it for your 6.17 PR. Yao Zi refactored mux clk registration which fixes an orphan mux clk issue which would occur when booting with mainline uboot.

Thanks,
Drew

The following changes since commit 19272b37aa4f83ca52bdf9c16d5d81bdd1354494:

Linux 6.16-rc1 (2025-06-08 13:44:43 -0700)

are available in the Git repository at:

Quoted 7 lines

git://git.kernel.org/pub/scm/linux/kernel/git/fustini/linux.git tags/thead-clk-for-v6.17-p2

for you to fetch changes up to 54edba916e2913b0893b0f6404b73155d48374ea:

clk: thead: th1520-ap: Describe mux clocks with clk_mux (2025-07-22 15:40:54 -0700)

----------------------------------------------------------------

Thanks. Pulled into to clk-next