| Message ID | 20191218202142.11717-8-digetx@gmail.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | NVIDIA Tegra20 CPUFreq driver major update | expand |
Hi Dmitry Finally, I came around giving this a try again on the Colibri T20 and Apalis/Colibri T30. Overall, works like a charm and you may add the following to the entire series: On Wed, 2019-12-18 at 23:21 +0300, Dmitry Osipenko wrote: > PLLX may be kept disabled if cpufreq driver selects some other clock > for > CPU. In that case PLLX will be disabled later in the resume path by > the > CLK driver, which also can enable PLLX if necessary by itself. Thus > there > is no need to enable PLLX early during resume. Tegra114/124 CLK > drivers do > not manage PLLX on resume and thus they are left untouched by this > patch. > > Signed-off-by: Dmitry Osipenko <digetx@gmail.com> Tested-by: Marcel Ziswiler <marcel@ziswiler.com> Tested-on: Colibri T20, Apalis/Colibri T30 on resp. EvalBoards > --- > arch/arm/mach-tegra/sleep-tegra30.S | 11 ++++++++++- > 1 file changed, 10 insertions(+), 1 deletion(-) > > diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach- > tegra/sleep-tegra30.S > index 9a20c93abe48..4f073869b8ac 100644 > --- a/arch/arm/mach-tegra/sleep-tegra30.S > +++ b/arch/arm/mach-tegra/sleep-tegra30.S > @@ -358,7 +358,6 @@ _no_pll_iddq_exit: > > pll_enable r1, r0, CLK_RESET_PLLM_BASE, CLK_RESET_PLLM_MISC > pll_enable r1, r0, CLK_RESET_PLLC_BASE, CLK_RESET_PLLC_MISC > - pll_enable r1, r0, CLK_RESET_PLLX_BASE, CLK_RESET_PLLX_MISC > > _pll_m_c_x_done: > pll_enable r1, r0, CLK_RESET_PLLP_BASE, CLK_RESET_PLLP_MISC > @@ -368,8 +367,18 @@ _pll_m_c_x_done: > pll_locked r1, r0, CLK_RESET_PLLP_BASE > pll_locked r1, r0, CLK_RESET_PLLA_BASE > pll_locked r1, r0, CLK_RESET_PLLC_BASE > + > + /* > + * CPUFreq driver could select other PLL for CPU. PLLX will be > + * enabled by the Tegra30 CLK driver on an as-needed basis, see > + * tegra30_cpu_clock_resume(). > + */ > + cmp r10, #TEGRA30 > + beq _pll_m_c_x_locked > + > pll_locked r1, r0, CLK_RESET_PLLX_BASE > > +_pll_m_c_x_locked: > mov32 r7, TEGRA_TMRUS_BASE > ldr r1, [r7] > add r1, r1, #LOCK_DELAY Unfortunately, that one does no longer apply after the following patch recently got applied on Wed Jan 8: commit 1a3388d506bf ("ARM: tegra: Enable PLLP bypass during Tegra124 LP1") Thanks! Cheers Marcel
02.02.2020 01:57, Marcel Ziswiler пишет: > Hi Dmitry > > Finally, I came around giving this a try again on the Colibri T20 and > Apalis/Colibri T30. > > Overall, works like a charm and you may add the following to the entire > series: > > On Wed, 2019-12-18 at 23:21 +0300, Dmitry Osipenko wrote: >> PLLX may be kept disabled if cpufreq driver selects some other clock >> for >> CPU. In that case PLLX will be disabled later in the resume path by >> the >> CLK driver, which also can enable PLLX if necessary by itself. Thus >> there >> is no need to enable PLLX early during resume. Tegra114/124 CLK >> drivers do >> not manage PLLX on resume and thus they are left untouched by this >> patch. >> >> Signed-off-by: Dmitry Osipenko <digetx@gmail.com> > > Tested-by: Marcel Ziswiler <marcel@ziswiler.com> > Tested-on: Colibri T20, Apalis/Colibri T30 on resp. EvalBoards Hello Marcel, Thank you very much for the testing! Very appreciate that! Could you please clarify whether you tried to enable the CPU DVFS by including OPPs/etc into device-tree files of the tested boards? >> --- >> arch/arm/mach-tegra/sleep-tegra30.S | 11 ++++++++++- >> 1 file changed, 10 insertions(+), 1 deletion(-) >> >> diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach- >> tegra/sleep-tegra30.S >> index 9a20c93abe48..4f073869b8ac 100644 >> --- a/arch/arm/mach-tegra/sleep-tegra30.S >> +++ b/arch/arm/mach-tegra/sleep-tegra30.S >> @@ -358,7 +358,6 @@ _no_pll_iddq_exit: >> >> pll_enable r1, r0, CLK_RESET_PLLM_BASE, CLK_RESET_PLLM_MISC >> pll_enable r1, r0, CLK_RESET_PLLC_BASE, CLK_RESET_PLLC_MISC >> - pll_enable r1, r0, CLK_RESET_PLLX_BASE, CLK_RESET_PLLX_MISC >> >> _pll_m_c_x_done: >> pll_enable r1, r0, CLK_RESET_PLLP_BASE, CLK_RESET_PLLP_MISC >> @@ -368,8 +367,18 @@ _pll_m_c_x_done: >> pll_locked r1, r0, CLK_RESET_PLLP_BASE >> pll_locked r1, r0, CLK_RESET_PLLA_BASE >> pll_locked r1, r0, CLK_RESET_PLLC_BASE >> + >> + /* >> + * CPUFreq driver could select other PLL for CPU. PLLX will be >> + * enabled by the Tegra30 CLK driver on an as-needed basis, see >> + * tegra30_cpu_clock_resume(). >> + */ >> + cmp r10, #TEGRA30 >> + beq _pll_m_c_x_locked >> + >> pll_locked r1, r0, CLK_RESET_PLLX_BASE >> >> +_pll_m_c_x_locked: >> mov32 r7, TEGRA_TMRUS_BASE >> ldr r1, [r7] >> add r1, r1, #LOCK_DELAY > > Unfortunately, that one does no longer apply after the following patch > recently got applied on Wed Jan 8: > > commit 1a3388d506bf ("ARM: tegra: Enable PLLP bypass during Tegra124 > LP1") > > Thanks! Yes, I'm aware about it and going to send out the updated version of the patches soon. Meanwhile you could find recent rebase of the patches here [1]. [1] https://github.com/grate-driver/linux/commits/master
diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach-tegra/sleep-tegra30.S index 9a20c93abe48..4f073869b8ac 100644 --- a/arch/arm/mach-tegra/sleep-tegra30.S +++ b/arch/arm/mach-tegra/sleep-tegra30.S @@ -358,7 +358,6 @@ _no_pll_iddq_exit: pll_enable r1, r0, CLK_RESET_PLLM_BASE, CLK_RESET_PLLM_MISC pll_enable r1, r0, CLK_RESET_PLLC_BASE, CLK_RESET_PLLC_MISC - pll_enable r1, r0, CLK_RESET_PLLX_BASE, CLK_RESET_PLLX_MISC _pll_m_c_x_done: pll_enable r1, r0, CLK_RESET_PLLP_BASE, CLK_RESET_PLLP_MISC @@ -368,8 +367,18 @@ _pll_m_c_x_done: pll_locked r1, r0, CLK_RESET_PLLP_BASE pll_locked r1, r0, CLK_RESET_PLLA_BASE pll_locked r1, r0, CLK_RESET_PLLC_BASE + + /* + * CPUFreq driver could select other PLL for CPU. PLLX will be + * enabled by the Tegra30 CLK driver on an as-needed basis, see + * tegra30_cpu_clock_resume(). + */ + cmp r10, #TEGRA30 + beq _pll_m_c_x_locked + pll_locked r1, r0, CLK_RESET_PLLX_BASE +_pll_m_c_x_locked: mov32 r7, TEGRA_TMRUS_BASE ldr r1, [r7] add r1, r1, #LOCK_DELAY
PLLX may be kept disabled if cpufreq driver selects some other clock for CPU. In that case PLLX will be disabled later in the resume path by the CLK driver, which also can enable PLLX if necessary by itself. Thus there is no need to enable PLLX early during resume. Tegra114/124 CLK drivers do not manage PLLX on resume and thus they are left untouched by this patch. Signed-off-by: Dmitry Osipenko <digetx@gmail.com> --- arch/arm/mach-tegra/sleep-tegra30.S | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)