[PATCH V3 1/4] clk: tegra: mark fuse clock as critical

[PATCH V3 1/4] clk: tegra: mark fuse clock as critical

[Stephen Warren]

From: Stephen Warren <swarren@nvidia.com>

For a little over a year, U-Boot on Tegra124 has configured the flow
controller to perform automatic RAM re-repair on off->on power transitions
of the CPU rail1]. This is mandatory for correct operation of Tegra124.
However, RAM re-repair relies on certain clocks, which the kernel must
enable and leave running. The fuse clock is one of those clocks. Mark this
clock as critical so that LP1 power mode (system suspend) operates
correctly.

[1] 3cc7942a4ae5 ARM: tegra: implement RAM repair

Reported-by: Jonathan Hunter <jonathanh@nvidia.com>
Cc: stable@vger.kernel.org
Signed-off-by: Stephen Warren <swarren@nvidia.com>
---
v3: Added comment to the clock table entry indicating why the clock is
    critical.
v2: Set CRITICAL flag on the clock, rather than enabling it in
    tegra124_init_table[].
---
 drivers/clk/tegra/clk-tegra-periph.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/clk/tegra/clk-tegra-periph.c b/drivers/clk/tegra/clk-tegra-periph.c
index 1ed85f120a1b..49b9f2f85bad 100644
--- a/drivers/clk/tegra/clk-tegra-periph.c
+++ b/drivers/clk/tegra/clk-tegra-periph.c
@@ -785,7 +785,11 @@ static struct tegra_periph_init_data gate_clks[] = {
 	GATE("ahbdma", "hclk", 33, 0, tegra_clk_ahbdma, 0),
 	GATE("apbdma", "pclk", 34, 0, tegra_clk_apbdma, 0),
 	GATE("kbc", "clk_32k", 36, TEGRA_PERIPH_ON_APB | TEGRA_PERIPH_NO_RESET, tegra_clk_kbc, 0),
-	GATE("fuse", "clk_m", 39, TEGRA_PERIPH_ON_APB, tegra_clk_fuse, 0),
+	/*
+	 * Critical for RAM re-repair operation, which must occur on resume
+	 * from LP1 system suspend and as part of CCPLEX cluster switching.
+	 */
+	GATE("fuse", "clk_m", 39, TEGRA_PERIPH_ON_APB, tegra_clk_fuse, CLK_IS_CRITICAL),
 	GATE("fuse_burn", "clk_m", 39, TEGRA_PERIPH_ON_APB, tegra_clk_fuse_burn, 0),
 	GATE("kfuse", "clk_m", 40, TEGRA_PERIPH_ON_APB, tegra_clk_kfuse, 0),
 	GATE("apbif", "clk_m", 107, TEGRA_PERIPH_ON_APB, tegra_clk_apbif, 0),
-- 
2.23.0

[PATCH V3 2/4] ARM: tegra: Enable PLLP bypass during Tegra124 LP1

[Stephen Warren]

From: Stephen Warren <swarren@nvidia.com>

For a little over a year, U-Boot has configured the flow controller to
perform automatic RAM re-repair on off->on power transitions of the CPU
rail1]. This is mandatory for correct operation of Tegra124. However, RAM
re-repair relies on certain clocks, which the kernel must enable and
leave running. PLLP is one of those clocks. This clock is shut down
during LP1 in order to save power. Enable bypass (which I believe routes
osc_div_clk, essentially the crystal clock, to the PLL output) so that
this clock signal toggles even though the PLL is not active. This is
required so that LP1 power mode (system suspend) operates correctly.

The bypass configuration must then be undone when resuming from LP1, so
that all peripheral clocks run at the expected rate. Without this, many
peripherals won't work correctly; for example, the UART baud rate would
be incorrect.

NVIDIA's downstream kernel code only does this if not compiled for
Tegra30, so the added code is made conditional upon the chip ID. NVIDIA's
downstream code makes this change conditional upon the active CPU
cluster. The upstream kernel currently doesn't support cluster switching,
so this patch doesn't test the active CPU cluster ID.

[1] 3cc7942a4ae5 ARM: tegra: implement RAM repair

Reported-by: Jonathan Hunter <jonathanh@nvidia.com>
Cc: stable@vger.kernel.org
Signed-off-by: Stephen Warren <swarren@nvidia.com>
---
v3: No change.
v2: No change.
---
 arch/arm/mach-tegra/sleep-tegra30.S | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach-tegra/sleep-tegra30.S
index b408fa56eb89..6922dd8d3e2d 100644
--- a/arch/arm/mach-tegra/sleep-tegra30.S
+++ b/arch/arm/mach-tegra/sleep-tegra30.S
@@ -370,6 +370,14 @@ _pll_m_c_x_done:
 	pll_locked r1, r0, CLK_RESET_PLLC_BASE
 	pll_locked r1, r0, CLK_RESET_PLLX_BASE
 
+	tegra_get_soc_id TEGRA_APB_MISC_BASE, r1
+	cmp	r1, #TEGRA30
+	beq	1f
+	ldr	r1, [r0, #CLK_RESET_PLLP_BASE]
+	bic	r1, r1, #(1<<31)	@ disable PllP bypass
+	str	r1, [r0, #CLK_RESET_PLLP_BASE]
+1:
+
 	mov32	r7, TEGRA_TMRUS_BASE
 	ldr	r1, [r7]
 	add	r1, r1, #LOCK_DELAY
@@ -630,7 +638,10 @@ tegra30_switch_cpu_to_clk32k:
 	str	r0, [r4, #PMC_PLLP_WB0_OVERRIDE]
 
 	/* disable PLLP, PLLA, PLLC and PLLX */
+	tegra_get_soc_id TEGRA_APB_MISC_BASE, r1
+	cmp	r1, #TEGRA30
 	ldr	r0, [r5, #CLK_RESET_PLLP_BASE]
+	orrne	r0, r0, #(1 << 31)	@ enable PllP bypass on fast cluster
 	bic	r0, r0, #(1 << 30)
 	str	r0, [r5, #CLK_RESET_PLLP_BASE]
 	ldr	r0, [r5, #CLK_RESET_PLLA_BASE]
-- 
2.23.0

[PATCH V3 3/4] ARM: tegra: modify reshift divider during LP1

[Stephen Warren]

From: Stephen Warren <swarren@nvidia.com>

The reshift hardware module implements the RAM re-repair process. This
module uses PLLP as an input clock during LP1 resume. The input divider
for this clock is typically set for PLLP's normal rate. During LP1
resume, PLLP is bypassed and so runs at the crystal rate, which is much
slower. Consequently, decrease the divider so that the reshift module
runs at a reasonable rate during LP1 resume.

NVIDIA's downstream kernel code only does this if not compiled for
Tegra30, so the added code is made conditional upon the chip ID.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
---
v3: No change.
v2: No change.
---
 arch/arm/mach-tegra/sleep-tegra30.S | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach-tegra/sleep-tegra30.S
index 6922dd8d3e2d..6191f9456288 100644
--- a/arch/arm/mach-tegra/sleep-tegra30.S
+++ b/arch/arm/mach-tegra/sleep-tegra30.S
@@ -59,6 +59,9 @@
 #define CLK_RESET_PLLX_MISC3_IDDQ	3
 #define CLK_RESET_PLLM_MISC_IDDQ	5
 #define CLK_RESET_PLLC_MISC_IDDQ	26
+#define CLK_RESET_PLLP_RESHIFT		0x528
+#define CLK_RESET_PLLP_RESHIFT_DEFAULT	0x3b
+#define CLK_RESET_PLLP_RESHIFT_ENABLE	0x3
 
 #define CLK_RESET_CLK_SOURCE_MSELECT	0x3b4
 
@@ -373,9 +376,13 @@ _pll_m_c_x_done:
 	tegra_get_soc_id TEGRA_APB_MISC_BASE, r1
 	cmp	r1, #TEGRA30
 	beq	1f
+
 	ldr	r1, [r0, #CLK_RESET_PLLP_BASE]
 	bic	r1, r1, #(1<<31)	@ disable PllP bypass
 	str	r1, [r0, #CLK_RESET_PLLP_BASE]
+
+	mov	r1, #CLK_RESET_PLLP_RESHIFT_DEFAULT
+	str	r1, [r0, #CLK_RESET_PLLP_RESHIFT]
 1:
 
 	mov32	r7, TEGRA_TMRUS_BASE
@@ -644,6 +651,10 @@ tegra30_switch_cpu_to_clk32k:
 	orrne	r0, r0, #(1 << 31)	@ enable PllP bypass on fast cluster
 	bic	r0, r0, #(1 << 30)
 	str	r0, [r5, #CLK_RESET_PLLP_BASE]
+	beq	1f
+	mov	r0, #CLK_RESET_PLLP_RESHIFT_ENABLE
+	str	r0, [r5, #CLK_RESET_PLLP_RESHIFT]
+1:
 	ldr	r0, [r5, #CLK_RESET_PLLA_BASE]
 	bic	r0, r0, #(1 << 30)
 	str	r0, [r5, #CLK_RESET_PLLA_BASE]
-- 
2.23.0

[PATCH V3 4/4] ARM: tegra: use clk_m CPU on Tegra124 LP1 resume

[Stephen Warren]

From: Stephen Warren <swarren@nvidia.com>

Configure the clock controller to set an alternate clock for the CPU when
it receives an IRQ during LP1 (system suspend). Specifically, use clk_m
(the crystal) rather than clk_s (a 32KHz clock). Such an IRQ will be the
LP1 wake event. This reduces the amount of time taken to resume from LP1.

NVIDIA's downstream kernel executes this code on both Tegra30 and
Tegra124, so it appears OK to make this change unconditionally.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
---
v3: No change.
v2: No change.
---
 arch/arm/mach-tegra/sleep-tegra30.S | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/arch/arm/mach-tegra/sleep-tegra30.S b/arch/arm/mach-tegra/sleep-tegra30.S
index 6191f9456288..ba5e9c07d1b6 100644
--- a/arch/arm/mach-tegra/sleep-tegra30.S
+++ b/arch/arm/mach-tegra/sleep-tegra30.S
@@ -670,8 +670,12 @@ tegra30_switch_cpu_to_clk32k:
 	pll_iddq_entry r1, r5, CLK_RESET_PLLX_MISC3, CLK_RESET_PLLX_MISC3_IDDQ
 _no_pll_in_iddq:
 
-	/* switch to CLKS */
-	mov	r0, #0	/* brust policy = 32KHz */
+	/*
+	 * Switch to clk_s (32KHz); bits 28:31=0
+	 * Enable burst on CPU IRQ; bit 24=1
+	 * Set IRQ burst clock source to clk_m; bits 10:8=0
+	 */
+	mov	r0, #(1 << 24)
 	str	r0, [r5, #CLK_RESET_SCLK_BURST]
 
 	ret	lr
-- 
2.23.0

Re: [PATCH V3 1/4] clk: tegra: mark fuse clock as critical

[Stephen Warren]

On 10/3/19 2:50 PM, Stephen Warren wrote:
> From: Stephen Warren <swarren@nvidia.com>
> 
> For a little over a year, U-Boot on Tegra124 has configured the flow
> controller to perform automatic RAM re-repair on off->on power transitions
> of the CPU rail1]. This is mandatory for correct operation of Tegra124.
> However, RAM re-repair relies on certain clocks, which the kernel must
> enable and leave running. The fuse clock is one of those clocks. Mark this
> clock as critical so that LP1 power mode (system suspend) operates
> correctly.
> 
> [1] 3cc7942a4ae5 ARM: tegra: implement RAM repair

Thierry, this series doesn't seem to be applied yet; could you please 
take a look? Thanks.

Re: [PATCH V3 1/4] clk: tegra: mark fuse clock as critical

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 1125 bytes --]

On Thu, Oct 03, 2019 at 02:50:30PM -0600, Stephen Warren wrote:
> From: Stephen Warren <swarren@nvidia.com>
> 
> For a little over a year, U-Boot on Tegra124 has configured the flow
> controller to perform automatic RAM re-repair on off->on power transitions
> of the CPU rail1]. This is mandatory for correct operation of Tegra124.
> However, RAM re-repair relies on certain clocks, which the kernel must
> enable and leave running. The fuse clock is one of those clocks. Mark this
> clock as critical so that LP1 power mode (system suspend) operates
> correctly.
> 
> [1] 3cc7942a4ae5 ARM: tegra: implement RAM repair
> 
> Reported-by: Jonathan Hunter <jonathanh@nvidia.com>
> Cc: stable@vger.kernel.org
> Signed-off-by: Stephen Warren <swarren@nvidia.com>
> ---
> v3: Added comment to the clock table entry indicating why the clock is
>     critical.
> v2: Set CRITICAL flag on the clock, rather than enabling it in
>     tegra124_init_table[].
> ---
>  drivers/clk/tegra/clk-tegra-periph.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)

Applied to for-5.6/clk, thanks.

Thierry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH V3 2/4] ARM: tegra: Enable PLLP bypass during Tegra124 LP1

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 1784 bytes --]

On Thu, Oct 03, 2019 at 02:50:31PM -0600, Stephen Warren wrote:
> From: Stephen Warren <swarren@nvidia.com>
> 
> For a little over a year, U-Boot has configured the flow controller to
> perform automatic RAM re-repair on off->on power transitions of the CPU
> rail1]. This is mandatory for correct operation of Tegra124. However, RAM
> re-repair relies on certain clocks, which the kernel must enable and
> leave running. PLLP is one of those clocks. This clock is shut down
> during LP1 in order to save power. Enable bypass (which I believe routes
> osc_div_clk, essentially the crystal clock, to the PLL output) so that
> this clock signal toggles even though the PLL is not active. This is
> required so that LP1 power mode (system suspend) operates correctly.
> 
> The bypass configuration must then be undone when resuming from LP1, so
> that all peripheral clocks run at the expected rate. Without this, many
> peripherals won't work correctly; for example, the UART baud rate would
> be incorrect.
> 
> NVIDIA's downstream kernel code only does this if not compiled for
> Tegra30, so the added code is made conditional upon the chip ID. NVIDIA's
> downstream code makes this change conditional upon the active CPU
> cluster. The upstream kernel currently doesn't support cluster switching,
> so this patch doesn't test the active CPU cluster ID.
> 
> [1] 3cc7942a4ae5 ARM: tegra: implement RAM repair
> 
> Reported-by: Jonathan Hunter <jonathanh@nvidia.com>
> Cc: stable@vger.kernel.org
> Signed-off-by: Stephen Warren <swarren@nvidia.com>
> ---
> v3: No change.
> v2: No change.
> ---
>  arch/arm/mach-tegra/sleep-tegra30.S | 11 +++++++++++
>  1 file changed, 11 insertions(+)

Patches 2-4 applied to for-5.6/arm/core, thanks.

Thierry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]