Skip to contentSkip to conversation

ARM: stm32: Initialize TAMP_SMCR BKP..PROT fields on STM32MP15xx

PATCH Superseded Marek Vasut 2 participants 3 messages Apr 14, 2024 – Apr 15, 2024

Subsystems (3)

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ARM: stm32: Initialize TAMP_SMCR BKP..PROT fields on STM32MP15xx

In case of an OTP-CLOSED STM32MP15xx system, the CPU core 1 cannot be released from endless loop in BootROM only by populating TAMP BKPxR 4 and 5 with magic and branch address and sending SGI0 interrupt from core 0 to core 1 twice. TAMP_SMCR BKP..PROT fields must be initialized as well to release the core 1 from endless loop during the second SGI0 handling on core 1. Initialize TAMP_SMCR to protect the first 16 backup registers, the ones which contain the core 1 magic, branch address and boot information.

This requirement seems to be undocumented, therefore it was necessary to trace and analyze the STM32MP15xx BootROM using OpenOCD and objdump. Ultimately, it turns out that a certain BootROM function reads out the TAMP_SMCR register and tests whether the BKP..PROT fields are non-zero. If they are zero, the BootROM code again waits for SGI0 using WFI, else the execution moves forward until it reaches handoff to the TAMP BKPxR 5 branch address.

This fixes CPU core 1 release using U-Boot PSCI implementation on an OTP-CLOSED system, i.e. system with fuse 0 bit 6 set.

---
Cc: Igor Opaniuk <[email protected]>
Cc: Patrice Chotard <[email protected]>
Cc: Patrick Delaunay <[email protected]>
Cc: Simon Glass <[email protected]
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +16 -0
index dd99150fbc2..138a6d6b614 100644
@@ -14,6 +14,7 @@
 #include <asm/arch/sys_proto.h>
 #include <dm/device.h>
 #include <dm/uclass.h>
+#include <linux/bitfield.h>
 
 /* RCC register */
 #define RCC_TZCR		(STM32_RCC_BASE + 0x00)
@@ -41,6 +42,9 @@
 #define TZC_REGION_ID_ACCESS0	(STM32_TZC_BASE + 0x114)
 
 #define TAMP_CR1		(STM32_TAMP_BASE + 0x00)
+#define TAMP_SMCR		(STM32_TAMP_BASE + 0x20)
+#define TAMP_SMCR_BKPRWDPROT	GENMASK(7, 0)
+#define TAMP_SMCR_BKPWDPROT	GENMASK(23, 16)
 
 #define PWR_CR1			(STM32_PWR_BASE + 0x00)
 #define PWR_MCUCR		(STM32_PWR_BASE + 0x14)
@@ -136,6 +140,18 @@ static void security_init(void)
 	 */
 	writel(0x0, TAMP_CR1);
 
+	/*
+	 * TAMP: Configure non-zero secure protection settings. This is
+	 * checked by BootROM function 35ac on OTP-CLOSED device during
+	 * CPU core 1 release from endless loop. If secure protection
+	 * fields are zero, the core 1 is not released from endless
+	 * loop on second SGI0.
+	 */
+	clrsetbits_le32(TAMP_SMCR,
+			TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPRWDPROT,
+			FIELD_PREP(TAMP_SMCR_BKPRWDPROT, 0x10) |
+			FIELD_PREP(TAMP_SMCR_BKPWDPROT, 0x10));
+
 	/* GPIOZ: deactivate the security */
 	writel(BIT(0), RCC_MP_AHB5ENSETR);
 	writel(0x0, GPIOZ_SECCFGR);
Signature · 1 line
2.43.0
#1 lore raw
Quoted 59 lines · Marek Vasut

In case of an OTP-CLOSED STM32MP15xx system, the CPU core 1 cannot be released from endless loop in BootROM only by populating TAMP BKPxR 4 and 5 with magic and branch address and sending SGI0 interrupt from core 0 to core 1 twice. TAMP_SMCR BKP..PROT fields must be initialized as well to release the core 1 from endless loop during the second SGI0 handling on core 1. Initialize TAMP_SMCR to protect the first 16 backup registers, the ones which contain the core 1 magic, branch address and boot information.

This requirement seems to be undocumented, therefore it was necessary to trace and analyze the STM32MP15xx BootROM using OpenOCD and objdump. Ultimately, it turns out that a certain BootROM function reads out the TAMP_SMCR register and tests whether the BKP..PROT fields are non-zero. If they are zero, the BootROM code again waits for SGI0 using WFI, else the execution moves forward until it reaches handoff to the TAMP BKPxR 5 branch address.

This fixes CPU core 1 release using U-Boot PSCI implementation on an OTP-CLOSED system, i.e. system with fuse 0 bit 6 set.

---
Cc: Igor Opaniuk <[email protected]>
Cc: Patrice Chotard <[email protected]>
Cc: Patrick Delaunay <[email protected]>
Cc: Simon Glass <[email protected]
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
 arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +4 -0
index dd99150fbc2..138a6d6b614 100644
@@ -14,6 +14,7 @@
 #include <asm/arch/sys_proto.h>
 #include <dm/device.h>
 #include <dm/uclass.h>
+#include <linux/bitfield.h>
 
 /* RCC register */
 #define RCC_TZCR		(STM32_RCC_BASE + 0x00)
@@ -41,6 +42,9 @@
 #define TZC_REGION_ID_ACCESS0	(STM32_TZC_BASE + 0x114)
 
 #define TAMP_CR1		(STM32_TAMP_BASE + 0x00)
+#define TAMP_SMCR		(STM32_TAMP_BASE + 0x20)
+#define TAMP_SMCR_BKPRWDPROT	GENMASK(7, 0)
+#define TAMP_SMCR_BKPWDPROT	GENMASK(23, 16)
 
 #define PWR_CR1			(STM32_PWR_BASE + 0x00)
 #define PWR_MCUCR		(STM32_PWR_BASE + 0x14)
@@ -136,6 +140,18 @@ static void security_init(void)
 	 */
 	writel(0x0, TAMP_CR1);

arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c static void security_init(void) · line 151

Quoted 10 lines
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +9 -0
 
+	/*
+	 * TAMP: Configure non-zero secure protection settings. This is
+	 * checked by BootROM function 35ac on OTP-CLOSED device during
+	 * CPU core 1 release from endless loop. If secure protection
+	 * fields are zero, the core 1 is not released from endless
+	 * loop on second SGI0.
+	 */
+	clrsetbits_le32(TAMP_SMCR,
+			TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPRWDPROT,

Hi Marek

there is a typo, you used twice TAMP_SMCR_BKPRWDPROT :

TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPRWDPROT => TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPWDPROT

                                    ^~~~~                 

Patrice

Quoted 6 lines
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +3 -0
+			FIELD_PREP(TAMP_SMCR_BKPRWDPROT, 0x10) |
+			FIELD_PREP(TAMP_SMCR_BKPWDPROT, 0x10));
+
 	/* GPIOZ: deactivate the security */
 	writel(BIT(0), RCC_MP_AHB5ENSETR);
 	writel(0x0, GPIOZ_SECCFGR);
#2 lore raw
Quoted 75 lines · Patrice CHOTARD
Quoted 59 lines · Marek Vasut

In case of an OTP-CLOSED STM32MP15xx system, the CPU core 1 cannot be released from endless loop in BootROM only by populating TAMP BKPxR 4 and 5 with magic and branch address and sending SGI0 interrupt from core 0 to core 1 twice. TAMP_SMCR BKP..PROT fields must be initialized as well to release the core 1 from endless loop during the second SGI0 handling on core 1. Initialize TAMP_SMCR to protect the first 16 backup registers, the ones which contain the core 1 magic, branch address and boot information.

This requirement seems to be undocumented, therefore it was necessary to trace and analyze the STM32MP15xx BootROM using OpenOCD and objdump. Ultimately, it turns out that a certain BootROM function reads out the TAMP_SMCR register and tests whether the BKP..PROT fields are non-zero. If they are zero, the BootROM code again waits for SGI0 using WFI, else the execution moves forward until it reaches handoff to the TAMP BKPxR 5 branch address.

This fixes CPU core 1 release using U-Boot PSCI implementation on an OTP-CLOSED system, i.e. system with fuse 0 bit 6 set.

---
Cc: Igor Opaniuk <[email protected]>
Cc: Patrice Chotard <[email protected]>
Cc: Patrick Delaunay <[email protected]>
Cc: Simon Glass <[email protected]
Cc: Simon Glass <[email protected]>
Cc: Tom Rini <[email protected]>
Cc: [email protected]
Cc: [email protected]

---
  arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c | 16 ++++++++++++++++
  1 file changed, 16 insertions(+)
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +4 -0
index dd99150fbc2..138a6d6b614 100644
@@ -14,6 +14,7 @@
  #include <asm/arch/sys_proto.h>
  #include <dm/device.h>
  #include <dm/uclass.h>
+#include <linux/bitfield.h>
 
  /* RCC register */
  #define RCC_TZCR		(STM32_RCC_BASE + 0x00)
@@ -41,6 +42,9 @@
  #define TZC_REGION_ID_ACCESS0	(STM32_TZC_BASE + 0x114)
 
  #define TAMP_CR1		(STM32_TAMP_BASE + 0x00)
+#define TAMP_SMCR		(STM32_TAMP_BASE + 0x20)
+#define TAMP_SMCR_BKPRWDPROT	GENMASK(7, 0)
+#define TAMP_SMCR_BKPWDPROT	GENMASK(23, 16)
 
  #define PWR_CR1			(STM32_PWR_BASE + 0x00)
  #define PWR_MCUCR		(STM32_PWR_BASE + 0x14)
@@ -136,6 +140,18 @@ static void security_init(void)
  	 */
  	writel(0x0, TAMP_CR1);

arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c static void security_init(void) · line 151

Quoted 10 lines
arch/arm/mach-stm32mp/stm32mp1/stm32mp15x.c +9 -0
 
+	/*
+	 * TAMP: Configure non-zero secure protection settings. This is
+	 * checked by BootROM function 35ac on OTP-CLOSED device during
+	 * CPU core 1 release from endless loop. If secure protection
+	 * fields are zero, the core 1 is not released from endless
+	 * loop on second SGI0.
+	 */
+	clrsetbits_le32(TAMP_SMCR,
+			TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPRWDPROT,

Hi Marek

Quoted 4 lines

there is a typo, you used twice TAMP_SMCR_BKPRWDPROT :

TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPRWDPROT => TAMP_SMCR_BKPRWDPROT | TAMP_SMCR_BKPWDPROT

                                     ^~~~~

Fixed in V2, thanks.

btw are there any other such undocumented surprises in the BootROM ?