Hi Vignesh,
Am 04.03.24 um 06:06 schrieb Vignesh Raghavendra:
Hi Wadim,
On 26/02/24 19:00, Wadim Egorov wrote:
Texas Instruments has begun enabling security settings on the SoCs it produces to instruct ROM and TIFS to begin protecting the Security Management Subsystem (SMS) from other binaries we load into the chip by default.
One way ROM and TIFS do this is by enabling firewalls to protect the OCSRAM and HSM RAM regions they're using during bootup.
The HSM RAM the wakeup SPL is in is firewalled by TIFS to protect itself from the main domain applications. This means the 'bootindex' value in HSM RAM, left by ROM to indicate if we're using the primary or secondary boot-method, must be moved to OCSRAM (that TIFS has open for us) before we make the jump to the main domain so the main domain's bootloaders can keep access to this information.
Based on commit b672e8581070 ("arm: mach-k3: copy bootindex to OCRAM for main domain SPL")
FYI, this is mostly a problem in non SPL flow (TI prosperity SBL for example) where HSM RAM would be used by HSM firmware. This should be a issue in R5 SPL flow. Do you see any issues today? If so, whats the TIFS firmware being used?
I remember I was losing the bootindex using ti/downstream u-boot. But can't figure out the exact version anymore. Just did a bit of testing and I can not see the Issue with the current u-boot. Boot index in 0x43c3f290 stays intact.
Would it be okay to drop this patch and keep only the 2nd patch that factors out into get_boot_device()?
Regards, Wadim
Signed-off-by: Wadim Egorov w.egorov@phytec.de
arch/arm/mach-k3/Kconfig | 3 ++- arch/arm/mach-k3/am625_init.c | 15 +++++++++++++-- arch/arm/mach-k3/include/mach/am62_hardware.h | 15 +++++++++++++++ 3 files changed, 30 insertions(+), 3 deletions(-)
diff --git a/arch/arm/mach-k3/Kconfig b/arch/arm/mach-k3/Kconfig index 03898424c9..f5d06593f7 100644 --- a/arch/arm/mach-k3/Kconfig +++ b/arch/arm/mach-k3/Kconfig @@ -75,7 +75,8 @@ config SYS_K3_BOOT_PARAM_TABLE_INDEX default 0x41cffbfc if SOC_K3_J721E default 0x41cfdbfc if SOC_K3_J721S2 default 0x701bebfc if SOC_K3_AM642
- default 0x43c3f290 if SOC_K3_AM625
- default 0x43c3f290 if SOC_K3_AM625 && CPU_V7R
- default 0x7000f290 if SOC_K3_AM625 && ARM64 default 0x43c3f290 if SOC_K3_AM62A7 && CPU_V7R default 0x7000f290 if SOC_K3_AM62A7 && ARM64 help
diff --git a/arch/arm/mach-k3/am625_init.c b/arch/arm/mach-k3/am625_init.c index 6c96e88114..67cf63b103 100644 --- a/arch/arm/mach-k3/am625_init.c +++ b/arch/arm/mach-k3/am625_init.c @@ -35,8 +35,10 @@ static struct rom_extended_boot_data bootdata __section(".data"); static void store_boot_info_from_rom(void) { bootindex = *(u32 *)(CONFIG_SYS_K3_BOOT_PARAM_TABLE_INDEX);
- memcpy(&bootdata, (uintptr_t *)ROM_EXTENDED_BOOT_DATA_INFO,
sizeof(struct rom_extended_boot_data));
if (IS_ENABLED(CONFIG_CPU_V7R)) {
memcpy(&bootdata, (uintptr_t *)ROM_EXTENDED_BOOT_DATA_INFO,sizeof(struct rom_extended_boot_data));} }
static void ctrl_mmr_unlock(void)
@@ -175,6 +177,15 @@ void board_init_f(ulong dummy) k3_sysfw_loader(true, NULL, NULL); }
+#if defined(CONFIG_CPU_V7R)
- /*
* Relocate boot information to OCRAM (after TIFS has opend this* region for us) so the next bootloader stages can keep access to* primary vs backup bootmodes.*/- writel(bootindex, K3_BOOT_PARAM_TABLE_INDEX_OCRAM);
+#endif
- /*
- Force probe of clk_k3 driver here to ensure basic default clock
- configuration is always done.
diff --git a/arch/arm/mach-k3/include/mach/am62_hardware.h b/arch/arm/mach-k3/include/mach/am62_hardware.h index 54380f36e1..9f504f4642 100644 --- a/arch/arm/mach-k3/include/mach/am62_hardware.h +++ b/arch/arm/mach-k3/include/mach/am62_hardware.h @@ -76,8 +76,23 @@ #define CTRLMMR_MCU_RST_CTRL (MCU_CTRL_MMR0_BASE + 0x18170)
#define ROM_EXTENDED_BOOT_DATA_INFO 0x43c3f1e0 +#define K3_BOOT_PARAM_TABLE_INDEX_OCRAM 0x7000F290
+/*
- During the boot process ROM will kill anything that writes to OCSRAM.
R5 ROM is long gone when R5 SPL starts, how would it kill anything?
- This means the wakeup SPL cannot use this region during boot. To
- complicate things, TIFS will set a firewall between HSM RAM and the
- main domain.
- So, during the wakeup SPL, we will need to store the EEPROM data
- somewhere in HSM RAM, and the main domain's SPL will need to store it
- somewhere in OCSRAM
- */
+#ifdef CONFIG_CPU_V7R #define TI_SRAM_SCRATCH_BOARD_EEPROM_START 0x43c30000 +#else
- #define TI_SRAM_SCRATCH_BOARD_EEPROM_START 0x70000001
Why not 0x70000000 ?
+#endif /* CONFIG_CPU_V7R */
Can't we store directly in OCRAM in both stages? This RAM should be accessible post TIFS is up (ie post k3_sysfw_loader() call)
static inline int k3_get_core_nr(void) {