stm32h5 run application in external flash memory XIP by FRASTM · Pull Request #88579 · zephyrproject-rtos/zephyr

FRASTM · 2025-04-14T09:51:35Z

the stm32h5 disco kit board can run an application in external flash

Requires the #88646 which removes the warning in the DTC

With this PR it enable the memorymapped mode on the external flash memory so that XIP is possible on the external NOR octoflash

Two samples are running in XIP on the external flash

samples/application_development/code_relocation_nocopy/
samples/boards/st/hello_world_xip/ built with mcu boot mcu-tools/mcuboot#2323

Fixes #87531

This PR completes the #88051 for the stm32h5 serie

FRASTM · 2025-04-14T09:53:19Z

west build -b stm32h573i_dk samples/application_development/code_relocation_nocopy/
west flash : stm32cubeprogrammer runner is taking the external loader to download into the external NOR (0x90000000)

*** Booting Zephyr OS build v4.1.0-2350-g86df576c4134 ***
Address of main function 0x80005e1
Address of function_in_ext_flash 0x90000001
Address of var_ext_sram_data 0x200000a0 (10)
Address of function_in_sram 0x20000001
Address of var_sram_data 0x200000a4 (10)
Hello World! stm32h573i_dk

FRASTM · 2025-04-14T09:57:38Z

with a special sample dedicated to st boards for XiP
west build -p -b stm32h573i_dk samples/boards/st/hello_world_xip/ --sysbuild -- -DSB_CONFIG_BOOTLOADER_MCUBOOT=y

requires mcu-tools/mcuboot#2154
download with stm32cubeprogrammer GUI

the build/mcuboot/zephyr/zephyr.bin at internal flash address 0x8000000
the build/hello_world_xip/zephyr/zephyr.signed.bin at external flash address 0x90000000

*** Booting MCUboot v2.1.0-rc1-276-g540654e87167 ***
*** Using Zephyr OS build v4.1.0-2352-gb567cdd8cbfb ***
I: Starting bootloader
I: Primary image: magic=unset, swap_type=0x1, copy_done=0x3, image_ok=0x3
I: Secondary image: magic=unset, swap_type=0x1, copy_done=0x3, image_ok=0x3
I: Boot source: none
I: Image index: 0, Swap type: none
I: Bootloader chainload address offset: 0x0
I: Image version: v0.0.0
I: Jumping to the first image slot
*** Booting Zephyr OS build v4.1.0-2352-gb567cdd8cbfb ***                       
Hello World! from external flash  stm32h573i_dk                                 
--> PC at 0x9000092a

FRASTM · 2025-04-14T12:33:52Z

Define larger partition in the external NOR flash (consequently adjusting the SECTOR size for the mcu boot)

FRASTM · 2025-04-14T15:28:31Z

CI failure "invalid length for memory region EXTFLASH" given by the samples/application_development/code_relocation_nocopy/linker_arm_nocopy.ld

map giving : EXTFLASH 0x0000000090000000 0xffffffffffffffff xr

Build operation displays wrong extflash too:

Memory region         Used Size  Region Size  %age Used
        EXTFLASH:          56 B 18446744073709551615 B      0.00%

compared to

Memory region         Used Size  Region Size  %age Used
        EXTFLASH:          56 B        64 MB      0.00%

when #define EXTFLASH_SIZE (0x4000000)

It looks like the DT_INST_REG_SIZE does not exist for the st,stm32-xspi-nor compat (DT_INST_REG_ADDR exists)

--> see issue #88404

tpambor · 2025-04-15T10:08:04Z

See #88392 for a fix.

FRASTM · 2025-04-24T14:09:07Z

set the DNM mabel until the #88646 is merged

FRASTM · 2025-04-29T08:45:28Z

Adding the commit to retrieve the CONFIG_FLASH_BASE_ADDRESS from the DTS by the soc/st/stm32/Kconfig.defconfig

FRASTM · 2025-05-05T09:41:27Z

set the DNM mabel until the #88051 is merged

etienne-lms

For consistency, I would suggest to squash 2 commits and reorder some:

Commit "soc: stm32: symbol for application XiP from xspi external memory";
A single commit that squashes:
commit "drivers: clock: stm32H5 clock control of the PLL in MemoryMapped mode"
and commit "drivers: clock control of stm32 devices CONFIG external xspi flash";
Commit "drivers: flash: stm32 xspi flash driver skip init when executing in place"
Commit "boards: arm: stm32h573 disco kit move partition to ext flash"
Commit "samples: sysbuild: with mcu_boot on the stm32h573i_dk"
Commit "Revert "samples: mgmt: hawkbit: force BOOT_MAX_IMG_SECTORS for stm32h573i_dk""

IIUC commit "drivers: clock control for stm32h5 driver implements control_get_status" can be placed anywhere in this series.

For commit "Revert "samples: mgmt: hawkbit: force BOOT_MAX_IMG_SECTORS for stm32h573i_dk"", I think it would be nice to add a few words in the commit message to exaplin why it's reverted. By the way, S-o-b tag is missing in that commit message.

maass-hamburg · 2025-07-10T08:40:00Z

+ */
+/ {
+	chosen {
+		zephyr,flash = &mx25lm51245;


maybe it would be better to use the gparent of the chosen zephyr,code-partition if USE_DT_CODE_PARTITION is enabled in general, as having this overlay needed when using a image with mcuboot.

Changes like:

diff --git a/soc/st/stm32/Kconfig.defconfig b/soc/st/stm32/Kconfig.defconfig index aafb321151a..8544215e390 100644 --- a/soc/st/stm32/Kconfig.defconfig +++ b/soc/st/stm32/Kconfig.defconfig @@ -33,6 +33,13 @@ DT_CHOSEN_FLASH_NODE := $(dt_chosen_path,$(DT_CHOSEN_Z_FLASH)) DT_CHOSEN_FLASH_PARENT := $(dt_node_parent,$(DT_CHOSEN_FLASH_NODE)) DT_FLASH_PARENT_IS_XSPI := $(dt_node_has_compat,$(DT_CHOSEN_FLASH_PARENT),$(DT_COMPAT_XSPI)) +DT_CHOSEN_Z_CODE_PARTITION := zephyr,code-partition + +DT_CHOSEN_Z_CODE_PARTITION_NODE := $(dt_chosen_path,$(DT_CHOSEN_Z_CODE_PARTITION)) +DT_CHOSEN_Z_CODE_PARTITION_FLASH_NODE := $(dt_node_parent,$(dt_node_parent, $(DT_CHOSEN_Z_CODE_PARTITION_NODE))) +DT_CHOSEN_Z_CODE_PARTITION_FLASH_PARENT := $(dt_node_parent,$(DT_CHOSEN_Z_CODE_PARTITION_FLASH_NODE)) +DT_CHOSEN_Z_CODE_PARTITION_FLASH_PARENT_IS_XSPI := $(dt_node_has_compat,$(DT_CHOSEN_Z_CODE_PARTITION_FLASH_PARENT),$(DT_COMPAT_XSPI)) + config SYS_CLOCK_HW_CYCLES_PER_SEC default "$(DT_STM32_RCC_CLOCK_FREQ)" if "$(dt_nodelabel_enabled,rcc)" @@ -81,6 +88,10 @@ config BUILD_WITH_TFM default y if TRUSTED_EXECUTION_NONSECURE config FLASH_BASE_ADDRESS + default $(dt_node_reg_addr_hex,$(DT_CHOSEN_Z_CODE_PARTITION_FLASH_PARENT),1) \ + if $(DT_CHOSEN_Z_CODE_PARTITION_FLASH_PARENT_IS_XSPI) && USE_DT_CODE_PARTITION + default $(dt_node_reg_addr_hex,$(DT_CHOSEN_Z_CODE_PARTITION_FLASH_NODE),1) \ + if USE_DT_CODE_PARTITION default $(dt_node_reg_addr_hex,$(DT_CHOSEN_FLASH_PARENT),1) \ if $(DT_FLASH_PARENT_IS_XSPI) default $(dt_chosen_reg_addr_hex,$(DT_CHOSEN_Z_FLASH))

yes, but I am not sure this is the purpose of this PR fixing the issue for the 4.2.0 release

okay then in another PR

Add a STM32_APP_IN_EXT_FLASH to determine that an application is eXecuting in Place on an external xspi flash. That will control the clock init of this external xspi controller. Signed-off-by: Francois Ramu <francois.ramu@st.com>

FRASTM · 2025-07-10T11:48:37Z

re-ordering commits as suggested

JarmouniA

Thank you!

maass-hamburg

the overlay file is also needed for the hawkbit sample

FRASTM · 2025-07-10T13:33:08Z

the overlay file is also needed for the hawkbit sample

do you mean adding a samples/subsys/mgmt/hawkbit/boards/stm32h573i_dk.overlay

/ {
	chosen {
		zephyr,flash = &mx25lm51245;
		zephyr,flash-controller = &mx25lm51245;
	};
};

maass-hamburg · 2025-07-10T14:16:21Z

the overlay file is also needed for the hawkbit sample

do you mean adding a samples/subsys/mgmt/hawkbit/boards/stm32h573i_dk.overlay
/ {
	chosen {
		zephyr,flash = &mx25lm51245;
		zephyr,flash-controller = &mx25lm51245;
	};
};

yes

FRASTM · 2025-07-10T14:19:56Z

with the samples/subsys/mgmt/hawkbit/boards/stm32h573i_dk.overlay included

FRASTM · 2025-07-10T14:22:43Z

the overlay file is also needed for the hawkbit sample

do you mean adding a samples/subsys/mgmt/hawkbit/boards/stm32h573i_dk.overlay
/ {
	chosen {
		zephyr,flash = &mx25lm51245;
		zephyr,flash-controller = &mx25lm51245;
	};
};
yes

done

etienne-lms

Non-blocking comments:

Nitpicking in commit message for commit "drivers: clock: stm32H5 clock control of the PLL in MemoryMapped mode".

-Do not disable the pll clock if it source the xspi and if the external
+Do not disable the PLL clock if it sources the XSIP and if the external
 flash is executing in Place. After mcuboot reset, the code is executed
 on the external flash, through the xspi.
-The CONFIG_STM32_APP_IN_EXT_FLASH tis set and will avoid re-config
+The CONFIG_STM32_APP_IN_EXT_FLASH is set and will avoid re-config
-of the pll which is sourcing the xspi peripheral. When excuting in Place
+of the PLL which is sourcing the XSPI peripheral. When executing in Place
-on this external NOR It must not disable its own clock source (PLL).
+on this external NOR, it must not disable its own clock source (PLL).

Or simplify:

Do not disable the PLL clock if it sources the XSIP and if the external
flash is executing in Place. CONFIG_STM32_APP_IN_EXT_FLASH 
provides this information.

In commit "drivers: flash: stm32 xspi flash driver skip init when executing in place",
the below sentence in the commit message is no more true since the previous commit implements that API function:

... (stm32h5 has no clock_control_get_status API)

For commit "Revert "samples: mgmt: hawkbit: force BOOT_MAX_IMG_SECTORS for stm32h573i_dk""
This commit does more than reverting a commit. IMHO the commit message header line should not be "Revert "samples: mgmt: ...".
Nitpicking on some commit message body: some terminal dots . are missing at end of sentences, see commits
"drivers: flash: stm32 xspi flash driver skip init when executing in place",
"boards: arm: stm32h573 disco kit move partition to ext flash" and
"samples: sysbuild: with mcu_boot on the stm32h573i_dk".

Do not disable the PLL clock if it sources the XSPI and if the external flash is executing in Place. After mcuboot reset, the code is executed on the external flash, through the xspi. The CONFIG_STM32_APP_IN_EXT_FLASH is set and will avoid re-config of the PLL which is sourcing the XSPI peripheral. When eXecuting in Place on this external NOR, it must not disable its own clock source (PLL). Signed-off-by: Francois Ramu <francois.ramu@st.com>

Add the control_get_status API function to the stm32h5 clock_control driver Signed-off-by: Francois Ramu <francois.ramu@st.com>

…lace The flash_stm32_xspi driver should not initialize the xspi, if this one is being use to execute in Place : the init is skipped. This mode is identified with the CONFIG_STM32_APP_IN_EXT_FLASH. Checking the memory mapped mode bit is possible when the xspi peripheral clock is not off (stm32h5 has no clock_control_get_status API) Signed-off-by: Francois Ramu <francois.ramu@st.com>

Add the XSPI 2 which is an octoSPI connection to a octo NOR flash mx25lm51245 on the stm32h573i_dk disco kit. Define the partition on the 64MBytes space of the external flash. Use the STM32Cube programmer to flash with the corresponding external loader for XiP mode. Removing usb_device will avoid test feature for that board. Signed-off-by: Francois Ramu <francois.ramu@st.com>

Run the sample to execute in place on the external flash of the stm32h573 disco kit. with_boot is build and linked for the external flash. Download the mcuboot zephyr.bin in internal memory. Download the with_mcuboot zephyr.signed.bin in external memory. Signed-off-by: Francois Ramu <francois.ramu@st.com>

Reverts commit 5700ffe. Not required anymore as the CONFIG_BOOT_MAX_IMG_SECTORS is fixed by the mcuboot at 4096. Add the corresponding overlay file. Signed-off-by: Francois Ramu <francois.ramu@st.com>

FRASTM · 2025-07-15T07:48:46Z

changes in some commit msg as recommended by @etienne-lms

sonarqubecloud · 2025-07-15T08:08:21Z

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
34.2% Duplication on New Code

See analysis details on SonarQube Cloud

erwango · 2025-07-16T08:03:15Z

+	bool
+	help
+	  Allows the SoC clock driver to correctly initialize the
+	  Q/O/XSPI controller clocks when the application is residing


So far, it only deals with XSPI. Q/O SPI will have to be updated to take it into account.

FRASTM force-pushed the stm32h5_xip branch 3 times, most recently from 0eed22a to 0a25f40 Compare April 14, 2025 12:32

FRASTM force-pushed the stm32h5_xip branch 2 times, most recently from c3998c8 to d0126b3 Compare April 14, 2025 13:47

FRASTM force-pushed the stm32h5_xip branch 3 times, most recently from 9bb0775 to 9d2579a Compare April 16, 2025 11:55

FRASTM force-pushed the stm32h5_xip branch 3 times, most recently from 16878f8 to 048cb76 Compare April 23, 2025 13:56

FRASTM added the DNM This PR should not be merged (Do Not Merge) label Apr 24, 2025

FRASTM force-pushed the stm32h5_xip branch 2 times, most recently from 5d3b191 to b496f43 Compare April 28, 2025 16:04

FRASTM mentioned this pull request Apr 29, 2025

mcuboot supports the stm32 external flash in XiP mode mcu-tools/mcuboot#2154

Closed

FRASTM mentioned this pull request Apr 29, 2025

stm32h7rs run application in external flash memory XIP #88052

Merged

FRASTM force-pushed the stm32h5_xip branch 2 times, most recently from 0103ad5 to 4b2bea0 Compare May 5, 2025 09:39

FRASTM added the platform: STM32 ST Micro STM32 label May 5, 2025

FRASTM force-pushed the stm32h5_xip branch 3 times, most recently from 3a614c3 to 80ed1b1 Compare May 5, 2025 14:21

etienne-lms reviewed Jul 10, 2025

View reviewed changes

Comment thread soc/st/stm32/Kconfig.defconfig

Comment thread drivers/flash/flash_stm32_xspi.c Outdated

maass-hamburg reviewed Jul 10, 2025

View reviewed changes

JarmouniA previously approved these changes Jul 10, 2025

View reviewed changes

maass-hamburg requested changes Jul 10, 2025

View reviewed changes

JarmouniA approved these changes Jul 10, 2025

View reviewed changes

maass-hamburg approved these changes Jul 10, 2025

View reviewed changes

etienne-lms reviewed Jul 10, 2025

View reviewed changes

Comment thread soc/st/stm32/Kconfig

Comment thread soc/st/stm32/Kconfig

Comment thread drivers/clock_control/clock_stm32_ll_h5.c

FRASTM mentioned this pull request Jul 11, 2025

STM32 OSPI NOR: run application from external flash memory #93032

Closed

FRASTM added 6 commits July 15, 2025 09:45

drivers: clock control for stm32h5 driver implements control_get_status

d6c4d25

Add the control_get_status API function to the stm32h5 clock_control driver Signed-off-by: Francois Ramu <francois.ramu@st.com>

samples: mgmt: hawkbit: no BOOT_MAX_IMG_SECTORS for stm32h573i_dk

edce0de

Reverts commit 5700ffe. Not required anymore as the CONFIG_BOOT_MAX_IMG_SECTORS is fixed by the mcuboot at 4096. Add the corresponding overlay file. Signed-off-by: Francois Ramu <francois.ramu@st.com>

etienne-lms approved these changes Jul 15, 2025

View reviewed changes

JarmouniA approved these changes Jul 15, 2025

View reviewed changes

gautierg-st approved these changes Jul 16, 2025

View reviewed changes

erwango reviewed Jul 16, 2025

View reviewed changes

erwango approved these changes Jul 16, 2025

View reviewed changes

This was referenced Oct 12, 2025

Flash partitions layout changed for stm32h573i_dk #97423

Closed

boards: st: stm32h573i_dk: Add board variant for app in external flash #97678

Merged

erwango mentioned this pull request Oct 22, 2025

STM32: H7RSx skip clock and xspi config when using XiP #97907

Closed

Conversation

FRASTM commented Apr 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

FRASTM commented Apr 14, 2025

Uh oh!

FRASTM commented Apr 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

FRASTM commented Apr 14, 2025

Uh oh!

FRASTM commented Apr 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

tpambor commented Apr 15, 2025

Uh oh!

FRASTM commented Apr 24, 2025

Uh oh!

FRASTM commented Apr 29, 2025

Uh oh!

FRASTM commented May 5, 2025

Uh oh!

etienne-lms left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

maass-hamburg Jul 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

FRASTM Jul 10, 2025

Choose a reason for hiding this comment

Uh oh!

maass-hamburg Jul 10, 2025

Choose a reason for hiding this comment

Uh oh!

FRASTM commented Jul 10, 2025

Uh oh!

JarmouniA left a comment

Choose a reason for hiding this comment

Uh oh!

maass-hamburg left a comment

Choose a reason for hiding this comment

Uh oh!

FRASTM commented Jul 10, 2025

Uh oh!

maass-hamburg commented Jul 10, 2025

Uh oh!

FRASTM commented Jul 10, 2025

Uh oh!

FRASTM commented Jul 10, 2025

Uh oh!

etienne-lms left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

FRASTM commented Jul 15, 2025

Uh oh!

sonarqubecloud Bot commented Jul 15, 2025

Quality Gate passed

Uh oh!

erwango Jul 16, 2025

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

12 participants

FRASTM commented Apr 14, 2025 •

edited

Loading

FRASTM commented Apr 14, 2025 •

edited

Loading

FRASTM commented Apr 14, 2025 •

edited

Loading

maass-hamburg Jul 10, 2025 •

edited

Loading

etienne-lms left a comment •

edited

Loading