drivers: flash: Introduce BFLB flash driver. #96680

VynDragon · 2025-09-28T15:32:53Z

Introduce flash driver for 1 bank of flash (all of them on bl60x and bl70x, half of them on bl61x)

Caveat: no bank2 support (feature of bl61x)
Not sure how to handle the flash naming and devices as they are either separate die flash or external flash but need the controller to work (and so be soc-nv-flash).

github-actions · 2025-09-28T15:34:00Z

The following west manifest projects have changed revision in this Pull Request:

Name	Old Revision	New Revision	Diff

✅ All manifest checks OK

Note: This message is automatically posted and updated by the Manifest GitHub Action.

VynDragon · 2025-10-24T10:51:43Z

@de-nordic PTAL? Thanks.

de-nordic · 2025-10-24T11:28:35Z

drivers/flash/flash_bflb.c

+	if (length == 0) {
+		return 0;
+	}


In all cases, lack of action should not generate an error. This means the length check should go above range check.

de-nordic · 2025-10-24T11:33:03Z

drivers/flash/flash_bflb.c

+	if (offset < 0) {
+		LOG_WRN("0x%lx: before start of flash", (long)offset);
+		return -EINVAL;
+	}
+	if ((offset + len) > TOTAL_SIZE) {
+		LOG_WRN("0x%lx: ends past the end of flash", (long)offset);
+		return -EINVAL;
+	}


No need to separate checks, imho.
The check should be:

offset < 0 || len > TOTAL_SIZE || ((TOTAL_SIZE - offset) < len)

above avoids integer overflows.

Separated errors for different problems however, are useful. Done

de-nordic · 2025-10-24T12:20:48Z

drivers/flash/flash_bflb.c

+		return ret;
+	}
+
+	if (len == 0) {


Len check above range valid check.

Sorry, counted 3, forgetting there's 2 different read... Done

VynDragon · 2025-11-14T12:44:24Z

@de-nordic PTAL

de-nordic · 2025-11-14T12:55:04Z

drivers/flash/flash_bflb.c

+static bool flash_bflb_is_in_xip(void *func)
+{
+	if ((uint32_t)func > BFLB_XIP_BASE && (uint32_t)func < BFLB_XIP_END) {
+		LOG_ERR("function at %d is in XIP and will crash the device", (uint32_t)func);


You probably want to use %p instead of %d and avoid casting the func.

de-nordic · 2025-11-14T12:56:58Z

drivers/flash/flash_bflb.c

+
+#if defined(CONFIG_SOC_SERIES_BL70X) || defined(CONFIG_SOC_SERIES_BL60X)
+
+static void flash_bflb_l1c_wrap(bool yes)


Not my driver, but yes for a bool variable is not descriptive of the purpose.

it's the same as 'writing?' yes/no, l1c_wrap? yes/no

de-nordic · 2025-11-14T12:58:29Z

drivers/flash/flash_bflb.c

+
+static void flash_bflb_l1c_wrap(bool yes)
+{
+	/* Do nothing on Bl61x: no L1C */


You will probably get here unused variable warning from static analysis, so it would be good to have ARG_UNUSED(yes) here

de-nordic · 2025-11-14T13:01:48Z

drivers/flash/flash_bflb.c

+		if (data_mode == 0) {
+			return 0;
+		} else if (data_mode == 1) {
+			return 1;
+		} else if (data_mode == 2) {
+			return 2;
+		}


You can just return data_mode here.
Shouldn't be here some asserts, for validation builds, to check if only values covered by ifs are processed?

de-nordic · 2025-11-14T13:02:53Z

drivers/flash/flash_bflb.c

+		tmp &= ~SF_CTRL_SF_IF_0_DAT_EN_MSK;
+	}
+
+	/* writing ? */


I do not know. Are we?

de-nordic · 2025-11-14T13:18:04Z

drivers/flash/flash_bflb.c

+	erase_cmd.rw = 0;
+	erase_cmd.addr_size = 3;
+
+	for (uint32_t i = start / ERASE_SIZE; i < ((len / ERASE_SIZE) + start / ERASE_SIZE); i++) {


On some CPUs it is better avoid division when possible, and this loop will have at least two of these every run, if fail to optimize. Consider pre calculating loop-independent ending condition before defining loop.

This only runs on CPUs that are fine with that and all have FPUs as well, will add the precalculations however.

de-nordic · 2025-11-14T13:18:34Z

drivers/flash/flash_bflb.c

+	ret = flash_bflb_restore_xip_state(data);
+	irq_unlock(locker);
+
+	return ret;
+
+exit_here:
+	flash_bflb_restore_xip_state(data);
+	irq_unlock(locker);


Same as flash_bflb_write

de-nordic · 2025-11-14T13:19:22Z

drivers/flash/flash_bflb.c

+/* from SDK because the zephyr CRC are no good for the bflb flash crc, this is supposedly a
+ * implementation of ZIP crc32
+ */


No good in what respect?

It is a different algorithm / gives different results

de-nordic · 2025-11-14T13:19:51Z

drivers/flash/flash_bflb.c

+	/* copy cfg reg to memory as cfg will not be in it and inaccessible */
+	data->reg_copy = cfg->reg;
+
+	/* get flash config using xip access */


? Some leftover?

No, it is designating all the code under it until the unlock

de-nordic · 2025-11-14T13:21:21Z

drivers/flash/flash_bflb.c

+		irq_unlock(locker);
+		return ret;
+	}
+	/* TODO: AES flash support goes here */


Ever gonna do that TODO thing?

It's for a future feature update

Introduce Bouffalolab Flash Controller with support for bank 1 Signed-off-by: Camille BAUD <[email protected]>

Enable the Flash Controller Signed-off-by: Camille BAUD <[email protected]>

Enable the flash controller. Relocation is required. Signed-off-by: Camille BAUD <[email protected]>

Adds the flash to supported for testing Signed-off-by: Camille BAUD <[email protected]>

VynDragon · 2025-11-14T16:38:52Z

GREAT again hitting some global changes that aren't related to my PR!

sonarqubecloud · 2025-11-14T16:52:26Z

Please retry analysis of this Pull-Request directly on SonarQube Cloud

VynDragon requested a review from josuah September 28, 2025 15:33

github-actions bot added manifest manifest-hal_bouffalolab DNM (manifest) This PR should not be merged (controlled by action-manifest) labels Sep 28, 2025

VynDragon added area: Flash platform: Bouffalo Lab and removed manifest manifest-hal_bouffalolab DNM (manifest) This PR should not be merged (controlled by action-manifest) labels Sep 28, 2025

VynDragon force-pushed the bflb_flash_work branch from cd0455e to e821c9d Compare September 28, 2025 15:37

github-actions bot added manifest manifest-hal_bouffalolab DNM (manifest) This PR should not be merged (controlled by action-manifest) labels Sep 28, 2025

VynDragon force-pushed the bflb_flash_work branch 3 times, most recently from 5db0053 to 2a4c5d9 Compare October 10, 2025 18:46

VynDragon marked this pull request as ready for review October 10, 2025 18:50

zephyrbot added area: Boards/SoCs area: RISCV RISCV Architecture (32-bit & 64-bit) labels Oct 10, 2025

zephyrbot requested review from de-nordic, edersondisouza, fkokosinski, katsuster, kgugala, masz-nordic, mgielda, npitre, rghaddab and tgorochowik October 10, 2025 18:52

VynDragon requested a review from josuah October 24, 2025 10:46

github-actions bot removed manifest manifest-hal_bouffalolab DNM (manifest) This PR should not be merged (controlled by action-manifest) labels Oct 24, 2025

josuah previously approved these changes Oct 24, 2025

View reviewed changes

de-nordic reviewed Oct 24, 2025

View reviewed changes

VynDragon dismissed josuah’s stale review via 5e98d2d October 24, 2025 11:58

VynDragon force-pushed the bflb_flash_work branch from 51322a4 to 5e98d2d Compare October 24, 2025 11:58

VynDragon requested review from de-nordic and josuah October 24, 2025 11:58

de-nordic reviewed Oct 24, 2025

View reviewed changes

VynDragon force-pushed the bflb_flash_work branch from 5e98d2d to 0a04e32 Compare October 24, 2025 12:25

VynDragon requested a review from de-nordic October 24, 2025 12:48

josuah previously approved these changes Oct 24, 2025

View reviewed changes

VynDragon modified the milestones: v4.3.0, v4.4.0 Nov 5, 2025

de-nordic reviewed Nov 14, 2025

View reviewed changes

VynDragon dismissed josuah’s stale review via 2cf8917 November 14, 2025 16:24

VynDragon force-pushed the bflb_flash_work branch from 0a04e32 to 2cf8917 Compare November 14, 2025 16:24

VynDragon requested a review from de-nordic November 14, 2025 16:27

VynDragon added 4 commits November 14, 2025 17:32

drivers: flash: Introduce BFLB flash controller

2aecf84

Introduce Bouffalolab Flash Controller with support for bank 1 Signed-off-by: Camille BAUD <[email protected]>

dts: bflb: Enable Flash Controller

23f510d

Enable the Flash Controller Signed-off-by: Camille BAUD <[email protected]>

soc: bflb: Enable Flash Controller

1cb3a8a

Enable the flash controller. Relocation is required. Signed-off-by: Camille BAUD <[email protected]>

boards: Add flash to supported for bflb boards

562f09f

Adds the flash to supported for testing Signed-off-by: Camille BAUD <[email protected]>

VynDragon force-pushed the bflb_flash_work branch from 2cf8917 to 562f09f Compare November 14, 2025 16:32


		#if defined(CONFIG_SOC_SERIES_BL70X) \|\| defined(CONFIG_SOC_SERIES_BL60X)

		static void flash_bflb_l1c_wrap(bool yes)

drivers: flash: Introduce BFLB flash driver. #96680

Are you sure you want to change the base?

drivers: flash: Introduce BFLB flash driver. #96680

Conversation

VynDragon commented Sep 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

github-actions bot commented Sep 28, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

VynDragon commented Oct 24, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

VynDragon commented Nov 14, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

VynDragon commented Nov 14, 2025

Uh oh!

sonarqubecloud bot commented Nov 14, 2025

Uh oh!

Reviewers

Assignees

VynDragon commented Sep 28, 2025 •

edited

Loading

github-actions bot commented Sep 28, 2025 •

edited

Loading