Skip to content

imx95-foundries-emmc.md

Latest commit

 

History

History
410 lines (300 loc) · 22.2 KB

imx95-foundries-emmc.md

File metadata and controls

410 lines (300 loc) · 22.2 KB

FRDM-IMX95 Foundries LmP — eMMC programming and boot

Board: FRDM-IMX95 15×15 LPDDR4x (imx95-frdm-evk in Foundries; NXP machine imx95-15x15-lpddr4x-frdm)
Factory: dynamic-devices, target 2707 (example)

Reference: Getting Started with FRDM-IMX95 (GS-FRDM-IMX95)

Hardware setup

Item Setting
Board FRDM-IMX95 15×15 LPDDR4x, 32 GB eMMC on USDHC1 (mmc0)
USB programming J3 (not J1 debug) → host
Serial (SPL / early boot) ttyACM0 @ 921600 — full SPL output during uuu
Serial (Linux runtime) ttyLP0 @ 115200 (Foundries LmP)
SW1 Serial Download (UUU) BOOT_MODE1=0, BOOT_MODE0=1
SW1 eMMC boot (after flash) BOOT_MODE1=1, BOOT_MODE0=0
SW1 SD boot BOOT_MODE1=1, BOOT_MODE0=1

Stop ModemManager before uuu (script does this automatically). Install passwordless sudo once: ./scripts/install-flash-sudoers.sh.

⚠️ Known blocker: SDPV second stage fails on i.MX95 (2026-05-31)

Hardware testing (board at SW1=Serial Download, ModemManager neutralised via udev) shows the SPL second-stage SDPV download does not reach fastboot on this i.MX95. After SDPS: boot the board enumerates at 1fc9:0151 (SPL in SDP mode) and waits; every SDPV variant tried then resets the board back to 1fc9:015d, looping 015d → 0151 → 015d without ever reaching fastboot (1fc9:0152). All three second-stage forms fail:

SDPV second stage tried Result
SDPV: write -f imx-boot-mfgtool -skipspl Unknown Image type on imx95 AHAB
SDPV: write -f imx-boot-mfgtool (no -skipspl) write 100%, jump → board resets to 015d (loop)
SDPV: write -f u-boot-mfgtool.itb (FIT) write 100%, jump → board resets to 015d (loop)

This matches NXP's own guidance (see flash-imx95-nxp-reference.sh): “NO SDPV + u-boot-mfgtool.itb — that imx8-style path fails on imx95 SPL.” uuu's built-in emmc_all script (uuu -bshow emmc_all) also uses a single combined _flash.bin for both SDPS: boot and SDPV: write … -skipspl, not a separate SPL + ITB pair.

Root cause: the Foundries lmp-mfgtool artifacts (imx-boot-mfgtool SPL + u-boot-mfgtool.itb) assume the imx8-style split SDPV flow, which i.MX95 SPL does not support. i.MX95 needs the NXP-style single full-container flash_all image at SDPS that boots straight to fastboot (no SDPV).

Resolution required (BSP): produce an imx95 flash_all-equivalent combined mfgtool container (SPL+ATF+U-Boot+fastboot in one AHAB image) and flash with SDPS: boot -f <flash_all> -scanterm then FB: — no SDPV. Track in meta-dynamicdevices-bsp mfgtool-files recipe — see DynamicDevices/meta-dynamicdevices-bsp#12. Until then, use the default flash path in flash-imx95-foundries.sh (full_image-nxp-boot.uuu: NXP LF flash_all at SDPS/SDPV + Foundries LmP FB).

Programming flow: NXP LF vs Foundries LmP

i.MX95 FRDM Foundries path (intended): SDPS → SDPV → fastboot (FB) — but see blocker above.

Phase USB ID Image Action
SDPS ROM 1fc9:015d imx-boot-mfgtool SPL loaded into DRAM via HID
SDPV / SPL 1fc9:0151 u-boot-mfgtool.itb U-Boot FIT sent to SPL → fastboot
Fastboot 1fc9:0152 — eMMC flash via fastboot protocol

SDPS image (imx-boot-mfgtool): SPL-only AHAB container (flash_a55 target, ~2.3 MB). Initialises DRAM, presents as USB gadget 0151.
SDPV image (u-boot-mfgtool.itb): U-Boot FIT image (~1.4 MB). SPL loads it, starts full U-Boot in fastboot mode (0152).
Never use imx-boot-mfgtool at SDPV — SPL cannot accept an AHAB container, causes reset loop back to 0151.
Never use -skipspl — "Unknown Image type on imx95 AHAB".

Item NXP Linux LF (LF_v6.18.2-1.0.0_images_IMX95) Foundries LmP (our path)
uuu script uuu.auto-imx95-15x15-lpddr4x-frdm Default full_image-nxp-boot.uuu (NXP flash_all + LmP FB); debug full_image.uuu with --foundries-boot
SDPS boot imx-boot-imx95-15x15-lpddr4x-frdm-sd.bin-flash_all imx-boot-mfgtool (lmp-mfgtool flash_a55)
SDPV image skipped (NXP uses full AHAB at SDPS, no SDPV) u-boot-mfgtool.itb (FIT, SPL loads U-Boot)
FB WIC imx-image-full-imx95evk.wic (uncompressed) .wic.gz in uuu by default; gunzip → .wic with --wic-uncompressed
FB bootloader same *-flash_all as SDPS imx-boot-imx95-frdm-evk (production)
FB mmc target ${emmc_dev} (= 0, USDHC1 eMMC) mmcdev 0 explicit (was ${emmc_dev}; default mmcdev=1 broke eMMC flash)
Invalid for SDPS — Android u-boot-imx95-15x15-frdm-uuu.imx, production imx-boot-imx95-frdm-evk

NXP reference flash (hardware validation)

Extract LF_v6.18.2 images to ~/Downloads/LF_v6.18.2-1.0.0_images_IMX95, then:

# SW1 → Serial Download (0,1), USB J3, power on
./scripts/flash-imx95-nxp-reference.sh

# Production WIC (~multi-GB): allow 30 min
./scripts/flash-imx95-nxp-reference.sh --timeout 1800

# Or from bundle directory:
cd ~/Downloads/LF_v6.18.2-1.0.0_images_IMX95
sudo uuu uuu.auto-imx95-15x15-lpddr4x-frdm

Or via Foundries script wrapper: ./scripts/flash-imx95-foundries.sh --reference-nxp

Foundries LmP flash (single command)

Default path uses NXP LF flash_all for USB boot phases until Foundries mfgtool is fixed (meta-dynamicdevices-bsp#12). Install NXP LF bundle to ~/Downloads/LF_v6.18.2-1.0.0_images_IMX95 (or set NXP_LF_DIR).

# SW1 → Serial Download (0,1), USB J3, power on
# Default: full_image-nxp-boot.uuu (NXP flash_all SDPS/SDPV + LmP WIC + prod bootloader)
./scripts/flash-imx95-foundries.sh --emmc 2707 dynamic-devices

# Production WIC flash (350 MB compressed → needs more time):
./scripts/flash-imx95-foundries.sh --emmc 2707 dynamic-devices --timeout 1800

# Debug Foundries-only mfgtool path (imx-boot-mfgtool at SDPS — currently blocked on hardware):
./scripts/flash-imx95-foundries.sh --foundries-boot --emmc 2707 dynamic-devices

# Power off, SW1 → eMMC boot (1,0), power on — login fio / fio

If Foundries artifacts lack imx-boot-mfgtool, build locally or pass --sdps-boot (see § Obtain imx-boot-mfgtool).

Re-flash after wrong mmc target (recovery)

If the board boots to U-Boot but shows No partition table on mmc0, the LmP WIC likely landed on mmc1 (empty SD slot) or eMMC boot hwpart instead of mmc0 user area. Re-flash:

# SW1 → Serial Download (0,1), USB J3, stop ModemManager if needed
./scripts/flash-imx95-foundries.sh --emmc 2707 dynamic-devices --timeout 1800

If the board is already at fastboot (1fc9:0152) from a prior attempt:

cd downloads/target-2707/flash-bundle
sudo uuu -pp 100 fb-only.uuu                    # compressed .wic.gz (default primary)
sudo uuu -pp 100 fb-only-wic-uncompressed.uuu     # gunzip .wic (NXP-style)

Power off, SW1 → eMMC boot (1,0), power on.

WIC compressed vs uncompressed (A/B verification)

Foundries ships lmp-factory-image-imx95-frdm-evk.wic.gz. NXP LF uses an uncompressed .wic.
flash-imx95-foundries.sh defaults to .wic.gz in uuu (--wic-compressed) so you can test whether imx95 fastboot flash -raw2sparse accepts gzip bytes. Use --wic-uncompressed to gunzip in flash-bundle and reference .wic in primary scripts (NXP-style).

After --prepare-only or --regen-scripts, the bundle contains both companion sets:

Script WIC file
fb-only.uuu / full_image-nxp-boot.uuu Primary (default: .wic.gz)
fb-only-wic-compressed.uuu .wic.gz
fb-only-wic-uncompressed.uuu .wic (requires gunzip; created during prepare)

Verify eMMC content in U-Boot (after flash)

Interrupt boot or use serial after re-flash. On mmc0 user area:

mmc dev 0
mmc read 0x80400000 0 1
md.b 0x80400000 16
mmc part
md.b at LBA0 Meaning
1f 8b 08 … gzip magic — compressed .wic.gz was written raw (BAD)
other bytes; md.b 0x80400200 16 shows 45 46 49 20 50 41 52 54 EFI PART GPT header at 0x200 (GOOD, NXP-style uncompressed WIC)

Also run mmc part — a good flash shows GPT partitions on mmc 0; gzip-at-LBA0 shows ** No partition table **.

Test commands (board at Serial Download or fastboot)

A — compressed .wic.gz (default primary):

./scripts/flash-imx95-foundries.sh --prepare-only --emmc 2707 dynamic-devices
cd downloads/target-2707/flash-bundle
# SW1 Serial Download (0,1), USB J3; or if already at 1fc9:0152:
sudo uuu -pp 100 fb-only-wic-compressed.uuu
# full SDPS path:
sudo uuu -pp 100 full_image-nxp-boot-wic-compressed.uuu

B — uncompressed .wic (NXP-style):

./scripts/flash-imx95-foundries.sh --prepare-only --wic-uncompressed --emmc 2707 dynamic-devices
cd downloads/target-2707/flash-bundle
sudo uuu -pp 100 fb-only.uuu
# or without changing primary scripts:
sudo uuu -pp 100 fb-only-wic-uncompressed.uuu

Compare md.b / mmc part after each test. Expect A to show gzip magic if fastboot does not gunzip; B should show GPT.

Serial output during uuu

Serial message USB phase Action
SDP: handle #N requests 1fc9:0151 SPL waiting for SDPV — normal during full_image.uuu run
Wrong container, no image found 0151/0152 Wrong image at SDPV — must be u-boot-mfgtool.itb (FIT); not imx-boot-mfgtool (AHAB)
Fastboot / Downloading in uuu log 1fc9:0152 FB phase — fb-only OK if retrying without power-cycle

Configure terminal: 921600 8N1 on ttyACM0 for SPL. Do not open serial on ttyACM0 while uuu runs (causes LIBUSB_ERROR_IO).

Generated full_image.uuu (Foundries)

scripts/flash-imx95-foundries.sh --emmc generates SDPS + SDPV + FB:

uuu_version 1.4.149

SDPS: boot -f imx-boot-mfgtool

SDPV: delay 1000
SDPV: write -f u-boot-mfgtool.itb
SDPV: jump

FB: ucmd setenv fastboot_dev mmc
FB: ucmd setenv emmc_dev 0
FB: ucmd setenv mmcdev 0
FB: ucmd mmc dev 0
FB: ucmd mmc dev 0 1; mmc erase 0 0x2000; mmc dev 0 0
FB[-t 600000]: flash -raw2sparse all lmp-factory-image-imx95-frdm-evk.wic.gz
FB: flash bootloader imx-boot-imx95-frdm-evk
FB: ucmd if env exists emmc_ack; then ; else setenv emmc_ack 0; fi;
FB: ucmd mmc partconf 0 ${emmc_ack} 1 0
FB: done

SDPS: imx-boot-mfgtool (SPL-only AHAB, flash_a55 target, ~2.3 MB). AHAB tag 0x87202302 at offset 0.
SDPV: u-boot-mfgtool.itb (U-Boot FIT image, ~1.4 MB) — SPL loads this to start full U-Boot.
FB bootloader: production imx-boot-imx95-frdm-evk only — never use for SDPS or SDPV.

Does FRDM-IMX95 have eMMC?

Yes. 32 GB eMMC 5.1 on USDHC1 (8-bit, non-removable). microSD is USDHC2. LmP boot.cmd uses devnum 0 (eMMC). Default NXP boot mode is eMMC (SW1: 1,0).

MMC numbering (critical for UUU fastboot)

Layer Onboard eMMC (USDHC1) microSD (USDHC2)
Device tree alias mmc0 = &usdhc1 mmc1 = &usdhc2
U-Boot mmc list FSL_SDHC: 0 (eMMC) FSL_SDHC: 1 (SD slot)
Linux block device /dev/mmcblk0 /dev/mmcblk1
LmP boot.cmd devnum 0 not used by default
NXP emmc_dev / fastboot target 0 SD path uses 1

NXP uuu.auto-imx95-15x15-lpddr4x-frdm uses FB: ucmd setenv mmcdev ${emmc_dev} with emmc_dev=0 in U-Boot default env (imx95_frdm.env).

Pitfall: NXP imx95_15x15_frdm_defconfig sets CONFIG_SYS_MMC_ENV_DEV=1 (U-Boot env stored on microSD). Default compiled env therefore has mmcdev=1. During Serial Download (USB) boot, mmc_get_env_dev() reads mmcdev and falls back to 1 if unset — so fastboot can flash microSD (mmc1) while production boot reads eMMC (mmc0).

Foundries UUU scripts now hardcode mmcdev 0 / emmc_dev 0 for --emmc and restore user hwpart after boot-partition erase (mmc dev 0 0) before WIC sparse flash. Without hwpart restore, GPT/WIC can land on the small eMMC boot partition instead of user area → boot shows ** No partition table - mmc 0 **.

Verify in fastboot (interrupt U-Boot, or use fb-only.uuu after reaching 1fc9:0152):

mmc list
mmc dev 0
mmc part

Expect mmc 0 = onboard eMMC with GPT after a good flash.

Troubleshooting

Symptom Likely cause Fix
** No partition table - mmc 0 ** / Couldn't find partition mmc 0:1 after flash WIC/GPT written to mmc1 (SD) or eMMC boot hwpart, not eMMC user area Re-flash with fixed scripts (./scripts/flash-imx95-foundries.sh --emmc 2707 --timeout 1800). Confirm mmc list in fastboot shows image on mmc 0.
SPL Trying to boot from MMC1 then U-Boot fails on mmc0 SPL MMC1 = USDHC1/eMMC (1-based boot-device name); empty mmc0 user area means flash targeted wrong device Same re-flash; see MMC table above
MMC: no card present on mmc1 Normal if no microSD inserted; boot should use mmc0 Insert SD only for --sd flash path
LIBUSB_ERROR_IO / LIBUSB_ERROR_NO_DEVICE / Fail HID(W) at SDPS ModemManager grabbing/resetting the 1fc9 USB device, or minicom on ttyACM0 Permanent fix: install config/udev/99-nxp-mm-ignore.rules (MM ignores 1fc9, no per-flash stop). Or stop MM each time: sudo systemctl stop ModemManager. Close serial terminals.
Wrong container, no image found (serial) Wrong image type at SDPS or SDPV SDPS: use imx-boot-mfgtool (flash_a55 SPL); SDPV: use u-boot-mfgtool.itb (FIT). Never use imx-boot-mfgtool at SDPV.
Unknown Image type on imx95 AHAB -skipspl used with AHAB container at SDPV Remove -skipspl; script now generates correct SDPV without it
Success 0 Failure 0 at 0151 full_image.uuu starts with SDPS, uuu waits for 015d that never comes Power-cycle to 015d; or use sdpv-resume-full.uuu when stuck at 0151
Success 0 Failure 0 with 10s timeout Timeout too short — SDPS transfer alone takes >10s at HID rates Default now 300s; use --timeout 1800 for full WIC
uuu exits at ~31% SDPS Wrong SDPS image (Android flash_all) or ttyACM0 open Script validates AHAB tag; close serial terminals
Stuck at 1fc9:0151 SPL waiting for SDPV; previous flash interrupted Power-cycle to 015d, re-run full_image.uuu; or run sdpv-resume-full.uuu directly
Phase mismatch after failed run Board left in fastboot (0152) Re-run (fb-only) or power-cycle to Serial Download (015d)
Test if Foundries imx-boot-mfgtool is the blocker NXP flash_all reaches fastboot; Foundries SDPS image does not Default: ./scripts/flash-imx95-foundries.sh --emmc 2707; debug Foundries path: --foundries-boot; track fix in meta-dynamicdevices-bsp#12
FAT p1 has tee.bin + DTB only — no boot.itb; U-Boot loops on distro_bootcmd meta-lmp has no mx95 sota wiring, so WKS_FILE:sota / IMAGE_BOOT_FILES:sota = "boot.itb" were unset; VFAT never got the OSTree boot script Rebuild with BSP sota wiring mirrored from mx93-generic-bsp (see below). Expect p1 to contain boot.itb; kernel + DTB load from the OSTree deploy on p2.
U-Boot autoboot runs bsp_bootcmd, fails loading boot.scr / Image (FAT p1 has only boot.itb) NXP imx95_15x15_frdm_defconfig sets CONFIG_BOOTCOMMAND="bootflow scan -l; run bsp_bootcmd"; LmP sota uses boot.itb (FIT script), not legacy Android layout Rebuild u-boot-fio + imx-boot with ostree-boot.cfg + lmp-spl-fit.cfg. Interim manual boot below.
Executing script at 83500000 then synchronous abort / reset loop ostree-boot.cfg used source 0x83500000 but imx95 had no SPL FIT (CONFIG_SPL_LOAD_FIT); nothing valid at that address Same rebuild; autoboot is now fatload mmc 0:1 ${loadaddr} boot.itb; bootm ${loadaddr}.
fdtfile=imx95-15x15-evk-adv7535-ap1302.dtb in U-Boot env NXP EVK defconfig default; wrong for FRDM Rebuild u-boot-fio + imx-boot with custom-dtb.cfg (CONFIG_DEFAULT_FDT_FILE=imx95-15x15-frdm.dtb). boot.cmd also sets fdtfile.

Passwordless sudo

scripts/flash-imx95-foundries.sh runs env -u TERMINFO sudo -n uuu -pp 100 …. One-time:

./scripts/install-flash-sudoers.sh
# optional bundled uuu symlink:
./scripts/install-flash-sudoers.sh --link-bundled-uuu

This also installs NOPASSWD for systemctl stop/start ModemManager (both /usr/bin/ and /bin/ paths). If MM stop fails, stop it manually before flashing: sudo systemctl stop ModemManager.

Android UUU (uuu_imx_android_flash.sh) is separate and not covered by this sudoers file.

Regenerating UUU scripts without Foundries auth

If artifacts are already cached but UUU scripts are stale:

# Regenerate primary + -wic-compressed / -wic-uncompressed companion scripts
./scripts/flash-imx95-foundries.sh --regen-scripts --emmc 2707

# Primary scripts reference .wic instead of .wic.gz
./scripts/flash-imx95-foundries.sh --regen-scripts --wic-uncompressed --emmc 2707

No fioctl authentication required — reads from downloads/target-2707/flash-bundle/.
Gunzip to .wic runs automatically for companion -wic-uncompressed scripts (and for --wic-uncompressed primary).

Obtain imx-boot-mfgtool

Target 2707 mfgtools may ship stub full_image.uuu and omit imx-boot-mfgtool until BSP mfgtool-files bbappend is merged.

Local kas build:

# From meta-dynamicdevices checkout
../android-container/scripts/kas-build-lowprio.sh build kas/imx95-frdm-evk-mfgtool.yml
cp build/tmp/deploy/images/imx95-frdm-evk/mfgtool-files/imx-boot-mfgtool \
   ../android-container/downloads/target-2707/imx-boot-mfgtool
./scripts/flash-imx95-foundries.sh --sdps-boot downloads/target-2707/imx-boot-mfgtool --emmc 2707

Foundries artifact (after mfg_tools rebuild): imx95-frdm-evk-mfgtools/other/mfgtool-files/imx-boot-mfgtool

Is a new Foundries build required?

Goal New build?
Flash 2707 to eMMC and boot No — ./scripts/flash-imx95-foundries.sh --emmc 2707
Larger root on 32 GB eMMC Yes — custom .wks + IMAGE_ROOTFS_EXTRA_SPACE
GPT / split-boot / LUKS Yes — custom WKS + factory CI params
Working full_image.uuu in mfgtool artifact Yes — mfgtool template / CI
Signed boot + SIT Yes — enable signing + imx8mm-style uuu

BSP / CI changes for production eMMC factory

Manifest pin (imx95 branch): lmp-manifest/dynamic-devices.xml → meta-dynamicdevices-bsp @ feature/imx95-frdm-evk-support

The Foundries (sota) boot flow is the same as imx93-jaguar-eink: u-boot-ostree-scr-fit compiles boot.cmd into a boot.itb FIT (boot script only) that lands on VFAT /boot; the kernel and DTB are loaded from the OSTree deployment on the rootfs. meta-lmp ships this wiring for mx93-generic-bsp but not for mx95, so the BSP mirrors it.

1. imx95-frdm-evk.conf (sota wiring, mirrors mx93-generic-bsp)

  • WKS_FILE:sota:imx95-frdm-evk = "emmc-imx-gpt-sota.wks" (reuses meta-lmp's GPT eMMC layout: VFAT /boot + ext4 OSTree root; imx-boot lives in eMMC boot0/boot1 hwpart, not the WIC)
  • IMAGE_BOOT_FILES:sota:imx95-frdm-evk = "boot.itb"
  • PREFERRED_PROVIDER_u-boot-default-script:sota:mx95-generic-bsp = "u-boot-ostree-scr-fit"
  • KERNEL_IMAGETYPE:sota / KERNEL_CLASSES:sota → fitImage / kernel-lmp-fitimage
  • SOTA_CLIENT_FEATURES:append = " ubootenv"
  • UBOOT_DTB_NAME:imx95-frdm-evk = "imx95-15x15-frdm.dtb"

2. U-Boot Kconfig fragments (u-boot-fio/imx95-frdm-evk/)

  • custom-dtb.cfg — CONFIG_DEFAULT_FDT_FILE=imx95-15x15-frdm.dtb
  • fix-environment-config.cfg — CONFIG_SYS_MMC_ENV_DEV=0 (onboard eMMC, not microSD)
  • lmp-spl-fit.cfg — SPL FIT boot (mirrors imx93 lmp.cfg; meta-lmp has no imx95 lmp.cfg)
  • ostree-boot.cfg — CONFIG_BOOTCOMMAND fatloads boot.itb from VFAT p1 and bootm ${loadaddr} (replaces NXP bsp_bootcmd)
  • enable-foundries-imx-commands.cfg — CONFIG_CMD_FIOVB, CONFIG_CMD_SECONDARY_BOOT (+ OPTEE/TEE deps) for Foundries boot.cmd (fiovb, imx_is_closed, imx_secondary_boot)
  • 0005-fdt-pack-reg-unaligned-access.patch — backport put_unaligned_be64() fix for fdt_pack_reg() (fixes synchronous abort at FIT FDT load during bootm, esr 0x96000046 @ 0x83000000)

Interim U-Boot manual boot (current flash, before prod imx-boot rebuild): at the U-Boot prompt after stopping autoboot:

setenv devnum 0
setenv bootpart 1
setenv rootpart 2
setenv fdtfile imx95-15x15-frdm.dtb
setenv fdt_file imx95-15x15-frdm.dtb
setenv fdt_file_final imx95-15x15-frdm.dtb
mmc dev 0
fatload mmc 0:1 ${loadaddr} boot.itb
setenv verify 1
bootm ${loadaddr}

After a rebuild with lmp-spl-fit.cfg, SPL may also preload the embedded script to 0x83500000; autoboot still uses fatload/bootm as the reliable path.

3. boot.cmd (u-boot-ostree-scr-fit/imx95-frdm-evk/)

Sets fdtfile / fdt_file / fdt_file_final to imx95-15x15-frdm.dtb; devnum 0, bootpart 1, rootpart 2.

4. foundries-build/ci-scripts/factory-config.yml

Under refs/heads/main-imx95-frdm-devel:

params:
  WKS_FILE:sota: "imx95-frdm-evk.wks"
mfg_tools:
  - machine: imx95-frdm-evk
    params:
      WKS_FILE:sota: "imx95-frdm-evk.wks"

5. mfgtool / U-Boot (required for UUU SDPS)

BSP fix in meta-dynamicdevices-bsp/recipes-support/mfgtool-files/:

  1. mfgtool-files_%.bbappend — deploy imx-boot-mfgtool, u-boot-mfgtool.itb, fitImage-*-mfgtool
  2. mfgtool-files/imx95-frdm-evk/full_image.uuu.in — SDPS + SDPV (u-boot-mfgtool.itb) + FB eMMC
  3. Rebuild mfg_tools; verify artifact imx-boot-mfgtool exists

imx-boot_%.bbappend sets IMXBOOT_TARGETS:lmp-mfgtool:imx95-frdm-evk = "flash_a55".

Related

  • Foundries flash: scripts/flash-imx95-foundries.sh
  • NXP reference flash: scripts/flash-imx95-nxp-reference.sh
  • Android eMMC flash (separate): README.md § Flash (15×15 FRDM)
  • NXP Getting Started: GS-FRDM-IMX95