u-boot-z.bin: build the LZMA-wrapped ~250K binary end-to-end - #10
Merged
Conversation
Three source-side fixes so `make u-boot-z.bin` works in this fork
without CI-time sed-patches:
* Top-level Makefile recipe (u-boot-z.bin): use $(srctree) when
$(KBUILD_SRC) is empty (in-tree builds), stage drivers/ddr/goke/
default/ddr_training_{impl,ctl,boot,console}.c and
drivers/ddr/goke/$(SOC)/ddr_training_custom.c next to the SoC
sources before the sub-make, and forward ENABLE_MINI_BOOT=y so the
arch/arm/cpu/armv7/$(SOC)/Makefile selects u-boot-mini.lds and
drops emmc_boot.o/div0.o (which live elsewhere) from COBJS. That
matches the bootrom path that decompresses u-boot.bin.
* arch/arm/cpu/armv7/{gk7202v300,gk7205v200,gk7205v300,gk7605v100}/
Makefile: replace the HiSi-internal
$(OUTDIR)/../../../tools/utils/bin/gzip path with system `gzip`.
The HiSi BSP layout the original path assumed doesn't exist in
this fork.
* include/config_distro_defaults.h: stop unconditionally defining
CONFIG_EFI_PARTITION. disk/part_efi.c's
`static efi_guid_t = PARTITION_SYSTEM_GUID;` trips this gcc with
"initializer element is not constant" because EFI_GUID() expands
to a compound literal. The gk SoCs boot from raw NOR/NAND, not
GPT, so the default never needed it. Boards that actually need it
can set CONFIG_EFI_PARTITION in their own header.
CI workflow simplifies accordingly: drops the EFI_PARTITION sed and
gains a `make u-boot-z.bin` step that ships the LZMA-wrapped
~250 KiB binary instead of raw u-boot.bin, matching the historical
OpenIPC/firmware release-asset format.
Two follow-ups to the prior recipe attempt: * Don't pass ENABLE_MINI_BOOT=y to the per-SoC sub-make. Doing so triggered an unresolved-symbol link error on gk7205v200/gk7202v300: arch/arm/cpu/armv7/<soc>/start.S inserts `bl timer_init` under `#ifdef ENABLE_MINI_BOOT`, but timer_init in startup.c is gated by `#ifdef CONFIG_MINI_BOOT` (the Kconfig symbol, only enabled by the separate `<soc>_mini_defconfig`). Mismatched gating === undefined reference. ev200's sibling u-boot-z.bin recipe also omits the flag, for the same reason. Drop it; start.S then compiles without the timer_init call, COBJS keeps emmc_boot.o + div0.o, and the link uses the full u-boot.lds. gk7205v300/gk7605v100 (which never had the ENABLE_MINI_BOOT guard or u-boot-mini.lds in their SoC Makefiles) also build cleanly under this path now. * Stage arch/arm/lib/div0.c into the sub-make build dir alongside the ddr_training_*.c files. Without it, `div0.o` in COBJS has no source visible to the sub-make's `%.o: %.c` rule. Re-add qemu_smoke for gk7205v200 / gk7202v300 — widgetii/qemu- hisilicon#60 (the HW-gzip decompressor stub for ev200-family) is now closed, which is the gap that previously blocked these mirrors. gk7205v300 / gk7605v100 still emit no UART output in QEMU and publish unguarded for now. Verified locally with arm-hisiv300-linux: all four SoCs produce ~250 KiB u-boot-<soc>.bin, matching the ev200/cv500 family format. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
3 tasks
widgetii
added a commit
that referenced
this pull request
May 8, 2026
Real Goke V4 silicon's HW gzip decompressor only parses streams compressed with WSIZE=8 KiB; system gzip emits 32 KiB-window streams that the decompressor rejects with "Uncompress Fail!". The QEMU model is more permissive (accepted both), so the qemu_smoke gate didn't catch this on PR #10. User reported the May 8 gk7205v300 binary fails this way on real hardware (vs the May 6 raw 500 KiB baseline that booted cleanly). Test on real hi3516ev300 silicon confirms ev200's wrapped binary (uses hi_gzip) boots — Goke is a HiSilicon clone with the same source tree, so the same compression tool applies. Changes: * Import tools/hi_gzip/ from u-boot-hi3516ev200. Same gzip-1.8 + HiSi window-size patch ev200 has shipped for years; not Goke-specific. * arch/arm/cpu/armv7/<soc>/Makefile (4 SoCs): use ./gzip (the staged hi_gzip binary) instead of system gzip, mirroring ev200's hw_compressed/Makefile pattern. * CI: build hi_gzip and stage it at arch/arm/cpu/armv7/<soc>/gzip before make u-boot-z.bin. The recipe's existing cp -raf carries it into the sub-make work dir as ./gzip. Same pattern u-boot- hi3516ev200's CI uses. Verified locally with arm-hisiv300-linux: produces 252 KiB u-boot-gk7205v200.bin (within 256 KiB partition gate). On-hardware verification still TODO (user has the rig). Co-authored-by: john-1 <john-1@users.noreply.github.com> Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Restores the LZMA-wrapped `u-boot-z.bin` build path for all four
gk SoCs in this repo, matching the ev200/cv500 family's release
format (~250 KiB versus the ~500 KiB raw `u-boot.bin` we've been
publishing). End-to-end CI gate, with a QEMU smoke test on the two
machines that have working flash boot.
Changes
fallback when `$(KBUILD_SRC)` is empty, stage the ddr_training
source files and `arch/arm/lib/div0.c` into the sub-make build
dir, and drop the `ENABLE_MINI_BOOT=y` flag (mismatched with
startup.c's `CONFIG_MINI_BOOT` gating — see commit message for
details). Mirrors the working ev200 sibling recipe.
HiSi-internal `tools/utils/bin/gzip` path with system `gzip`.
`CONFIG_EFI_PARTITION` (compound-literal `static` init breaks on
this gcc; gk SoCs boot from raw flash, not GPT).
re-added `qemu_smoke` matrix for `gk7205v200` / `gk7202v300`
(gated by hi3516ev200 / hi3518ev300: u-boot "hardware decompress overtime" — HW decompressor unimplemented qemu-hisilicon#60, now fixed).
Test plan
`u-boot-.bin` at ~250 KiB.
assets refresh and shrink from ~500 KiB to ~250 KiB.
🤖 Generated with Claude Code