hisilicon-opensdk: bump 55b72ec → 74d8a65 (cv200 + av100 V2/V2A neo, issue #51)#2121
Merged
Merged
Conversation
…issue #51) Pulls in the V2 (hi3516cv200, ARM926EJ-S) and V2A (hi3516av100, Cortex-A7) modern-kernel bring-up landed in openhisilicon #162 and #165. Both PRs follow the cv300 pattern from #148 / #150 — all version branching lives in kernel/compat/kernel_compat.h, never #ifdef LINUX_VERSION_CODE in driver source. Why bump now: openhisilicon kept the cv200/av100 lite (4.9 kernel) production paths byte-equivalent to the prior tip — independently verified by rebuilding each against this commit with HISILICON_OPENSDK_OVERRIDE_SRCDIR and diffing rootfs against nightly: - hi3516cv200_lite: 697/697 files, 34/34 hisilicon .ko, QEMU lsmod identical, dmesg diff 20 benign lines only. open_v2_shim.ko correctly NOT shipped (gated to no-op on 4.9). - hi3516av100_lite: 290/290 files, 61/61 hisilicon .ko, same rationale. So this bump is safe for existing lite cameras (verifiable mechanically via a sysupgrade smoke test), while unlocking V2/V2A neo on the openipc/linux upstream-patches base for follow-up firmware-side PRs. What the four commits in this range touch: 74d8a65 kernel/hi3516av100: V2A bring-up — open_v2a_shim.ko exports do_gettimeofday / register_sysctl_table / init_timer_key for av100 blobs on >= 5.0; 11 av100 source modules routed through existing compat helpers + one new no_llseek shim (already added by #162 in the same line; merged cleanly). abadb7cf kernel/hi3516cv200: V2 bring-up — open_v2_shim.ko exports the same plus strlcpy (cv200-specific); 12 cv200 source modules patched analogously. New no_llseek -> NULL shim in kernel/compat/kernel_compat.h on >= 6.12. 91cada81 ci: drop hi3516av300 row from QEMU NNIE forward regression (av300 ships only ev300_neo / av300_neo combos — no NNIE). 824ed48f nnie_neo: multi-segment .wk parser + ForwardWithBbox validation harness (cv500-only; no firmware-build impact for other SoCs). Follow-up firmware PRs (not in this bump): - hi3516cv200_neo: defconfig + kernel config + post-image script to actually consume open_v2_shim.ko under a 7.0 kernel - hi3516av100_neo: ditto for V2A Both neo bring-ups carry the same struct timer_list.data drift caveat as cv300_neo for chnl/viu/vpss blobs — broken video pipeline expected until either a kernel-side .data restore patch lands or the blobs are recompiled. Documented in both shim sources. 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
1 participant
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
Bumps
HISILICON_OPENSDK_VERSIONfrom55b72ec→74d8a65, pulling in V2 (hi3516cv200) + V2A (hi3516av100) modern-kernel bring-up from openhisilicon #162 and #165 — both squash-merged today (issue #51).Why this is safe for existing lite cameras
cv200 and av100 *_lite firmware variants (kernel 4.9) were rebuilt against the new openhisilicon tip via
HISILICON_OPENSDK_OVERRIDE_SRCDIRand byte-equivalence-tested against nightly:.kocountlsmodhi3516cv200_lite(4.9)hi3516av100_lite(4.9)dmesgdiffs were ≤ 20 benign lines (BogoMIPS jitter, random MAC, ramdisk rounding). Neither shim module (open_v2_shim.ko,open_v2a_shim.ko) ships on 4.9 — they're gated to no-op and the av100/cv200 firmware install blocks don't reference them.What's in the range
Only #162 and #165 affect firmware-built binaries; #166 is cv500-only NNIE host-side work, #167 is openhisilicon-CI cleanup.
What's NOT in this PR (follow-ups)
hi3516cv200_neodefconfig + neo kernel config + post-image script — needed to actually consumeopen_v2_shim.kounder a 7.0 kernel. Separate firmware PR.hi3516av100_neoditto — separate firmware PR.struct timer_list.datadrift caveat as cv300_neo (chnl/viu/vpss blobs) — broken video pipeline expected until either a kernel-side.datarestore patch lands or the blobs are recompiled. Documented inside both shim sources.Test plan
hi3516cv200_litebyte-equivalence vs nightly (no regression)hi3516av100_litebyte-equivalence vs nightly (no regression)🤖 Generated with Claude Code