hisilicon-opensdk: bump to 2a38a27, so a CMA gk7205v500 can stream

[widgetii]

Problem

GK7205V510 (Zenointel SD-2N-4G, MIS2008) on mmz_allocator=cma, the gk7205v500 default since #2526: majestic logged Timeout from venc channel 0, and once the sensor delivered frames the board rebooted with nothing on the console.

It was a kernel panic, hidden by the image's console loglevel of 0 and followed by a reboot panic= seconds later. With printk raised:

Alignment trap: not handling instruction f8d31045 at VENC_VpssSend+0x10f [open_venc]
Unhandled fault: alignment exception (0x001) at 0xc99b6045   *ppte=42008012

The closed venc object reads its VPSS-bind descriptors with an unaligned load. osal's CMA allocator mapped non-cached MMBs strongly-ordered, where that load faults. The carve-out allocator (ioremap_wc, Normal non-cacheable) never hit it. OpenIPC/openhisilicon#238 maps them write-combined, and also passes __dma_clear_buffer() the coherent flag the 4.9 kernel expects. dfc3a81..2a38a27 is exactly that one commit.

Hardware tested on

GK7205V510, Zenointel SD-2N-4G, MIS2008: gk7205v500_ultimate NAND image built from this branch (master 77c0d82 + this commit), flashed with sysupgrade --archive. Bootargs at the image default, mem=128M mmz_allocator=cma mmz=anonymous,0,0x42000000,96M.

Video on this board also needs its sensor bring-up: lanes 0|1 in the MIS2008 ini, and GPIO50 low, which stock does from init.sh. Both were applied by hand for the test; neither is part of this PR.

Evidence

Before (image at dfc3a81): kernel panic within 1 s of the first bound frame, from majestic and from the vendor SPC020 sample_vio:

Alignment trap: not handling instruction f8d31045 at [<bfa025d8>]
Unhandled fault: alignment exception (0x001) at 0xc99b6045
PC is at VENC_VpssSend+0x10f/0x164 [open_venc]
(VENC_VpssSend [open_venc]) from (VpssFakeSend+0x1aa/0x294 [open_vpss])

After (this branch's image, xm_osal.ko md5 8805f719 from the image, no overlay):

mmz_allocator=cma mmz=anonymous,0,0x42000000,96M osal=8805f719
up=40.84 VI=387 jpg:200:101796
...
up=76.46 VI=2167 jpg:200:102361
alignment_traps=0
stream|codec_name=hevc|width=1920|height=1080|avg_frame_rate=20/1

Image sizes: uImage: [1909KB/2048KB], fitImage: [1911KB/4096KB], rootfs.ubi: [15104KB/16384KB].

Scope

• No kernel patches under general/package/all-patches/linux/ (those go to OpenIPC/linux)
• No files specific to a single retail camera model (those go to OpenIPC/builder)
• No probing or bring-up tooling (that goes to OpenIPC/ipctool)
• Nothing under general/overlay/ or in a shared load_<vendor> script hardcodes a value specific to my board
• Package sources come from an OpenIPC repository, and any version bump keeps at least the specificity of the pin it replaces (a new package should pin a full 40-character SHA)
• No LD_PRELOAD, and no binaries that cannot be rebuilt from source
• New code is selected by a defconfig, so CI actually builds it

[qodo-free-for-open-source-projects]

PR Summary by Qodo

Bump openhisilicon to fix CMA streaming on GK7205V500

🐞 Bug fix ⚙️ Configuration changes 🕐 10-20 Minutes

AI Description

• Pin openhisilicon to the upstream CMA mapping fix for GK7205V500 devices.
• Prevent an alignment fault that panicked the kernel when VPSS first delivered a frame to VENC.

Diagram

graph TD
  Pin["Package pin"] --> Source["Openhisilicon source"] --> OSAL["XM OSAL"] --> MMB["CMA memory"] --> VENC["VENC"] --> Video["Video stream"]
  VPSS["VPSS frames"] --> VENC

Loading

High-Level Assessment

Pinning the focused upstream fix keeps the allocator correction in its source repository and preserves the existing full-SHA dependency pin. A local kernel or package patch would duplicate that fix without a clear benefit.

Files changed (1) +1 / -1

Other (1) +1 / -1

hisilicon-opensdk.mkPin openhisilicon to the CMA mapping fix +1/-1

Pin openhisilicon to the CMA mapping fix

• Replaces the full openhisilicon commit SHA with 2a38a27ba4764d4e2265a6f1753ecca0ef8a773f. According to the PR evidence, the upstream commit fixes the CMA mapping that caused an alignment fault during GK7205V500 streaming; no local source or build logic changes.

general/package/hisilicon-opensdk/hisilicon-opensdk.mk

[qodo-free-for-open-source-projects]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can tweak Display settings with a live preview to see your comment before it ships

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗