ci: drop cv300 qemu-boot allow-failure (issue chain complete)

[widgetii]

Summary

Two upstream fixes unblock the cv300 qemu-boot row:

1. hi3516cv300: load_hisilicon — insmod hi_osal.ko in CMA branch too (closes #2061) firmware#2105 — `load_hisilicon` now uses `hi_osal.ko` in both branches (was: `cma_osal.ko`, which the firmware never builds).
2. hi3516cv300 machine: extra_cmdline 'mmz=' breaks load_hisilicon on non-CMA V3 kernel widgetii/qemu-hisilicon#101 (commit `42dea48a`) — `hi3516cv300` machine model no longer appends `mmz=…` to bootargs, matching real V3 board u-boot env. The script now takes the hisi allocator branch as it always did on real hardware.

Test plan

Local QEMU boot with both fixes in place:

```
$ ./run-qemu-cv300.sh fixed
.../uImage.hi3516cv300
.../rootfs.squashfs.hi3516cv300

===CI:HISI_MOD_COUNT===31===
===CI:LSMOD_BEGIN===
open_acodec open_adec open_aenc open_ai open_aio open_ao
open_base open_chnl open_h264e open_h265e open_hwrng open_isp
open_ive open_jpege open_mipi_rx open_osal open_pwm open_rc
open_rgn open_sensor_i2c open_sensor_spi open_sys open_sys_config open_vedu
open_venc open_vgs open_vi open_vpss open_wdt
hi3516cv300_vou hi_piris (vendor blobs from osdrv)
===CI:LSMOD_END===

Welcome to OpenIPC
openipc-hi3516cv300 login:
```

• 31 HISI modules (vs 34 on `openipc-hi3516cv300.dlab.torturelabs.com` real hardware — diff is benign QEMU env)
• No Oops/panic/BUG/Trace
• login prompt reached

Drop `allow-failure: true`, gate on `min_modules: 20` (well below observed 31, with margin for benign drift).

Closes the CI-hardcoding cleanup item originally tracked in OpenIPC/firmware#2061.