MT76x8 kernel driver crash with fail safe mode desactivated

Dear all,
I'm using openwrt 21, and for security reasons, i want to disable the fail safe mode.
In the menuconfig, i did the following

  • checked "image configuration" box
  • checked "preinit config option" box
  • checked "disable fail safe mode" box
    The image generation works perfectly and the kernel boot as well, nevertheless, there is a major issue. After a few seconds, at the end of the boot, the is a kernel crash. Here is the back trace:
[   27.398683] IPv6: ADDRCONF(NETDEV_CHANGE): br-lan: link becomes ready
[   33.079919] ------------[ cut here ]------------
[   33.089121] WARNING: CPU: 0 PID: 0 at net/sched/sch_generic.c:473 dev_watchdog+0x27c/0x284
[   33.105506] NETDEV WATCHDOG: eth0 (mtk_soc_eth): transmit queue 0 timed out
[   33.119289] Modules linked in: pppoe ppp_async iptable_nat xt_state xt_nat xt_conntrack xt_REDIRECT xt_MASQUERADE xt_FLOWOFFLOAD pppox ppp_generic nf_nat nf_flow_table_hw nf_flow_table nf_conntrack mn
[   33.244774] CPU: 0 PID: 0 Comm: swapper Not tainted 5.4.215 #0
[   33.256325] Stack : 805f0000 8058df04 00000000 00000000 8058cffc 87c0be14 805c6188 805c5d63
[   33.272878]         80539584 00000000 807233b8 805e0000 00200000 00000001 87c0bdc8 ceed23f6
[   33.289429]         00000000 00000000 80750000 00000000 00000030 000000b7 322e342e 23203531
[   33.305986]         00000000 00000001 00000000 0003bc26 00000000 00000009 00000000 8037d088
[   33.322540]         00000009 805e0000 00200000 00000122 00000001 802b3250 00000000 80720000
[   33.339094]         ...
[   33.343935] Call Trace:
[   33.348804] [<80009c90>] show_stack+0x30/0x100
[   33.357612] [<80023ee8>] __warn+0xc0/0x10c
[   33.365725] [<80023fc0>] warn_slowpath_fmt+0x8c/0xac
[   33.375560] [<8037d088>] dev_watchdog+0x27c/0x284
[   33.384891] [<800696c4>] call_timer_fn.isra.34+0x20/0x90
[   33.395412] [<800698d8>] run_timer_softirq+0x1a4/0x1ec
[   33.405611] [<804b2878>] __do_softirq+0x110/0x298
[   33.414932] [<80005670>] except_vec_vi_end+0xb8/0xc4
[   33.424766] [<804b2644>] r4k_wait_irqoff+0x18/0x24
[   33.434255] ---[ end trace 29b041315ae4478e ]---

After the crash, the ethernet interface seems in an inconsistent state, the following log is printed to the console periodically:

 33.443409] mtk_soc_eth 10100000.ethernet eth0: transmit timed out
[   33.455655] mtk_soc_eth 10100000.ethernet eth0: dma_cfg:00000057
[   33.467559] mtk_soc_eth 10100000.ethernet eth0: tx_ring=0, base=07094000, max=1024, ctx=3, dtx=1, fdx=1, next=3
[   33.487556] mtk_soc_eth 10100000.ethernet eth0: rx_ring=0, base=07270000, max=1024, calc=1023, drx=0
[   44.087937] mtk_soc_eth 10100000.ethernet eth0: transmit timed out

Does anybody knows where it comes from ?
If i enable again the failsafe mode, as i said before, everything works well.

Best regards and thank's in advance for help.

Nicolas

Edit: It seems that it is a known bug, fixed several years ago ... it is strange that it comes back with the fail safe mode off ... what is the relationship?