Hi Sven, I met below crash on v5.2-rc1 and found it's related with commit 728e0fbf263e, after googled, the latest Linux-next should fix the issue, please check https://lkml.org/lkml/2019/6/3/1405. After I rebased to the latest linux-next, no such issue but still another crash issue after kernel bootup, I believe that's Caused by NFS, because the issue is gone if I mount to mmc rootfs instead of NFS. I don't think it's related with Sdma firmware download, but seems your issue is different with mine. Anyway, could you try the latest linux-next? [ 17.794449] Internal error: Oops - undefined instruction: 0 [#1] SMP ARM [ 17.794459] Modules linked in: [ 17.794473] CPU: 0 PID: 29 Comm: kworker/0:1 Not tainted 5.2.0-rc2-00431-gcaaadc9 #285 [ 17.794478] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree) [ 17.794499] Workqueue: events bpf_prog_free_deferred [ 17.794515] PC is at __free_vmap_area+0x7c/0x390 [ 17.794523] LR is at 0xe83a2be0 [ 17.794528] pc : [<c02626bc>] lr : [<e83a2be0>] psr: 20000013 [ 17.794533] sp : e81f5e60 ip : f0da1000 fp : eafac378 [ 17.794537] r10: c12089ec r9 : c12300e4 r8 : c198b0cc [ 17.794542] r7 : c198b0dc r6 : e87f5ee0 r5 : e87f5ed0 r4 : e87f5ee0 [ 17.794547] r3 : e83a2bb0 r2 : e83a2c60 r1 : e83a2bb0 r0 : f0d9d000 [ 17.794555] Flags: nzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment none [ 17.794561] Control: 10c5387d Table: 38b9c04a DAC: 00000051 [ 17.794569] Process kworker/0:1 (pid: 29, stack limit = 0x042ac40f) [ 17.794574] Stack: (0xe81f5e60 to 0xe81f6000) [ 17.794584] 5e60: c198b0cc c198b0dc e87f5ea0 00000004 c198b0cc c198b0dc 0000c000 c12300e4 [ 17.794593] 5e80: c12089ec c0262a9c c12e3205 ffffffff 00000000 c12e3205 00000001 00000004 [ 17.794602] 5ea0: c12089ec c0263644 00000000 00000000 c02634f0 c0bf8e70 00000000 c11b638c [ 17.794610] 5ec0: 00000075 e97164c0 00000001 f0b5d000 00000000 ffffffff c1208928 c12e43f0 [ 17.794618] 5ee0: c12e30a9 c0265efc e9142ea4 e81e9080 eaf6f4c0 eaf72700 e81f5f1c c01468e4 [ 17.794627] 5f00: 00000001 00000000 c014682c eaf6f4d0 00000000 00000000 c0147b54 c1986060 [ 17.794635] 5f20: c15949d0 00000000 c0ec3a08 a010c325 c1205900 e81e9080 e81e9094 eaf6f4c0 [ 17.794644] 5f40: 00000008 eaf6f4f4 e81f4000 eaf6f4c0 c1205900 c0147ae0 e813a10c c0bf8e70 [ 17.794652] 5f60: e813a100 e813a100 00000000 e81e4dc0 e81f4000 e81e9080 c0147aac e813a138 [ 17.794660] 5f80: e80d9e90 c014d564 e8130c80 e81e4dc0 c014d458 00000000 00000000 00000000 [ 17.794668] 5fa0: 00000000 00000000 00000000 c01010b4 00000000 00000000 00000000 00000000 [ 17.794675] 5fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 [ 17.794683] 5fe0: 00000000 00000000 00000000 00000000 00000013 00000000 00000000 00000000 [ 17.794698] [<c02626bc>] (__free_vmap_area) from [<c0262a9c>] (__purge_vmap_area_lazy+0xcc/0x15c) [ 17.794710] [<c0262a9c>] (__purge_vmap_area_lazy) from [<c0263644>] (_vm_unmap_aliases+0x1a8/0x220) [ 17.794720] [<c0263644>] (_vm_unmap_aliases) from [<c0265efc>] (__vunmap+0x174/0x218) [ 17.794736] [<c0265efc>] (__vunmap) from [<c01468e4>] (process_one_work+0x2d0/0x704) [ 17.794747] [<c01468e4>] (process_one_work) from [<c0147ae0>] (worker_thread+0x34/0x560) [ 17.794759] [<c0147ae0>] (worker_thread) from [<c014d564>] (kthread+0x10c/0x148) [ 17.794770] [<c014d564>] (kthread) from [<c01010b4>] (ret_from_fork+0x14/0x20) [ 17.794775] Exception stack(0xe81f5fb0 to 0xe81f5ff8) [ 17.794782] 5fa0: 00000000 00000000 00000000 00000000 [ 17.794790] 5fc0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000 [ 17.794797] 5fe0: 00000000 00000000 00000000 00000000 00000013 00000000 [ 17.794806] Code: e5132010 9a000002 e15c0002 9a000008 (e7f001f2) [ 17.794816] ---[ end trace 478473ef71849173 ]--- [ 44.153270] rcu: INFO: rcu_sched self-detected stall on CPU [ 44.158868] rcu: 3-....: (2599 ticks this GP) idle=f42/1/0x40000002 softirq=2523/2523 fqs=1300 [ 44.167659] (t=2600 jiffies g=585 q=509) [ 44.171676] NMI backtrace for cpu 3 [ 44.175176] CPU: 3 PID: 1 Comm: systemd Tainted: G D 5.2.0-rc2-00431-gcaaadc9 #285 [ 44.184051] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree) [ 44.190609] [<c0112680>] (unwind_backtrace) from [<c010ceb4>] (show_stack+0x10/0x14) [ 44.198369] [<c010ceb4>] (show_stack) from [<c0bd623c>] (dump_stack+0xd4/0x108) [ 44.205690] [<c0bd623c>] (dump_stack) from [<c0bdd608>] (nmi_cpu_backtrace+0xac/0xbc) [ 44.213534] [<c0bdd608>] (nmi_cpu_backtrace) from [<c0bdd6f8>] (nmi_trigger_cpumask_backtrace+0xe0/0x134) [ 44.223114] [<c0bdd6f8>] (nmi_trigger_cpumask_backtrace) from [<c01a8390>] (rcu_dump_cpu_stacks+0xac/0xf0) [ 44.232780] [<c01a8390>] (rcu_dump_cpu_stacks) from [<c01a77b0>] (rcu_sched_clock_irq+0x7d0/0xa30) [ 44.241752] [<c01a77b0>] (rcu_sched_clock_irq) from [<c01af444>] (update_process_times+0x30/0x5c) [ 44.250638] [<c01af444>] (update_process_times) from [<c01c3e50>] (tick_sched_timer+0x5c/0xc0) [ 44.259263] [<c01c3e50>] (tick_sched_timer) from [<c01b00f8>] (__hrtimer_run_queues+0x170/0x500) [ 44.268058] [<c01b00f8>] (__hrtimer_run_queues) from [<c01b1444>] (hrtimer_interrupt+0x154/0x2d0) [ 44.276941] [<c01b1444>] (hrtimer_interrupt) from [<c01119a4>] (twd_handler+0x2c/0x38) [ 44.284869] [<c01119a4>] (twd_handler) from [<c0192554>] (handle_percpu_devid_irq+0xd4/0x384) [ 44.293405] [<c0192554>] (handle_percpu_devid_irq) from [<c018c28c>] (generic_handle_irq+0x20/0x34) [ 44.302462] [<c018c28c>] (generic_handle_irq) from [<c018c88c>] (__handle_domain_irq+0x64/0xe0) [ 44.311177] [<c018c88c>] (__handle_domain_irq) from [<c04ebb40>] (gic_handle_irq+0x58/0xb8) [ 44.319542] [<c04ebb40>] (gic_handle_irq) from [<c0101a70>] (__irq_svc+0x70/0x98) [ 44.327030] Exception stack(0xe80b3bb0 to 0xe80b3bf8) [ 44.332089] 3ba0: c12300e4 00000000 000003c4 000003c3 [ 44.340275] 3bc0: c12300e4 ffffe000 f0800000 00004000 00000000 f0800000 ffffffff 00000001 [ 44.348459] 3be0: e80b2000 e80b3c00 c0264c6c c01834b0 800d0013 ffffffff > -----Original Message----- > From: Sven Van Asbroeck <thesven73@xxxxxxxxx> > Sent: 2019年6月5日 1:57> > Hi Robin, > > On Tue, Jun 4, 2019 at 4:51 AM Robin Gong <yibin.gong@xxxxxxx> wrote: > > > > Hi Sven, > > I saw similar kernel crash issue on the latest linux-next > > during kernel boot up, but not on the 'Linux 5.2-rc1 ' tag. Will do bisect later. > But v5.0 should be okay, could you help double check? > > Of course. My tests show: > > v5.0 bad > v5.2-rc1 bad > v5.2-rc2 bad > v5.2-rc3 bad > v4.20 good > > where "bad" means: kernel crashes on boot if sdma firmware is loaded. > kernel works ok if we use sdma firmware already in rom. > > Are you able to reproduce this issue? I started bisecting last week, but ran out > of time. I could try again here, if you cannot reliably reproduce it.
