Hi Prabhakar,
On Wed, Jul 1, 2020 at 5:42 PM Lad Prabhakar
<prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx> wrote:
> serial core expects the spinlock to be initialized by the controller
> driver for serial console, this patch makes sure the spinlock is
> initialized, fixing the below issue:
>
> [ 0.865928] BUG: spinlock bad magic on CPU#0, swapper/0/1
> [ 0.865945] lock: sci_ports+0x0/0x4c80, .magic: 00000000, .owner: <none>/-1, .owner_cpu: 0
> [ 0.865955] CPU: 0 PID: 1 Comm: swapper/0 Not tainted 5.8.0-rc1+ #112
> [ 0.865961] Hardware name: HopeRun HiHope RZ/G2H with sub board (DT)
> [ 0.865968] Call trace:
> [ 0.865979] dump_backtrace+0x0/0x1d8
> [ 0.865985] show_stack+0x14/0x20
> [ 0.865996] dump_stack+0xe8/0x130
> [ 0.866006] spin_dump+0x6c/0x88
> [ 0.866012] do_raw_spin_lock+0xb0/0xf8
> [ 0.866023] _raw_spin_lock_irqsave+0x80/0xa0
> [ 0.866032] uart_add_one_port+0x3a4/0x4e0
> [ 0.866039] sci_probe+0x504/0x7c8
> [ 0.866048] platform_drv_probe+0x50/0xa0
> [ 0.866059] really_probe+0xdc/0x330
> [ 0.866066] driver_probe_device+0x58/0xb8
> [ 0.866072] device_driver_attach+0x6c/0x90
> [ 0.866078] __driver_attach+0x88/0xd0
> [ 0.866085] bus_for_each_dev+0x74/0xc8
> [ 0.866091] driver_attach+0x20/0x28
> [ 0.866098] bus_add_driver+0x14c/0x1f8
> [ 0.866104] driver_register+0x60/0x110
> [ 0.866109] __platform_driver_register+0x40/0x48
> [ 0.866119] sci_init+0x2c/0x34
> [ 0.866127] do_one_initcall+0x88/0x428
> [ 0.866137] kernel_init_freeable+0x2c0/0x328
> [ 0.866143] kernel_init+0x10/0x108
> [ 0.866150] ret_from_fork+0x10/0x18
Interesting...
How can I reproduce that? I do have CONFIG_DEBUG_SPINLOCK=y.
I'm wondering why haven't we seen this before...
Thanks!
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
