Hi,

On  3.05.2016 19:25, Tony Lindgren wrote:
> * Ivaylo Dimitrov <ivo.g.dimitrov.75-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org> [160430 23:34]:
>>
>> Unfortunately that makes my n900 unbootable with USB cable detached, see the
>> attached file (at [  152.121673]) when there is USB cable. I guess we're
>> missing a couple of pm_runtime_get()/put() calls.
>
> OK so it seems. I did some brief testing with n900 too, but never
> hit this so far. Care to email me the kernel .config you're using
> so I can try to reproduce it here?
>

.config is attached. There is stuff in it still to be upstreamed, just 
ignore it :). Or, you may pull 
https://github.com/freemangordon/linux-n900/commits/v4.6-rc4-n900-camera, apply 
the $subject series on top of it and build rx51_defconfig. There 
shouldn't be any patches touching USB.

Does bqXXX and isp1707_charger modules got loaded when you tested it? Or 
those were built-in? or not built at all? As it seems the problem arises 
when those try to touch USB ULPI.

>> [  152.121673] Unhandled fault: external abort on non-linefetch (0x1028) at 0xfa0ab001
> ...
>
>> [  152.403594] [<c0401dc4>] (musb_default_readb) from [<c0402ce8>] (musb_ulpi_write+0x34/0xf8)
>> [  152.409454] [<c0402ce8>] (musb_ulpi_write) from [<bf2471f8>] (isp1704_charger_probe+0x15c/0x4e0 [isp1704_charger])
>> [  152.421325] [<bf2471f8>] (isp1704_charger_probe [isp1704_charger]) from [<c03a4578>] (platform_drv_probe+0x58/0xa0)
>> [  152.433624] [<c03a4578>] (platform_drv_probe) from [<c03a2d64>] (driver_probe_device+0x120/0x2b0)
>> [  152.440093] [<c03a2d64>] (driver_probe_device) from [<c03a2f74>] (__driver_attach+0x80/0xa4)
>> [  152.446624] [<c03a2f74>] (__driver_attach) from [<c03a14d4>] (bus_for_each_dev+0x50/0x88)
>> [  152.453155] [<c03a14d4>] (bus_for_each_dev) from [<c03a2444>] (bus_add_driver+0xc4/0x1e4)
>> [  152.459716] [<c03a2444>] (bus_add_driver) from [<c03a37e4>] (driver_register+0x9c/0xe0)
>> [  152.466247] [<c03a37e4>] (driver_register) from [<c0101668>] (do_one_initcall+0xf8/0x1c4)
>> [  152.472839] [<c0101668>] (do_one_initcall) from [<c019b6e8>] (do_init_module+0x54/0x1ac)
>> [  152.479400] [<c019b6e8>] (do_init_module) from [<c017904c>] (load_module+0x185c/0x1b14)
>> [  152.485961] [<c017904c>] (load_module) from [<c0179444>] (SyS_init_module+0x140/0x158)
>> [  152.492553] [<c0179444>] (SyS_init_module) from [<c0107160>] (ret_fast_syscall+0x0/0x3c)
>
> Yeah looks like a case of missing PM runtime get/put.
>
> Also, I'm aware of some remaining phy-twl4030-usb issues with
> having cable plugged in on the boot. Will send patches for those
> after the MUSB core issues are out of the way.
>

Hmm, maybe I was unclear - with no cable attached I am unable to boot 
the device at all, is it goes in an oops loop. Unfortunately I can't 
provide logs as I don't have serial port cable attached to my device(it 
is production device). The logs I provided are taken with USB cable 
attached at boot.

Thanks,
Ivo