Re: Issue with isp1704 loaded too soon

From: Felipe Contreras
Date: Thu Dec 15 2011 - 12:51:48 EST


On Thu, Dec 15, 2011 at 1:15 AM, Felipe Contreras
<felipe.contreras@xxxxxxxxx> wrote:
> On Tue, Dec 13, 2011 at 11:19 PM, Felipe Contreras
> <felipe.contreras@xxxxxxxxx> wrote:
>> On Mon, Dec 12, 2011 at 9:09 AM, Jarkko Nikula <jarkko.nikula@xxxxxxxxxx> wrote:
>>> Indeed yes. I checked that in 3.0 it still works but not in 3.1 so some
>>> non isp1704_charger change has broke it as there hasn't been changes on it.
>>
>> Actually it's broken in 3.0 as well, try this configuration:
>>
>> Â Â Â Â# CONFIG_USB_MUSB_HOST is not set
>> Â Â Â Â# CONFIG_USB_MUSB_PERIPHERAL is not set
>> Â Â Â ÂCONFIG_USB_MUSB_OTG=y
>> Â Â Â ÂCONFIG_USB_GADGET_MUSB_HDRC=y
>> Â Â Â ÂCONFIG_USB_MUSB_HDRC_HCD=y
>>
>> Â Â Â ÂCONFIG_USB_GADGET_SELECTED=y
>> Â Â Â Â# CONFIG_USB_GADGET_FUSB300 is not set
>> Â Â Â Â# CONFIG_USB_GADGET_OMAP is not set
>> Â Â Â Â# CONFIG_USB_GADGET_R8A66597 is not set
>> Â Â Â Â# CONFIG_USB_GADGET_PXA_U2O is not set
>> Â Â Â Â# CONFIG_USB_GADGET_M66592 is not set
>> Â Â Â Â# CONFIG_USB_GADGET_DUMMY_HCD is not set
>>
>> I will try to find where the issues started to happen, but it's a bit
>> difficult because all this USB Kconfig stuff is completely messed up.
>>
>> *Sigh*
>
> Ok, I made some progress.
>
> The issue appeared on 2.6.39, specifically, on commit 0df0914, which
> is a merge from Linus. The issue doesn't appear in any of the two
> parents of that merge.
>
> I manually merged various commits, and the first one that triggered
> the problem is 18a2689, but the commit itself is not the problem but
> something that was already merged to Linus' tree, specifically at
> point 6899608.
>
> IOW, it's a mixture of two changes that appear together at point,
> 0df0914, one of them is 18a2689, the other one I still don't know.
>
> Unless you can find what's wrong with this commit:
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=blobdiff;f=arch/arm/mach-omap2/usb-musb.c;h=a9d4d143086d73463b97c093cc9f022de8e1d679;hp=9788b4941857f40d50d48fa6470ee14e179bc373;hb=18a2689;hpb=273ff8c3bc04cf358c131f49ee7a11efa7ec73d7
>
> I guess the safest route is to continue with the manual merges to find
> the second relevant commit =/

Found it.

So, the first commit is this one:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=18a2689

Which needs this one:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=870ea2b

But it's only triggered with this other one:
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=commitdiff;h=7acc619

Seems to be some power management stuff, and it's only triggered with
certain configurations.

And this concludes the most tedious git bisect in history =/

Cheers.

--
Felipe Contreras
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/