Re: [PATCH 1/3] usb: gadget: function: phonet: balance usb_ep_disable calls

[Pavel Machek <pavel@ucw.cz>]

> > > > > > In current state I review all 3 patches as:
> > > > > > 
> > > > > > Rejected-by: Pali Rohár <pali.rohar@gmail.com>
> > > > > > [It breaks booting Nokia N900 device]
> > > > > 
> > > > > next step, figure why it's broken. Working just fine here
> > > > > on AM335x which has the same musb IP.
> > > > 
> > > > Why is broken? That is easy. You send 3 patches which broke
> > > > it.
> > > 
> > > Actually when I reverted only that patch which adds line:
> > > 
> > > pm_runtime_irq_safe(musb->controller)
> > > 
> > > then early boot crash disappeared.
> > > 
> > > But other two patches did not fixed support for external .ko
> > > gadget modules. State is same -- crash after modprobe.
> > 
> > Here is crash from qemu when musb is compiled into kernel:
> > 
> > [    0.641662] Unable to handle kernel NULL pointer dereference at virtual address 00000000
> > [    0.642211] pgd = c0004000
...
> > [    0.672882] ---[ end Kernel panic - not syncing: Attempted to kill init! exitcode=0x0000000b
> > [    0.672882]
> > 
> > Reason why it crashes is because when function
> > omap2430_runtime_resume() is called pointer to functions 
> > musb_readl and musb_writel are both NULL. And so NULL pointer
> > dereference.
> 
> https://git.kernel.org/cgit/linux/kernel/git/balbi/usb.git/commit/?h=testing/next&id=1861a2c60351a390272b3395f4d88480cdfd9e58
>

So you commit a fix to your internal branch without telling Pali, who
is actively debugging same problem at the same time, and then you just
send him url without at least saying "sorry"?

									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html