usb: musb: break the huge isr musb_stage0_irq() into small functions

Quoted 3 lines · Bin Liu

Hi Dan,

I appreciate you scaning the patches and reporting the issues.

Quoted 1 line

These are kbuild stuff so I basically just forward them. It's no effort.

Quoted 25 lines · Dan Carpenter

Quoted 8 lines

Hi Bin,

I love your patch! Perhaps something to improve:

[auto build test WARNING on balbi-usb/next]
[also build test WARNING on v4.17-rc1 next-20180417]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

Quoted 5 lines

url:    https://github.com/0day-ci/linux/commits/Bin-Liu/usb-musb-cleanup/20180417-133633
base:   https://git.kernel.org/pub/scm/linux/kernel/git/balbi/usb.git next

smatch warnings:
drivers/usb/musb/musb_core.c:797 musb_handle_intr_connect() error: we previously assumed 'musb->hcd' could be null (see line 783)

It appears the condition check was introduced back in 2013 in an attempt to separate host-only and gadget-only configurations for musb [1], but it seems the work is not complete, because the musb->hcd allocation is unconditional. So in the current code, musb->hcd won't be NULL.

But I am now hesitated to remove this 'if(musb->hcd)' check to solve this smatch warning, because I am still debating on either continue the work to separate the two configurations or clean up the uncompleted code left by the attempt to not separate them.

Quoted 2 lines

So I want to leave the patch as is for now, is it reasonable?

Fine by me. Right now it's sort of hard for Smatch to parse the code well enough for it to know that ->hcd is always non-NULL but I'm hopeful that maybe in a couple years that will be possible...