From mboxrd@z Thu Jan  1 00:00:00 1970
From: Marek Vasut <marex@denx.de>
Date: Wed, 27 Aug 2014 19:45:57 +0200
Subject: [U-Boot] [RFC PATCH] USB: get rid of warning when compile with
	debug enabled
In-Reply-To: <53FD8B18.5060901@atmel.com>
References: <1408958599-13518-1-git-send-email-voice.shen@atmel.com>
	<201408260846.59463.marex@denx.de> <53FD8B18.5060901@atmel.com>
Message-ID: <201408271945.57807.marex@denx.de>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

On Wednesday, August 27, 2014 at 09:39:04 AM, Bo Shen wrote:
> Hi Marek,
> 
> On 08/26/2014 02:46 PM, Marek Vasut wrote:
> > On Tuesday, August 26, 2014 at 03:11:13 AM, Bo Shen wrote:
> >> Hi Marek,
> >> 
> >> On 08/25/2014 06:43 PM, Marek Vasut wrote:
> >>> On Monday, August 25, 2014 at 11:23:19 AM, Bo Shen wrote:
> >>>> When compile with debug information is enabled, if call
> >>>> spin_lock_irqsave, it will give following warning information.
> >>>> This patch is used to get rid of it.
> >>>> --->8---
> >>>> warning: 'flags' is used uninitialized in this function
> >>>> [-Wuninitialized] ---8<---
> >>> 
> >>> The patch is wrong. The compiler complains that flags might be used
> >>> uninited because that is the case -- you call spin_lock_irqsave() with
> >>> uninited flags and because debug() is expanded to printf() I guess, the
> >>> compiler spews.
> >>> 
> >>> So which file does this warning come from ?
> >> 
> >> I enable debug information (In common.h, define the DEBUG).
> >> 
> >> It comes out from usb gadget driver, for example, atmel_usba_udc.c and
> >> also s3c_udc_otg.c and etc.
> > 
> > Then those need fixing. There is no problem with the macro.
> > 
> >> I see the spin_lock_irqsave() function in
> >> <drivers/usb/musb-new/linux-compat.h> defined as:
> >> --->8---
> >> #define spin_lock_irqsave(lock, flags) do {} while (0)
> >> ---8<---
> >> 
> >> As in u-boot, there is no lock actually, so I think we can fix it as
> >> <drivers/usb/musb-new/linux-compat.h>, or use the patch as I suggest, or
> >> anything else, what about your opinion?
> > 
> > Fix all the offending drivers which pass uninited variable into macro
> > which might use it. That's the fix.
> 
> Thanks for your opinion.
> 
> I will send out the patch as you suggested to fix this issue.

Thanks!

Best regards,
Marek Vasut