Re: [PATCH net-next] net: mvpp2: Initialize link in mvpp2_isr_handle_{xlg,gmac_internal}

[Russell King - ARM Linux admin <linux@armlinux.org.uk>]

On Thu, Sep 10, 2020 at 05:31:42PM -0700, Nathan Chancellor wrote:
> On Thu, Sep 10, 2020 at 03:28:11PM -0700, David Miller wrote:
> > From: Nathan Chancellor <natechancellor@gmail.com>
> > Date: Thu, 10 Sep 2020 10:48:27 -0700
> > 
> > > Clang warns (trimmed for brevity):
> > > 
> > > drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c:3073:7: warning:
> > > variable 'link' is used uninitialized whenever 'if' condition is false
> > > [-Wsometimes-uninitialized]
> > >                 if (val & MVPP22_XLG_STATUS_LINK_UP)
> > >                     ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c:3075:31: note:
> > > uninitialized use occurs here
> > >                 mvpp2_isr_handle_link(port, link);
> > >                                             ^~~~
> > > ...
> > > drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c:3090:8: warning:
> > > variable 'link' is used uninitialized whenever 'if' condition is false
> > > [-Wsometimes-uninitialized]
> > >                         if (val & MVPP2_GMAC_STATUS0_LINK_UP)
> > >                             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > drivers/net/ethernet/marvell/mvpp2/mvpp2_main.c:3092:32: note:
> > > uninitialized use occurs here
> > >                         mvpp2_isr_handle_link(port, link);
> > >                                                     ^~~~
> > > 
> > > Initialize link to false like it was before the refactoring that
> > > happened around link status so that a valid valid is always passed into
> > > mvpp2_isr_handle_link.
> > > 
> > > Fixes: 36cfd3a6e52b ("net: mvpp2: restructure "link status" interrupt handling")
> > > Link: https://github.com/ClangBuiltLinux/linux/issues/1151
> > > Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
> > 
> > This got fixed via another change, a much mode simply one in fact,
> > changing the existing assignments to be unconditional and of the
> > form "link = (bits & MASK);"
> 
> Ah great, that is indeed cleaner, thank you for letting me know!

Hmm, I'm not sure why gcc didn't find that. Strangely, the 0-day bot
seems to have only picked up on it with clang, not gcc.

Thanks for fixing.

-- 
RMK's Patch system: https://www.armlinux.org.uk/developer/patches/
FTTP is here! 40Mbps down 10Mbps up. Decent connectivity at last!