On Thu, 2012-05-24 at 18:22 +0200, Uwe Kleine-K?nig wrote:

On Thu, May 24, 2012 at 04:38:23PM +0200, Paul Bolle wrote:

quoted

1) This is apparently legal. Doesn't gcc issue a warning for this?

I wondered about that too, gcc doesn't even warn when the two values are
different. See for example commit
bd9e310dca15c9987256f67af19f9f42426e7493.

In that case it were two strings ("rtc-pcf8563" versus "pcf8563"). Why
should gcc care? Or is there something going on behind the scenes which
somehow depends on the contents of the string?
 

quoted

 arch/arm/mach-imx/mach-cpuimx35.c     |    1 -
 arch/arm/mach-imx/mach-cpuimx51sd.c   |    1 -

I already submitted a patch for the imx bits with Message id
1335989287-10094-1-git-send-email-u.kleine-koenig at pengutronix.de.
(e.g.
http://mid.gmane.org/1335989287-10094-1-git-send-email-u.kleine-koenig at pengutronix.de)

Thanks. I missed that one, obviously. Should I just drop that part of
the patch and resend?

By the way, together with my previous patch (Message-ID
1337869829.22505.15.camel at x61.thuisdomein , probably not yet archived on
the web) we have found all instances of this issue, haven't we?

Don't know if Sascha took it already. If he did, he didn't tell me.

Paul Bolle