From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Mon, 15 Sep 2014 15:56:52 +0200 Subject: [Buildroot] [PATCH v7] openvmtools: new package In-Reply-To: References: <1409674416-33961-1-git-send-email-kaszak@gmail.com> <20140915092122.7582b3f9@free-electrons.com> <20140915133811.166295a3@free-electrons.com> Message-ID: <20140915155652.5a575f39@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear K?roly Kasza, On Mon, 15 Sep 2014 15:51:50 +0200, K?roly Kasza wrote: > Found it. > This is because of gcc 4.8's intended behaviour on sizeof() function in > cases when calling on a pointer. > See https://gcc.gnu.org/gcc-4.8/porting_to.html "New warnings for pointer > access". > It also exists as an unconfirmed bug: > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=54381. > This behaviour is only gcc 4.8, while 4.9 and pre-4.8 versions shouldn't do > this (Sourcery uses 4.7 and 4.6, I used the own toolchain w/ 4.9) > > On the open-vm-tools mail list, it came up > http://sourceforge.net/p/open-vm-tools/mailman/message/32550520/ , but they > used -Wno-sizeof-pointer-memaccess for compilation - basically silencing > the warning. > > Would you recommend using -Wno-sizeof-pointer-memaccess in the .mk file or > just use a 1 line patch to syncDriverPosix.c - dereferencing > gRemoteFSTypes[i] with an asterisk? Thanks for the investigation! As the solution, I would change the package to not use -Werror. Maybe it has a configure option to do so? If not, adding one would be good (so that the patch could potentially be upstreamed), or just changing the package to not use -Werror at all. Thanks, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com