]> git.itanic.dy.fi Git - linux-stable/commit
tools: usb: ffs-test: Fix build on big endian systems
authorPeter Senna Tschudin <peter.senna@gmail.com>
Tue, 10 Jul 2018 14:01:45 +0000 (16:01 +0200)
committerFelipe Balbi <felipe.balbi@linux.intel.com>
Tue, 17 Jul 2018 07:12:51 +0000 (10:12 +0300)
commita2b22dddc7bb6110ac3b5ed1a60aa9279836fadb
treed0f9846283d388b0b6fb0f9af168efe2b231fb22
parentbb2863369562b3f0320fc7d8a96a5cde4b50aaea
tools: usb: ffs-test: Fix build on big endian systems

The tools/usb/ffs-test.c file defines cpu_to_le16/32 by using the C
library htole16/32 function calls. However, cpu_to_le16/32 are used when
initializing structures, i.e in a context where a function call is not
allowed.

It works fine on little endian systems because htole16/32 are defined by
the C library as no-ops. But on big-endian systems, they are actually
doing something, which might involve calling a function, causing build
failures, such as:

   ffs-test.c:48:25: error: initializer element is not constant
    #define cpu_to_le32(x)  htole32(x)
                            ^~~~~~~
   ffs-test.c:128:12: note: in expansion of macro ‘cpu_to_le32’
      .magic = cpu_to_le32(FUNCTIONFS_DESCRIPTORS_MAGIC_V2),
               ^~~~~~~~~~~

To solve this, we code cpu_to_le16/32 in a way that allows them to be
used when initializing structures. This fix was imported from
meta-openembedded/android-tools/fix-big-endian-build.patch written by
Thomas Petazzoni <thomas.petazzoni@free-electrons.com>.

CC: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Peter Senna Tschudin <peter.senna@gmail.com>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
tools/usb/ffs-test.c