[Buildroot] [PATCH] libusb: disable on avr32

[Buildroot] [PATCH] libusb: disable on avr32

[spdawson at gmail.com]

From: Simon Dawson <spdawson@gmail.com>

For target architecture avr32, the gpsd package fails to build when
libusb is selected. The following autobuild failure illustrates the problem.

  http://autobuild.buildroot.net/results/aa3d4763a0e90c995bc8431cf345730ee22e65a1/build-end.log

The issue is that libusb requires timerfd_settime and timerfd_create, but
the underlying syscalls are not implemented on avr32.

This patch disables the libusb package for the avr32 target architecture.

Signed-off-by: Simon Dawson <spdawson@gmail.com>
---
 package/libusb/Config.in |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/package/libusb/Config.in b/package/libusb/Config.in
index 2eed3fa..cbaeac6 100644
--- a/package/libusb/Config.in
+++ b/package/libusb/Config.in
@@ -1,5 +1,6 @@
 config BR2_PACKAGE_LIBUSB
 	bool "libusb"
+	depends on !BR2_avr32 # timerfd not in uClibc-0.9.31
 	depends on BR2_TOOLCHAIN_HAS_THREADS
 	help
 	  Userspace library for accessing USB devices
@@ -7,4 +8,4 @@ config BR2_PACKAGE_LIBUSB
 	  http://libusb.sourceforge.net/
 
 comment "libusb needs a toolchain with thread support"
-	depends on !BR2_TOOLCHAIN_HAS_THREADS
+	depends on !BR2_TOOLCHAIN_HAS_THREADS && !BR2_avr32
-- 
1.7.10.4

[Buildroot] [PATCH] libusb: disable on avr32

[Peter Korsgaard]

>>>>> "spdawson" == spdawson  <spdawson@gmail.com> writes:

 spdawson> From: Simon Dawson <spdawson@gmail.com>
 spdawson> For target architecture avr32, the gpsd package fails to build when
 spdawson> libusb is selected. The following autobuild failure illustrates the problem.

 spdawson>   http://autobuild.buildroot.net/results/aa3d4763a0e90c995bc8431cf345730ee22e65a1/build-end.log

 spdawson> The issue is that libusb requires timerfd_settime and timerfd_create, but
 spdawson> the underlying syscalls are not implemented on avr32.

 spdawson> This patch disables the libusb package for the avr32 target architecture.

What about libusb's reverse dependencies?

git grep 'select BR2_PACKAGE_LIBUSB'
package/bluez_utils/Config.in:  select BR2_PACKAGE_LIBUSB
package/ccid/Config.in: select BR2_PACKAGE_LIBUSB
package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB
package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB
package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
package/libiqrf/Config.in:      select BR2_PACKAGE_LIBUSB
package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB
package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB
package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
package/pcsc-lite/Config.in:    select BR2_PACKAGE_LIBUSB
package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB
package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB_COMPAT
package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB
package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
package/usbutils/Config.in:     select BR2_PACKAGE_LIBUSB

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH] libusb: disable on avr32

[Simon Dawson]

Hi Peter; thanks for the feedback.

On 17 November 2012 19:58, Peter Korsgaard <jacmet@uclibc.org> wrote:
> What about libusb's reverse dependencies?
>
> git grep 'select BR2_PACKAGE_LIBUSB'
> package/bluez_utils/Config.in:  select BR2_PACKAGE_LIBUSB
> package/ccid/Config.in: select BR2_PACKAGE_LIBUSB
> package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB
> package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
> package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB
> package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
> package/libiqrf/Config.in:      select BR2_PACKAGE_LIBUSB
> package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB
> package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
> package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB
> package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
> package/pcsc-lite/Config.in:    select BR2_PACKAGE_LIBUSB
> package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB
> package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB_COMPAT
> package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB
> package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
> package/usbutils/Config.in:     select BR2_PACKAGE_LIBUSB

Ah, yes. What's the right thing to do here? I don't really like the
thought of that "depends on !BR2_avr32" all over the place, but I
guess that's the cleanest solution, right?

Simon.

[Buildroot] [PATCH] libusb: disable on avr32

[Peter Korsgaard]

>>>>> "Simon" == Simon Dawson <spdawson@gmail.com> writes:

 Simon> Hi Peter; thanks for the feedback.
 Simon> On 17 November 2012 19:58, Peter Korsgaard <jacmet@uclibc.org> wrote:
 >> What about libusb's reverse dependencies?
 >> 
 >> git grep 'select BR2_PACKAGE_LIBUSB'
 >> package/bluez_utils/Config.in:  select BR2_PACKAGE_LIBUSB
 >> package/ccid/Config.in: select BR2_PACKAGE_LIBUSB
 >> package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB
 >> package/libftdi/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB
 >> package/libhid/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/libiqrf/Config.in:      select BR2_PACKAGE_LIBUSB
 >> package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB
 >> package/libnfc/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB
 >> package/openocd/Config.in:      select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/pcsc-lite/Config.in:    select BR2_PACKAGE_LIBUSB
 >> package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB
 >> package/python-nfc/Config.in:   select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB
 >> package/usb_modeswitch/Config.in:       select BR2_PACKAGE_LIBUSB_COMPAT
 >> package/usbutils/Config.in:     select BR2_PACKAGE_LIBUSB

 Simon> Ah, yes. What's the right thing to do here? I don't really like the
 Simon> thought of that "depends on !BR2_avr32" all over the place, but I
 Simon> guess that's the cleanest solution, right?

Yes, that's the only solution we have right now.

Remember to look recursively upwards. These libusb users might be
selected by other packages.

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH] libusb: disable on avr32

[Simon Dawson]

On 17 November 2012 20:13, Peter Korsgaard <jacmet@uclibc.org> wrote:
> Yes, that's the only solution we have right now.
>
> Remember to look recursively upwards. These libusb users might be
> selected by other packages.

Okay; thanks Peter.

Simon.