Next release of uclibc

[Chris Zankel]

Hi,

It would be great if I could add the Xtensa patch series that's
currently only in the master branch to this release. I 'cherry picked'
those patches into the 0.9.33 branch and it builds and runs without
problems.

That series of patches affect almost only files in the 'xtensa'
directory. The only change to a common file is to utils/ldd.c which adds
the ELFCLASSM definition for Xtensa, so looks harmless to other
architectures (see below).

What do you guys think?

Thanks,
-Chris

diffstat of all changes:

 libc/sysdeps/linux/xtensa/bits/atomic.h                 |  233
++++++++++++++++
 libc/sysdeps/linux/xtensa/bits/sigcontext.h             |   41 ++
 libc/sysdeps/linux/xtensa/setjmp.S                      |  150 +++++-----
 libc/sysdeps/linux/xtensa/sys/user.h                    |   20 +
 libc/sysdeps/linux/xtensa/sysdep.h                      |    2
 libpthread/linuxthreads.old/sysdeps/xtensa/pt-machine.h |   42 ++
 test/math/libm-test-ulps-xtensa                         |  129 ++++++++
 utils/ldd.c                                             |    5
 8 files changed, 553 insertions(+), 69 deletions(-)

The change to a file shared by other architectures:

--- a/utils/ldd.c
+++ b/utils/ldd.c
@@ -106,6 +106,11 @@
 #define ELFCLASSM      ELFCLASS32
 #endif
 
+#if defined(__xtensa__)
+#define MATCH_MACHINE(x) (x == EM_XTENSA)
+#define ELFCLASSM      ELFCLASS32
+#endif
+
 #ifndef MATCH_MACHINE
 # ifdef __linux__
 #  include <asm/elf.h>


On 01/23/2013 03:44 AM, Bernhard Reutner-Fischer wrote:
> On 23 January 2013 12:27, Florian Fainelli <florian at openwrt.org> wrote:
>> Bernhard,
>>
>> Do you want to backport more fixes into 0.9.33.3 or do you think it is ready
>> for a release now?
> I have pushed these 3 additional patches from master onto the 0.9.33
> branch, fwiw:
> 641a5356a021f90ee922229bd8e1aa6eafe152bc mount.h: update
> 8eccce991d08960d135b97066621c8d3248a79b7 socket.h: pull socket_type.h
> from eglibc
> 479f8407c4822d2b872afb8bb14e5ab596714744 statfs: support f_frsize
>
> It would be good to diagnose and fix the pread_pwrite thing that
> Natanael mentioned to have seen on x86_64.
> Volunteers?
>
> I will take care of changing MAP_UNINITIALIZE -> MAP_UNINITIALIZED on
> master and the release-branch ASAP.
> _______________________________________________
> uClibc mailing list
> uClibc at uclibc.org
> http://lists.busybox.net/mailman/listinfo/uclibc