[PATCH] Post-b34 Fixes

[Andrew Hughes]

This is a set of minor fixes we found for issues with the b34 release.
They were too late for the b34 release, but were included in IcedTea 1.13.6.

The webrev is:

http://cr.openjdk.java.net/~andrew/openjdk6/secfix/

[ignore the .hgtags change; it's because a merge was necessary as these
changes pre-date b34 being pushed and tagged]

Changes:

* OPENJDK6-51: Sync patch for 4873188 with 7 version

The backport of 4873188 (TLS 1.1 support) accidentally incorporated part
of 7031830 (see below) and placed a chunk of code in a slightly different
position to where it is in 7. Neither issue results in a problem, but
this changeset just makes the diff between 6 & 7 smaller.

* 7031830: bad_record_mac failure on TLSv1.2 enabled connection with SSLEngine

We finish the backport of 7031830 which was partially started in b34.
This fix makes sure that byte buffers work with the TLS provider when
the offset is != 0. The TLSv1.2 reference can be ignored as the issue
is present with code introduced for TLSv1.1.

* OPENJDK6-53: strict-aliasing warnings issued on PPC32

For some reason, -fno-strict-aliasing is not passed to gcc on ppc32.
This was recently fixed in 9 and this does the same for 6. We also
move the PPC32-only -fsigned-char addition to CFLAGS_REQUIRED as it
is in 7 on (that change is part of a huge changeset for supporting
the proprietary Oracle PPC32 & ARM32 ports which we don't want to
backport as a whole)

* OPENJDK6-52: Application of 6786276 introduces compatibility issue

Serbia & Montenegro split apart in 2006, but, prior to b34, getISOLocales()
would still return the ISO code of the country, "CS". This changed
when we brought in 6786276, which was combined with 6627549 in one
changeset. 6627549 is marked as being backported to the proprietary
Oracle 6 JDK but 6786276 is not, so we revert it from OpenJDK 6 to
retain compatibility.

* 6730740: Fix for 6729881 has apparently broken several 64 bit tests:  "Bad address"

Omair recommended this change after 6729881 was part of the networking
changes in b34.

* OPENJDK6-54: 6911104 reintroduces test fragment removed in existing 6964018 backport

In 7, 6911104 was applied prior to 6964018. In 6, the changes were backported
in the opposite order, leading to the backport of 6911104 introducing a
CYGWIN fragment which is removed by 6964018 in 7, but not in 6 (as it didn't
exist). This removes the duplicate block.

Ok to push?
-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

PGP Key: ed25519/35964222 (hkp://keys.gnupg.net)
Fingerprint = 5132 579D D154 0ED2 3E04  C5A0 CFDA 0F9B 3596 4222

PGP Key: rsa4096/248BDC07 (hkp://keys.gnupg.net)
Fingerprint = EC5A 1F5E C0AD 1D15 8F1F  8F91 3B96 A578 248B DC07