Handle -fstack-protector prologue code

[Ulrich Weigand]

FAIL: gdb.mi/gdb792.exp: list children of class A
FAIL: gdb.mi/gdb792.exp: list children of A.public
FAIL: gdb.mi/gdb792.exp: list children of A.private
FAIL: gdb.mi/gdb792.exp: list children of A.protected
FAIL: gdb.mi/gdb792.exp: list children of A.protected.b
FAIL: gdb.mi/gdb792.exp: list children of A.protected.b.public
FAIL: gdb.mi/gdb792.exp: list children of A.protected.b.private
FAIL: gdb.mi/gdb792.exp: list children of class C

Needs further analysis.

[Yao Qi (yao-codesourcery)]

Buildlogs on https://launchpad.net/ubuntu/+source/gdb/7.1-1ubuntu4 shows this case fails on armel/x86/amd64.

[Ulrich Weigand (uweigand)]

The Ubuntu compiler will automatically build everything (including this testcase) with -fstack-protector.
It seems GDB prologue parsing gets confused by stack protector code both on i386 and ARM ...

[Michael Hope (michaelh1)]

Note that the testsuite is run with the latest Linaro compiler which may not (does not?) enable stack protector or fortify.

See the head of the configure log, such as:
 http://ex.seabright.co.nz/build/gcc-linaro-4.5-2010.10-0/logs/armv7l-maverick-cbuild11-pavo4/gcc-configure.txt

for the current configure flags.

The gdb-maverick-testsuite run is built using the Ubuntu-supplied host compiler and will have both.

[Ulrich Weigand (uweigand)]

Good point. And indeed, if you look at the test run with the Linaro compiler, these tests succeed:
http://ex.seabright.co.nz/build/gdb-linaro-7.2-2010.10-0/logs/armv7l-maverick-cbuild11-pavo1/gdb-testsuite.txt
while in the test run with the Maverick compiler, they fail:
http://ex.seabright.co.nz/build/gdb-linaro-7.2-2010.10-0/logs/armv7l-maverick-cbuild11-pavo1/gdb-maverick-testsuite.txt

[Yao Qi (yao-codesourcery)]

Patch for ARM is submitted upstreams. http://sourceware.org/ml/gdb-patches/2010-12/msg00110.html

[Yao Qi (yao-codesourcery)]

Patch for ARM is approved upstreams and committed. Will backport to Linaro tree.

Patch for i386 is submitted upstreams. http://sourceware.org/ml/gdb-patches/2010-12/msg00451.html

[Ulrich Weigand (uweigand)]

These tests still fail for me on current Natty ... Investigating.

[Ulrich Weigand (uweigand)]

OK, there was an overly strict check in arm_skip_stack_protector that caused the fix to fail on my system.

Upstream fix checked in here:
http://sourceware.org/ml/gdb-patches/2011-02/msg00047.html

Fix also checked into Linaro GDB 7.2 for the 2011.02 release.