Re: [PATCH 4/4] testsuite, gdb.btrace: fix 32-bit PIE false positives

[Simon Marchi via Gdb-patches <gdb-patches@sourceware.org>]

On 2020-12-17 10:12 a.m., Markus Metzger via Gdb-patches wrote:
> For 32-bit position independent executables, GCC generates an extra call to
> 
>     __x86.get_pc_thunk.<reg>
> 
> which appears in the function call history.  It is correct to appear there
> but this confuses the tests, which check for an expected sequence of
> functions.
> 
> Build with -no-pie -fno-pie to avoid this complication.
> 
> gdb/testsuite/ChangeLog:
> 2020-12-04  Markus Metzger  <markus.t.metzger@intel.com>
> 
> 	* gdb.btrace/exception.exp: Build with -no-pie -fno-pie.
> 	* gdb.btrace/function_call_history.exp: Likewise.
> 	* gdb.btrace/unknown_functions.exp: Likewise.
> ---
>  gdb/testsuite/gdb.btrace/exception.exp             | 5 ++++-
>  gdb/testsuite/gdb.btrace/function_call_history.exp | 5 ++++-
>  gdb/testsuite/gdb.btrace/unknown_functions.exp     | 9 ++++++---
>  3 files changed, 14 insertions(+), 5 deletions(-)
> 
> diff --git a/gdb/testsuite/gdb.btrace/exception.exp b/gdb/testsuite/gdb.btrace/exception.exp
> index 0d242227938..b515de87b01 100755
> --- a/gdb/testsuite/gdb.btrace/exception.exp
> +++ b/gdb/testsuite/gdb.btrace/exception.exp
> @@ -22,8 +22,11 @@ if { [skip_btrace_tests] } {
>      return -1
>  }
>  
> +# We expect a specific function call history.  This gets messed up with
> +# PIE on 32-bit.
>  standard_testfile exception.cc
> -if [prepare_for_testing "failed to prepare" $testfile $srcfile {c++ debug}] {
> +if [prepare_for_testing "failed to prepare" $testfile $srcfile \
> +	{additional_flags=-fno-pie ldflags=-no-pie c++ debug}] {
>      return -1
>  }

There's the "nopie" option you can pass to gdb_compile to disable PIE.  Although from
what I see, it doesn't add -fno-pie to compilation options, just adds -no-pie to link
options.  I suppose you need both?  If so, it could be preferable in the long run to
teach gdb_compile to pass -fno-pie to the compilation step as well.

Simon