On Wed, Jul 02, 2025 at 11:43:13AM +0200, Stefan Schulze Frielinghaus wrote:
> On Tue, Jul 01, 2025 at 06:33:12PM +0200, Jakub Jelinek wrote:
> > On Tue, Jul 01, 2025 at 03:47:53PM +0200, Stefan Schulze Frielinghaus wrote:
> > > In the past years I have started to use more and more function body
> > > checks whenever gcc emits optimal code for a function. With that I
> > > wanted to make sure that we do not regress like introducing unnecessary
> > > extends or whatever which might not have been caught by only testing the
> > > "interesting"/actual part of a patch. Thus, as long as those function
> > > body checks are stable enough, i.e., not subject to insn reordering or
> > > the like, I would like to make use of them in the future, too. That
> > > being said I'm wondering whether it would make sense to automatically
> > > add option -fno-stack-protector for tests which make use function-body
> > > checks? If the testsuite infrastructure doesn't provide this
> > > functionality trivially, I will try to keep this in mind and always add
> > > the option manually.
> >
> > I think even better would be to make the check-function-body UNSUPPORTED
> > if there is no explicit -f{,no-}stack-protector* among
> > dg-options/dg-additional-options and the option is still used.
> > Because the test can be also dg-do run and it might be useful to run the
> > test.
>
> So maybe something along the lines?
>
> diff --git a/gcc/testsuite/lib/scanasm.exp b/gcc/testsuite/lib/scanasm.exp
> index 97935cb23c3..bb16ae897c1 100644
> --- a/gcc/testsuite/lib/scanasm.exp
> +++ b/gcc/testsuite/lib/scanasm.exp
> @@ -1042,6 +1042,15 @@ proc check-function-bodies { args } {
> # The name might include a list of options; extract the file name.
> set filename [lindex $testcase 0]
>
> + global compiler_flags
> + set current_compiler_flags [current_compiler_flags]
> + if { [string match "* -fstack-protector* *" " ${compiler_flags} "]
> + && ![string match "* -fstack-protector* *" "
> ${current_compiler_flags} "]
> + && ![string match "* -fno-stack-protector* *" "
> ${current_compiler_flags} "] } {
> + unsupported "$testcase: skip check-function-bodies due to stack
> protector"
> + return
> + }
> +
> global srcdir
> set input_filename "$srcdir/$filename"
> set output_filename "[file rootname [file tail $filename]]"
>
> I'm pretty new to tcl and didn't do extensive testing but for my few
> experiments it worked so far. I guess `string match` uses globbing so
> something like "* -f(no-)?stack-protector* *" doesn't work which is why
> I used two matches.
Yes. Though maybe it could/should include other options known to change
the prologue/epilogue substantially, I think e.g. -fstack-clash-protection
or -fhardened or -fstack-check* might be other candidates.
I admit I haven't looked out what compiler_flags and current_compiler_flags
are at that point, which one is the effective command line after all
RUNTESTFLAGS and dg-options/dg-additional-options options and what is just
from the latter two.
And we'd need to document it.
That said, I think some distros (Debian?) configure with -fhardened on by
default or similar options, not sure if those tests fail in all those cases
and if those distros have been just ignoring it or what.
Jakub
