[Bug ports/23967] [2.28 Regression]: New sigaction implementation breaks m68k

[adhemerval.zanella at linaro dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=23967

--- Comment #7 from Adhemerval Zanella <adhemerval.zanella at linaro dot org> ---
(In reply to John Paul Adrian Glaubitz from comment #6)
> (In reply to Adhemerval Zanella from comment #3)
> > (In reply to joseph@codesourcery.com from comment #2)
> > > Does an existing test in the glibc testsuite fail on m68k because of this 
> > > issue?  If not, we need to add a new test.
> > > 
> > 
> > I don't have access to a m68k since 2.26, so I can't really tell if these
> > fail on a real or emulated system. I am working on debian developers to get
> > a working system where I can actually check this.
> 
> Yes, we're working on this :-). I actually have one Amiga 4000 with an 68060
> up and running 24/7 which is being used as a porterbox. It's a tad bit slow
> but it runs reliably and it's being used for a lot of testing. I will create
> an account on this box for you and also send you a guide on how to install
> Debian inside Aranym and qemu-m68k-system.

I could get a Aranym with debian and kernel 3.16 working and I could verify
that at least one signal testcase indeed fails on the system:

./testrun.sh signal/tst-sigaction; echo $?                                      
error: tst-sigaction.c:46: not true: !(act.sa_flags & SA_RESETHAND)
error: 1 test failures
1

And it is fixed by using the generic kernel_sigaction.h. I will prepare a
patch.

-- 
You are receiving this mail because:
You are on the CC list for the bug.