Re: HEADS UP: macppc gcc8 fallout in -current

[Valery Ushakov <uwe%stderr.spb.ru@localhost>]

On Thu, Jan 02, 2020 at 09:59:17 -0500, Christos Zoulas wrote:

> On Jan 2,  4:09pm, uwe%stderr.spb.ru@localhost (Valery Ushakov) wrote:
> -- Subject: Re: HEADS UP: macppc gcc8 fallout in -current
> 
> | On Thu, Jan 02, 2020 at 10:37:43 +0100, Martin Husemann wrote:
> | 
> | > Program terminated with signal SIGABRT, Aborted.
> | > #0  0xfdba20c4 in std::thread::_M_start_thread(std::shared_ptr<std::thread::_Impl_base>) () from /usr/lib/libstdc++.so.9
> | 
> | Thanks!
> | 
> | So it's probably this _M_start_thread:

Hmm, actually, that's not what I see.  Note also, that it's SIGABRT,
not SIGSEGV that the tests encounter.

I grabbed gdb binary from an old comp.tgz and I see:

$ ogdb -q --args /usr/libexec/cc1 -dH empty.c 
Reading symbols from /usr/libexec/cc1...(no debugging symbols found)...done.
(gdb) r
Starting program: /usr/libexec/cc1 -dH empty.c
[New LWP 1 of process 1223]

Thread 2 received signal SIGSEGV, Segmentation fault.
0x024b2a60 in iterative_hash ()
(gdb) bt
#0  0x024b2a60 in iterative_hash ()
#1  0x01957efc in type_hash_canon_hash(tree_node*) ()
#2  0x0196ad5c in build_function_type(tree_node*, tree_node*) ()
#3  0x01870c3c in def_fn_type(c_builtin_type, c_builtin_type, bool, int, ...)
    ()
#4  0x0187c460 in c_common_nodes_and_builtins() ()
#5  0x0180e220 in c_init_decl_processing() ()
#6  0x0184137c in c_objc_common_init() ()
#7  0x0258c9d8 in toplev::main(int, char**) ()
#8  0x0258b844 in main ()
(gdb) x/i $pc
=> 0x24b2a60 <iterative_hash+480>:      lbz     r9,3(r3)
(gdb) p/x $r3
$1 = 0x7ef543c

Which is consistent with addr=0x7ef543f I got from ktrace earlier:

$ kdump | grep PSIG
   797      1 cc1      PSIG  SIGSEGV caught handler=0x1af0c5c mask=(11): code=SEGV_MAPERR, addr=0x7ef543f, trap=768)

-uwe