This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
RE: [PING] [RFC] Sanitize rtx_addr_can_trap_p_1
- From: Bernd Edlinger <bernd dot edlinger at hotmail dot de>
- To: "gcc-patches at gcc dot gnu dot org" <gcc-patches at gcc dot gnu dot org>
- Cc: Richard Biener <richard dot guenther at gmail dot com>, Jakub Jelinek <jakub at redhat dot com>, Jeff Law <law at redhat dot com>, Eric Botcazou <ebotcazou at adacore dot com>
- Date: Wed, 1 Jul 2015 14:31:41 +0200
- Subject: RE: [PING] [RFC] Sanitize rtx_addr_can_trap_p_1
- Authentication-results: sourceware.org; auth=none
- References: <DUB118-W34295B153A59B9B953311CE4BC0 at phx dot gbl>,<DUB118-W12AB87BBFF140A5FAE7090E4B80 at phx dot gbl>
Hi,
the bogus offsets-issue is now entered in bugzilla, see PR66614.
It is however only a very minor issue, and does probably have
no impact on the generated code at all.
How should I continue with the rtx_addr_can_trap-patch?
Is it OK to commit?
Thanks
Bernd.
----------------------------------------
> From: bernd.edlinger@hotmail.de
> To: gcc-patches@gcc.gnu.org
> CC: richard.guenther@gmail.com; jakub@redhat.com; law@redhat.com; ebotcazou@adacore.com
> Subject: RE: [RFC] Sanitize rtx_addr_can_trap_p_1
> Date: Mon, 15 Jun 2015 05:50:46 +0200
>
> Hi,
>
> I have here an updated patch, which improves two things:
>
> First it moves debug code out to an extra patch, as suggested by Jakub.
>
> Secondly, it fixes the checks on STACK_GROWS_DOWNWARD and
> ARGS_GROW_DOWNWARD. Previously these used to be conditionally defined
> symbols, but recently they were changed to be always defined, but with 0 or 1.
>
> That made all #ifndef checks on those two flags work the wrong way, and it caused
> most of the false positives in the previous version.
>
> Now the number of false positives in the stage2 drops significantly to only 4 new ones:
>
> *** argp can trap: function=uw_init_context_1, pass=reload, offset=236, size=4, low_bound=-16, high_bound=16
> *** argp can trap: function=uw_init_context_1, pass=reload, offset=236, size=4, low_bound=-16, high_bound=16
> *** argp can trap: function=uw_init_context_1, pass=reload, offset=440, size=8, low_bound=-16, high_bound=16
> *** argp can trap: function=uw_init_context_1, pass=reload, offset=440, size=8, low_bound=-16, high_bound=16
>
> These came from libgcc/unwind-dw2.c
>
> They seem to have the same reason as the 2930 "frame can trap" warnings that were already
> there before the patch.
>
>
> All these seem to happen due to some hidden bug. In the debugger the call stack looks always like this:
>
> #0 rtx_addr_can_trap_p_1 (x=0x7ffff6ecb378, offset=440, size=8, mode=DImode, unaligned_mems=false) at ../../gcc-trunk/gcc/rtlanal.c:671
> #1 0x0000000000d90f4a in rtx_addr_can_trap_p_1 (x=0x7ffff67ac7f8, offset=0, size=8, mode=DImode, unaligned_mems=false) at ../../gcc-trunk/gcc/rtlanal.c:699
> #2 0x0000000000d953c4 in may_trap_p_1 (x=0x7ffff67ac810, flags=0) at ../../gcc-trunk/gcc/rtlanal.c:2760
> #3 0x0000000000d95619 in may_trap_p_1 (x=0x7ffff671ac90, flags=0) at ../../gcc-trunk/gcc/rtlanal.c:2838
> #4 0x0000000000d956cc in may_trap_p (x=0x7ffff671ac90) at ../../gcc-trunk/gcc/rtlanal.c:2857
> #5 0x0000000000c6b2ab in process_bb_lives (bb=0x7ffff6c3a958, curr_point=@0x7fffffffd8e4: 23, dead_insn_p=true) at ../../gcc-trunk/gcc/lra-lives.c:698
> #6 0x0000000000c6d19a in lra_create_live_ranges_1 (all_p=true, dead_insn_p=true) at ../../gcc-trunk/gcc/lra-lives.c:1262
> #7 0x0000000000c6d47c in lra_create_live_ranges (all_p=true, dead_insn_p=true) at ../../gcc-trunk/gcc/lra-lives.c:1327
> #8 0x0000000000c4a253 in lra (f=0x24f7940) at ../../gcc-trunk/gcc/lra.c:2309
> #9 0x0000000000bf3d25 in do_reload () at ../../gcc-trunk/gcc/ira.c:5401
> #10 0x0000000000bf40d3 in (anonymous namespace)::pass_reload::execute (this=0x23fce20) at ../../gcc-trunk/gcc/ira.c:5572
> #11 0x0000000000d08e37 in execute_one_pass (pass=0x23fce20) at ../../gcc-trunk/gcc/passes.c:2359
> #12 0x0000000000d09081 in execute_pass_list_1 (pass=0x23fce20) at ../../gcc-trunk/gcc/passes.c:2412
> #13 0x0000000000d090b2 in execute_pass_list_1 (pass=0x23fbda0) at ../../gcc-trunk/gcc/passes.c:2413
> #14 0x0000000000d090ef in execute_pass_list (fn=0x7ffff7044c78, pass=0x23f8bc0) at ../../gcc-trunk/gcc/passes.c:2423
> #15 0x00000000008de80c in cgraph_node::expand (this=0x7ffff6c09498) at ../../gcc-trunk/gcc/cgraphunit.c:1937
> #16 0x00000000008dee3d in expand_all_functions () at ../../gcc-trunk/gcc/cgraphunit.c:2073
> #17 0x00000000008df954 in symbol_table::compile (this=0x7ffff6ecf000) at ../../gcc-trunk/gcc/cgraphunit.c:2426
> #18 0x00000000008dfb68 in symbol_table::finalize_compilation_unit (this=0x7ffff6ecf000) at ../../gcc-trunk/gcc/cgraphunit.c:2513
> #19 0x0000000000e094ba in compile_file () at ../../gcc-trunk/gcc/toplev.c:580
> #20 0x0000000000e0b9fa in do_compile () at ../../gcc-trunk/gcc/toplev.c:2070
> #21 0x0000000000e0bc46 in toplev::main (this=0x7fffffffdc50, argc=35, argv=0x7fffffffdd58) at ../../gcc-trunk/gcc/toplev.c:2171
> #22 0x00000000016b656d in main (argc=35, argv=0x7fffffffdd58) at ../../gcc-trunk/gcc/main.c:39
>
> I cannot find any instruction in the rtl dumps that corresponds to this large argp offset. So I think there must be
> something wrong along the call stack above, which looks identically even on the bogus frame pointer references.
>
>
> Is this patch OK for trunk now?
>
> At least Jakub and I are in favour of it, Eric is against it.
> That makes 2:1 ...
>
>
> Thanks
> Bernd.
>