[Bug lto/56061] New: [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[Bug lto/56061] New: [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

             Bug #: 56061
           Summary: [4.8 Regression] ICE in lto1 (in inline_call, at
                    ipa-inline-transform.c:267)
    Classification: Unclassified
           Product: gcc
           Version: 4.8.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: d.g.gorbachev@gmail.com


Created attachment 29231
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=29231
Testcase

GCC 4.8.0 20130120 (experimental).

lto1: internal compiler error: in inline_call, at ipa-inline-transform.c:267
0x8367899 inline_call(cgraph_edge*, bool, vec<cgraph_edge*, va_heap, vl_ptr>*,
int*, bool)
    ../../gcc-4.8/gcc/ipa-inline-transform.c:263
0x8357d10 inline_small_functions
    ../../gcc-4.8/gcc/ipa-inline.c:1613
0x835817f ipa_inline
    ../../gcc-4.8/gcc/ipa-inline.c:1794
Please submit a full bug report,

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.8.0

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-01-28
                 CC|                            |hubicka at gcc dot gnu.org
     Ever Confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> 2013-01-28 14:35:29 UTC ---
Confirmed.  Compile-time -O1 vs. link/WPA-time -O0 (I don't think we
necessarily
want to support this fully... we do want to support compile-time -O0 vs.
link-time -O1)

Not sure why size estimates go wrong in this case though.

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #2 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2013-01-30 10:54:22 UTC ---
(In reply to comment #1)

> I don't think we necessarily want to support this fully...

Are such bug reports useful, or it's just a nuisance to you, gcc devs? :)

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> 2013-01-30 11:21:11 UTC ---
(In reply to comment #2)
> (In reply to comment #1)
> 
> > I don't think we necessarily want to support this fully...
> 
> Are such bug reports useful, or it's just a nuisance to you, gcc devs? :)

They are useful.  It should be technically possible to support -O1 vs. -O0,
and if not, we have means to forcefully enable -O1 at link-time (which we
should do then).

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Manuel López-Ibáñez <manu at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |manu at gcc dot gnu.org

--- Comment #3 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2013-01-30 11:20:39 UTC ---
Does it make sense to allow "-O0 -flto" at all?

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> 2013-01-30 11:25:50 UTC ---
(In reply to comment #3)
> Does it make sense to allow "-O0 -flto" at all?

The classical example why we want to support this is a static library
which you'd compile and install with -O0 -g -fno-fat-lto-objects.

At link time you can then either get a fully optimized program
(use -On) or a fully debug enabled program (use -O0 -g).

The other way around, compiling and installing with -O2 but then
at link time use -O0 -g to get a debug build is more questionable
(and most of the time a user error, unless -O0 was explicit at the
link command line).

Thus, eventually we'd want to provide an optimization level for link-time
if there was none specified by doing some magic in combining the optimization
levels used at compile-time.  Possible combinations are endless, thus
nobody has yet come up with a formal description of what would happen at
link-time.

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[manu at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #6 from Manuel López-Ibáñez <manu at gcc dot gnu.org> 2013-01-30 11:29:02 UTC ---
(In reply to comment #3)
> Does it make sense to allow "-O0 -flto" at all?

Answering myself, the docs have this example:

Additionally, the optimization flags used to compile individual files are not
necessarily related to those used at link time. For instance,

          gcc -c -O0 -flto foo.c
          gcc -c -O0 -flto bar.c
          gcc -o myprog -flto -O3 foo.o bar.o


This produces individual object files with unoptimized assembler code, but the
resulting binary myprog is optimized at -O3. If, instead, the final binary is
generated without -flto, then myprog is not optimized.

However, I still don't see the point of "-O0 -flto" at link-time. We should
either force it to be at least -O1 or (in my opinion better) give an error.

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #7 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2013-01-30 13:39:37 UTC ---
> The other way around, compiling and installing with
> -O2 but then at link time use -O0 -g to get a debug
> build is more questionable

> However, I still don't see the point of "-O0 -flto"
> at link-time. We should either force it to be at
> least -O1 or (in my opinion better) give an error.

I think it could make sense. For example, I compile a certain static library
with `-O3 -flto', without `-g' (it's considered already debugged) and without
`-fno-fat-lto-objects' (so it can be used either with or without LTO); it is
shared between several programs. The programs have their own static libs, which
are compiled with `-g -O1 -flto -fno-fat-lto-objects' (could be `-O0', but
doesn't work, PR55102). Then, they are linked using `-g -O0 -flto' to get a
(partially) debug enabled program, or with `-O3 -flto' for a fully optimized
program.

> the docs have this example

Yes, but there is an aforementioned bug 55102.

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[hubicka at ucw dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #8 from Jan Hubicka <hubicka at ucw dot cz> 2013-01-30 17:16:25 UTC ---
> They are useful.  It should be technically possible to support -O1 vs. -O0,
> and if not, we have means to forcefully enable -O1 at link-time (which we
> should do then).
I personally lean towards making all early passes -O level agnostic and
prohibit
mixing of LTO -O0 and optimized objects..
I do not see much of other options for making LTO .o files ready for
distribution
in libraries in longer run.

Honza

[Bug lto/56061] [4.8 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug lto/56061] [4.8/4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.0                       |4.8.1

--- Comment #9 from Jakub Jelinek <jakub at gcc dot gnu.org> 2013-03-22 14:42:45 UTC ---
GCC 4.8.0 is being released, adjusting target milestone.

[Bug lto/56061] [4.8/4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.1                       |4.8.2

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.8.1 has been released.

[Bug lto/56061] [4.8/4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.2                       |4.8.3

--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.8.2 has been released.

[Bug lto/56061] [4.8/4.9/4.10 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.3                       |4.8.4

--- Comment #12 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 4.8.3 is being released, adjusting target milestone.

[Bug lto/56061] [4.8/4.9/5 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.4                       |4.8.5

--- Comment #13 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.8.4 has been released.

[Bug lto/56061] [4.8/4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.8/4.9/5 Regression] ICE  |[4.8/4.9 Regression] ICE in
                   |in lto1 (in inline_call, at |lto1 (in inline_call, at
                   |ipa-inline-transform.c:267) |ipa-inline-transform.c:267)

--- Comment #14 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
This is fixed on mainline by streaming optimization nodes.

[Bug lto/56061] [4.8/4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.8.5                       |4.9.3

--- Comment #15 from Richard Biener <rguenth at gcc dot gnu.org> ---
The gcc-4_8-branch is being closed, re-targeting regressions to 4.9.3.

[Bug lto/56061] [4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

--- Comment #16 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 4.9.3 has been released.

[Bug lto/56061] [4.9 Regression] ICE in lto1 (in inline_call, at ipa-inline-transform.c:267)

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=56061

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|4.9.3                       |4.9.4