Linaro GCC Tracking

armel ICE gcc 4.5

Bug #689887 reported by Jani Monoses on 2010-12-13

This bug affects 4 people

	Status	Importance	Assigned to	Milestone
Linaro GCC	Fix Released	Medium	Richard Sandiford	Linaro GCC 4.6-2011.10
Linaro GCC Tracking	Fix Committed	Undecided	Unassigned	Linaro GCC Tracking 4.7.0
gcc	Fix Released	Medium	gcc-bugzilla #49030
gcc-4.5 (Ubuntu)	Won't Fix	Medium	Unassigned
gcc-4.6 (Debian)	Fix Released	Unknown	debbugs #633479
gcc-4.6 (Ubuntu)	Fix Released	Medium	Unassigned
libisofs (Ubuntu)	Fix Released	Low	Unassigned
traverso (Ubuntu)	Fix Released	High	Unassigned

Bug Description

Binary package hint: gcc-4.5

The traverso package FTBFS on natty because of an ICE

http://launchpadlibrarian.net/60143405/buildlog_ubuntu-natty-armel.traverso_0.49.2-2ubuntu1_FAILEDTOBUILD.txt.gz

Tags:

Related branches

lp://staging/~linaro-toolchain-dev/gcc-linaro/lp-689887-armel-ice

Rejected for merging into lp://staging/gcc-linaro/4.5

Ramana Radhakrishnan (community): Needs Fixing on 2011-04-11

Linaro Toolchain Builder: Needs Fixing on 2011-04-01

Jani Monoses (jani) on 2010-12-13

tags:

added: armel

Revision history for this message

Chung-Lin Tang (cltang) wrote on 2010-12-20:

The ICE looks familiar, can you provide a preprocessed .i file as a testcase? Thanks

Revision history for this message

Jani Monoses (jani) wrote on 2011-01-03:

Here is a small program based on the traverso code, that exhibits the error when built with -O2

gcc (Ubuntu/Linaro 4.5.2-1ubuntu2) 4.5.2

internal compiler error: in get_arm_condition_code, at config/arm/arm.c:16985

any of these changes leads to successful build
- using long instead of 'long long'
- not using the shift operator
- not wrapping in a loop

#include <limits.h>

void sample_move_d32u24_sS (char *dst, char *src, unsigned long nsamples, unsigned long dst_skip)

{
long long y;

        while (nsamples--) {
                y = (long long)(*src) << 8;
                if (y < INT_MIN) {
                        *((int *) dst) = INT_MIN;
                } else {
                        *((int *) dst) = (int)y;
                }
        }
}

Revision history for this message

Jani Monoses (jani) wrote on 2011-01-03:

An even smaller testcase

Changing INT_MIN to LONG_LONG_MIN in the conditional fixes the build. (using it this way is arguably a bug in the codebase)

#include <limits.h>

void small(char *dst)

{
long long y;

        while (1) {
                y = (long long)(*dst) << 8;
                if (y < INT_MIN) {
                        *((int *) dst) = 0;
                } else {
                        *((int *) dst) = 1;
                }
        }
}

Revision history for this message

Linaro Toolchain Builder (cbuild) wrote on 2011-01-05:

Confirmed in gcc-linaro-4.5-2010.12-0:

michaelh@pavo1:~/linaro/bugs$ /tools/toolchains/gcc-linaro-4.5-2010.12-0-armv7l-maverick-cbuild19-pavo4/bin/gcc -O2 -c traverso.c
traverso.c: In function 'small':
traverso.c:16:1: internal compiler error: in get_arm_condition_code, at config/arm/arm.c:16990

The code in question is:

case CC_NCVmode:
switch (comp_code)
{
case GE: return ARM_GE;
case LT: return ARM_LT;
case GEU: return ARM_CS;
case LTU: return ARM_CC;
default: gcc_unreachable ();
}

Michael Hope (michaelh1) on 2011-01-05

Changed in gcc-linaro:
status:	New → Triaged
importance:	Undecided → Medium

Revision history for this message

Chung-Lin Tang (cltang) wrote on 2011-01-13:

Patch posted here, awaiting review:
http://gcc.gnu.org/ml/gcc-patches/2011-01/msg00794.html

Chung-Lin Tang (cltang) on 2011-01-13

Changed in gcc-linaro:
status:	Triaged → In Progress

Revision history for this message

Jani Monoses (jani) wrote on 2011-02-17:

I see there is no upstream feedback on this. Is this blocking getting it in linaro-gcc as well?

Michael Hope (michaelh1) on 2011-02-17

Changed in gcc-linaro:
assignee:	nobody → Chung-Lin Tang (cltang)

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-02-18:

Three-week old ping here:
http://gcc.gnu.org/ml/gcc-patches/2011-01/msg01847.html

Chung-Lin, could you ping this again please?

Revision history for this message

Chung-Lin Tang (cltang) wrote on 2011-02-18:

I'll ping again when Richard E. returns.

Revision history for this message

Jani Monoses (jani) wrote on 2011-03-08:

ping?

Revision history for this message

Chung-Lin Tang (cltang) wrote on 2011-03-08:

#10

Pinged upstream yesterday: http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00308.html

Revision history for this message

Loïc Minier (lool) wrote on 2011-03-29:

#11

Also caused libisofs to FTBFS; see bug #742961.

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-04-01:

#12

The proposed fixed seems to cause a bootstrap failure when configured with --with-cpu=cortex-a9 --with-fpu=neon. See the merge request.

Revision history for this message

Chung-Lin Tang (cltang) wrote on 2011-04-17:

#13

I tried a native bootstrap of upstream trunk with the patch applied, and succeeded.
This is getting more elusive...

Revision history for this message

In GCC Bugzilla #49030, Rmansfield-t (rmansfield-t) wrote on 2011-05-17:

#19

Created attachment 24268
preprocessed source

$ ./xgcc -v
Using built-in specs.
COLLECT_GCC=./xgcc
Target: arm-unknown-linux-gnueabi
Configured with: ../configure --target=arm-unknown-linux-gnueabi --prefix=/home/ryan/x-tools/arm-unknown-linux-gnueabi --with-sysroot=/home/ryan/x-tools/arm-unknown-linux-gnueabi/arm-unknown-linux-gnueabi//sys-root --disable-multilib --with-local-prefix=/home/ryan/x-tools/arm-unknown-linux-gnueabi/arm-unknown-linux-gnueabi/sys-root --disable-nls --enable-threads=posix --enable-symvers=gnu --enable-c99 --enable-long-long --enable-target-optspace target_alias=arm-unknown-linux-gnueabi --enable-languages=c++ --disable-shared --disable-libmudflap --disable-libssp
Thread model: posix
gcc version 4.7.0 20110517 (experimental) [trunk revision 173832] (GCC)
$ ./xgcc -B. -O1 ~/ice.i
/home/ryan/ice.i: In function 'bar':
/home/ryan/ice.i:33:1: internal compiler error: in get_arm_condition_code, at config/arm/arm.c:17180
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Revision history for this message

In GCC Bugzilla #49030, Ibolton (ibolton) wrote on 2011-05-20:

#20

I haven't been able to reproduce this. Please can you include the args that are being sent to cc1.

Revision history for this message

In GCC Bugzilla #49030, Rmansfield-t (rmansfield-t) wrote on 2011-05-20:

#21

With rev173940

$ ./cc1 -fpreprocessed ice.i -quiet -O1
ice.i: In function 'bar':
ice.i:33:1: internal compiler error: in get_arm_condition_code, at config/arm/arm.c:17021
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Revision history for this message

In GCC Bugzilla #49030, Ibolton (ibolton) wrote on 2011-05-20:

#22

I was hoping the compiler command would tell me what -mcpu it was defaulting to.

Anyway, I have managed to reproduce it now, by compiling as follows:

cc1 -fpreprocessed 49030.i -quiet -O1 -o 49030.s -mcpu=arm7tdmi -marm

Note that it all works with -mthumb and/or -mcpu=cortex-a9.

I have tried this with my old toolchain and it is broken as far back as r164700.

Revision history for this message

In GCC Bugzilla #49030, Rmansfield-t (rmansfield-t) wrote on 2011-05-20:

#23

(In reply to comment #3)
> I was hoping the compiler command would tell me what -mcpu it was defaulting
> to.

I'm using an arm-unknown-linux-gnueabi configuration which defaults an arm10tdmi.

linux-eabi.h:#define SUBTARGET_CPU_DEFAULT TARGET_CPU_arm10tdmi

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-06-09:

#14

For reference, this test fails in gcc-linaro-{4.5,4.6}-2011.05-0, and gcc trunk r174795.

It succeeds in gcc-4.6.0 and gcc-4.5.3.

All 4.4 based compilers are fine.

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-06-09:

#15

The work-around is to compile at -O0 or -O3. -O1 and -O2 fail.

Revision history for this message

In GCC Bugzilla #49030, Michael Hope (michaelh1) wrote on 2011-06-09:

#24

The same assert is seen in https://bugs.launchpad.net/ubuntu/+source/gcc-4.5/+bug/689887 in Thumb-2 mode on trunk r174795.

Chung-Lin posted a patch at http://gcc.gnu.org/ml/gcc-patches/2011-03/msg00790.html but there are bootstrap issues.

Adding -fno-if-conversion works around the problem.

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-06-09:

#16

Related: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49030

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-06-09:

#17

Adding -fno-if-conversion works around the problem.

Revision history for this message

In GCC Bugzilla #49030, Mikpe (mikpe) wrote on 2011-06-10:

#25

It's caused by r161764:

Author: sandra
Date: Sat Jul 3 01:00:37 2010
New Revision: 161764

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=161764
Log:
2010-07-02 Daniel Jacobowitz <email address hidden>
Julian Brown <email address hidden>
Sandra Loosemore <email address hidden>

gcc/
* config/arm/arm.c (arm_canonicalize_comparison): Canonicalize DImode
comparisons. Adjust to take both operands.
(arm_select_cc_mode): Handle DImode comparisons.
(arm_gen_compare_reg): Generate a scratch register for DImode
comparisons which require one. Use xor for Thumb equality checks.
(arm_const_double_by_immediates): New.
(arm_print_operand): Allow 'Q' and 'R' for constants.
(get_arm_condition_code): Handle new CC_CZmode and CC_NCVmode.
...

Revision history for this message

In GCC Bugzilla #49030, Mikpe (mikpe) wrote on 2011-06-12:

#26

Created attachment 24503
more reduced test case, ICEs at -O1 for armv5te and armv6

Using this slightly more reduced test case I've found that gcc transforms

(set (reg:CC_NCV ...) (compare:CC_NCV (reg:DI ...) (const_int 1 ...)))
(if_then_else (ge (reg:CC_NCV ...) ...) ...)

(set (reg:CC_NCV ...) (compare:CC_NCV (reg:DI ...) (const_int 0 ...)))
(if_then_else (gt (reg:CC_NCV ...) ...)

because ifcvt calls noce_get_condition which calls canonicalize_condition, and the latter has a general rule to transform (GE x const) to (GT x const-1).

So the careful code in e.g. cbranchdi4 to not generate unimplemented comparisons is neutralized by generic code.

Do DImode comparisons really have to be asymmetric? I.e., could CC_NCV be removed? I see that LT is apparently supported so why can't GT be too?

Revision history for this message

In GCC Bugzilla #49030, Mikpe (mikpe) wrote on 2011-06-12:

#27

I have a preliminary patch which creates a clone of *arm_cmpdi_insn that does a reverse comparison, via a new CC mode for must-be-reversed comparisons, and corresponding updates to arm_select_cc_mode and get_arm_condition_code. It's the same transformation from eg (GT x y) to (LT y x) that cstoredi4 does, only later so it can handle problematic comparisons introduced by the middle-end.

This fixes the ICEs on the two test cases in this PR. I'll do a full bootstrap and regtest as soon as my test machine is done with its current bootstrap and regtest cycle.

I'll send the FSF Yet Another Ping about my employer's copyright disclaimer first thing tomorrow. They've had the papers since January or February...

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-07-11:

#18

Note that Mikael Pettersson has a patch against the upstream ticket but has copyright assignment issues.

Changed in gcc-linaro:
status:	In Progress → Triaged
assignee:	Chung-Lin Tang (cltang) → nobody

Matthias Klose (doko) on 2011-07-15

Changed in gcc-4.5 (Ubuntu):
importance:	Undecided → Medium
status:	New → Triaged
Changed in gcc-4.6 (Ubuntu):
importance:	Undecided → Medium
status:	New → Triaged

Bug Watch Updater (bug-watch-updater) on 2011-08-11

Changed in gcc-4.6 (Debian):
status:	Unknown → Confirmed

Richard Sandiford (rsandifo) on 2011-09-02

Changed in gcc-linaro:
assignee:	nobody → Richard Sandiford (rsandifo)

Bug Watch Updater (bug-watch-updater) on 2011-09-05

Changed in gcc:
importance:	Unknown → Medium
status:	Unknown → Confirmed

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-09-07:

#28

Related to https://bugs.launchpad.net/ubuntu/+source/oss4/+bug/809761

Revision history for this message

In GCC Bugzilla #49030, Rsandifo-gcc (rsandifo-gcc) wrote on 2011-09-07:

#29

Author: rsandifo
Date: Wed Sep 7 13:48:03 2011
New Revision: 178636

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=178636
Log:
gcc/
PR target/49030
* config/arm/arm-protos.h (maybe_get_arm_condition_code): Declare.
* config/arm/arm.c (maybe_get_arm_condition_code): New function,
reusing the old code from get_arm_condition_code. Return ARM_NV
for invalid comparison codes.
(get_arm_condition_code): Redefine in terms of
maybe_get_arm_condition_code.
* config/arm/predicates.md (arm_comparison_operator): Use
maybe_get_arm_condition_code.

gcc/testsuite/
PR target/49030
* gcc.dg/torture/pr49030.c: New test.

Added:
    trunk/gcc/testsuite/gcc.dg/torture/pr49030.c
Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/arm/arm-protos.h
    trunk/gcc/config/arm/arm.c
    trunk/gcc/config/arm/predicates.md
    trunk/gcc/testsuite/ChangeLog

Revision history for this message

In GCC Bugzilla #49030, Rmansfield-t (rmansfield-t) wrote on 2011-09-07:

#30

I am out of the office until September 19th. For emergencies please contact Fred Plante.

Michael Hope (michaelh1) on 2011-09-08

Changed in gcc-linaro:
milestone:	none → 4.6-2011.09
status:	Triaged → In Progress

Revision history for this message

In GCC Bugzilla #49030, Gjl (gjl) wrote on 2011-09-09:

#31

Author: gjl
Date: Fri Sep 9 17:00:26 2011
New Revision: 178736

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=178736
Log:
PR target/49030
* gcc.dg/torture/pr49030.c: Run only if target int32plus.

Modified:
trunk/gcc/testsuite/ChangeLog
trunk/gcc/testsuite/gcc.dg/torture/pr49030.c

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-09-14:

#32

Confirmed that the fix fixes the original test case.

Changed in gcc-linaro:
status:	In Progress → Fix Committed

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-09-16:

#33

The fix was reverted due to LP: #850099

Changed in gcc-linaro:
status:	Fix Committed → In Progress
milestone:	4.6-2011.09 → 4.6-2011.10

Matthias Klose (doko) on 2011-09-27

tags:	added: ftbfs oneiric
Changed in traverso (Ubuntu):
importance:	Undecided → High
status:	New → Confirmed

Revision history for this message

Michael Hope (michaelh1) wrote on 2011-09-27:

#34

Note the follow on patch that fixes LP: #850099 is here:
http://gcc.gnu.org/ml/gcc-patches/2011-09/msg01128.html

Richard will be backporting both to gcc-linaro in due course.

Michael Hope (michaelh1) on 2011-10-10

Changed in gcc-linaro:
status:	In Progress → Fix Committed

Revision history for this message

Loïc Minier (lool) wrote on 2011-10-10:

#35

Once the gcc-linaro fix lands in Ubuntu, we should revert the -O0 workaround in libisofs and retry a traverso build.

Changed in libisofs (Ubuntu):
status:	New → Confirmed
importance:	Undecided → Low

Michael Hope (michaelh1) on 2011-10-13

Changed in gcc-linaro:
status:	Fix Committed → Fix Released

Revision history for this message

Matthias Klose (doko) wrote on 2011-12-03:

#36

won't fix anymore in 4.5

Changed in gcc-4.5 (Ubuntu):
status:	Triaged → Won't Fix

Revision history for this message

Matthias Klose (doko) wrote on 2011-12-03:

#37

fixed in gcc-4.6 (precise)

Changed in gcc-4.6 (Ubuntu):
status:	Triaged → Fix Released

Revision history for this message

Matthias Klose (doko) wrote on 2011-12-04:

#38

traverso did build in precise

Changed in traverso (Ubuntu):
status:	Confirmed → Fix Released

Revision history for this message

Matthias Klose (doko) wrote on 2011-12-04:

#39

libisofs 1.1.6-1ubuntu1 built in precise without work-around

Changed in libisofs (Ubuntu):
status:	Confirmed → Fix Released

Bug Watch Updater (bug-watch-updater) on 2011-12-09

Changed in gcc-4.6 (Debian):
status:	Confirmed → Fix Released

Revision history for this message

In GCC Bugzilla #49030, Rsandifo-gcc (rsandifo-gcc) wrote on 2012-01-18:

#40

Fixed on trunk, and not marked as a regression.

Bug Watch Updater (bug-watch-updater) on 2012-01-18

Changed in gcc:
status:	Confirmed → Fix Released

Revision history for this message

Andrew Stubbs (ams-codesourcery) wrote on 2012-02-15:

#41

Related: lp:gcc-linaro/4.6,revno=106819

Changed in gcc-linaro-tracking:
milestone:	none → 4.7.0
status:	New → Fix Committed

Report a bug

This report contains Public information

Everyone can see this information.

Duplicates of this bug

You are

Subscribing...

Edit bug mail

Other bug subscribers

Remote bug watches

debbugs #633479
[done serious upstream patch] Edit
gcc-bugzilla #49030
[RESOLVED FIXED] Edit

Bug watches keep track of this bug in other bug trackers.