Re: [racket-dev] Compilation issues of racketcs with GCC7.4.0

['Paulo Matos' via Racket Developers]

On 05/06/2019 15:22, Matthew Flatt wrote:
> Thanks!
> 
> It looks like this happens only with -O3, so most users won't run into
> it (with the default -O2).
> 

Correct.

> 
> I see that changing the declaration
> 
>   uptr vspace_offsets[vspaces_count+1];
> 
> to
> 
>   uptr vspace_offsets[2*vspaces_count+1];
> 
> makes the error go away, but I don't know whether it's just a bug in
> array-bounds checking (in which case extending the array size should be
> harmless) or reflects a miscompilation (in which case extending the
> array size just masks a problem).
> 

It is afaict, a bug in the array bounds calculation. So, increasing it
masks the problem.

I have tested latest commit from the gcc 7 branch and it still has the
problem.

/home/pmatos/Projects/gcc-build/gcc/cc1 -fpreprocessed vfasl.i -quiet
-dumpbase vfasl.i -m64 -msse2 -mtune=generic -march=x86-64 -auxbase
vfasl -O3 -Wpointer-arith -Wextra -Werror -Wimplicit-fallthrough=0 -Wall
-version -o /tmp/ccYGabNR.s
GNU C11 (GCC) version 7.4.1 20190605 (x86_64-pc-linux-gnu)
        compiled by GNU C version 8.2.1 20181127, GMP version 6.1.2,
MPFR version 4.0.2, MPC version 1.1.0, isl version none
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
GNU C11 (GCC) version 7.4.1 20190605 (x86_64-pc-linux-gnu)
        compiled by GNU C version 8.2.1 20181127, GMP version 6.1.2,
MPFR version 4.0.2, MPC version 1.1.0, isl version none
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
Compiler executable checksum: 4c52f0564d8672c9a189665c732e246d
vfasl.c: In function ‘S_vfasl’:
vfasl.c:352:134: error: array subscript is above array bounds
[-Werror=array-bounds]

I will bisect this to see what's going on just out of curiosity.

Will let you know.

> At Wed, 5 Jun 2019 15:04:10 +0200, "'Paulo Matos' via Racket Developers" 
> wrote:
>> Hi,
>>
>> This is just to let you know of a GCC bug in GCC7.4.0 (default in Ubuntu
>> 18.04 LTS) when compiling racketcs.
>>
>> GCC complains about
>> vfasl.c:358:134: error: array subscript is above array bounds
>> [-Werror=array-bounds]
>>
>> on vfasl.c of Chez sources. However, this is a false positive fixed on
>> more recent GCC versions. On ubuntu 18.04:
>> racket/racket/src/ChezScheme/ta6le/c$ apt-get install -y gcc-8
>> racket/racket/src/ChezScheme/ta6le/c$ gcc  -m64 -msse2 -Wpointer-arith
>> -Wall -Wextra -Werror -Wno-implicit-fallthrough -O2 -D_REENTRANT
>> -pthread -march=native -O3 -Wall -DELF_FIND_BOOT_SECTION -pthread -c
>> -DX86_64 -I../boot/ta6le -I../zlib -I../lz4/lib vfasl.c
>> vfasl.c: In function ‘S_vfasl’:
>> vfasl.c:352:134: error: array subscript is above array bounds
>> [-Werror=array-bounds]
>>        MAYBE_FIXUP(0);
>> ...
>> cc1: all warnings being treated as errors
>> racket/racket/src/ChezScheme/ta6le/c$ gcc-8  -m64 -msse2 -Wpointer-arith
>> -Wall -Wextra -Werror -Wno-implicit-fallthrough -O2 -D_REENTRANT
>> -pthread -march=native -O3 -Wall -DELF_FIND_BOOT_SECTION -pthread -c
>> -DX86_64 -I../boot/ta6le -I../zlib -I../lz4/lib vfasl.c
>> $
>>
>> The broken gcc version:
>> $ gcc --version
>> gcc (Ubuntu 7.4.0-1ubuntu1~18.04) 7.4.0
>> Copyright (C) 2017 Free Software Foundation, Inc.
>> This is free software; see the source for copying conditions.  There is NO
>> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>>
>> Fixed:
>> $ gcc-8 --version
>> gcc-8 (Ubuntu 8.3.0-6ubuntu1~18.04) 8.3.0
>> Copyright (C) 2018 Free Software Foundation, Inc.
>> This is free software; see the source for copying conditions.  There is NO
>> warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
>>
>> The might have come in long before 8.3.0 but I haven't run further
>> tests. This is relevant to know because the bug is in the default GCC
>> for Ubuntu LTS.
>>
>> If I find more, I will let you know.
>>
>> Regards,
>> -- 
>> Paulo Matos
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "Racket Developers" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to racket-dev+unsubscr...@googlegroups.com.
>> To post to this group, send email to racket-dev@googlegroups.com.
>> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/racket-dev/72642853-8f01-61a3-033d-b38be5e712
>> 07%40linki.tools.
>> For more options, visit https://groups.google.com/d/optout.
> 

-- 
Paulo Matos

-- 
You received this message because you are subscribed to the Google Groups 
"Racket Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to racket-dev+unsubscr...@googlegroups.com.
To post to this group, send email to racket-dev@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/racket-dev/768e5035-4b54-ad47-5052-de8453ee06a1%40linki.tools.
For more options, visit https://groups.google.com/d/optout.