Re: Warning: conditional outside an IT block for Thumb.

public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed

From: Richard Earnshaw <Richard.Earnshaw@foss.arm.com>
To: Mathieu Malaterre <mathieu.malaterre@gmail.com>
Cc: gcc-help <gcc-help@gcc.gnu.org>
Subject: Re: Warning: conditional outside an IT block for Thumb.
Date: Thu, 25 Aug 2022 15:09:41 +0100	[thread overview]
Message-ID: <6292c9e1-cbc7-e16b-e368-6ab37d2426e8@foss.arm.com> (raw)
In-Reply-To: <CA+7wUsxpXR0sYeJQZV70km4FuTHGtJdsoE+px1T0YGATM9D10w@mail.gmail.com>



On 25/08/2022 14:58, Mathieu Malaterre wrote:
> On Thu, Aug 25, 2022 at 3:48 PM Mathieu Malaterre
> <mathieu.malaterre@gmail.com> wrote:
>>
>> Hi Richard,
>>
>> Thanks for the kind help.
>>
>> On Thu, Aug 25, 2022 at 12:28 PM Richard Earnshaw
>> <Richard.Earnshaw@foss.arm.com> wrote:
>>>
>>>
>>>
>>> On 25/08/2022 07:40, Mathieu Malaterre via Gcc-help wrote:
>>>> Hi all,
>>>>
>>>> I am seeing a flood of warnings when compiling openvdb+tbb on Debian/armel:
>>>>
>>>> [...]
>>>>    [  2%] Building CXX object
>>>> openvdb/openvdb/CMakeFiles/openvdb_shared.dir/io/Archive.cc.o
>>>>    cd /<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb &&
>>>> /usr/bin/c++ -DBOOST_ALL_NO_LIB -DBOOST_IOSTREAMS_DYN_LINK
>>>> -DOPENVDB_DLL -DOPENVDB_PRIVATE -DOPENVDB_USE_LOG4CPLUS
>>>> -Dopenvdb_shared_EXPORTS -I/<<PK
>>>> GBUILDDIR>>/openvdb/openvdb/..
>>>> -I/<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb
>>>> -I/<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb/openvdb
>>>> -I/<<PKGBUILDDIR>>/openvdb/openvdb/. -I/usr/include/Imat
>>>> h -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong
>>>> -Wformat -Werror=format-security -Wno-psabi -fvisibility=hidden
>>>> -fvisibility-inlines-hidden -Wa,-mimplicit-it=thumb -Wdate-time
>>>> -D_FORTIFY
>>>> _SOURCE=2 -fPIC -std=c++17 -MD -MT
>>>> openvdb/openvdb/CMakeFiles/openvdb_shared.dir/Grid.cc.o -MF
>>>> CMakeFiles/openvdb_shared.dir/Grid.cc.o.d -o
>>>> CMakeFiles/openvdb_shared.dir/Grid.cc.o -c /<<PKGBUILDDIR>>/openvdb/o
>>>> penvdb/Grid.cc
>>>>    cd /<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb &&
>>>> /usr/bin/c++ -DBOOST_ALL_NO_LIB -DBOOST_IOSTREAMS_DYN_LINK
>>>> -DOPENVDB_DLL -DOPENVDB_PRIVATE -DOPENVDB_USE_LOG4CPLUS
>>>> -Dopenvdb_shared_EXPORTS -I/<<PK
>>>> GBUILDDIR>>/openvdb/openvdb/..
>>>> -I/<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb
>>>> -I/<<PKGBUILDDIR>>/obj-arm-linux-gnueabi/openvdb/openvdb/openvdb
>>>> -I/<<PKGBUILDDIR>>/openvdb/openvdb/. -I/usr/include/Imat
>>>> h -g -O2 -ffile-prefix-map=/<<PKGBUILDDIR>>=. -fstack-protector-strong
>>>> -Wformat -Werror=format-security -Wno-psabi -fvisibility=hidden
>>>> -fvisibility-inlines-hidden -Wa,-mimplicit-it=thumb -Wdate-time
>>>> -D_FORTIFY
>>>> _SOURCE=2 -fPIC -std=c++17 -MD -MT
>>>> openvdb/openvdb/CMakeFiles/openvdb_shared.dir/io/Archive.cc.o -MF
>>>> CMakeFiles/openvdb_shared.dir/io/Archive.cc.o.d -o
>>>> CMakeFiles/openvdb_shared.dir/io/Archive.cc.o -c /<<PKGBU
>>>> ILDDIR>>/openvdb/openvdb/io/Archive.cc
>>>> /tmp/ccFlliCl.s: Assembler messages:
>>>> /tmp/ccFlliCl.s:294: Warning: conditional outside an IT block for Thumb.
>>>> /tmp/ccFlliCl.s:338: Warning: conditional outside an IT block for Thumb.
>>>> [...]
>>>>
>>>> There are so many warnings that it eventually times out the compilation.
>>>>
>>>> I could not find a way to tell gcc not to print a warning. Is there a
>>>> way to remove this warning ?
>>>
>>> The warning is coming from the assembler.  But I don't think it should
>>> be just 'ignored'.  You'll need to track down why it's happening at all.
>>>    The most likely case is that there is some inlined assembly code
>>> somewhere that was written on the assumption that the compiler would be
>>> targeting the A32 state rather than the T32 state (thumb), but there's
>>> not enough information here to work out what's happened for sure.
>>>
>>> What we need is a snippet of the assembler output when compiled with
>>> debugging information, so that we can track back to the source code that
>>> is generating this.
>>>
>>>>
>>>> For reference the very large log:
>>>>
>>>> * https://buildd.debian.org/status/fetch.php?pkg=openvdb&arch=armel&ver=9.1.0-4&stamp=1661358853&raw=0
>>>>
>>>> Thanks much,
>>>
>>> The log isn't much help, it's the contents of the file that's being
>>> assembled that will tell us what we need to know.
>>
>> I did add '-save-temps' to one TU compilation:
>>
>>   % cd /home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb
>> && /usr/bin/c++ -DBOOST_ALL_NO_LIB -DBOOST_IOSTREAMS_DYN_LINK
>> -DOPENVDB_DLL -DOPENVDB_PRIVATE -DOPENVDB_USE_LOG4CPLUS
>> -Dopenvdb_shared_EXPORTS
>> -I/home/malat/openvdb-9.1.0/openvdb/openvdb/..
>> -I/home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb
>> -I/home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb/openvdb
>> -I/home/malat/openvdb-9.1.0/openvdb/openvdb/. -I/usr/include/Imath -g
>> -O2 -ffile-prefix-map=/home/malat/openvdb-9.1.0=.
>> -fstack-protector-strong -Wformat -Werror=format-security -Wno-psabi
>> -fvisibility=hidden -fvisibility-inlines-hidden
>> -Wa,-mimplicit-it=thumb -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC
>> -std=c++17 -MD -MT
>> openvdb/openvdb/CMakeFiles/openvdb_shared.dir/io/Compression.cc.o -MF
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.o.d -o
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.o -save-temps -c
>> /home/malat/openvdb-9.1.0/openvdb/openvdb/io/Compression.cc
>>
>> Output is:
>>
>>   % cd /home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb
>> && /usr/bin/c++ -DBOOST_ALL_NO_LIB -DBOOST_IOSTREAMS_DYN_LINK
>> -DOPENVDB_DLL -DOPENVDB_PRIVATE -DOPENVDB_USE_LOG4CPLUS
>> -Dopenvdb_shared_EXPORTS
>> -I/home/malat/openvdb-9.1.0/openvdb/openvdb/..
>> -I/home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb
>> -I/home/malat/openvdb-9.1.0/obj-arm-linux-gnueabi/openvdb/openvdb/openvdb
>> -I/home/malat/openvdb-9.1.0/openvdb/openvdb/. -I/usr/include/Imath -g
>> -O2 -ffile-prefix-map=/home/malat/openvdb-9.1.0=.
>> -fstack-protector-strong -Wformat -Werror=format-security -Wno-psabi
>> -fvisibility=hidden -fvisibility-inlines-hidden
>> -Wa,-mimplicit-it=thumb -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC
>> -std=c++17 -MD -MT
>> openvdb/openvdb/CMakeFiles/openvdb_shared.dir/io/Compression.cc.o -MF
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.o.d -o
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.o -save-temps -c
>> /home/malat/openvdb-9.1.0/openvdb/openvdb/io/Compression.cc
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s: Assembler messages:
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:146: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:162: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:215: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:330: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:344: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:352: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:598: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:768: Warning:
>> conditional outside an IT block for Thumb.
>> CMakeFiles/openvdb_shared.dir/io/Compression.cc.s:934: Warning:
>> conditional outside an IT block for Thumb.
>> [truncated]
>>
>> My gcc version is:
>>
>>   % /usr/bin/c++ -v
>> Using built-in specs.
>> COLLECT_GCC=/usr/bin/c++
>> COLLECT_LTO_WRAPPER=/usr/lib/gcc/arm-linux-gnueabi/12/lto-wrapper
>> Target: arm-linux-gnueabi
>> Configured with: ../src/configure -v --with-pkgversion='Debian
>> 12.2.0-1' --with-bugurl=file:///usr/share/doc/gcc-12/README.Bugs
>> --enable-languages=c,ada,c++,go,d,fortran,objc,obj-c++,m2
>> --prefix=/usr --with-gcc-major-version-only --program-suffix=-12
>> --program-prefix=arm-linux-gnueabi- --enable-shared
>> --enable-linker-build-id --libexecdir=/usr/lib
>> --without-included-gettext --enable-threads=posix --libdir=/usr/lib
>> --enable-nls --enable-clocale=gnu --enable-libstdcxx-debug
>> --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new
>> --enable-gnu-unique-object --disable-libitm --disable-libquadmath
>> --disable-libquadmath-support --enable-plugin --enable-default-pie
>> --with-system-zlib --enable-libphobos-checking=release
>> --with-target-system-zlib=auto --enable-objc-gc=auto
>> --enable-multiarch --disable-sjlj-exceptions --with-arch=armv5te
>> --with-float=soft --disable-werror --enable-checking=release
>> --build=arm-linux-gnueabi --host=arm-linux-gnueabi
>> --target=arm-linux-gnueabi
>> Thread model: posix
>> Supported LTO compression algorithms: zlib zstd
>> gcc version 12.2.0 (Debian 12.2.0-1)
>>
>> *.ii file:
>> https://people.debian.org/~malat/openvdb/Compression.cc.ii.xz
>>
>>
>> Once again thanks again for your time,
> 
> I've removed:
> 
> -Wa,-mimplicit-it=thumb
> 
> from compilation line and it seems to be going much better
> 
> sorry for the noise. This seems to have been a hack for armhf/Debian
> back in 2002...

Ah, I think Debian/Ubuntu normally use a compiler targetting Thumb code 
(--with-mode=thumb during configure), but that appears to be missing 
from your compiler, so you're probably getting Arm code generated.  GCC 
doesn't generate Arm code with the expectation that IT instructions will 
guard conditional instructions, so yes, you'd get a lot of warnings from 
the assembler in that case.

R.

     prev parent reply	other threads:[~2022-08-25 14:09 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-08-25  6:40 Mathieu Malaterre
2022-08-25 10:28 ` Richard Earnshaw
2022-08-25 13:48   ` Mathieu Malaterre
2022-08-25 13:58     ` Mathieu Malaterre
2022-08-25 14:09       ` Richard Earnshaw [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=6292c9e1-cbc7-e16b-e368-6ab37d2426e8@foss.arm.com \
    --to=richard.earnshaw@foss.arm.com \
    --cc=gcc-help@gcc.gnu.org \
    --cc=mathieu.malaterre@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Be sure your reply has a Subject: header at the top and a blank line before the message body.

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).