llvm dev - Mar 2017 - Use of host/target compiler when building compiler-rt

On Wed, Mar 8, 2017 at 3:23 PM Chris Bieneman <beanz at apple.com> wrote:
> On Mar 8, 2017, at 3:16 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
>
> On Wed, Mar 8, 2017 at 2:55 PM Chris Bieneman <beanz at apple.com>
wrote:
>
> David,
>
> This is an area that has had a lot of development over the last two years.
>
> There are two supported ways in the LLVM build system to build compiler-rt
> with the just-built compiler.
>
> 1) The legacy way is for if compiler-rt is under LLVM/projects. You can
> specify -DLLVM_BUILD_EXTERNAL_COMPILER_RT=On, which will configure
> compiler-rt using the just-built clang after clang is built.
>
>
> I thought the BUILD_EXTERNAL variables were for use when projects were not
> embedded within the llvm source tree (mostly in use by Takumi's flat
> buildbots that checout the top-level project without embedding, say, clang
> or compiler-rt within the llvm source tree)?
>
>
> You are confusing this with the similarly
> named LLVM_EXTERNAL_${nameUPPER}_SOURCE_DIR variables.
>
Ah, right - indeed.
>
> 2) The new way, is to place compiler-rt under LLVM/runtimes. In this path
> the build system will automatically build with the just-built compiler.
> This path also splits compiler-rt into two separate build steps, one that
> configures and builds the builtins with the just-built compiler, and a
> second that configures and builds the sanitizer libraries.
>
>
> Huh, OK - could someone remove the legacy format, then? If it's a trap.
>
>
> I'm not sure the new path is fully supported in every workflow, so
> removing it seems like a not great idea at the moment.
>
>
> That said, I tried putting compiler-rt in runtimes instead of projects and
> I got a bunch of cmake errors starting with:
>
> CMake Error at
>
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
> (add_dependencies):
>   The dependency target "GotsanRuntimeCheck" of target
"check-runtimes"
> does
>   not exist.
> Call Stack (most recent call first):
>   CMakeLists.txt:110 (add_lit_target)
>
> Any ideas?
>
>
> I have never encountered that issue. It looks like the tsan test code is
> out of sync. If you go into tsan/test/CMakeLists.txt and on Line 2 add this
> to the if statement "AND TARGET GotsanRuntimeCheck" that should
fix the
> issue.
>
Hrm - not sure which CMakeLists.txt you're referring to? In my
runtimes/compiler-rt/lib/tsan/tests/CMakeLists.txt the first few lines are:

include_directories(../rtl)

add_custom_target(TsanUnitTests)
set_target_properties(TsanUnitTests PROPERTIES
  FOLDER "TSan unittests")

no if condition I could modify?

>
> -Chris
>
>
>
>
> The second path also works for many (but not all) of our other runtime
> library projects. I know it works for libcxx, libcxxabi, and
> libunwind. Petr Hosek (CC'd) has also been working on support for
> multi-arch builtin and runtime library builds so that you can generate full
> cross-compilers from a single cmake invocation.
>
> -Chris
>
>
> On Mar 8, 2017, at 2:35 PM, David Blaikie via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
>
>
> On Wed, Mar 8, 2017 at 2:03 PM Sterling Augustine <saugustine at
google.com>
> wrote:
>
> Yes, this is a aspect of the larger problem that clang bootstrap
doesn't
> work for a cross-compiler. The build (mostly?) assumes that host==target
> during the build of clang itself, and then if you want another architecture
> also, you run a second build of the target libraries, and manually merge
> the trees.
>
>
> I kind of roughly follow that, but not too well.
>
>
> If you think about compiler-rt as being compiled for the target rather
> than the host, the problem you describe here is exactly the same one, and
> we have been getting lucky.
>
>
> Sure - if a PPC clang is being built from an x86 host, how would
> compiler-rt be built (OK, it could be built with the just-built clang,
> which it isn't at the moment) and tested (can't really be tested
because
> the host can't run PPC binaries).
>
>
> At the moment, the blaze builds of clang do exactly the procedure
> described above, so this hasn't been a terrible problem for Google, but
I
> do think it is something that should be fixed (I'm working on another
> aspect of compiler-rt bringup at the moment, so won't solve this in the
> immediate future.)
>
>
> Rightio
>
>
>
> gnu systems have a make variable, "CC_FOR_TARGET" that addresses
this
> problem. I imagine llvm should adopt a similar mechanism inside cmake.
>
>
> Not sure I follow on the need/use of CC_FOR_TARGET compared to using the
> just-built clang as the CC_FOR_TARGET (which it seems we have some plumbing
> for already - the just-built clang is used for building the compiler-rt
> tests, but not for building the library. I /think/ it should be used for
> both)
>
> - Dave
>
>
>
> On Wed, Mar 8, 2017 at 1:54 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
> I stumbled across what seems to be a bug (to me) in the compiler-rt build:
>
> The compiler-rt libraries themselves are built with the host compiler
> while the tests are built and then linked with the just-built clang.
>
> It was my understanding that the goal/intent/need was to have the
> compiler-rt library build with the just-built clang? Did I misunderstand
> that?*
>
> Sterling: Chandler seemed to think you might be interested in this issue
&
> possibly addressing it given you're working on compiler-rt bring-up?
It'd
> probably be useful to have compiler-rt built with the just-built clang for
> performance reasons.
>
> Evgeniy - not sure if you're interested in this or have much context?
Know
> if this is right/wrong/neutral, etc?
>
> * reasons include performance, ABI compatibility, etc (I thought this was
> necessary for correctness in some way) - also, otherwise it seems excessive
> to hold up the whole build on waiting for just-built clang to finish, then
> use that to compile some tests. (well, I realize some of the tests are
> end-to-end, so they do need the just-built compiler)
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20170309/3978b000/attachment-0001.html>

> On Mar 8, 2017, at 4:42 PM, David Blaikie <dblaikie at gmail.com>
wrote:
> 
> 
> 
> On Wed, Mar 8, 2017 at 3:23 PM Chris Bieneman <beanz at apple.com
<mailto:beanz at apple.com>> wrote:
>> On Mar 8, 2017, at 3:16 PM, David Blaikie <dblaikie at gmail.com
<mailto:dblaikie at gmail.com>> wrote:
>> 
>> 
>> 
>> On Wed, Mar 8, 2017 at 2:55 PM Chris Bieneman <beanz at apple.com
<mailto:beanz at apple.com>> wrote:
>> David,
>> 
>> This is an area that has had a lot of development over the last two
years.
>> 
>> There are two supported ways in the LLVM build system to build
compiler-rt with the just-built compiler.
>> 
>> 1) The legacy way is for if compiler-rt is under LLVM/projects. You can
specify -DLLVM_BUILD_EXTERNAL_COMPILER_RT=On, which will configure compiler-rt
using the just-built clang after clang is built.
>> 
>> I thought the BUILD_EXTERNAL variables were for use when projects were
not embedded within the llvm source tree (mostly in use by Takumi's flat
buildbots that checout the top-level project without embedding, say, clang or
compiler-rt within the llvm source tree)?
> 
> You are confusing this with the similarly named
LLVM_EXTERNAL_${nameUPPER}_SOURCE_DIR variables.
> 
> Ah, right - indeed.
>> 
>> 2) The new way, is to place compiler-rt under LLVM/runtimes. In this
path the build system will automatically build with the just-built compiler.
This path also splits compiler-rt into two separate build steps, one that
configures and builds the builtins with the just-built compiler, and a second
that configures and builds the sanitizer libraries.
>> 
>> Huh, OK - could someone remove the legacy format, then? If it's a
trap.
> 
> I'm not sure the new path is fully supported in every workflow, so
removing it seems like a not great idea at the moment.
> 
>> 
>> That said, I tried putting compiler-rt in runtimes instead of projects
and I got a bunch of cmake errors starting with:
>> 
>> CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
>>   The dependency target "GotsanRuntimeCheck" of target
"check-runtimes" does
>>   not exist.
>> Call Stack (most recent call first):
>>   CMakeLists.txt:110 (add_lit_target)
>> 
>> Any ideas?
> 
> I have never encountered that issue. It looks like the tsan test code is
out of sync. If you go into tsan/test/CMakeLists.txt and on Line 2 add this to
the if statement "AND TARGET GotsanRuntimeCheck" that should fix the
issue.
> 
> Hrm - not sure which CMakeLists.txt you're referring to? In my
runtimes/compiler-rt/lib/tsan/tests/CMakeLists.txt the first few lines are:
Sorry, I meant <compiler-rt>/test/tsan/CMakeLists.txt, not tsan/test.

-Chris
> 
> include_directories(../rtl)
> 
> add_custom_target(TsanUnitTests)
> set_target_properties(TsanUnitTests PROPERTIES
>   FOLDER "TSan unittests")
> 
> no if condition I could modify?
>  
> 
> -Chris
> 
>>  
>> 
>> The second path also works for many (but not all) of our other runtime
library projects. I know it works for libcxx, libcxxabi, and libunwind. Petr
Hosek (CC'd) has also been working on support for multi-arch builtin and
runtime library builds so that you can generate full cross-compilers from a
single cmake invocation.
>> 
>> -Chris
>> 
>> 
>> 
>>> On Mar 8, 2017, at 2:35 PM, David Blaikie via llvm-dev <llvm-dev
at lists.llvm.org <mailto:llvm-dev at lists.llvm.org>> wrote:
>>> 
>> 
>>> 
>>> 
>>> On Wed, Mar 8, 2017 at 2:03 PM Sterling Augustine <saugustine at
google.com <mailto:saugustine at google.com>> wrote:
>>> Yes, this is a aspect of the larger problem that clang bootstrap
doesn't work for a cross-compiler. The build (mostly?) assumes that
host==target during the build of clang itself, and then if you want another
architecture also, you run a second build of the target libraries, and manually
merge the trees.
>>> 
>>> I kind of roughly follow that, but not too well.
>>>  
>>> If you think about compiler-rt as being compiled for the target
rather than the host, the problem you describe here is exactly the same one, and
we have been getting lucky.
>>> 
>>> Sure - if a PPC clang is being built from an x86 host, how would
compiler-rt be built (OK, it could be built with the just-built clang, which it
isn't at the moment) and tested (can't really be tested because the host
can't run PPC binaries).
>>>  
>>> At the moment, the blaze builds of clang do exactly the procedure
described above, so this hasn't been a terrible problem for Google, but I do
think it is something that should be fixed (I'm working on another aspect of
compiler-rt bringup at the moment, so won't solve this in the immediate
future.)
>>> 
>>> Rightio
>>>  
>>> 
>>> gnu systems have a make variable, "CC_FOR_TARGET" that
addresses this problem. I imagine llvm should adopt a similar mechanism inside
cmake.
>>> 
>>> Not sure I follow on the need/use of CC_FOR_TARGET compared to
using the just-built clang as the CC_FOR_TARGET (which it seems we have some
plumbing for already - the just-built clang is used for building the compiler-rt
tests, but not for building the library. I /think/ it should be used for both)
>>> 
>>> - Dave
>>>  
>>> 
>>> On Wed, Mar 8, 2017 at 1:54 PM, David Blaikie <dblaikie at
gmail.com <mailto:dblaikie at gmail.com>> wrote:
>>> I stumbled across what seems to be a bug (to me) in the compiler-rt
build:
>>> 
>>> The compiler-rt libraries themselves are built with the host
compiler while the tests are built and then linked with the just-built clang.
>>> 
>>> It was my understanding that the goal/intent/need was to have the
compiler-rt library build with the just-built clang? Did I misunderstand that?*
>>> 
>>> Sterling: Chandler seemed to think you might be interested in this
issue & possibly addressing it given you're working on compiler-rt
bring-up? It'd probably be useful to have compiler-rt built with the
just-built clang for performance reasons.
>>> 
>>> Evgeniy - not sure if you're interested in this or have much
context? Know if this is right/wrong/neutral, etc?
>>> 
>>> * reasons include performance, ABI compatibility, etc (I thought
this was necessary for correctness in some way) - also, otherwise it seems
excessive to hold up the whole build on waiting for just-built clang to finish,
then use that to compile some tests. (well, I realize some of the tests are
end-to-end, so they do need the just-built compiler)
>>> 
>> 
>>> _______________________________________________
>>> LLVM Developers mailing list
>>> llvm-dev at lists.llvm.org <mailto:llvm-dev at
lists.llvm.org>
>>> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
<http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20170309/d07e8eb1/attachment.html>

On Thu, Mar 9, 2017 at 11:25 AM Chris Bieneman <beanz at apple.com> wrote:
> On Mar 8, 2017, at 4:42 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
>
> On Wed, Mar 8, 2017 at 3:23 PM Chris Bieneman <beanz at apple.com>
wrote:
>
> On Mar 8, 2017, at 3:16 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
>
>
> On Wed, Mar 8, 2017 at 2:55 PM Chris Bieneman <beanz at apple.com>
wrote:
>
> David,
>
> This is an area that has had a lot of development over the last two years.
>
> There are two supported ways in the LLVM build system to build compiler-rt
> with the just-built compiler.
>
> 1) The legacy way is for if compiler-rt is under LLVM/projects. You can
> specify -DLLVM_BUILD_EXTERNAL_COMPILER_RT=On, which will configure
> compiler-rt using the just-built clang after clang is built.
>
>
> I thought the BUILD_EXTERNAL variables were for use when projects were not
> embedded within the llvm source tree (mostly in use by Takumi's flat
> buildbots that checout the top-level project without embedding, say, clang
> or compiler-rt within the llvm source tree)?
>
>
> You are confusing this with the similarly
> named LLVM_EXTERNAL_${nameUPPER}_SOURCE_DIR variables.
>
>
> Ah, right - indeed.
>
>
> 2) The new way, is to place compiler-rt under LLVM/runtimes. In this path
> the build system will automatically build with the just-built compiler.
> This path also splits compiler-rt into two separate build steps, one that
> configures and builds the builtins with the just-built compiler, and a
> second that configures and builds the sanitizer libraries.
>
>
> Huh, OK - could someone remove the legacy format, then? If it's a trap.
>
>
> I'm not sure the new path is fully supported in every workflow, so
> removing it seems like a not great idea at the moment.
>
>
> That said, I tried putting compiler-rt in runtimes instead of projects and
> I got a bunch of cmake errors starting with:
>
> CMake Error at
>
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
> (add_dependencies):
>   The dependency target "GotsanRuntimeCheck" of target
"check-runtimes"
> does
>   not exist.
> Call Stack (most recent call first):
>   CMakeLists.txt:110 (add_lit_target)
>
> Any ideas?
>
>
> I have never encountered that issue. It looks like the tsan test code is
> out of sync. If you go into tsan/test/CMakeLists.txt and on Line 2 add this
> to the if statement "AND TARGET GotsanRuntimeCheck" that should
fix the
> issue.
>
>
> Hrm - not sure which CMakeLists.txt you're referring to? In my
> runtimes/compiler-rt/lib/tsan/tests/CMakeLists.txt the first few lines are:
>
>
> Sorry, I meant <compiler-rt>/test/tsan/CMakeLists.txt, not tsan/test.
>
Think that got me past that error, but (sorry I wasn't clear) it was only
one of many errors. Here are the next... 5, say...

CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
  The dependency target "TsanUnitTests" of target
"check-runtimes" does not
  exist.
Call Stack (most recent call first):
  CMakeLists.txt:110 (add_lit_target)


CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
  The dependency target "cfi" of target "check-runtimes"
does not exist.
Call Stack (most recent call first):
  CMakeLists.txt:110 (add_lit_target)


CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
  The dependency target "TsanUnitTests" of target
"check-compiler-rt" does
  not exist.
Call Stack (most recent call first):
  compiler-rt/test/CMakeLists.txt:94 (add_lit_target)


CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
  The dependency target "cfi" of target "check-compiler-rt"
does not exist.
Call Stack (most recent call first):
  compiler-rt/test/CMakeLists.txt:94 (add_lit_target)


CMake Error at
/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1174
(add_dependencies):
  The dependency target "TsanUnitTests" of target
"check-tsan" does not
  exist.
Call Stack (most recent call first):

/usr/local/google/home/blaikie/dev/llvm/build/clang/debug/split/notypes/nostandalone/lib/cmake/llvm/AddLLVM.cmake:1195
(add_lit_target)
  compiler-rt/test/tsan/CMakeLists.txt:46 (add_lit_testsuite)


I'm assuming there's some systemic problem with the way I'm holding
this
(if it's working for other people)?

- Dave

>
> -Chris
>
>
> include_directories(../rtl)
>
> add_custom_target(TsanUnitTests)
> set_target_properties(TsanUnitTests PROPERTIES
>   FOLDER "TSan unittests")
>
> no if condition I could modify?
>
>
>
> -Chris
>
>
>
>
> The second path also works for many (but not all) of our other runtime
> library projects. I know it works for libcxx, libcxxabi, and
> libunwind. Petr Hosek (CC'd) has also been working on support for
> multi-arch builtin and runtime library builds so that you can generate full
> cross-compilers from a single cmake invocation.
>
> -Chris
>
>
> On Mar 8, 2017, at 2:35 PM, David Blaikie via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
>
>
> On Wed, Mar 8, 2017 at 2:03 PM Sterling Augustine <saugustine at
google.com>
> wrote:
>
> Yes, this is a aspect of the larger problem that clang bootstrap
doesn't
> work for a cross-compiler. The build (mostly?) assumes that host==target
> during the build of clang itself, and then if you want another architecture
> also, you run a second build of the target libraries, and manually merge
> the trees.
>
>
> I kind of roughly follow that, but not too well.
>
>
> If you think about compiler-rt as being compiled for the target rather
> than the host, the problem you describe here is exactly the same one, and
> we have been getting lucky.
>
>
> Sure - if a PPC clang is being built from an x86 host, how would
> compiler-rt be built (OK, it could be built with the just-built clang,
> which it isn't at the moment) and tested (can't really be tested
because
> the host can't run PPC binaries).
>
>
> At the moment, the blaze builds of clang do exactly the procedure
> described above, so this hasn't been a terrible problem for Google, but
I
> do think it is something that should be fixed (I'm working on another
> aspect of compiler-rt bringup at the moment, so won't solve this in the
> immediate future.)
>
>
> Rightio
>
>
>
> gnu systems have a make variable, "CC_FOR_TARGET" that addresses
this
> problem. I imagine llvm should adopt a similar mechanism inside cmake.
>
>
> Not sure I follow on the need/use of CC_FOR_TARGET compared to using the
> just-built clang as the CC_FOR_TARGET (which it seems we have some plumbing
> for already - the just-built clang is used for building the compiler-rt
> tests, but not for building the library. I /think/ it should be used for
> both)
>
> - Dave
>
>
>
> On Wed, Mar 8, 2017 at 1:54 PM, David Blaikie <dblaikie at gmail.com>
wrote:
>
> I stumbled across what seems to be a bug (to me) in the compiler-rt build:
>
> The compiler-rt libraries themselves are built with the host compiler
> while the tests are built and then linked with the just-built clang.
>
> It was my understanding that the goal/intent/need was to have the
> compiler-rt library build with the just-built clang? Did I misunderstand
> that?*
>
> Sterling: Chandler seemed to think you might be interested in this issue
&
> possibly addressing it given you're working on compiler-rt bring-up?
It'd
> probably be useful to have compiler-rt built with the just-built clang for
> performance reasons.
>
> Evgeniy - not sure if you're interested in this or have much context?
Know
> if this is right/wrong/neutral, etc?
>
> * reasons include performance, ABI compatibility, etc (I thought this was
> necessary for correctness in some way) - also, otherwise it seems excessive
> to hold up the whole build on waiting for just-built clang to finish, then
> use that to compile some tests. (well, I realize some of the tests are
> end-to-end, so they do need the just-built compiler)
>
>
> _______________________________________________
> LLVM Developers mailing list
> llvm-dev at lists.llvm.org
> http://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.llvm.org/pipermail/llvm-dev/attachments/20170309/d6adc12c/attachment.html>