[llvm-dev] Please dogfood LLD

Thu Mar 16 14:40:41 PDT 2017

On Thu, Mar 16, 2017 at 2:36 PM Rui Ueyama <ruiu at google.com> wrote:

> What program did you use to test the feature, and what was missing
> information? I'd like to file that as a bug so that we can fix this later.
>

I was using GDB 7.9 (a google internal build, not sure how much that
matters - there might be some things that didn't make it upstream) to debug
LLVM's 'llc' binary on an assert-failure.

The backtrace was missing source information from some frames:

Not the best representation of it, but here's a chunk of my debugging
session:

#9  0x00000000037d9278 in llvm::DwarfCompileUnit::constructVariableDIEImpl
(this=0x540a390, DV=..., Abstract=false) at /usr/local/google/hom
e/blaikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:543
(gdb)
#10 0x00000000037d8c6c in llvm::DwarfCompileUnit::constructVariableDIE
(this=0x540a390, DV=..., Abstract=false) at /usr/local/google/home/bl
aikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:479
(gdb)
#11 0x00000000037d97b6 in llvm::DwarfCompileUnit::constructVariableDIE
(this=0x540a390, DV=..., Scope=..., ObjectPointer=@0x7fffffffb488: 0x
0) at
/usr/local/google/home/blaikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:560
(gdb)
#12 0x00000000037d8231 in llvm::DwarfCompileUnit::createScopeChildrenDIE
(this=0x540a390, Scope=0x5482df0, Children=llvm::SmallVector of len
gth 0, capacity 8, ChildScopeCount=0x0) at
/usr/local/google/home/blaikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:572
(gdb)
#13 0x00000000037d99b7 in llvm::DwarfCompileUnit::createAndAddScopeChildren
(this=0x540a390, Scope=0x5482df0, ScopeDIE=...) at /usr/local/go
ogle/home/blaikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:614
(gdb)
#14 0x00000000037d98a4 in
llvm::DwarfCompileUnit::constructSubprogramScopeDIE (this=0x540a390,
Sub=0x53cd568, Scope=0x5482df0) at /usr/local
/google/home/blaikie/dev/llvm/src/lib/CodeGen/AsmPrinter/DwarfCompileUnit.cpp:594
(gdb)
#15 0x000000000376a087 in llvm::DwarfDebug::endFunctionImpl
(this=0x54082c0, MF=0x5439b60) at
/usr/local/google/home/blaikie/dev/llvm/src/li
b/CodeGen/AsmPrinter/DwarfDebug.cpp:1240
(gdb)
#16 0x00000000037c93c9 in
llvm::DebugHandlerBase::endFunction(llvm::MachineFunction const*) ()
(gdb)
#17 0x0000000003733d6a in llvm::AsmPrinter::EmitFunctionBody
(this=0x53fa9a0) at
/usr/local/google/home/blaikie/dev/llvm/src/lib/CodeGen/Asm
Printer/AsmPrinter.cpp:1083
(gdb)
#18 0x0000000003107e61 in
llvm::X86AsmPrinter::runOnMachineFunction(llvm::MachineFunction&) ()
(gdb)
#19 0x0000000003a263b1 in
llvm::MachineFunctionPass::runOnFunction(llvm::Function&) ()

>
> On Thu, Mar 16, 2017 at 2:34 PM, David Blaikie <dblaikie at gmail.com> wrote:
>
> FWIW - selfhosting I did find that GDB wasn't able to find the source code
> for some functions when using LLD's gdb_index, so I've switched back to
> gold+gdb_index for now (given the performance problems you mentioned,
> sounds like I wasn't gaining much/anything in terms of link time by using
> lld anyway).
>
>
> On Thu, Mar 16, 2017 at 11:17 AM Rui Ueyama <ruiu at google.com> wrote:
>
> I personally haven't tried gdb_index, and I don't know the quality of the
> produced index. Most of the code was written by George.
>
> One thing I noticed about the feature (and filed as
> http://bugs.llvm.org/show_bug.cgi?id=32228) is that our gdb_index feature
> is much slower than the gold. Apparently there's room for improvement.
>
> On Thu, Mar 16, 2017 at 8:35 AM, David Blaikie <dblaikie at gmail.com> wrote:
>
> Any idea of the quality/completeness of the gdb_index support? I've
> started using gdb_index recently (reduces GDB startup time for Clang from
> 50 seconds to 3 seconds - though I haven't measure how much it increases
> link time*) & it's really handy. Would love to have some assurance it's
> expected to work for LLD.
>
> Simple experiments seem to indicate that it's OK/good. Though a quick
> llvm-dwarfdump of gdb_index between the two linkers shows about a 10%
> disparity in size (LLD's is smaller) given the same inputs. (219MB v 196MB
> of text dumped - I didn't compare the actual section sizes)
>
> *though honestly I'd be willing to trade startup time for link time 1:1
> even, because builds/links are already long enough that I go off and do
> something else - whereas getting GDB start down to 3 seconds, I can start
> writing my breakpoint command, etc, in those 3 seconds and be all but
> unaffected by it - at 50 seconds I basically have to go off and do
> something else/context switch/etc.
>
> On Tue, Mar 14, 2017 at 11:40 AM Rui Ueyama via llvm-dev <
> llvm-dev at lists.llvm.org> wrote:
>
> Hi all,
>
> LLVM 4.0.0 is out, and I can say that LLD/ELF is now ready for production
> use at least for x86-64 (and probably for AArch64 and MIPS). I believe
> you've heard a few good news about the linker -- it just works
> <http://lld.llvm.org/#features> and is very fast
> <http://lld.llvm.org/#performance>, clean, compact and supported by the
> active community. I don't think I need to reiterate why having a good
> linker is important for us as the LLVM community.
>
> There's one thing you can help us without writing even a line of code.
> Please use it!
>
> Using LLD to link LLVM/clang/etc is easy. You need to check out the LLD
> repository just like you probably already did for clang, build it, and then
> install it. LLD will be installed as ld.lld (and it won't be used by
> default.) After that, re-run cmake with -DLLVM_ENABLE_LLD=On along with
> your usual options so that LLD will be used to build LLVM.
>
> For the details of the build process, please see
> http://lld.llvm.org/#build.
>
> Thanks,
> Rui
> _______________________________________________
> 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/20170316/e4abc60a/attachment.html>