similar to: LLVM behavior different depending on function symbol name

On Mon, Jun 19, 2017 at 12:06 PM, Mehdi AMINI <joker.eph at gmail.com> wrote: > Hi, > > 2017-06-19 8:45 GMT-07:00 Andrew Kelley via llvm-dev < > llvm-dev at lists.llvm.org>: > >> Greetings, >> >> I have a Zig implementation of ceil which is emitted into LLVM IR like >> this: >> >> ; Function Attrs: nobuiltin nounwind >> define

using `opt --print-after-all -O3` I see that EarlyCSE is interpreting the call to `ceil` and constant fold: *** IR Dump After Early CSE *** ; Function Attrs: nobuiltin nounwind define i1 @do_test() #2 { Entry: %0 = call fastcc float @ceil(float 0.000000e+00) #6 %1 = call fastcc float @ceil32(float 0.000000e+00) #6 %2 = fcmp fast oeq float 0.000000e+00, %1 ret i1 %2 } So just running `opt

When zig updated to llvm 5 we started hitting this assertion: zig: /home/andy/downloads/llvm-project/llvm/include/llvm/Support/Casting.h:106: static bool llvm::isa_impl_cl<To, const From*>::doit(const From*) [with To = llvm::Instruction; From = llvm::Value]: Assertion `Val && "isa<> used on a null pointer"' failed. I wonder if however this was caused by an

So, 90% of the time I've seen this, it was memory corruption, usually use after free. I know I fixed one after 5.0 branched. You should compile with address sanitizer enabled, and I suspect you will find the issue quicky. If not, we really need ir that reproduces it. On Sun, Sep 17, 2017, 12:27 PM Andrew Kelley via llvm-dev < llvm-dev at lists.llvm.org> wrote: > I think I forgot to

Can you please open a bug on bugzilla and attach the ir testcase? Your fix doesn't look right (just hiding the assertion failure) On Sep 17, 2017 10:45 AM, "Andrew Kelley via llvm-dev" < llvm-dev at lists.llvm.org> wrote: > What do you think about this patch? > > --- a/llvm/lib/Transforms/Scalar/NewGVN.cpp > +++ b/llvm/lib/Transforms/Scalar/NewGVN.cpp > @@

Below is attached a full IR module that can reproduce this issue, but the part to notice is this: %Foo96Bits = type <{ i24, i24, i24, i24 }> define internal fastcc i16 @main.0.1() unnamed_addr #2 !dbg !113 { Entry: %value = alloca %Foo96Bits, align 1 %b = alloca i24, align 4 %0 = bitcast %Foo96Bits* %value to i8*, !dbg !129 call void @llvm.memcpy.p0i8.p0i8.i64(i8* %0, i8* bitcast

Dump the IR before newgvn and run through opt -newgvn. It should crash. I'm not familiar with your frontend but you might want to use -mllvm -opt-bisect-limit. On Sep 17, 2017 1:06 PM, "Andrew Kelley" <superjoe30 at gmail.com> wrote: Valgrind is strictly better than address sanitizer, is that right? It runs valgrind-clean: [nix-shell:~/dev/zig/build-llvm5-debug]$ valgrind

Hi, I see things like this. What does it mean? Is it documented somewhere? Thanks. ; preds = %for.body https://llvm.org/docs/LangRef.html ; <label>:91: ; preds = %88 %92 = load i8**, i8*** @glob_complete_word.matches, align 8, !dbg !99798 %93 = load i32, i32* @glob_complete_word.ind, align 4, !dbg !99799 %94 = sext i32 %93 to i64, !dbg !99798

I just reported this bug: https://bugs.llvm.org/show_bug.cgi?id=35314 This very simple IR is causing infinite recursion in llvm::DwarfUnit::getOrCreateTypeDIE in llvm 5.0.0. Is there a workaround? Is anyone willing to save me a recompile and check if this is fixed in 5.0.1 or trunk? `clang -c test.ll` repros the issue. ; ModuleID = 'test' source_filename = "test" target

I have some ideas to allow the BitcodeReader to lazy-load debug info metadata, and wanted to air this on llvm-dev before getting too deep into the code. Motivation ========== Based on some analysis Mehdi ran (ping him for details), there are three (related) compile-time bottlenecks we're seeing with `-flto=thin -g`: a) Reading the large number of Metadata bitcode records in the global

> On 2016-Mar-23, at 07:17, Teresa Johnson <tejohnson at google.com> wrote: > > > > On Tue, Mar 22, 2016 at 7:28 PM, Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote: > I have some ideas to allow the BitcodeReader to lazy-load debug info > metadata, and wanted to air this on llvm-dev before getting too deep > into the code. > > Motivation >

So, based on the info that I got from Devang, I was able to solve the problem with the duplicate assembly symbols, although I still don't quite understand why it was generating erroneous metadata in the first place. For example, through trial and error I discovered that the problem disappeared if I passed my DICompileUnit to CreateLocation() rather than my DISubprogram, but I don't know

On Tue, Mar 22, 2016 at 7:28 PM, Duncan P. N. Exon Smith < dexonsmith at apple.com> wrote: > I have some ideas to allow the BitcodeReader to lazy-load debug info > metadata, and wanted to air this on llvm-dev before getting too deep > into the code. > > Motivation > ========== > > Based on some analysis Mehdi ran (ping him for details), there are three > (related)

+pcc, who had some other ideas/patch out for improving memory usage of debug info +Reid, who's responsible for the windows/CodeView/PDB debug info which is motivating some of the ideas about changes to type emission So how does this relate, or not, to Peter's (pcc) work trying to reduce the DIE overhead during code gen? Are you folks chasing different memory bottlenecks? Are they both

I figured it out. I was using this implementation of __chkstk from compiler-rt: DEFINE_COMPILERRT_FUNCTION(___chkstk) push %rcx cmp $0x1000,%rax lea 16(%rsp),%rcx // rsp before calling this routine -> rcx jb 1f 2: sub $0x1000,%rcx test %rcx,(%rcx) sub $0x1000,%rax cmp $0x1000,%rax ja 2b 1:

On Wed, Mar 23, 2016 at 9:31 AM, Duncan P. N. Exon Smith < dexonsmith at apple.com> wrote: > > > On 2016-Mar-23, at 07:17, Teresa Johnson <tejohnson at google.com> wrote: > > > > > > > > On Tue, Mar 22, 2016 at 7:28 PM, Duncan P. N. Exon Smith < > dexonsmith at apple.com> wrote: > > I have some ideas to allow the BitcodeReader to

> On 2016-Mar-23, at 09:42, Teresa Johnson <tejohnson at google.com> wrote: > > > > On Wed, Mar 23, 2016 at 9:31 AM, Duncan P. N. Exon Smith <dexonsmith at apple.com> wrote: > > > On 2016-Mar-23, at 07:17, Teresa Johnson <tejohnson at google.com> wrote: > > > > > > > > On Tue, Mar 22, 2016 at 7:28 PM, Duncan P. N. Exon Smith

On Tue, Dec 1, 2009 at 10:56 PM, Talin <viridia at gmail.com> wrote: > So, based on the info that I got from Devang, I was able to solve the > problem with the duplicate assembly symbols, although I still don't quite > understand why it was generating erroneous metadata in the first place. For > example, through trial and error I discovered that the problem disappeared >

> On Mar 22, 2016, at 8:04 PM, David Blaikie <dblaikie at gmail.com> wrote: > > +pcc, who had some other ideas/patch out for improving memory usage of debug info > +Reid, who's responsible for the windows/CodeView/PDB debug info which is motivating some of the ideas about changes to type emission > > So how does this relate, or not, to Peter's (pcc) work trying to

I'm going to show the same function, first normally, and then as a coroutine, and show how gdb can see the variable when it's a normal function, but not when it's a coroutine. I'd like to understand if this can be improved. I'm trying to debug a real world problem, but the lack of debug info on variables in coroutines is making it difficult. Should I file a bug? Is this a