thr3ads.net - llvm dev - [LLVMdev] 2.1 Pre-Release Available (testers needed) [Sep 2007]

If this information is useful, please help other people find it:
Share via:

Emil Mikulic

2007-Sep-19 07:24 UTC

[LLVMdev] 2.1 Pre-Release Available (testers needed)

On Tue, Sep 18, 2007 at 06:41:38PM +1000, Emil Mikulic
wrote:> The "make check" produced:
>                 ===  Summary ==> 
> # of expected passes            2209
> # of unexpected failures        41
> # of expected failures          5
> gmake[1]: *** [check-local] Error 1
> gmake[1]: Leaving directory `/home/emil/ll/objdir-llvm/test'
> gmake: *** [check] Error 2
At Duncan's suggestion, I re-built LLVM (and llvm-gcc4, and then llvm
again) with Debug+Checks and I now get:

		===  Summary ==
# of expected passes		2207
# of unexpected failures	43
# of expected failures		5
gmake[1]: *** [check-local] Error 1
gmake[1]: Leaving directory `/home/emil/ll/objdir-llvm/test'
gmake: *** [check] Error 2

Typescript here, which looks different and trips at least one STL check
that I can see:

http://goanna.cs.rmit.edu.au/~emil/llvm2.1-check-debug.txt

(old typescript here:)> http://goanna.cs.rmit.edu.au/~emil/llvm2.1-check.txt
--Emil

Emil Mikulic

2007-Sep-19 07:45 UTC

head link

[LLVMdev] 2.1 Pre-Release Available (testers needed)

On Wed, Sep 19, 2007 at 05:24:12PM +1000, Emil Mikulic
wrote:> http://goanna.cs.rmit.edu.au/~emil/llvm2.1-check-debug.txt
Here's an ARM test that cores:

$ llvm-as <
/home/emil/ll/llvm-2.1/test/CodeGen/ARM/2007-01-19-InfiniteLoop.ll | llc
-march=arm -mattr=+v6,+vfp2
Segmentation fault (core dumped)

$ gdb `which llc` llc.core
[...]
(gdb) where
#0  0x0853d606 in llvm::ilist_traits<llvm::MachineInstr>::getNext (
    N=0x5a5a5a5a) at MachineBasicBlock.h:37
#1  0x0854d807 in llvm::ilist_iterator<llvm::MachineInstr>::operator-- (
    this=0xbfbfe880) at ilist:137
#2  0x085502c5 in
std::__advance<llvm::ilist_iterator<llvm::MachineInstr>, int>
    (__i=@0xbfbfe880, __n=0) at stl_iterator_base_funcs.h:145
#3  0x08550311 in
std::advance<llvm::ilist_iterator<llvm::MachineInstr>, int> (
    __i=@0xbfbfe880, __n=-1) at stl_iterator_base_funcs.h:177
#4  0x08550339 in llvm::prior<llvm::ilist_iterator<llvm::MachineInstr>
> (it      {<bidirectional_iterator<llvm::MachineInstr,int>> =
{<std::iterator<std::bidirectional_iterator_tag,llvm::MachineInstr,int,llvm::MachineInstr*,llvm::MachineInstr&>>
= {<No data fields>}, <No data fields>}, NodePtr = 0x5a5a5a5a})
    at STLExtras.h:154
#5  0x085c08c8 in LoadStoreMultipleOpti (this=0x29005ec0, MBB=@0x2904c400)
    at /home/emil/ll/llvm-2.1/lib/Target/ARM/ARMLoadStoreOptimizer.cpp:683
#6  0x085c0a8e in runOnMachineFunction (this=0x29005ec0, Fn=@0x290282c0)
    at /home/emil/ll/llvm-2.1/lib/Target/ARM/ARMLoadStoreOptimizer.cpp:744
#7  0x0853d4b8 in llvm::MachineFunctionPass::runOnFunction (this=0x29005ec0, 
    F=@0x29007420) at MachineFunctionPass.h:41
#8  0x08b48a5d in llvm::FPPassManager::runOnFunction (this=0x29037060, 
    F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/PassManager.cpp:1168
#9  0x08b48cc4 in llvm::FunctionPassManagerImpl::run (this=0x2902a040, 
    F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/PassManager.cpp:1126
#10 0x08b48e48 in llvm::FunctionPassManager::run (this=0xbfbfebcc, 
    F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/PassManager.cpp:1071
#11 0x0851b0da in main (argc=3, argv=0xbfbfec84)
    at /home/emil/ll/llvm-2.1/tools/llc/llc.cpp:295

line 37 is:
37        static MachineInstr* getNext(MachineInstr* N) { return N->next; }

(gdb) print N
$1 = (llvm::MachineInstr *) 0x5a5a5a5a
(gdb) print *N
Cannot access memory at address 0x5a5a5a5a

Should I take this (and the other crashes?) off the list and start
filing bugs or something?

--Emil

Evan Cheng

2007-Sep-19 18:04 UTC

head link

[LLVMdev] 2.1 Pre-Release Available (testers needed)

Is this the same issue as PR1686? If so, please add to it. What kind  
of host machine are you using?

Thanks,

Evan

On Sep 19, 2007, at 12:45 AM, Emil Mikulic wrote:
> On Wed, Sep 19, 2007 at 05:24:12PM +1000, Emil Mikulic wrote:
>> http://goanna.cs.rmit.edu.au/~emil/llvm2.1-check-debug.txt
>
> Here's an ARM test that cores:
>
> $ llvm-as < /home/emil/ll/llvm-2.1/test/CodeGen/ARM/2007-01-19- 
> InfiniteLoop.ll | llc -march=arm -mattr=+v6,+vfp2
> Segmentation fault (core dumped)
>
> $ gdb `which llc` llc.core
> [...]
> (gdb) where
> #0  0x0853d606 in llvm::ilist_traits<llvm::MachineInstr>::getNext (
>     N=0x5a5a5a5a) at MachineBasicBlock.h:37
> #1  0x0854d807 in  
> llvm::ilist_iterator<llvm::MachineInstr>::operator-- (
>     this=0xbfbfe880) at ilist:137
> #2  0x085502c5 in  
> std::__advance<llvm::ilist_iterator<llvm::MachineInstr>, int>
>     (__i=@0xbfbfe880, __n=0) at stl_iterator_base_funcs.h:145
> #3  0x08550311 in  
> std::advance<llvm::ilist_iterator<llvm::MachineInstr>, int> (
>     __i=@0xbfbfe880, __n=-1) at stl_iterator_base_funcs.h:177
> #4  0x08550339 in  
> llvm::prior<llvm::ilist_iterator<llvm::MachineInstr> > (it> 
{<bidirectional_iterator<llvm::MachineInstr,int>> =
>
{<std::iterator<std::bidirectional_iterator_tag,llvm::MachineInstr,int
> ,llvm::MachineInstr*,llvm::MachineInstr&>> = {<No data
fields>},
> <No data fields>}, NodePtr = 0x5a5a5a5a})
>     at STLExtras.h:154
> #5  0x085c08c8 in LoadStoreMultipleOpti (this=0x29005ec0,  
> MBB=@0x2904c400)
>     at /home/emil/ll/llvm-2.1/lib/Target/ARM/ 
> ARMLoadStoreOptimizer.cpp:683
> #6  0x085c0a8e in runOnMachineFunction (this=0x29005ec0,  
> Fn=@0x290282c0)
>     at /home/emil/ll/llvm-2.1/lib/Target/ARM/ 
> ARMLoadStoreOptimizer.cpp:744
> #7  0x0853d4b8 in llvm::MachineFunctionPass::runOnFunction  
> (this=0x29005ec0,
>     F=@0x29007420) at MachineFunctionPass.h:41
> #8  0x08b48a5d in llvm::FPPassManager::runOnFunction (this=0x29037060,
>     F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/ 
> PassManager.cpp:1168
> #9  0x08b48cc4 in llvm::FunctionPassManagerImpl::run (this=0x2902a040,
>     F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/ 
> PassManager.cpp:1126
> #10 0x08b48e48 in llvm::FunctionPassManager::run (this=0xbfbfebcc,
>     F=@0x29007420) at /home/emil/ll/llvm-2.1/lib/VMCore/ 
> PassManager.cpp:1071
> #11 0x0851b0da in main (argc=3, argv=0xbfbfec84)
>     at /home/emil/ll/llvm-2.1/tools/llc/llc.cpp:295
>
> line 37 is:
> 37        static MachineInstr* getNext(MachineInstr* N) { return N- 
> >next; }
>
> (gdb) print N
> $1 = (llvm::MachineInstr *) 0x5a5a5a5a
> (gdb) print *N
> Cannot access memory at address 0x5a5a5a5a
>
> Should I take this (and the other crashes?) off the list and start
> filing bugs or something?
>
> --Emil
> _______________________________________________
> LLVM Developers mailing list
> LLVMdev at cs.uiuc.edu         http://llvm.cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/llvmdev

Anton Korobeynikov

2007-Sep-20 16:36 UTC

head link

[LLVMdev] 2.1 Pre-Release Available (testers needed)

Emil,
> 	# of expected passes            2243
> 	# of unexpected failures        7
> 	# of expected failures          5Could you please provide a log here? It's interesting to see, what is
failing now.

-- 
With best regards, Anton Korobeynikov.

Faculty of Mathematics & Mechanics, Saint Petersburg State University.

Emil Mikulic

2007-Sep-21 00:43 UTC

head link

[LLVMdev] 2.1 Pre-Release Available (testers needed)

On Thu, Sep 20, 2007 at 08:36:39PM +0400, Anton Korobeynikov
wrote:> Emil,
> 
> > 	# of expected passes            2243
> > 	# of unexpected failures        7
> > 	# of expected failures          5
>
> Could you please provide a log here? It's interesting to see, what is
> failing now.
Sure: http://goanna.cs.rmit.edu.au/~emil/llvm2.1-check-armfix.txt

I think it's the same as my previous attempt, except without the ARM
problems.

--Emil

Maybe Matching Threads

Search for more seemingly similar threads

llvm dev - Sep 2007 - [LLVMdev] 2.1 Pre-Release Available (testers needed)

[LLVMdev] 2.1 Pre-Release Available (testers needed)

[LLVMdev] 2.1 Pre-Release Available (testers needed)

[LLVMdev] 2.1 Pre-Release Available (testers needed)

[LLVMdev] 2.1 Pre-Release Available (testers needed)

[LLVMdev] 2.1 Pre-Release Available (testers needed)

Maybe Matching Threads