similar to: [LLVMdev] Is LowerInvoke's "-enable-correct-eh-support" option unused?

The PNaCl project has implemented various IR simplification passes that simplify LLVM IR by lowering complex features to simpler features. We'd like to upstream some of these IR passes to LLVM. We'd like to explore if this acceptable, and if so, how we should go about doing this. The immediate reason is that Emscripten is reusing PNaCl's IR passes for its new "fastcomp"

On Tue, Mar 4, 2014 at 6:17 PM, Chandler Carruth <chandlerc at google.com>wrote: > On Tue, Mar 4, 2014 at 1:04 PM, Mark Seaborn <mseaborn at chromium.org>wrote: > >> The PNaCl project has implemented various IR simplification passes that >> simplify LLVM IR by lowering complex features to simpler features. We'd >> like to upstream some of these IR passes

When testing with LLVM 3.0 I noticed a crash in LLC with a test case we are running. After examining the source, I found that the crash coming from SjLjEHPrepare.cpp, specifically from "ReplaceLandingPadVal". I can explain about the crash if needed, but I will omit that for now. Testing with and LLC built from the top of trunk showed no LLC crash, and the code in SjLhEHPrepare.cpp has

All, I'm looking at the lowerinvoke pass as a starting point for getting SJLJ based exception handling working on ARM, but am having some troubles with it. When I run a simple testcase (attached) through llc and specify -enable-correct-eh-support, llc asserts on me. It appears there's been some bitrot somewhere. SelectDAGBuild and SelectionDAGISel cooperate to track landing pads

On Tue, Mar 4, 2014 at 3:11 PM, Sean Silva <chisophugis at gmail.com> wrote: > On Tue, Mar 4, 2014 at 4:04 PM, Mark Seaborn <mseaborn at chromium.org>wrote: > >> The PNaCl project has implemented various IR simplification passes that >> simplify LLVM IR by lowering complex features to simpler features. We'd >> like to upstream some of these IR passes to

Hi Jim, > I'm looking at the lowerinvoke pass as a starting point for getting > SJLJ based exception handling working on ARM, but am having some > troubles with it. When I run a simple testcase (attached) through llc > and specify -enable-correct-eh-support, llc asserts on me. It appears > there's been some bitrot somewhere. SelectDAGBuild and >

Thanks Mark! That'd be a useful starting point. For the rest of the people here: to be a bit more specific, I am having doubts about the following: - Can later optimization passes or code generation still create machine code that takes and stores (in a register perhaps) an address of something on the stack, even if not semantically visible to the programmer? Can this somehow be detected?

On Sun, Dec 18, 2016 at 11:58 PM, Jonas Maebe <jonas-devlists at watlock.be> wrote: > > Actually, there's another —even more fundamental— problem: the longjmp > will always restore the non-volatile registers to the contents they had > at the start of the try-block, which is not what LLVM expects when > entering an SEH-based landing pad. > The SjLjEHPrepare pass tries

Hi, I am toying with an idea of having LLVM generate code that has position-independent stacks. This would be a very useful property for implementing all sorts of micro-thread libraries (I am thinking something similar to Python greenlets <http://stackoverflow.com/a/17447308>), because you'd be able to easily save threadlet state from one OS thread and later restore it into another.

On Apr 21, 2009, at 1:53 PM, Jim Grosbach wrote: > All, > > I'm looking at the lowerinvoke pass as a starting point for getting > SJLJ based exception handling working on ARM, but am having some > troubles with it. When I run a simple testcase (attached) through > llc and specify -enable-correct-eh-support, llc asserts on me. It > appears there's been some

Although most of the leaks I detected in LLVM were from singleton objects, there also seem to be some real leaks. One such leak (which is creating problems for me when I try to get rid of the constant singletons) seems to be a GlobalVariable created on line 145 of Transforms/Scalar/lowerInvoke.cpp -- any suggestions how I can make sure this GlobalVariable gets deleted? Actually I'm a bit

On Fri, 2005-05-13 at 08:06 +0200, Markus F.X.J. Oberhumer wrote: > Actually that's the only case I stumbled over this problem in a somewhat > larger C++ program, and it's clearly the wrong type in LowerInvoke.cpp - > it really should be IntPtrTy. But maybe we could use just IntTy here to > avoid target dependencies. Wait a minute. You want to lower a 64 bit thing to a 32

Hello, Jim > -enable-correct-eh-support, llc asserts on me. It appears there's been some > bitrot somewhere. Well, correct. Because many places expects exceptions to be dwarf-style. > Is it reasonable to expect that lowerinvoke is a good place to start for > doing what I'm after? I really don't think so. Since you're trying to map dwarf-based structures into sjlj

Alastair Lynn wrote: > > From what I understand, the unwind instruction is implemented only > for the interpreter: there is a -lowerunwind pass for compiling to > other systems which will either lower unwind and invoke to setjmp/ > longjmp (slow) or turn invokes into calls and unwinds into abort()s. > Ah cheers, Alastair. That could be very useful (at least for me

The functions that I placed in ScalarTargetTransformInfo are functions what were used by LSR and LowerInvoke. getJumpSize and getJumpAlignment are used by LowerInvoke. Do you suggest that I remove them from TargetLowering and keep them in ScalarTargetTransformInfo ? Thanks, Nadav On Oct 9, 2012, at 5:47 PM, Evan Cheng <evan.cheng at apple.com> wrote: > Hi Nadav, > > The

im compiling llvm with visual studio 10. I have used cmake build app. visual studio 10 is complaining about error c2039: 'setjmp' : is not a member of llvm::Intrinsic its line 154, LowerInvoke.cpp any ideas? -- View this message in context: http://old.nabble.com/errors-when-compiling-with-visual-studio-10-tp29427176p29427176.html Sent from the LLVM - Dev mailing list archive at

Chris would like to turn on -enable-eh rather than -enable-correct-eh- support in the llvm testsuite for those targets that support it. The following patch is intended to turn it on for x86 and ppc. Anton, Duncan, are you OK with this? -------------- next part -------------- A non-text attachment was scrubbed... Name: mf.patch Type: application/octet-stream Size: 669 bytes Desc: not

Hello Clang- and LLVM-Experts, I was not sure which list is the right one, so I tried both - sorry for any inconvenient >o< Lately I was working a lot with exceptions under Windows and especially with the Clang compiler. Out of curiosity I came along "Exception Handling in LLVM" and tried to understand the SJLJ exception handling. At first glance this made total sense to me! Store

> probably there should be a switch to choose whether codegen should turn > unwind/invoke into dwarf or setjmp/longjmp style code. There is, but it happens before codegen and is slow. -enable-correct-eh-support will translate invoke/unwind into setjmp/longjmp pairs for the correct behavior. See: http://llvm.org/docs/Passes.html#lowerinvoke Nick > > Ciao, > > Duncan. >

From what I understand, the unwind instruction is implemented only for the interpreter: there is a -lowerunwind pass for compiling to other systems which will either lower unwind and invoke to setjmp/ longjmp (slow) or turn invokes into calls and unwinds into abort()s. On 26 Dec 2008, at 15:19, Matt Giuca wrote: > > > Jon Harrop wrote: >> >> Is it? I was just reading