similar to: No remapping of clone instruction in CloneBasicBlock

Hi Henrique, Thanks for the quick reply! On Tue, Nov 12, 2013 at 9:13 PM, Henrique Santos < henrique.nazare.santos at gmail.com> wrote: > PRE normally uses a latest placement algorithm to do something of the sort. > I don't know about GVN/PRE, but older version of PRE might have it. > Just placing the calls to free at the predecessors (dominated by BB12) of > the dominance

Hi all, I have been writing a pass to heapify some alloca's (it is pessimistization, not optimization). For example, in the following control flow graph, there is a call to malloc inserted in block BB12. In order to avoid memory leak, free's are needed. The free cannot be inserted in BB23 because BB23 is not dominated by BB12. There are two ways to go I can think of here. One way is to

Hi Henrique, I have tried using -mergereturn and inserting a free into the predecessors of dominance frontier of malloc block and it caused double free. It is possible for multiple free's to be inserted on the path from malloc to an exit. For example, in the following CFG: BB10 (malloc) / \ BB11 BB12 ... / \ / \

> > It seems that placing the calls to free at the predecessors of dominance > frontier is inadequate. It is possible that there are exit blocks that are > dominated by BB12 (calls to malloc). I guess we can also insert calls to > free at these exit blocks too. That crossed my mind a few minutes later. : ) If you're interested, PRE.cpp existed last at r25315. It calculates the

Try breaking the critical edges (-break-crit-edges). This way, a new block will be created between BB13 and BB11 (call this BB11.break) and BB15 and BB12 (call this BB12.break). The predecessors of the dominance frontier will, thus, be BB11.break, BB12.break, and BB14. When we enter through a block with a call to malloc(), we will end up in one of the blocks in the dominance frontier (kind of).

Thanks! I will try that and see whether it works. On Wed, Nov 13, 2013 at 5:01 AM, Henrique Santos < henrique.nazare.santos at gmail.com> wrote: > It seems that placing the calls to free at the predecessors of dominance >> frontier is inadequate. It is possible that there are exit blocks that are >> dominated by BB12 (calls to malloc). I guess we can also insert calls to

PRE normally uses a latest placement algorithm to do something of the sort. I don't know about GVN/PRE, but older version of PRE might have it. Just placing the calls to free at the predecessors (dominated by BB12) of the dominance frontier of BB12 seems to work, however. Is there anything wrong with this? H. On Tue, Nov 12, 2013 at 11:30 PM, Bin Tzeng <bintzeng at gmail.com> wrote:

Bill, Just in case this is relevant for you: If you're working with C++ code, or otherwise have any functions that might throw exceptions, you might also need to catch those exceptions in order to free the allocated memory. This will involve looking for calls to functions that mayThrow(), changing their calls to invokes, and freeing the memory before resuming the unwinding. -Hal -----

Dear all, after optimizing a small LLVM example program (i.e., with opt -O3), the resulting code contains several basic blocks, which I consider identical or redundant. For instance, return: ; preds = %min.exit ret i32 0 bb15: ; preds = %min.exit ret i32 0 or, bb7.i:

Would it make sense to have a similar pass that operates on llvm main IR? On Sat, May 8, 2010 at 5:15 PM, Dale Johannesen <dalej at apple.com> wrote: > The branch folding pass does this, but it operates later, on the > target-dependent form in llc. > > On May 8, 2010, at 8:48 AM, Heinz Riener wrote: > >> Dear all, >> >> after optimizing a small LLVM example

The branch folding pass does this, but it operates later, on the target-dependent form in llc. On May 8, 2010, at 8:48 AM, Heinz Riener wrote: > Dear all, > > after optimizing a small LLVM example program (i.e., with opt -O3), > the > resulting code contains several basic blocks, which I consider > identical > or redundant. For instance, > > return:

Hi, I've been working on a Blackfin backend (llvm-3.6.0) based on the previous one that was removed in llvm-3.1. llc generates codes like this: 29 p1 = r2; 30 r5 = [p1]; 31 p1 = r2; 32 r6 = [p1 + 4]; 33 r5 = r6 + r5; 34 r6 = [p0 + -4]; 35 r5 *= r6; 36 p1 = r2; 37 r6 = [p1 + 8]; 38 p1 = r2; p1 and r2 are in different register classes. A p*

Eugene Toder wrote: > Would it make sense to have a similar pass that operates on llvm main IR? > The unify exit return node (-mergereturn?) pass should take care of the first example. The -simplifycfg option might take care of the second example. Both work on LLVM IR. -- John T. > On Sat, May 8, 2010 at 5:15 PM, Dale Johannesen <dalej at apple.com> wrote: > >>

I am working on a transform that uses an scc_iterator on a program's CallGraph. However, I am running into a problem in which not all callees are being processed before callers, leading me to believe there might be a bug. In particular, this is happening in a case where a callee is a function pointer. I ran bugpoint and have included the bugpoint-reduced-simplified.ll below. If

Hi Sanjay, unfortunately that patch does not help in my case. Here's the IR that fails to get fully optimized: target datalayout = "e-m:e-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-unknown-linux-gnu" define fastcc zeroext i1 @testfunc(i8** noalias nocapture readonly dereferenceable(8)) unnamed_addr { entry-block: %1 = load i8*, i8**

Hi, AFAICT there are two places where zext instructions may get folded into PHI nodes. One is FoldPHIArgZextsIntoPHI and the other is the more generic FoldPHIArgOpIntoPHI. Now, the former only handles PHIs with more than 2 incoming values, while the latter only handles casts where the source type is legal. This means that for an PHI node with two incoming i8 values, both resulting from `zext i1

On Sun, 3 Dec 2006, Ryan M. Lefever wrote: > I am working on a transform that uses an scc_iterator on a program's > CallGraph. However, I am running into a problem in which not all > callees are being processed before callers, leading me to believe there > might be a bug. In particular, this is happening in a case where a > callee is a function pointer. I ran bugpoint and

Shuguang Feng wrote: >> What does "llc -debug-pass=Structure" say? Is the ProfileLoaderPass >> really the last pass to touch the ProfileInfo before you are using it? > > Below is the sequence of passes that I see. Although the > NoProfileInfo pass is being run, it should be subsequently overridden > by ProfileInfoLoaderPass (LoaderPass) correct? Yes. >

I printed the call graph as you suggested. The bugpoint bytecode that I sent below prints the following call graph: Indirect call node / | | \ / | | \ V | V V execute | prog_char clear_func | | V V push_constant | V Node0x8d7b370 i.e., Indirect call node -> execute

Hi Quentin, I have rerun the test case on a recent commit, so the numbers have changed. There are also now a few more basic blocks very small basic blocks in the function, and therefore there are some slight differences. I tried to go back to earlier commits, without success for some reason... This is however very similar, except that there becomes two COPYs back to sibling value after the loop.