similar to: [LLVMdev] gdb + clang/llvm

Hi Roland, Is Foo.cpp a Clang source file? If so, make sure you run the compiler itself with "clang -cc1", not the compiler *driver* with bare "clang". See here ( http://lists.cs.uiuc.edu/pipermail/cfe-dev/2012-March/020275.html ) for a similar question. -Matt On 04/13/2012 04:19 PM, Roland Leißa wrote: > Hi all, > > I currently have a problem setting up gdb

On Apr 13, 2012, at 2:19 PM, Roland Leißa <leissa at cs.uni-saarland.de> wrote: > Hi all, > > I currently have a problem setting up gdb with clang/llvm. > > I configured llvm/clang (trunk) with: > ../llvm/configure --prefix=PROJECT_DIR/install_debug --enable-expensive-checks --enable-targets=host --enable-debug-runtime --enable-shared > > Now, when I invoke gdb

> > ok, I now have a problem in the code generator. > > Let's say I want to set a breakpoint somewhere in > > clang/lib/CodeGen/CGExprScalar.cpp > > > > What do I have to do in order to trigger breakpoints, see > > backtraces and so on? > > You set a breakpoint and ask to see a backtrace? > > -eric haha, OK, I think I should be more specific.

Hi Eric, I'm using gcc. My make command is make -j9 as mentioned earlier, the -cc1 switch did the trick for me. Thanks for helping out. -- Roland ----- Original Message ----- > From: "Eric Christopher" <echristo at apple.com> > To: "Roland Leißa" <leissa at cs.uni-saarland.de> > Cc: LLVMdev at cs.uiuc.edu > Sent: Freitag, 13. April 2012 14:39:53

Hi Jim, thanks for replying. As mentioned in my earlier mail, I have to specify -emit-obj along with -cc1 in order to trigger codegen. This solves the problem for me. As I learned from the last email, specifying -v is a good starting point in order to see what actually is happening and tackle such problems. -- Roland ----- Original Message ----- > From: "Jim Grosbach"

ok, I now have a problem in the code generator. Let's say I want to set a breakpoint somewhere in clang/lib/CodeGen/CGExprScalar.cpp What do I have to do in order to trigger breakpoints, see backtraces and so on? -- Roland ----- Original Message ----- > From: "Eric Christopher" <echristo at apple.com> > To: "Roland Leißa" <leissa at cs.uni-saarland.de>

On Apr 23, 2012, at 2:14 PM, Roland Leißa <leissa at cs.uni-saarland.de> wrote: >>> ok, I now have a problem in the code generator. >>> Let's say I want to set a breakpoint somewhere in >>> clang/lib/CodeGen/CGExprScalar.cpp >>> >>> What do I have to do in order to trigger breakpoints, see >>> backtraces and so on? >>

On Apr 13, 2012, at 3:45 PM, Roland Leißa <leissa at cs.uni-saarland.de> wrote: > Hi Eric, > > I'm using gcc. My make command is > make -j9 > > as mentioned earlier, the -cc1 switch did the trick for me. Thanks for helping out. Aha. Yes, that'd do it. Glad you're up and debugging. -eric

On Apr 23, 2012, at 2:14 PM, Roland Leißa <leissa at cs.uni-saarland.de> wrote: > As you can see the line is question is executed but the breakpoint is not triggered. Apparently, clang executes the code generation in a different process. What do I have to do in order to debug this process? What you did before, run it under -cc1. The best way is to use -v to get the full -cc1 line and

On Apr 20, 2012, at 6:35 PM, Roland Leißa <leissa at cs.uni-saarland.de> wrote: > ok, I now have a problem in the code generator. > Let's say I want to set a breakpoint somewhere in > clang/lib/CodeGen/CGExprScalar.cpp > > What do I have to do in order to trigger breakpoints, see backtraces and so on? You set a breakpoint and ask to see a backtrace? -eric

Hi all, I'm currently working on an extension for clang. Apparently, I'm doing sth incorrectly currently. The Verify complains with Wrong type for attribute signext <8 x i32> (<256 x i1>, i16, <2 x double>)* @_Z4testsDv2_d Now, this information is not that useful. Is it possible to emit this "broken" llvm code in order to better understand what went wrong? --

Hi all, from llvm-trunk I'm getting this failure: llvm[0]: Running test suite make[1]: Entering directory `/home/rleissa/projects/llvm/build_debug/test' Making a new site.exp file... Making LLVM 'lit.site.cfg' file... Making LLVM unittest 'lit.site.cfg' file... make -C /home/rleissa/projects/llvm/build_debug/test/../tools/clang/test lit.site.cfg Unit/lit.site.cfg

Hi all, I've just written a small enhancement to SimplifyPHINode. The idea is the following: If we have this: a = phi(X, X, undef, undef, X, undef) X = select cond, sth, a or this: a = phi(X, X, undef, undef, X, undef) X = select cond, a, sth we can replace the phi by 'a' and the select by 'sth'. Why does this work? Well, in those cases where control-flow happens to

Hi Leslie, I suggest adding these 3 papers to your reading list. Register allocation for programs in SSA-form Sebastian Hack, Daniel Grund, and Gerhard Goos http://www.rw.cdl.uni-saarland.de/~grund/papers/cc06-ra_ssa.pdf Simple and Efficient Construction of Static Single Assignment Form Matthias Braun , Sebastian Buchwald , Sebastian Hack , Roland Leißa , Christoph Mallon , and Andreas

Hi Duncan, thanks for the hint. I tried both variants: define <4 x float> @masked_add_1(<4 x i1> signext %mask, <4 x float> %a, <4 x float> %b) define <4 x float> @masked_add_32(<4 x i32> %mask, <4 x float> %a, <4 x float> %b) Unfortunately, this will raise an assertion: Wrong types for attribute: zeroext signext noalias nocapture sret byval nest

I consider my patch not mature enough to be committed. I just wanted to hear what others are saying. I'm not quite sure how to exactly deal with these cascades of selects which induce a cycle via a Phi; I'd like to implement that as well. Also, I'm not sure whether InstructionSimplify.cpp is the proper place for this optimization. -- Roland ----- Original Message ----- > From:

Hey guys, I can't seem to figure out how to debug a COM32 app with GDB under qemu. I start "qemu -s" and then "gdb hello.elf ; target remote:1234" but can't set any breakpoints that work. I"m using syslinux4 (pathbased from the repo). Can anyone verify that they can use qemu+com32+gdb with something as simple as com32/samples/hello.c32 ? Also, is it

> Hello, > I have rsync process which is a crontab job: > 0 0-23 * * * /bin/rsync -rvg /net/server1/local/filesys1/userhomedir/project_dir /local/filesys1/userhomedir/project_dir > which sync the user project dir on localsystem from another machine called server1 via nfs, but the weird thing its coping all files every time it syncs am I using the wrong flags? is it possible to copy

Hi Andi, Here's a couple of patches to fix up COMPAT_VDSO: The first is a straightforward implementation of Jan's original idea of relocating the VDSO to match its mapped location. Unlike Jan and Zach's version, I changed it to relocate based on the phdrs rather than the sections; the result is pleasantly compact. The second patch takes advantage of the fact that all the

Hi Andi, Here's a couple of patches to fix up COMPAT_VDSO: The first is a straightforward implementation of Jan's original idea of relocating the VDSO to match its mapped location. Unlike Jan and Zach's version, I changed it to relocate based on the phdrs rather than the sections; the result is pleasantly compact. The second patch takes advantage of the fact that all the