similar to: [LLVMdev] Getting rid of phi instructions?

On Tue, Aug 30, 2011 at 9:12 AM, Teemu Rinta-aho <teemu.rinta-aho at nomadiclab.com> wrote: > Hi all, > > is there a pass to get rid of phi-instructions in a function? There's no loop involved. reg2mem. > I have a function approx. like this: > > void @func() { > entry: >        … > bb1: >        … > bb2: >        … >        %tmp100 = phi i32 [ 0,

Hi, We are writing a utility whose input includes the names of some C++ classes, and we want to locate the corresponding LLVM types in the current module. I'm trying to look through LLVM and clang sources but cannot figure out how the LLVM type names are constructed for the C++ class names. Any help appreciated (like the name of a library call to do it or even a pointer to code where this

On Wed, Aug 31, 2011 at 2:35 AM, Teemu Rinta-aho <teemu.rinta-aho at nomadiclab.com> wrote: > On 31.8.2011, at 11.41, Eli Friedman wrote: > >> Mmm... reg2mem will transform IR with PHI's into IR without them, but >> it generates a bunch of alloca's, which I would assume are not cheap >> to lower to VHDL.  You might have to write your own pass to get the

> the next tool reading the IR does not like phis when it's generating VHDL. If you're doing a conversion from LLVM IR to some other non-SSA IR (like the tool's), you can do the phi node removal yourself as you convert. Basically, every predecessor block referenced by a phi node will have an assignment to that variable before branching. There are techniques to make the resultant

I have a problem with long user names. I was thinking if I could use Firstname.Lastname notation also in user names to be used in [Homes]. We are using Alpha Linux as one of our servers. I tried it, but if the username is over 12 characters long, NT 4.0 client (at least) does not co-operate with [Homes]-share. It gives you an error: "Network name can not be found" I tried it also with a

On 30.8.2011, at 19.19, Eli Friedman wrote: > reg2mem won't do quite this transformation... not sure exactly what you need. I need to get rid of phis. This code is compiled from C++ and for some functions there are no phis, but multiple call instructions. I am targeting hardware in the end, and the next tool reading the IR does not like phis when it's generating VHDL. My questions may

On 31.8.2011, at 11.41, Eli Friedman wrote: > Mmm... reg2mem will transform IR with PHI's into IR without them, but > it generates a bunch of alloca's, which I would assume are not cheap > to lower to VHDL. You might have to write your own pass to get the > precise transformation you're looking for. Right. Thanks. I need to see the reg2mem source code. Teemu

On Wed, Aug 31, 2011 at 1:06 AM, Teemu Rinta-aho <teemu.rinta-aho at nomadiclab.com> wrote: > On 30.8.2011, at 19.19, Eli Friedman wrote: > >> reg2mem won't do quite this transformation... not sure exactly what you need. > > I need to get rid of phis. This code is compiled from C++ and for some functions > there are no phis, but multiple call instructions. I am

Hello everybody, I have stumbled upon a test case (the attached module is a slightly reduced version) that shows extremely reduced performance on linux compared to windows when executed using LLVM's JIT. We narrowed the problem down to the actual code being generated, the source IR on both systems is the same. Try compiling the attached module: llc -O3 -filetype=asm -o BAD.s BAD.ll Under

I'm attempting to modify a parser generator to emit LLVM code instead of C. So far the experience has been trivial, but I am now running into an error regarding labels that I can't seem to solve. Situation 1: A label is used immediately after a void function call (l6 in this case): <snip> %tmp26 = load i32* @yybegin, align 4 %tmp27 = load i32* @yyend, align 4 call void

On Sep 29, 2010, at 8:35 AMPDT, Ralf Karrenberg wrote: > Hello everybody, > > I have stumbled upon a test case (the attached module is a slightly > reduced version) that shows extremely reduced performance on linux > compared to windows when executed using LLVM's JIT. > > We narrowed the problem down to the actual code being generated, the > source IR on both systems

it were a Microsoft box... (Windoze is so stupid it'll never know the difference!) -----Original Message----- From: samba-admin@lists.samba.org [mailto:samba-admin@lists.samba.org] On Behalf Of Fred Jackson Sent: Monday, July 30, 2001 10:39 PM To: samba@lists.samba.org Subject: a question? Hi does the command line smbpasswd -j <domain> -r <PDC> apply to clients joining a

cheers, jerry --------------------------------------------------------------------- www.valinux.com VA Linux Systems gcarter_at_valinux.com www.samba.org SAMBA Team jerry_at_samba.org www.plainjoe.org jerry_at_plainjoe.org --"I never saved anything for the swim back." Ethan Hawk in Gattaca-- Return-Path:

apply to this (security=domain) case, since auth is via the PDC. So, shouldn't winbind_lookup_sid() be succeeding. Should I be trying to track the failure in that code? In case it's helpful, ldd /usr/sbin/smbd shows: libdl.so.2 => /usr/lib/libdl.so.2 (0x40017000) libnsl.so.1 => /lib/libnsl.so.1 (0x4001a000) libpam.so.0 => /lib/libpam.so.0 (0x40030000) libc.so.6 =>

have a hosts allow or hosts deny line in your smb.conf. If you don't want samba to do reverse lookups, then comment out any allow/deny hosts lines in your smb.conf file. Don -----Original Message----- From: future@yxtc.edu.cn [mailto:future@yxtc.edu.cn] Sent: Sunday, August 05, 2001 3:00 AM To: samba@lists.samba.org Subject: dns and samba Hi, I find that my samba server always does

Is there an official Git mirror for Dragonegg? Cannot find the URL with Google nor with guesswork. Teemu

Hi all, I used to use llvm-g++ to transform C++ source to LLVM IR. Now I am trying to switch to using DragonEgg, but I faced some problems because of the way DragonEgg is naming types in the resulting IR file. An example: LLVM-GCC: %"struct.Vector<ARPResponder::Entry>" = type { %"struct.ARPResponder::Entry"*, i32, i32 } %"struct.Vector<String>" = type {