search for: m_zero

...", foo(INT_MIN)); > } This will print -1 or 1, depending on the optimization level. This appears to be the relevant code: InstCombineAddSub.cpp:1556 > // 0 - (X sdiv C) -> (X sdiv -C) > if (match(Op1, m_SDiv(m_Value(X), m_Constant(C))) && > match(Op0, m_Zero())) > return BinaryOperator::CreateSDiv(X, ConstantExpr::getNeg(C)); - David Menendez

...m_CombineAnd(m_BinOp(m_Value(A), m_Value(B)), m_Value(Mul)), m_AllOnes()), m_Deferred(Mul), m_CombineAnd( m_c_BinOp(m_CombineAnd(m_Sub(m_Zero(), m_Deferred (A)), m_Value(Sub)), m_Deferred(B)), m_Value(Mul2)))) && Pred == ICmpInst::Predicate::ICMP_SGT) { } > Could you guys give some comment? Is there any better i...

Hi, There is an opportunity in instCombine for following instruction pattern: %mul = mul nsw i32 %b, %a %cmp = icmp sgt i32 %mul, -1 %sub = sub i32 0, %a %mul2 = mul nsw i32 %sub, %b %cond = select i1 %cmp, i32 %mul, i32 %mul2 Source code for above pattern: return (a*b) >=0 ? (a*b) : -a*b; Currently, llvm(-O3) can not recognize this as abs(a*b). I initially think we could do this in

This is fundamentally different than the problems we’re trying to handle with the constrained FP intrinsics. The constrained FP intrinsics were necessary because LLVM IR otherwise assumes that FP instructions do not have side effects. In the case of integer divide-by-zero, the optimizer generally recognizes this as a possibility and avoids introducing it (through speculation, for instance). But if

...ot_verbose", RB_VERBOSE}, + {"boot_multicons", RB_MULTIPLE}, + {NULL, 0} +}; + +static struct bios_smap xen_smap[MAX_E820_ENTRIES]; + +static int +start_xen_ap(int cpu) +{ + struct vcpu_guest_context *ctxt; + int ms, cpus = mp_naps; + + ctxt = malloc(sizeof(*ctxt), M_TEMP, M_NOWAIT | M_ZERO); + if (ctxt == NULL) + panic("unable to allocate memory"); + + ctxt->flags = VGCF_IN_KERNEL; + ctxt->user_regs.rip = (unsigned long) init_secondary; + ctxt->user_regs.rsp = (unsigned long) bootSTK; + + /* Set the CPU to use the same page tables and CR4 value */ + ctxt->ctrlr...

...a `Constant`. For scalable vectors this cannot be done and the original sequence is maintained. To mitigate this, `PatternMatch` is extended to better support `ConstantExpr`s along with new helpers like: ```cpp #define m_SplatVector(X) \ m_ShuffleVector( \ m_InsertElement(m_Undef(), X, m_Zero()), \ m_Value(), \ m_Zero()) \ ``` Zero is the exception with `zeroinitializer` applying equally well to scalable and non-scalable vectors. # Predicated floating point arithmetic {#predfparith} When a loop is fully predicated it becomes necessary to have masked versions of faulting i...

Hi, my FreeBSD-9/stable machine (FreeBSD freebsd-tower.goebo.site 9.1-PRERELEASE FreeBSD 9.1-PRERELEASE #2 r241044M: Sat Sep 29 12:52:01 CEST 2012 lbo@freebsd-tower.goebo.site:/usr/obj/usr/src/sys/GENERIC i386) crashes reproducibly when rsync-ing files to an NFSv4 share on the FreeBSD machine. The crash makes the system reboot. The crash creates files in /var/crash which may be obtained

Hi all, I was intending on trying out gjournal on a new disk i've added in my desktop. I had a look to see what the most recent patch provided by Pawel and found http://people.freebsd.org/~pjd/patches/gjournal6_20061024.patch I created the directories as per Pawel's original post (http://lists.freebsd.org/pipermail/freebsd-fs/2006-June/001962.html) and the patch succeeded with no failed