search for: cdq

...version of this function is quite OK, the code for the "long long" version misses an obvious optimisation: lsign: # @lsign mov eax, dword ptr [esp + 4] | mov eax, dword ptr [esp + 4] xor ecx, ecx | test eax, eax | cdq setg cl | neg eax sar eax, 31 | adc edx, edx add eax, ecx | mov eax, edx ret | ret llsign: # @llsign xor ecx, ecx | xor...

For the simple C program below I show the output of clang and the output of the VS compiler (I am on windows). Maybe this is obvious to you, but is it really faster to do 2 multiplications, 3 movl instructions, 2 shifts, 1 add, and 1 substract than to do 1 mov, 1 cdq, and 1 idiv? I run into this while trying to understand why my code runs slower with llvm than a comparable program on windows. Thanks for any help, Brent int f(int n) { return (n + 1) % 18; } "clang -O2 -S" produces this code: _f: # @f # BB#0:...

I downloaded current version of FLAC sources and compiled it with: * GCC 4.8.1 (MSYS from http://xhmikosr.1f0.de/tools/) * Intel C++ Composer XE 2013 update 5 * MSVS 2010 SP1 * MSVS 2012 update 3 (SSSE3 and SSE4.1 code was disabled for all compilers) Stereo 24-bit WAV file was encoded with -8 preset. Encoding time, in seconds: GCC 32-bit: 209 ICC 32-bit: 130 VS10 32-bit: 116 VS12 32-bit: 114

...it_thread_union+THREAD_SIZE, %_ASM_SP #ifdef CONFIG_X86_64 /* Set up %gs. @@ -46,7 +47,7 @@ ENTRY(startup_xen) * init data section till per cpu areas are set up. */ movl $MSR_GS_BASE,%ecx - movq $INIT_PER_CPU_VAR(irq_stack_union),%rax + movabsq $INIT_PER_CPU_VAR(irq_stack_union),%rax cdq wrmsr #endif diff --git a/arch/x86/xen/xen-pvh.S b/arch/x86/xen/xen-pvh.S index e1a5fbeae08d..43e234c7c2de 100644 --- a/arch/x86/xen/xen-pvh.S +++ b/arch/x86/xen/xen-pvh.S @@ -101,8 +101,8 @@ ENTRY(pvh_start_xen) call xen_prepare_pvh /* startup_64 expects boot_params in %rsi. */ - mov $_pa...

http://llvm.org/docs/CodeGenerator.html#machine-code-description-classes Section starting: Fixed (preassigned) registers It talks about converting: define i32 @test(i32 %X, i32 %Y) { %Z = udiv i32 %X, %Y ret i32 %Z } into ;; X is in EAX, Y is in ECX mov %EAX, %EDX sar %EDX, 31 idiv %ECX ret BUT, where does the "sar" come from? Kind Regards James

...uot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

On CentOS5 with the latest updates applied, the logwatch filter for postfix returns way too many lines from the log. I get an "unmatched entries" message and all messages that have gone through the system is listed. Here is an example: 8F930A8092: to=<morten at foo.bar>, orig_to=<morten at localhost>, relay=local, delay=0.19, delays=0.06/0.01/0/0.12, dsn=2.0.0,

...bit sar + */ + .text + .align 4 + .globl __ashrdi3 + .type __ashrdi3, at function +__ashrdi3: +#ifndef _REGPARM + movl 4(%esp),%eax + movl 8(%esp),%edx + movb 12(%esp),%cl +#endif + cmpb $32,%cl + jae 1f + + shrdl %cl,%edx,%eax + sarl %cl,%edx + ret + +1: + sarl %cl,%edx + movl %edx,%eax + cdq + ret + + .size __ashrdi3,.-__ashrdi3 diff --git a/usr/klibc/arch/i386/libgcc/__lshrdi3.S b/usr/klibc/arch/i386/libgcc/__lshrdi3.S new file mode 100644 index 0000000..6a63c52 --- /dev/null +++ b/usr/klibc/arch/i386/libgcc/__lshrdi3.S @@ -0,0 +1,29 @@ +/* + * arch/i386/libgcc/__lshrdi3.S + * + * 64-...

...&%R+7-U8G-C<FEP=', at +5=F;W)M M870],B M5V)A9"UF=6YC=&EO;BUC87-T("!C;VYF=&5S="YC(#$^)C4*8V]N M9FEG=7)E.B!);B!F=6YC=&EO;B!@;6%I;B<Z"F-O;F9I9W5R93HX,CDP.B!S M=&]R86=E('-I>F4@;V8 at 8&AD='(G(&ES;B=T(&MN;W=N"F-O;F9I9W5R93HX M,CDQ.B!W87)N:6YG.B!I;7!L:6-I="!D96-L87)A=&EO;B!O9B!F=6YC=&EO M;B!@<V5N9&9I;&4G"F-O;F9I9W5R93HX,CDP.B!W87)N:6YG.B!U;G5S960@ M=F%R:6%B;&4 at 8&AD='(G"F-O;F9I9W5R93H at 9F%I;&5D('!R;V=R86T@=V%S M. at HC;&EN92 X,C at R(")C;VYF:6=U<F4B"B-...

Changes: - patch v2: - Adapt patch to work post KPTI and compiler changes - Redo all performance testing with latest configs and compilers - Simplify mov macro on PIE (MOVABS now) - Reduce GOT footprint - patch v1: - Simplify ftrace implementation. - Use gcc mstack-protector-guard-reg=%gs with PIE when possible. - rfc v3: - Use --emit-relocs instead of -pie to reduce

Changes: - patch v2: - Adapt patch to work post KPTI and compiler changes - Redo all performance testing with latest configs and compilers - Simplify mov macro on PIE (MOVABS now) - Reduce GOT footprint - patch v1: - Simplify ftrace implementation. - Use gcc mstack-protector-guard-reg=%gs with PIE when possible. - rfc v3: - Use --emit-relocs instead of -pie to reduce

Changes: - patch v3: - Update on message to describe longer term PIE goal. - Minor change on ftrace if condition. - Changed code using xchgq. - patch v2: - Adapt patch to work post KPTI and compiler changes - Redo all performance testing with latest configs and compilers - Simplify mov macro on PIE (MOVABS now) - Reduce GOT footprint - patch v1: - Simplify ftrace

I have updated the klibc patchset based on feedback received. In particular, the patchset has been reorganized so as not to break git-bisect. Additionally, this updates the patch base to 2.6.17-git12 (d38b69689c349f35502b92e20dafb30c62d49d63) and klibc 1.4.8; the main difference on the klibc side is removal of obsolete code. This is also available as a git tree at: