On Tue, May 10, 2011 at 2:10 PM, H.J. Lu wrote: > On Tue, May 10, 2011 at 11:28 AM, Quentin Neill > wrote: >> On Tue, May 10, 2011 at 12:55 PM, H.J. Lu wrote: >>> On Tue, May 10, 2011 at 9:51 AM, Quentin Neill >>> wrote: >>>> On Fri, May 6, 2011 at 7:40 AM, H.J. Lu wrote: >>>>> On Thu, May 5, 2011 at 9:09 PM, Quentin Neill >>>>> wrote: >>>>>> On Wed, May 4, 2011 at 11:45 AM, H.J. Lu wrote: >>>>>>> On Wed, May 4, 2011 at 9:16 AM, Quentin Neill >>>>>>> wrote: >>>>>>>> This patch adds the bdver2 processor flag.  It passes make check on x86-64. >>>>>>>> >>>>>>>> Okay to commit? >>>>>>>> -- >>>>>>>> Quentin >>>>>>>> >>>>>>>> >>>>>>>> gas/ >>>>>>>>        * config/tc-i386.c (cpu_arch): Add bdver2. >>>>>>>>        (i386_align_code): Add PROCESSOR_BDVER2 cases. >>>>>>>>        * config/tc-i386.h (processor_type): Add PROCESSOR_BDVER2. >>>>>>>>        * doc/c-i386.texi: Add bdver2. >>>>>>>> >>>>>>>> opcodes/ >>>>>>>>        * i386-gen.c (cpu_flag_init): Add new CPU_BDVER2_FLAGS. >>>>>>>>        * i386-init.h: Regenerated. >>>>>>>> >>>>>>>> gas/testsuite/ >>>>>>>>        * gas/i386/i386.exp: Add new bdver2 test cases. >>>>>>>>        * gas/i386/nops-1-bdver2.d: New. >>>>>>>>        * gas/i386/x86-64-nops-1-bdver2.d: New. >>>>>>>> >>>>>>> >>>>>>> Do you nreally nned PROCESSOR_BDVER2? Can you use PROCESSOR_BDVER1? instead? >>>>>>> >>>>>>> -- >>>>>>> H.J. >>>>>>> >>>>>> >>>>>> BDVER2 adds BMI and TBM. >>>>> >>>>> That is controlled by CPU_BDVER2_FLAGS. >>>>> >>>> >>>> [ Sorry for the delay,  I was out for a couple of days.  Re-adding >>>> binutils@sourceware.org (I meant to copy on my 2nd reply). ] >>>> >>>> So without PROCESSOR_BDVER2, that means no -march=bdver2 flag support, >>>> but what else? >>>> >>> >>> Can you replace PROCESSOR_BDVER2 with  PROCESSOR_BDVER1 >>> to see if it makes a difference? >>> >> Now I understand, I was thinking you meant remove "bdver2" flag as well. >> >> Should I rename the PROCESSOR_BDVER1 variable to perhaps PROCESSOR_BDVER? > > Sounds good to me. Okay, the attached follows H.J.'s suggestions, and passes "make check" in x86-64. Okay to commit? -- Quentin