On Mon, Mar 15, 2021 at 01:45:21PM +0530, Naresh Kamboju wrote: > On Mon, 15 Mar 2021 at 05:56, Guenter Roeck <linux@xxxxxxxxxxxx> wrote: > > > > Building arm:axm55xx_defconfig ... failed > > -------------- > > Error log: > > /tmp/cc2ylxxJ.s: Assembler messages: > > /tmp/cc2ylxxJ.s:87: Error: co-processor register expected -- `mrc p10,7,r7,FPEXC,cr0,0' > > /tmp/cc2ylxxJ.s:103: Error: co-processor register expected -- `mcr p10,7,r3,FPEXC,cr0,0' > > /tmp/cc2ylxxJ.s:537: Error: co-processor register expected -- `mcr p10,7,r7,FPEXC,cr0,0' > > make[3]: *** [arch/arm/kvm/hyp/switch.o] Error 1 > > make[2]: *** [arch/arm/kvm/hyp] Error 2 > > make[2]: *** Waiting for unfinished jobs.... > > make[1]: *** [arch/arm/kvm] Error 2 > > make[1]: *** Waiting for unfinished jobs.... > > make: *** [sub-make] Error 2 > > These issues were noticed on the stable rc 5.4 branch also while building for > arm axm55xx_defconfig. > > /tmp/ccKgdkaN.s: Assembler messages: > /tmp/ccKgdkaN.s:87: Error: co-processor register expected -- `mrc > p10,7,r7,FPEXC,cr0,0' > /tmp/ccKgdkaN.s:103: Error: co-processor register expected -- `mcr > p10,7,r3,FPEXC,cr0,0' > /tmp/ccKgdkaN.s:556: Error: co-processor register expected -- `mcr > p10,7,r7,FPEXC,cr0,0' > make[3]: *** [/builds/1piRajjPILoGGDzi5cHI5ZMuCJZ/scripts/Makefile.build:261: > arch/arm/kvm/hyp/switch.o] Error 1 > > > > -------------- > > > > I didn't find an obvious candidate so I bisected it. > > > > # bad: [a233c6b3f6de88ca62da8fde45f330b104827851] Linux 4.19.181-rc1 > > # good: [030194a5b292bb7613407668d85af0b987bb9839] Linux 4.19.180 > > git bisect start 'HEAD' 'v4.19.180' > > # good: [ecee76d4b15b8431827e910589edfb4c12a589f9] powerpc/perf: Record counter overflow always if SAMPLE_IP is unset > > git bisect good ecee76d4b15b8431827e910589edfb4c12a589f9 > > # good: [722ce092b23ae91337694d40e6ac216b16962788] ARM: 8929/1: use APSR_nzcv instead of r15 as mrc operand > > git bisect good 722ce092b23ae91337694d40e6ac216b16962788 > > # bad: [2e6919206bb0bcac507b7905fc7c9b3dd861ab4b] ARM: 9025/1: Kconfig: CPU_BIG_ENDIAN depends on !LD_IS_LLD > > git bisect bad 2e6919206bb0bcac507b7905fc7c9b3dd861ab4b > > # good: [831e354481111c30d68c980434e2cfe42590f189] kbuild: add CONFIG_LD_IS_LLD > > git bisect good 831e354481111c30d68c980434e2cfe42590f189 > > # bad: [9b99f469087843c9216976865a97da96f9cdcbbc] ARM: 8991/1: use VFP assembler mnemonics if available > > git bisect bad 9b99f469087843c9216976865a97da96f9cdcbbc > > # good: [41ad45cb9ecb66f76abc77d938b3693839fb5e20] ARM: 8990/1: use VFP assembler mnemonics in register load/store macros > > git bisect good 41ad45cb9ecb66f76abc77d938b3693839fb5e20 > > # first bad commit: [9b99f469087843c9216976865a97da96f9cdcbbc] ARM: 8991/1: use VFP assembler mnemonics if available > > > > Reverting the offending patch from v4.19.y-queue fixes the problem. > > I didn't check v5.4.y-queue. > > https://gitlab.com/Linaro/lkft/mirrors/stable/linux-stable-rc/-/jobs/1095669013#L346 Thanks, both trees should now be fixed. greg k-h
