by robert.swiecki:

Before filing a bug, please check whether it has been fixed since
the latest release: run "hg pull -u" and retry what you did to
reproduce the problem.  Thanks.

What steps will reproduce the problem?
1. Type - 6g SIGSEGV.PC.0x7ffff785a052.CODE.128.ADDR.(nil).INSTR.pcmpeqb_xmm2,_[rdi].go

What is the expected output?
Output file created

What do you see instead?
Crash - SIGSEGV at address 0, with instruction pcmpeqb_xmm2,_[rdi] (rdi == 0)

Sometimes it doesn't crash (probably depends on layout of memory), then valgrind output
is useful:

$ valgrind /usr/local/google/src/go/bin/6g
SIGSEGV.PC.0x7ffff785a052.CODE.128.ADDR.\(nil\).INSTR.pcmpeqb_xmm2\,_\[rdi\].go
==31360== Invalid read of size 1
==31360==    at 0x4C286D2: __GI_strlen (mc_replace_strmem.c:284)
==31360==    by 0x43F4D7: fmtstrcpy (dofmt.c:281)
==31360==    by 0x43F8F1: __fmtdispatch (fmt.c:229)
==31360==    by 0x43F72B: dofmt (dofmt.c:80)
==31360==    by 0x43FEF8: fmtvprint (fmtvprint.c:41)
==31360==    by 0x423AE9: adderr (subr.c:50)
==31360==    by 0x423F7D: yyerror (subr.c:175)
==31360==    by 0x42AE4A: typecheck (typecheck.c:409)
==31360==    by 0x429E74: typecheck (typecheck.c:363)
==31360==    by 0x429E58: typecheck (typecheck.c:362)
==31360==    by 0x42997E: typecheck (typecheck.c:1234)
==31360==    by 0x42CBA9: typechecklist (typecheck.c:56)
==31360==  Address 0x20676e697373696d is not stack'd, malloc'd or (recently) free'd
==31360== 

Which compiler are you using (5g, 6g, 8g, gccgo)?
6g

Which operating system are you using?
Linux/amd64


Which revision are you using?  (hg identify)
c5c62aeb6267 tip

Please provide any additional information below.

Attachments:

1. SIGSEGV.PC.0x7ffff785a052.CODE.128.ADDR.(nil).INSTR.pcmpeqb_xmm2,_[rdi].go (1934 bytes)