From patchwork Wed Aug 5 03:26:43 2009 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Vince Weaver X-Patchwork-Id: 30763 Return-Path: X-Original-To: incoming@patchwork.ozlabs.org Delivered-To: patchwork-incoming@bilbo.ozlabs.org Received: from lists.gnu.org (lists.gnu.org [199.232.76.165]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by bilbo.ozlabs.org (Postfix) with ESMTPS id 800E2B7067 for ; Wed, 5 Aug 2009 13:27:46 +1000 (EST) Received: from localhost ([127.0.0.1]:39356 helo=lists.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MYXAD-0001Ip-Js for incoming@patchwork.ozlabs.org; Tue, 04 Aug 2009 23:27:41 -0400 Received: from mailman by lists.gnu.org with tmda-scanned (Exim 4.43) id 1MYX9k-0001Ik-Gl for qemu-devel@nongnu.org; Tue, 04 Aug 2009 23:27:12 -0400 Received: from exim by lists.gnu.org with spam-scanned (Exim 4.43) id 1MYX9f-0001IR-2E for qemu-devel@nongnu.org; Tue, 04 Aug 2009 23:27:11 -0400 Received: from [199.232.76.173] (port=33074 helo=monty-python.gnu.org) by lists.gnu.org with esmtp (Exim 4.43) id 1MYX9e-0001IO-T4 for qemu-devel@nongnu.org; Tue, 04 Aug 2009 23:27:06 -0400 Received: from mx20.gnu.org ([199.232.41.8]:31731) by monty-python.gnu.org with esmtps (TLS-1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.60) (envelope-from ) id 1MYX9e-0002DI-Dl for qemu-devel@nongnu.org; Tue, 04 Aug 2009 23:27:06 -0400 Received: from csl.cornell.edu ([128.84.224.10] helo=vlsi.csl.cornell.edu) by mx20.gnu.org with esmtp (Exim 4.60) (envelope-from ) id 1MYX9X-00079V-Tk for qemu-devel@nongnu.org; Tue, 04 Aug 2009 23:27:00 -0400 Received: from stanley.csl.cornell.edu (stanley.csl.cornell.edu [128.84.224.15]) by vlsi.csl.cornell.edu (8.13.4/8.13.4) with ESMTP id n753Qhd2070745 for ; Tue, 4 Aug 2009 23:26:48 -0400 (EDT) Date: Tue, 4 Aug 2009 23:26:43 -0400 (EDT) From: Vince Weaver To: qemu-devel@nongnu.org Message-ID: <20090804231934.G93384@stanley.csl.cornell.edu> MIME-Version: 1.0 X-Detected-Operating-System: by mx20.gnu.org: FreeBSD 2.0-4.2 X-detected-operating-system: by monty-python.gnu.org: GNU/Linux 2.6, seldom 2.4 (older, 4) Subject: [Qemu-devel] [patch] Fix extlh instruction on Alpha X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: qemu-devel.nongnu.org List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Errors-To: qemu-devel-bounces+incoming=patchwork.ozlabs.org@nongnu.org Hello The extlh instruction on Alpha currently doesn't work properly. It's a combination of a cut/paste bug (16 where it should be 32) as well as a "shift by 64" bug. Below is a patch that fixes the problem, and attached is a test case that exhibits the bug. The program should print a 4-char wide sliding window across the test string; without the patch this fails. Vince Signed-off-by: Vince Weaver # uldl.s by Vince Weaver # This shows a bug with Qemu in handling of the extlh instruction # assemble with "as -o uldl.o uldl.s" # link with "ld -o uldl uldl.o" # syscall numbers .equ SYSCALL_EXIT,1 .equ SYSCALL_WRITE,4 .equ STDIN,0 .equ STDOUT,1 .equ STDERR,2 .globl _start _start: br $27,0 # fake branch, to grab the location # of our entry point ldgp $gp,0($27) # load the GP proper for our entry point # this does automagic stuff... # gp is used for 64-bit jumps and constants # so if you use "la" and the like it will # load from gp for you. lda $17,title # load title br $26,write_stdout # print it lda $17,test_string # load test string br $26,write_stdout # print it lda $13,four_bytes # point $13 to our 32-bit wide # test location lda $11,test_string # point $11 to beginning of test string addq $11,20,$14 # repeat 20 times load_loop: # uldl $12,0($11) # load 32-bits from it # This expands to the following .set noat lda $28,0($11) ldq_u $23,0($28) ldq_u $24,3($28) extll $23,$28,$23 extlh $24,$28,$24 or $23,$24,$12 sextl $12,$12 .set at stl $12,0($13) # store to 4-byte location lda $17,four_bytes # point to it br $26,write_stdout # print 4 chars addq $11,1,$11 cmpeq $11,$14,$1 beq $1,load_loop #================================ # Exit #================================ exit: clr $16 # 0 exit value mov SYSCALL_EXIT,$0 # put the exit syscall number in v0 callsys # and exit #================================ # WRITE_STDOUT #================================ # $17 has string # $1 is trashed write_stdout: ldil $0,SYSCALL_WRITE # Write syscall in $0 ldil $16,STDOUT # 1 in $16 (stdout) clr $18 # 0 (count) in $18 str_loop1: addq $17,$18,$1 # offset in $1 ldbu $1,0($1) # load byte addq $18,1,$18 # increment pointer bne $1,str_loop1 # if not nul, repeat subq $18,1,$18 # correct count callsys # Make syscall ret $26 # return .data .align 3 four_bytes: .ascii "RPLC\n\0" .align 3 eight_bytes: .ascii "REPLACE!\n\0" title: .ascii "ULDL Test\n\0" linefeed: .ascii "\n\0" .align 3 test_string: .ascii "The quick brown fox jumped over the lazy dog\n\0" diff --git a/target-alpha/translate.c b/target-alpha/translate.c index 1fc5119..2a681b0 100644 --- a/target-alpha/translate.c +++ b/target-alpha/translate.c @@ -526,14 +526,24 @@ static always_inline void gen_ext_h(void (*tcg_gen_ext_i64)(TCGv t0, TCGv t1), else tcg_gen_mov_i64(cpu_ir[rc], cpu_ir[ra]); } else { + int l1; TCGv tmp1, tmp2; tmp1 = tcg_temp_new(); + l1 = gen_new_label(); + tcg_gen_andi_i64(tmp1, cpu_ir[rb], 7); tcg_gen_shli_i64(tmp1, tmp1, 3); + + tcg_gen_mov_i64(cpu_ir[rc], cpu_ir[ra]); + tcg_gen_brcondi_i64(TCG_COND_EQ, tmp1, 0, l1); + tmp2 = tcg_const_i64(64); tcg_gen_sub_i64(tmp1, tmp2, tmp1); tcg_temp_free(tmp2); tcg_gen_shl_i64(cpu_ir[rc], cpu_ir[ra], tmp1); + + gen_set_label(l1); + tcg_temp_free(tmp1); } if (tcg_gen_ext_i64) @@ -1320,7 +1330,7 @@ static always_inline int translate_one (DisasContext *ctx, uint32_t insn) break; case 0x6A: /* EXTLH */ - gen_ext_h(&tcg_gen_ext16u_i64, ra, rb, rc, islit, lit); + gen_ext_h(&tcg_gen_ext32u_i64, ra, rb, rc, islit, lit); break; case 0x72: /* MSKQH */