Date: 12 Apr 2003 09:06:57 -0000 From: Dave <friend@vortex4.net> To: FreeBSD-gnats-submit@freebsd.org Cc: alpha@freebsd.org Subject: -march=pca56 blows up nethack34-nox11 Message-ID: <20030412090657.40307.qmail@frane.vortex4.net>
next in thread | raw e-mail | index | archive | help
>Submitter-Id: current-users
>Originator: Dave <friend@vortex4.net>
>Organization: vortex4.net
>Confidential: no
>Synopsis: -march=pca56 blows up nethack34-nox11
>Severity: non-critical
>Priority: low
>Category: ports
>Class: sw-bug
>Release: FreeBSD 5.0-RELEASE-p6 alpha
>Environment:
System: FreeBSD frane.vortex4.net 5.0-RELEASE-p6 FreeBSD 5.0-RELEASE-p6 #2: Tue Mar 25 04:13:12 PST 2003 root@frane.vortex4.net:/usr/src/sys/alpha/compile/FRANE alpha
PC164SX Alpha system
cvsup'ed to current
built & installed world with the following options set in /etc/make.conf:
CPUTYPE?=pca56
CFLAGS= -O -pipe
CXXFLAGS+= -fmemoize-lookups -fsave-memoized
MAKE_SHELL?=sh
COPTFLAGS= -O -pipe
WANT_FORCE_OPTIMIZATION_DOWNGRADE=1
Most notably. Send me mail if you need the rest of my make.conf
the FRANE kernel pretty much GENERIC minus the hardware that isn't in this box.
>Description:
Built the nethack34-nox11 port. Builds cleanly, appears to run fine until
after the class/race/alignment selection process (or letting the game
choose), at which point the program exits with "illegal instruction".
Posted this problem to alpha@freebsd.org 3/27/2003. Ran gdb at the promping
of Andrew Gallatin with the following results:
Shall I pick a character's race, role, gender and alignment for you? [ynq] y(no
debugging symbols found)...
Program received signal SIGILL, Illegal instruction.
0x1201046d4 in intersect ()
(gdb) bt
#0 0x1201046d4 in intersect ()
#1 0x1201048c4 in split_rects ()
#2 0x120120550 in create_room ()
#3 0x12009c8b8 in makerooms ()
#4 0x12009deac in makelevel ()
#5 0x12009f0d0 in mklev ()
#6 0x120003f34 in newgame ()
#7 0x12017288c in main ()
#8 0x120002704 in _start ()
#9 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#10 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#11 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#12 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#13 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#14 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#15 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#16 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1 #17 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#18 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#19 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#20 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#21 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
#22 0x1601c9944 in _rtld_start () from /usr/libexec/ld-elf.so.1
---Type <return> to continue, or q <return> to quit---
...it goes on like that
(gdb) disassemble 0x1201046d0
Dump of assembler code for function intersect:
0x120104660 <intersect>: ldbu t1,0(a1)
0x120104664 <intersect+4>: sextb t1,t1
0x120104668 <intersect+8>: ldbu t0,2(a0)
0x12010466c <intersect+12>: sextb t0,t0
0x120104670 <intersect+16>: cmple t1,t0,t1
0x120104674 <intersect+20>: beq t1,0x1201046c0 <intersect+96>
0x120104678 <intersect+24>: ldbu t1,1(a1)
0x12010467c <intersect+28>: sextb t1,t1
0x120104680 <intersect+32>: ldbu t0,3(a0)
0x120104684 <intersect+36>: sextb t0,t0
0x120104688 <intersect+40>: cmple t1,t0,t1
0x12010468c <intersect+44>: beq t1,0x1201046c0 <intersect+96>
0x120104690 <intersect+48>: ldbu t1,2(a1)
0x120104694 <intersect+52>: sextb t1,t1
0x120104698 <intersect+56>: ldbu t0,0(a0)
0x12010469c <intersect+60>: sextb t0,t0
0x1201046a0 <intersect+64>: cmplt t1,t0,t1
0x1201046a4 <intersect+68>: bne t1,0x1201046c0 <intersect+96>
0x1201046a8 <intersect+72>: ldbu t1,3(a1)
0x1201046ac <intersect+76>: sextb t1,t1 0x1201046b0 <intersect+80>: ldbu t0,1(a0)
0x1201046b4 <intersect+84>: sextb t0,t0
0x1201046b8 <intersect+88>: cmplt t1,t0,t1
0x1201046bc <intersect+92>: beq t1,0x1201046c8 <intersect+104>
0x1201046c0 <intersect+96>: clr v0
0x1201046c4 <intersect+100>: ret
0x1201046c8 <intersect+104>: ldbu t1,0(a0)
0x1201046cc <intersect+108>: ldbu t0,0(a1)
0x1201046d0 <intersect+112>: maxsb8 t1,t0,t1
0x1201046d4 <intersect+116>: stb t1,0(a2)
0x1201046d8 <intersect+120>: ldbu t2,1(a0)
0x1201046dc <intersect+124>: ldbu t0,1(a1)
0x1201046e0 <intersect+128>: maxsb8 t2,t0,t4
0x1201046e4 <intersect+132>: stb t4,1(a2)
0x1201046e8 <intersect+136>: ldbu t2,2(a1)
0x1201046ec <intersect+140>: ldbu t0,2(a0)
0x1201046f0 <intersect+144>: minsb8 t2,t0,t2
0x1201046f4 <intersect+148>: stb t2,2(a2)
0x1201046f8 <intersect+152>: ldbu t3,3(a1)
0x1201046fc <intersect+156>: ldbu t0,3(a0)
0x120104700 <intersect+160>: minsb8 t3,t0,t3
0x120104704 <intersect+164>: stb t3,3(a2) 0x120104708 <intersect+168>: sextb t1,t1
0x12010470c <intersect+172>: sextb t2,t2
0x120104710 <intersect+176>: cmple t1,t2,t1
0x120104714 <intersect+180>: beq t1,0x12010472c <intersect+204>
0x120104718 <intersect+184>: sextb t4,t0
0x12010471c <intersect+188>: sextb t3,t1
0x120104720 <intersect+192>: cmple t0,t1,t0
0x120104724 <intersect+196>: lda v0,1
0x120104728 <intersect+200>: bne t0,0x120104730 <intersect+208>
0x12010472c <intersect+204>: clr v0
0x120104730 <intersect+208>: ret
0x120104734 <intersect+212>: unop
0x120104738 <intersect+216>: nop
0x12010473c <intersect+220>: unop
End of assembler dump.
At Andrew's prompting, I compiled and ran the following program on my system:
> #include <sys/cdefs.h>
> #include <sys/types.h>
> #include <stdio.h>
> #include <machine/bwx.h>
> #include <machine/alpha_cpu.h>
>
> main()
> {
> u_int8_t byte;
> u_int16_t word;
> u_int32_t lng;
> u_int64_t hack = 0x123456789abcdef;
> u_int64_t va = ((u_int64_t)&hack)+4UL;
> > printf("implver = 0x%lx\n", alpha_implver());
> printf("amask = 0x%lx\n", alpha_amask(-1L));
> printf ("0x%lx\n", hack);
> byte = ldbu(va);
> printf ("0x%x\n", byte);
> stb(va, 0xff);
> printf ("0x%lx\n", hack);
> }
>
with the following results:
$ cc -O -pipe -mcpu=pca56 test.c
$ ./a.out
implver = 0x1
amask = 0xfffffffffffffffe
0x123456789abcdef
0x67
0x12345ff89abcdef
Andrew's guess was that nethack was trapping on a BWX instruction, but
indicated that from the program output my box could handle those
instructions. Here is my dmesg output, if that's at all helpful:
Copyright (c) 1992-2003 The FreeBSD Project.
Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994
The Regents of the University of California. All rights reserved.
FreeBSD 5.0-RELEASE-p6 #2: Tue Mar 25 04:13:12 PST 2003
root@frane.vortex4.net:/usr/src/sys/alpha/compile/FRANE
Preloaded elf kernel "/boot/kernel/kernel" at 0xfffffc000062c000.
EB164
Digital AlphaPC 164SX 533 MHz, 531MHz
8192 byte page size, 1 processor.
CPU: PCA56 (21164PC) major=9 minor=2 extensions=0x1<BWX>
OSF PAL rev: 0x1000600020116
real memory = 534921216 (510 MB)
avail memory = 514654208 (490 MB)
Initializing GEOMetry subsystem
cia0: <2117x Core Logic chipset>
cia0: Pyxis, pass 1
cia0: extended capabilities: 1<BWEN>
pcib0: <2117x PCI host bus adapter> on cia0
pci0: <PCI bus> on pcib0
sym0: <875> port 0x10000-0x100ff mem 0x82040000-0x82040fff,0x82042000-0x820420ff
irq 9 at device 5.0 on pci0
sym0: Tekram NVRAM, ID 7, Fast-20, SE, parity checking
sym0: interrupting at CIA irq 9
isab0: <PCI-ISA bridge> at device 8.0 on pci0
isa0: <ISA bus> on isab0
pci0: <mass storage, ATA> at device 8.1 (no driver attached)
pci0: <mass storage, ATA> at device 8.2 (no driver attached)
pci0: <serial bus, USB> at device 8.3 (no driver attached)
xl0: <3Com 3c905B-TX Fast Etherlink XL> port 0x10100-0x1017f mem 0x82042100-0x82
04217f irq 8 at device 9.0 on pci0
xl0: interrupting at CIA irq 8
xl0: Ethernet address: 00:50:da:80:2c:11
miibus0: <MII bus> on xl0
xlphy0: <3Com internal media interface> on miibus0
xlphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
atkbdc0: <Keyboard controller (i8042)> at port 0x64,0x60 on isa0
atkbd0: <AT Keyboard> irq 1 on atkbdc0
atkbd0: interrupting at ISA irq 1
fdc0: <Enhanced floppy controller (i82077, NE72065 or clone)> at port 0x3f7,0x3f
0-0x3f5 irq 6 drq 2 on isa0
fdc0: interrupting at ISA irq 6
mcclock0: <MC146818A real time clock> at port 0x70-0x71 on isa0
sio0 at port 0x3f8-0x3ff irq 4 on isa0
sio0: type 16550A, console
sio0: interrupting at ISA irq 4
sio1 at port 0x2f8-0x2ff irq 3 on isa0
sio1: type 16550A
sio1: interrupting at ISA irq 3
Timecounter "i8254" frequency 1193182 Hz
Timecounter "alpha" frequency 533159245 Hz
Timecounters tick every 0.976 msec
Waiting 10 seconds for SCSI devices to settle
da0 at sym0 bus 0 target 0 lun 0
da0: <QUANTUM ATLAS IV 9 WLS 0808> Fixed Direct Access SCSI-3 device
da0: 40.000MB/s transfers (20.000MHz, offset 16, 16bit), Tagged Queueing Enabled
da0: 8683MB (17783249 512 byte sectors: 255H 63S/T 1106C)
Mounting root from ufs:/dev/da0a
It is with a heavy heart that I mark this low/non-critical. It's nethack,
after all :>
>How-To-Repeat:
Problem is consistent on each execution of nethack built with the -march=pca56
flag
>Fix:
pkg_delete nethack-nox11-3.4.0_1
edit make.conf, comment out CPUTYPE?=pca56
cd /usr/ports/games/nethack34-nox11 ; make all install clean
without -mcpu=pca56, the game runs without problems
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20030412090657.40307.qmail>