clang/ARM regression

[Warner Losh]

On Apr 28, 2013, at 5:23 PM, Tim Kientzle wrote:

> 
> On Apr 27, 2013, at 7:16 PM, Oleksandr Tymoshenko wrote:
> 
>> Hi,
>> 
>> There seems to be some kind of regression in recently imported clang on ARM.
>> Beaglebone kernel built using clang  with WITNESS enabled panics
>> on boot. If I change optimization flags from -O to -O0 it works fine.
>> I'm trying to isolate test case but it panics fairly early in boot process
>> so it's not easy. If somebody tracks clang development and aware of
>> known issues that might cause it - please share.
>> 
>> So far I tried fix from this bug but it didn't help:
>> http://llvm.org/bugs/show_bug.cgi?id=15581
> 
> I'm seeing a crash early on as well.  Is this the same one?
> 
> /boot/kernel/kernel data=0x404564+0x17c9f8 syms=[0x4+0x76a60+0x4+0x4bf98]
> Hit [Enter] to boot immediately, or any other key for command prompt.
> Booting [/boot/kernel/kernel]...               
> Using DTB provided by U-Boot.
> Kernel entry at 0x80200100...
> Kernel args: (null)
> KDB: debugger backends: ddb
> KDB: current backend: ddb
> Copyright (c) 1992-2013 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 is a registered trademark of The FreeBSD Foundation.
> FreeBSD 10.0-CURRENT #0: Sun Apr 28 05:25:04 PDT 2013
>    root at fci386.localdomain:/usr/home/tim/projects/crochet-rpi/work/obj/arm.armv6/usr/src/sys/BEAGLEBONE arm
> FreeBSD clang version 3.3 (trunk 178860) 20130405
> WARNING: WITNESS option enabled, expect reduced performance.
> panic: acquiring blockable sleep lock with spinlock or critical section held (rw) pmap pv global @ /usr/src/sys/arm/arm/pmap-v6.c:1187
> KDB: enter: panic
> [ thread pid 0 tid 0 ]
> Stopped at      $d:     ldrb    r15, [r15, r15, ror r15]!
> db> 

That's similar to what I'm seeing on my Allwinner board.

Warner