Date: Fri, 14 Sep 2012 21:50:40 -0700 From: John-Mark Gurney <jmg@funkthat.com> To: Alan Cox <alc@rice.edu> Cc: "arm@freebsd.org" <arm@freebsd.org> Subject: Re: arm pmap locking Message-ID: <20120915045040.GZ58312@funkthat.com> In-Reply-To: <504F8BAC.4040902@rice.edu> References: <502FD67A.7030003@rice.edu> <1345315508.27688.260.camel@revolution.hippie.lan> <503D12AE.1050705@rice.edu> <1346350374.1140.525.camel@revolution.hippie.lan> <5045351F.6060201@rice.edu> <1346723041.1140.602.camel@revolution.hippie.lan> <504B85BE.3030101@rice.edu> <1347316458.1137.41.camel@revolution.hippie.lan> <504F8BAC.4040902@rice.edu>
next in thread | previous in thread | raw e-mail | index | archive | help
Alan Cox wrote this message on Tue, Sep 11, 2012 at 14:06 -0500:
> On 09/10/2012 17:34, Ian Lepore wrote:
> >On Sat, 2012-09-08 at 12:51 -0500, Alan Cox wrote:
> >>Here is another patch. This simplifies the kernel pmap locking in
> >>pmap_enter_pv() and corrects some comments.
> >>
> >>Thanks in advance,
> >>Alan
> >>
> >I'm afraid I'm not going to be able to do this any time soon. I bricked
> >my DreamPlug last Friday (the bright side: the nandfs_newfs command in
> >-current apparently works just fine when applied to real hardware rather
> >than the simulator device). I haven't had any success in getting
> >openocd to transfer a new uboot image to it. So I'm probably going to
> >be without arm hardware that can run anything newer than 8.2 for a few
> >weeks (until my new atmel eval board arrives).
> >
>
> Thanks for letting me know.
>
> Could someone else here please test the attached patch?
I figure since you've been looking at arm's pmap, that I should report
to you a LOR that I observered recently on my armv6 board:
lock order reversal:
1st 0xc1cf70b0 pmap (pmap) @ /usr/src.HEAD/sys/arm/arm/pmap-v6.c:673
2nd 0xc091e608 PV ENTRY (UMA zone) @ /usr/src.HEAD/sys/vm/uma_core.c:2084
KDB: stack backtrace:
db_trace_self() at db_trace_self+0xc
scp=0xc05a294c rlv=0xc025b298 (X_db_sym_numargs+0x1bc)
rsp=0xc9d1a8fc rfp=0xc9d1aa18
X_db_sym_numargs() at X_db_sym_numargs+0x194
scp=0xc025b270 rlv=0xc0398340 (kdb_backtrace+0x3c)
rsp=0xc9d1aa1c rfp=0xc9d1aa2c
r4=0xc06bda44
kdb_backtrace() at kdb_backtrace+0xc
scp=0xc0398310 rlv=0xc03ad3f8 (witness_display_spinlock+0x80)
rsp=0xc9d1aa30 rfp=0xc9d1aa44
r4=0x00000001
witness_display_spinlock() at witness_display_spinlock+0x5c
scp=0xc03ad3d4 rlv=0xc03ae6d8 (witness_checkorder+0x884)
rsp=0xc9d1aa48 rfp=0xc9d1aa98
r5=0xc1cf70b0 r4=0xc06263d4
witness_checkorder() at witness_checkorder+0xc
scp=0xc03ade60 rlv=0xc0355b9c (_mtx_lock_flags+0xcc)
rsp=0xc9d1aa9c rfp=0xc9d1aabc
r10=0xc1cf70b0 r9=0x00000000
r8=0xc091d6e0 r7=0xc0620730 r6=0x00000824 r5=0x00000000
r4=0xc091e608
_mtx_lock_flags() at _mtx_lock_flags+0xc
scp=0xc0355adc rlv=0xc057d290 (uma_zalloc_arg+0x1a8)
rsp=0xc9d1aac0 rfp=0xc9d1aafc
r7=0xc091d748 r6=0x00000000
r5=0xc08fc0b8 r4=0xc0620730
uma_zalloc_arg() at uma_zalloc_arg+0xc
scp=0xc057d0f4 rlv=0xc05abb04 (pmap_growkernel+0xf20)
rsp=0xc9d1ab00 rfp=0xc9d1ab70
r10=0xc1cf70b0 r9=0x00000000
r8=0x00000000 r7=0x8122e032 r6=0x00000000 r5=0xc08fc0b8
r4=0x00000001
pmap_growkernel() at pmap_growkernel+0x3fc
scp=0xc05aafe0 rlv=0xc05ac14c (pmap_enter+0x70)
rsp=0xc9d1ab74 rfp=0xc9d1aba0
r10=0x00000007 r9=0x00000000
r8=0xc09885a8 r7=0xbffff000 r6=0xc08278c0 r5=0xc1cf70b0
r4=0xc06261e0
pmap_enter() at pmap_enter+0xc
scp=0xc05ac0e8 rlv=0xc057ff44 (vm_fault_hold+0x1638)
rsp=0xc9d1aba4 rfp=0xc9d1ad14
r10=0xc9d1ade4 r8=0x00000000
r7=0x00000002 r6=0x00000000 r5=0xc1cf7000 r4=0xc09885a8
vm_fault_hold() at vm_fault_hold+0xc
scp=0xc057e918 rlv=0xc058054c (vm_fault+0x8c)
rsp=0xc9d1ad18 rfp=0xc9d1ad3c
r10=0xc9d1ade4 r9=0x00000002
r8=0x00000000 r7=0x00000002 r6=0xbffff000 r5=0xc1cf7000
r4=0xc1cf5000
vm_fault() at vm_fault+0xc
scp=0xc05804cc rlv=0xc05b0ac8 (data_abort_handler+0x35c)
rsp=0xc9d1ad40 rfp=0xc9d1ade0
r8=0xbffff000 r7=0xc1cf5000
r6=0x00000000 r5=0xc0626844 r4=0xc1cf3088
data_abort_handler() at data_abort_handler+0xc
scp=0xc05b0778 rlv=0xc05a4150 (address_exception_entry+0x50)
rsp=0xc9d1ade4 rfp=0xc9d1ae84
r10=0xc0682bde r9=0xc1cf3000
r8=0xc9d1aeac r7=0xc0682bde r6=0xc0682bd4 r5=0xc05f1648
r4=0xbfffffff
exec_shell_imgact() at exec_shell_imgact+0x75c
scp=0xc031d348 rlv=0xc033b68c (fork_exit+0x94)
rsp=0xc9d1ae88 rfp=0xc9d1aea8
r10=0x00000000 r9=0x00000000
r8=0xc9d1aeac r7=0x00000000 r6=0xc031d33c r5=0xc1cf3000
r4=0xc1cf5000
fork_exit() at fork_exit+0xc
scp=0xc033b604 rlv=0xc05afe44 (fork_trampoline+0x14)
rsp=0xc9d1aeac rfp=0x00000000
r8=0x00000000 r7=0x8ffbf8ef
r6=0xf5f7f4a9 r5=0x00000000 r4=0xc031d33c
FreeBSD beaglebone 10.0-CURRENT FreeBSD 10.0-CURRENT #1 r240480: Thu Nov 3 14:57:01 PDT 2011 jmg@pcbsd-779:/usr/obj/arm.armv6/usr/src.HEAD/sys/helium arm
Do you need any more information?
--
John-Mark Gurney Voice: +1 415 225 5579
"All that I will do, has been done, All that I have, has not."
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20120915045040.GZ58312>