Date: Thu, 20 Dec 2001 13:47:16 -0800 (PST) From: Matthew Dillon <dillon@apollo.backplane.com> To: Brady Montz <bradym@balestra.org> Cc: Chad David <davidc@acns.ab.ca>, Richard Nyberg <rnyberg@it.su.se>, David Wolfskill <david@catwhisker.org>, freebsd-stable@FreeBSD.ORG Subject: Re: 4.4-STABLE crashes - suspects new ata-driver over wd-drivers Message-ID: <200112202147.fBKLlGg78745@apollo.backplane.com> References: <20011217142115.A45797@colnta.acns.ab.ca> <200112172133.fBHLXJg39627@bunrab.catwhisker.org> <20011217144541.A45973@colnta.acns.ab.ca> <20011218094330.A10601@gromit.it.su.se> <200112180857.fBI8vNh68173@apollo.backplane.com> <20011218023630.A5190@ranger.acns.ab.ca> <200112181920.fBIJKRt70523@apollo.backplane.com> <200112190757.fBJ7veY02579@mail.hydrologue.com> <200112190732.fBJ7WFQ86868@apollo.backplane.com> <87u1ulq1a8.fsf@balestra.org>
next in thread | previous in thread | raw e-mail | index | archive | help
(ok, I'm adding freebsd-stable back in).
I've looked at Brady's core dump. The crash:
#13 0xc02ac6ff in trap (frame={tf_fs = 16, tf_es = 16, tf_ds = 16,
tf_edi = 32, tf_esi = 3, tf_ebp = -926287844, tf_isp = -926287868,
tf_ebx = -1069087664, tf_edx = 31, tf_ecx = -1064198904,
tf_eax = -1063509624, tf_trapno = 12, tf_err = 2, tf_eip = -1071396729,
tf_cs = 8, tf_eflags = 66118, tf_esp = -1064198904, tf_ss = -926287816})
at /vol/src.stable/sys/i386/i386/trap.c:448
#14 0xc023c887 in vm_page_unqueue_nowakeup (m=0xc0919d08)
at /vol/src.stable/sys/vm/vm_page.c:587
#15 0xc023cb62 in vm_page_alloc (object=0xc8a08540, pindex=32, page_req=3)
at /vol/src.stable/sys/vm/vm_page.c:831
#16 0xc023d50c in vm_page_grab (object=0xc8a08540, pindex=32, allocflags=131)
at /vol/src.stable/sys/vm/vm_page.c:1525
#17 0xc02a941a in _pmap_allocpte (pmap=0xc89bc52c, ptepindex=32)
at /vol/src.stable/sys/i386/i386/pmap.c:1238
#18 0xc02aa127 in pmap_enter_quick (pmap=0xc89bc52c, va=134512640,
m=0xc08c13c4, mpte=0x0) at /vol/src.stable/sys/i386/i386/pmap.c:2230
---Type <return> to continue, or q <return> to quit---
#19 0xc02aa4d8 in pmap_object_init_pt (pmap=0xc89bc52c, addr=134512640,
object=0xc0361980, pindex=0, size=421888, limit=0)
at /vol/src.stable/sys/i386/i386/pmap.c:2416
#20 0xc0236964 in vm_map_insert (map=0xc89bc4c0, object=0xc0361980, offset=0,
start=134512640, end=134934528, prot=5 '\005', max=7 '\a', cow=10)
at /vol/src.stable/sys/vm/vm_map.c:612
#21 0xc0162e91 in elf_load_section (p=0xc89784e0, vmspace=0xc89bc4c0,
vp=0xc8387740, offset=0,
vmaddr=0x8048000 <Address 0x8048000 out of bounds>, memsz=419820,
filsz=419820, prot=5 '\005') at /vol/src.stable/sys/kern/imgact_elf.c:226
#22 0xc0163540 in exec_elf_imgact (imgp=0xc8c9fe5c)
at /vol/src.stable/sys/kern/imgact_elf.c:507
#23 0xc016c6e0 in execve (p=0xc89784e0, uap=0xc8c9ff80)
at /vol/src.stable/sys/kern/kern_exec.c:198
#24 0xc02ad189 in syscall2 (frame={tf_fs = 47, tf_es = 47, tf_ds = 47,
tf_edi = 134808064, tf_esi = 0, tf_ebp = -1077944488,
tf_isp = -926285868, tf_ebx = -1077944416, tf_edx = 1,
tf_ecx = 134550141, tf_eax = 59, tf_trapno = 12, tf_err = 2,
tf_eip = 134618516, tf_cs = 31, tf_eflags = 643, tf_esp = -1077944516,
tf_ss = 47}) at /vol/src.stable/sys/i386/i386/trap.c:1155
#25 0xc02a05f5 in Xint0x80_syscall ()
Cannot access memory at address 0xbfbfdf58.
And this is scary. Here's the VM_PAGE that it crashed on:
(kgdb) print &vm_page_array[12958]
$36 = (struct vm_page *) 0xc0919d08
(kgdb) print &vm_page_array[12959]
$37 = (struct vm_page *) 0xc0919d44
(kgdb) x/s &vm_page_array[12958]
0xc0919d08: ""
(kgdb)
0xc0919d09: ""
(kgdb)
0xc0919d0a: ""
(kgdb)
0xc0919d0b: ""
(kgdb)
0xc0919d0c: "\210!\234À"
(kgdb)
0xc0919d11: ""
(kgdb)
0xc0919d12: ""
(kgdb)
0xc0919d13: ""
(kgdb)
0xc0919d14: ""
(kgdb)
0xc0919d15: ""
(kgdb)
0xc0919d16: ""
(kgdb)
0xc0919d17: ""
(kgdb)
0xc0919d18: "ØV°È"
(kgdb)
0xc0919d1d: ""
(kgdb)
0xc0919d1e: ""
(kgdb)
0xc0919d1f: ""
(kgdb)
0xc0919d20: "pre-fetch"
(kgdb)
0xc0919d2a: ""
The string 'pre-fetch' is sitting right smack in the middle of the
VM_PAGE!!!!!!!. The entire vm_page is corrupt, though the vm_page's
surrounding it appear to be ok.
When I look for 'pre-fetch' in the raw vmcore file I see it occuring
all over the place. I think this may be related to disk I/O as well
but I am not the IDE driver guy so I don't think I can take it beyond
this point.
-Matt
Matthew Dillon
<dillon@backplane.com>
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-stable" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200112202147.fBKLlGg78745>