Re: 2.6.26-git: NULL pointer deref in __switch_to

["Vegard Nossum" <vegard.nossum@gmail.com>]

On Fri, Jun 13, 2008 at 7:42 PM, Patrick McHardy <kaber@trash.net> wrote:
> I get this oops once a day, its apparently triggered by something
> run by cron, but the process is a different one each time.
>
> Kernel is -git from yesterday shortly before the -rc6 release
> (last commit is the usb-2.6 merge, the x86 patches are missing),
> .config is attached.
>
> I'll retry with current -git, but the patches that have gone in
> since I last updated don't look related.
>

Thanks for the report.

>
> [62060.043009] BUG: unable to handle kernel NULL pointer dereference at
> 000001ff
> [62060.043009] IP: [<c0102a9b>] __switch_to+0x2f/0x118
> [62060.043009] *pde = 00000000
> [62060.043009] Oops: 0002 [#1] PREEMPT
> [62060.043009] Modules linked in: nfsd lockd nfs_acl auth_rpcgss sunrpc
> exportfs sch_red cls_fw cls_flow tun sit tunnel4 sch_drr sch_hfsc af_packet
> xt_statistic xt_CONNMARK xt_connmark xt_length xt_owner xt_MARK
> ip6table_mangle ipt_MASQUERADE ipt_REDIRECT ipt_TTL iptable_mangle
> iptable_nat nf_nat_sip nf_nat_irc nf_conntrack_irc nf_nat_ftp nf_nat
> nf_conntrack_ftp ip6t_hl ip6t_REJECT ip6t_ah ip6table_filter ipt_ttl
> ipt_REJECT xt_limit ipt_ah xt_esp xt_state xt_TCPMSS xt_tcpmss xt_helper
> xt_tcpudp xt_hashlimit iptable_filter ip6table_raw ip6_tables xt_policy
> xt_NFLOG iptable_raw ip_tables x_tables nfnetlink_log nfnetlink
> nf_conntrack_ipv6 nf_conntrack_ipv4 nf_conntrack_sip nf_conntrack deflate
> zlib_deflate zlib_inflate ctr twofish twofish_common camellia serpent
> blowfish des_generic xcbc sha256_generic sha1_generic crypto_null af_key cbc
> dm_crypt crypto_blkcipher dm_snapshot dm_mod lg cpufreq_ondemand p4_clockmod
> speedstep_lib aes_i586 aes_generic esp6 esp4 aead usblp parport_pc parport
> ehci_hcd ohci_hcd rtc e1000 sata_promise usbcore unix
> [62060.043009]
> [62060.043009] Pid: 18031, comm: find Not tainted (2.6.26-rc5 #5)
> [62060.043009] EIP: 0060:[<c0102a9b>] EFLAGS: 00010002 CPU: 0
> [62060.043009] EIP is at __switch_to+0x2f/0x118
> [62060.043009] EAX: 00000000 EBX: f7cf6c38 ECX: f6cfd0e0 EDX: f7cf6a20
> [62060.043009] ESI: f7cf6a20 EDI: f6cfd0e0 EBP: f7c41f04 ESP: f7c41ef4
> [62060.043009]  DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 0068
> [62060.043009] Process find (pid: 18031, ti=f7c41000 task=f6cfd0e0
> task.ti=f571c000)
> [62060.043009] Stack: f6cfd2f8 f7cf6a20 00000000 f6040d80 f571cde0 c0321c3c
> f7c41f34 00000046
> [62060.043009]        f7c41f98 f7c41fcc f6ac90e0 f7cf6a20 f7cf6b74 00000001
> c04153c0 f7c41f98
> [62060.043009]        f7c41fcc c015159a f7cf6a48 c047f934 f7c41f70 c047f918
> c0415e68 00000000
> [62060.043009] Call Trace:
> [62060.043009]  [<c0321c3c>] ? schedule+0x1a6/0x2e5
> [62060.043009]  [<c015159a>] ? kswapd+0x387/0x3f3
> [62060.043009]  [<c01164d0>] ? __dequeue_entity+0x24/0x95
> [62060.043009]  [<c014fb1a>] ? isolate_pages_global+0x0/0x46
> [62060.043009]  [<c012e582>] ? autoremove_wake_function+0x0/0x3a
> [62060.043009]  [<c0151213>] ? kswapd+0x0/0x3f3
> [62060.043009]  [<c0151213>] ? kswapd+0x0/0x3f3
> [62060.043009]  [<c012e285>] ? kthread+0x36/0x5a
> [62060.043009]  [<c012e24f>] ? kthread+0x0/0x5a
> [62060.043009]  [<c01047ef>] ? kernel_thread_helper+0x7/0x18
> [62060.043009]  =======================
> [62060.043009] Code: 56 53 83 ec 04 89 c7 89 d6 8d 80 18 02 00 00 89 45 f0
> 8d 9a 18 02 00 00 8b 47 04 f6 40 0c 01 0f 84 c9 00 00 00 8b 87 6c 02 00 00
> <0f> ae 00 0f ba 60 02 07 73 02 db e2 0f 1f 00 90 8d b4 26 00 00
> [62060.043009] EIP: [<c0102a9b>] __switch_to+0x2f/0x118 SS:ESP 0068:f7c41ef4
> [62060.043009] ---[ end trace b024364060382aa3 ]---
> [62060.043009] note: find[18031] exited with preempt_count 2
>

This decodes to

   0:   0f ae 00                fxsave (%eax)

so it's related to the floating-point context. This is the exact
location of the crash:

$ addr2line -e arch/x86/kernel/process_32.o -i ab0
include/asm/i387.h:232
include/asm/i387.h:262
arch/x86/kernel/process_32.c:595

...so it looks like prev_task->thread.xstate->fxsave has become NULL.
Or maybe it never had any other value.

Last FPU-related commit was:

commit 870568b39064cab2dd971fe57969916036982862
Author: Suresh Siddha <suresh.b.siddha@intel.com>
Date:   Mon Jun 2 15:57:27 2008 -0700

    x86, fpu: fix CONFIG_PREEMPT=y corruption of application's FPU stack


...I'm adding some Ccs.

If you simply want to boot your kernel without crashes, you can try
adding "no387 nofxsr" to the kernel parameters.


Vegard

-- 
"The animistic metaphor of the bug that maliciously sneaked in while
the programmer was not looking is intellectually dishonest as it
disguises that the error is the programmer's own creation."
	-- E. W. Dijkstra, EWD1036