Kernel 2.6.38.4 "Oops" under ESXi 4.1 when iscsiadm login to IET target

[iilluzion]

Kernel 2.6.38.4 "Oops" under ESXi 4.1 when iscsiadm login to IET
target:

# iscsiadm --version
iscsiadm version 2.0-872

# iscsiadm --mode node --targetname iqn.2011-05.test:01 --portal
10.10.10.10 --login
Logging in to [iface: default, target: iqn.2011-05.test:01, portal:
10.10.10.10,3260]
iscsiadm: got read error (0/0), daemon died?
iscsiadm: Could not login to [iface: default, target: iqn.2011-05.test:
01, portal: 10.10.10.10,3260].
iscsiadm: initiator reported error (18 - could not communicate to
iscsid)

# dmesg

...
[251507.006803] BUG: unable to handle kernel NULL pointer dereference
at 00000068
[251507.007103] IP: [<f803e393>] show_proc_name+0x9/0x1b [scsi_mod]
[251507.007294] *pde = 00000000
[251507.007430] Oops: 0000 [#18] SMP
[251507.007577] last sysfs file: /sys/devices/platform/host5/scsi_host/
host5/proc_name
[251507.007868] Modules linked in: md4 hmac nls_iso8859_1 cifs crc32c
iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi xt_time
xt_connlimit xt_realm iptable_raw xt_comment xt_policy ipt_ULOG
ipt_REJECT ipt_REDIRECT ipt_NETMAP ipt
_MASQUERADE ipt_LOG ipt_ECN ipt_ecn ipt_CLUSTERIP ipt_ah ipt_addrtype
nf_nat_tftp nf_nat_snmp_basic nf_nat_sip nf_nat_pptp nf_nat_proto_gre
nf_nat_irc nf_nat_h323 nf_nat_ftp nf_nat_amanda ts_kmp
nf_conntrack_amanda nf_conntrack_sane
nf_conntrack_tftp nf_conntrack_sip nf_conntrack_proto_sctp
nf_conntrack_pptp nf_conntrack_proto_gre nf_conntrack_netlink
nf_conntrack_netbios_ns nf_conntrack_irc nf_conntrack_h323
nf_conntrack_ftp xt_tcpmss xt_recent xt_pkttype xt_ph
ysdev xt_owner xt_NFQUEUE xt_NFLOG nfnetlink_log xt_multiport xt_mark
xt_mac xt_limit xt_length xt_iprange xt_helper xt_hashlimit xt_DSCP
xt_dscp xt_dccp xt_conntrack xt_connmark xt_CLASSIFY xt_tcpudp
xt_state iptable_nat nf_nat nf_c
onntrack_ipv4 nf_defrag_ipv4 nf_conntrack iptable_mangle nfnetlink
iptable_filter ip_tables x_tables xfs exportfs af_packet psmouse
serio_raw snd_pcsp snd_pcm snd_page_alloc snd_timer snd soundcore
shpchp pci_hotplug pcnet32 mii evde
v rtc_cmos rtc_core rtc_lib container parport_pc parport processor ac
fbcon font bitblit fbcon_rotate fbcon_ccw fbcon_ud fbcon_cw softcursor
tileblit fb fbdev sr_mod cdrom intel_agp intel_gtt agpgart i2c_piix4
i2c_core ata_piix ata_g
eneric pata_acpi libata mptspi mptscsih mptbase scsi_transport_spi
floppy button cramfs loop ext3 mbcache jbd usb_storage usb_libusual
usbcore nls_base sd_mod scsi_mod
[251507.011753]
[251507.011844] Pid: 6278, comm: iscsiadm Tainted: G D
2.6.38.4 #2 VMware, Inc. VMware Virtual Platform
[251507.012234] EIP: 0060:[<f803e393>] EFLAGS: 00210292 CPU: 0
[251507.012380] EIP is at show_proc_name+0x9/0x1b [scsi_mod]
[251507.012512] EAX: 00000000 EBX: f80452e4 ECX: f589b000 EDX:
f80452e4
[251507.012653] ESI: fffffffb EDI: f803e38a EBP: f529bf2c ESP:
f529bf2c
[251507.012798] DS: 007b ES: 007b FS: 00d8 GS: 0033 SS: 0068
[251507.012930] Process iscsiadm (pid: 6278, ti=f529a000 task=f5801c80
task.ti=f529a000)
[251507.013161] Stack:
[251507.013256] f529bf40 c115e723 f596e500 f50f62a0 c124e2dc f529bf6c
c10d2900 00000100
[251507.013501] 00000100 bf96e28c c1093f40 c124e2dc f596e514 00000100
f50d8b40 bf96e28c
[251507.013745] f529bf8c c109427e f529bf98 c10d2881 f4d799e0 f50d8b40
fffffff7 b7d93e8c
[251507.013985] Call Trace:
[251507.014093] [<c115e723>] dev_attr_show+0x19/0x36
[251507.014219] [<c10d2900>] sysfs_read_file+0x7f/0xf1
[251507.014348] [<c1093f40>] ? rw_verify_area+0xc4/0xe7
[251507.014479] [<c109427e>] vfs_read+0x86/0xf1
[251507.014598] [<c10d2881>] ? sysfs_read_file+0x0/0xf1
[251507.014729] [<c1094324>] sys_read+0x3b/0x60
[251507.014852] [<c122ee11>] syscall_call+0x7/0xb
[251507.014972] Code: 6a 14 51 e8 68 11 0c c9 c9 c3 55 89 e5 ff b0 18
ff ff ff 68 1a 28 04 f8 6a 14 51 e8 50 11 0c c9 c9 c3 55 8b 80 a4 fe
ff ff 89 e5 <ff> 70 68 68 1e 28 04 f8 6a 14 51 e8 35 11 0c c9 c9 c3 55
8a 80
[251507.015502] EIP: [<f803e393>] show_proc_name+0x9/0x1b [scsi_mod]
SS:ESP 0068:f529bf2c
[251507.015744] CR2: 0000000000000068
[251507.016179] ---[ end trace 33d79edf59f24953 ]---

[Mike Christie]

On 05/10/2011 08:48 AM, iilluzion wrote:
> [251507.006803] BUG: unable to handle kernel NULL pointer dereference
> at 00000068
> [251507.007103] IP: [<f803e393>] show_proc_name+0x9/0x1b [scsi_mod]

I am not hitting the problem here. The sht and proc name should be set
when we allocate the host, so before it is even exported in sysfs.

Is this just a kernel.org 2.6.38.4 kernel but with some vmware modules?
Has this worked with older kernels?

For the other scsi/sata drivers in your setup if you just do

cat /sys/class/scsi_host/hostXYX/proc_name does it work?

Could you also send me your kernel .config.

[iilluzion :)]

Thank you for the response!

1. Looks like it's just 2.6.38.4 kernel problem. It successfully works
on kernel 2.6.35.7 in same ESXi environment with no special vmware
module loaded.
2. cat /sys/class/scsi_host/hostXYX/proc_name for other sata/scsi
drives (where host5 related to iSCSI):

hostname:~# ls /sys/class/scsi_host/
host0 host1 host2 host5

hostname:~# cat /sys/class/scsi_host/host0/proc_name
mptspi

hostname:~# cat /sys/class/scsi_host/host1/proc_name
ata_piix

hostname:~# cat /sys/class/scsi_host/host2/proc_name
ata_piix

hostname:~# cat /sys/class/scsi_host/host5/proc_name
Killed

3. .config of successful kernel 2.6.35.7 see in attached file '2.6.35.7.config'

4. .config of failure kernel 2.6.38.4 see in attached file '2.6.38.4.config'