open-iscsi, qla4xxx, iscli and kernel 2.6.17
Joni
I'm not sure this is the right group for this question, but because I
see some other postings regarding qla4xxx I suppose there are some
people here that can help me.
I'm trying to have a custom Linux server (Linux From Scratch) with
kernel 2.6.17 to connect through iSCSI to a Dell EMC AX-150i. The
server has a QLogic QLA-4050C HBA. The qla4xxx driver I downloaded from
the QLogic website, which is version 5.00.04, does not compile with
kernels 2.6.16 and 2.6.17. Somewhere on this group I found a posting
with a link to 5.00.04-d5 which fixes that. So I compiled this driver
and it seems to work because the driver does loads and
/proc/scsi/qla4xxx/0 gives me all kind of information (including IP
address that I entered in the HBA BIOS), see below.
However the iscli utility that came with the package does not recognise
the HBA:
-------------------------------------------------------------
Current HBA/Port Information:
HBA Ser. Num.: Not Available HBA Port: -1
IP Address: Not Available Link: Not Available
Port Name: Not Available
Port Alias: Not Available
-------------------------------------------------------------
So I called the QLogic tech support, but they say they cannot help me
because 5.00.04-d5 is not a supported/released driver version and I am
not using RHEL or SUSE. Did anyone get this to work anyway? Is there
an alternative way to set the targets? Can open-iscsi be of any help in
this? Since I already bought the HBA I would really like to use them to
offload the CPU's on the servers, instead of using open-iscsi. However
I got the impression that maybe open-iscsi is able to use the qla4xxx
driver!?
Please advice!
Thanks,
Jonathan
$ modinfo qla4xxx
filename:
/lib/modules/2.6.17.8/kernel/drivers/scsi/qla4xxx/qla4xxx.ko
license: GPL
description: QLogic ISP4XXX iSCSI Host Bus Adapter driver
author: QLogic Corporation
alias: pci:v00001077d00004022sv*sd*bc*sc*i*
alias: pci:v00001077d00004010sv*sd*bc*sc*i*
depends:
vermagic: 2.6.17.8 SMP PENTIUM4 gcc-4.0
parm: ql4xdiscoverywait:Discovery wait time (int)
parm: ql4xdontresethba:Dont reset the HBA when the driver
gets 0x8002 AEN default it will reset hba :0 set to 1 to avoid
resetting HBA (int)
parm: ql4xcmdretrycount:Maximum number of mid-layer retries
allowed for a command. Default value is 20 (int)
parm: ql4xmaxqdepth:Maximum queue depth to report for target
devices. (int)
parm: extended_error_logging:Option to enable extended error
logging, Default is 0 - no logging. 1 - log errors. 2 - debug logging
(int)
parm: displayConfig:If 1 then display the configuration used
in /etc/modules.conf. (int)
----------
$ cat /proc/scsi/qla4xxx/0
QLogic iSCSI Adapter for ISP 4022:
Driver version 5.00.04-d5
Firmware version 2.00.00.07
Code starts at address = f8883000
Memory I/O = 0xf3efe000
IP Address = 10.1.0.11
Number of free request entries = 1024 of 1024
Number of Mailbox Timeouts = 0
Interrupt Status = 1280
ReqQptr=ea0c0000, ReqIn=0, ReqOut=0
Device queue depth = 0x10
Adapter flags = 0xd43, DPC flags = 0x0
Number of commands in retry_srb_q = 0
Number of commands in done_srb_q = 0
Keep Alive Timeout = 30
Number of active commands = 0
Total number of IOCBs (used/max) = (0/208)
Number of free srbs = 1024 of 1024
SCSI Device Information:
scsi-qla0-adapter-port="iqn.2000-04.com.qlogic:qla4050c.fs10612a04846.1";
scsi-qla0-adapter-address=10.1.0.11: 255.255.255.0;
scsi-qla0-adapter-gateway=0.0.0.0;
scsi-qla0-target-0=ddb 0x0000: "": 0.0.0.0;
SCSI LUN Information:
(H: B: T: L) * - indicates lun is not registered with the OS.
----------
dmesg:
QLogic iSCSI HBA Driver (f8883000)
ACPI: PCI Interrupt 0000:0a:01.1[B] -> GSI 33 (level, low) -> IRQ 19
qla4xxx 0000:0a:01.1: Found an ISP4022, irq 19, iobase 0xf880c000
qla4xxx 0000:0a:01.1: Configuring PCI space...
qla4xxx 0000:0a:01.1: Configuring NVRAM ...
qla4xxx 0000:0a:01.1: Initializing firmware..
qla4xxx 0000:0a:01.1: Waiting for Firmware Ready..
qla4xxx 0000:0a:01.1: Firmware Ready..
qla4xxx 0000:0a:01.1: Initializing DDBs ...
qla4xxx 0000:0a:01.1: DDB[0] MB0 4000 Tot 0 Next 0 State 0000 ConnErr
00000000 0.0.0.0:3260 ""
qla4xxx 0000:0a:01.1: DDB[0] MB0 4000 Tot 0 Next 0 State 0000 ConnErr
00000000 0.0.0.0:3260 ""
qla4xxx 0000:0a:01.1: DDB list done..
scsi-qla0-mac=00c0dd0898a6\;
scsi0 : qla4xxx
QLogic iSCSI HBA Driver version: 5.00.04-d5
QLogic ISP4022 @ 0000:0a:01.1 hdma+, host#=0, fw=02.00.00.07
Ravi Anand
You will need to load the ioctl module for QLogic App's to work.
Please download the latest driver and ioctl module from the following
link :
IOCTL Module :
ftp://ftp.qlogic.com/outgoing/linux/beta/iscsi/5.x/qisioctl-v1.00.00.tar.gz
Driver comptabile with above ioctl module :
ftp://ftp.qlogic.com/outgoing/linux/beta/iscsi/5.x/qla4xxx-5.00.07-d2.tar.gz
Btw in the latest kernel following member of scsi_cmd has been renamed :
data_cmnd -> cmnd
So you will have to massage the ioctl module and driver in couple of places
to fix the compile issue. Other than that I think it should work fine
and you should be able to run the apps.
Btw ioctl module need to be loaded after loading the driver module.
Let us know if there is any issue or concern.
Thanx
Ravi Anand
QLogic Corp
Joni
Thanks for your reply! I am using kernel 2.6.17.8 and the qla4xxx.ko
module has been succesfully built and installed (without messing around
with the data_cmnd).
May I suggest a modification for extras/build.sh script? If you are
compiling the kernel module on a machine running a different kernel
than the kernel you are compiling for, you will get some errors.
Therefore I edited the extras/build.sh and replaced the setting of the
K_VERSION variable to 'K_VERSION=2.6.17.8' instead of 'K_VERSION=`uname
-r`'. Also the depmod in the install phase failed, because I needed to
run 'depmod 2.6.17.8'. So maybe it would be possible to specify the
kernel version as a argument for the build script, which would make it
a bit easier.
Anyway I am not really experienced in building kernel modules, and I am
having some trouble building the ioctl. After setting the K_VERSION
statically in the build.sh script, it did build the qisioctl.ko
modules, although I did get some warnings:
# ./build.sh
make: Entering directory `/lfs-sources/linux-2.6.17.8'
Building modules, stage 2.
MODPOST
WARNING: /lfs-sources/qisioctl-v1.00.00/qisioctl.o - Section mismatch:
reference to .exit.text: from .smp_locks after '' (at offset 0x1c)
WARNING: /lfs-sources/qisioctl-v1.00.00/qisioctl.o - Section mismatch:
reference to .exit.text: from .smp_locks after '' (at offset 0x20)
However when I want to install the modules (# make install) I get
errors and looking at the Makefile I see that the environment variable
KERNEL_SOURCES is needed, so I set that but I am still getting errors:
# make install
make -C /lfs-sources/linux-2.6.17.8/ modules_add
SUBDIRS=/lfs-sources/qisioctl-v1.00.00
make[1]: Entering directory `/lfs-sources/linux-2.6.17.8'
make[1]: *** No rule to make target `modules_add'. Stop.
make[1]: Leaving directory `/lfs-sources/linux-2.6.17.8'
make: *** [modules_add] Error 2
Can I just copy the module in a specific directory and run depmod?
Please advice!
Thanks!
Jonathan
Alberto Cammozzo - mmzz
data_cmnd statements.
The qla4xxx module works, and even SANsurferiCLI/iscli.
There's only one problem: if I try to list LUN information (menu 7/7) I
get a kernel panic inside qla4xxx module.
I can manage to catch the messages enabling the serial console, if it's
important...
Kernel: 2.6.18-rc4
Iscli:
Program Version : 1.0.36.2
iSDMAPI : 00.05.03.17 QLSDM.DLL 06 IOCTL
HBA: QLA4010
Alberto
Joni
If you managed to get the qisioctl.ko to work, could you please give me
some advice on how you managed to do that? See my problem above.
Unfortunately I cannot help you with your problem, because I have not
reached that step yet. Maybe post a new thread instead, it might draw
more attention.
Thanks!
Jonathan
Alberto Cammozzo - mmzz
iscsi-target running on a PC , version 0.4.13.
Alberto
ravianand
>
> I compiled and loaded the qisioctl.ko on 2.6.18-rc4, changing the
> data_cmnd statements.
>
> The qla4xxx module works, and even SANsurferiCLI/iscli.
> There's only one problem: if I try to list LUN information (menu 7/7) I
> get a kernel panic inside qla4xxx module.
> I can manage to catch the messages enabling the serial console, if it's
> important...
Please capture the panic log and forward it to us. Definitely we want to
address this issue. Enable the debug in the driver as follows before
you replicate the problem
# insmod ./qla4xxx.ko extended_error_logging=1
Thanx
Ravi
ravianand
> Hi Ravi,
>
> Thanks for your reply! I am using kernel 2.6.17.8 and the qla4xxx.ko
> module has been succesfully built and installed (without messing around
> with the data_cmnd).
>
> May I suggest a modification for extras/build.sh script? If you are
> compiling the kernel module on a machine running a different kernel
> than the kernel you are compiling for, you will get some errors.
> Therefore I edited the extras/build.sh and replaced the setting of the
> K_VERSION variable to 'K_VERSION=2.6.17.8' instead of 'K_VERSION=`uname
> -r`'. Also the depmod in the install phase failed, because I needed to
> run 'depmod 2.6.17.8'. So maybe it would be possible to specify the
> kernel version as a argument for the build script, which would make it
> a bit easier.
Thanx for your suggestion. We will incorporate this change in the future
release.
>
> Anyway I am not really experienced in building kernel modules, and I am
> having some trouble building the ioctl. After setting the K_VERSION
> statically in the build.sh script, it did build the qisioctl.ko
> modules, although I did get some warnings:
>
> # ./build.sh
> make: Entering directory `/lfs-sources/linux-2.6.17.8'
> Building modules, stage 2.
> MODPOST
> WARNING: /lfs-sources/qisioctl-v1.00.00/qisioctl.o - Section mismatch:
> reference to .exit.text: from .smp_locks after '' (at offset 0x1c)
> WARNING: /lfs-sources/qisioctl-v1.00.00/qisioctl.o - Section mismatch:
> reference to .exit.text: from .smp_locks after '' (at offset 0x20)
>
> However when I want to install the modules (# make install) I get
> errors and looking at the Makefile I see that the environment variable
> KERNEL_SOURCES is needed, so I set that but I am still getting errors:
>
> # make install
> make -C /lfs-sources/linux-2.6.17.8/ modules_add
> SUBDIRS=/lfs-sources/qisioctl-v1.00.00
> make[1]: Entering directory `/lfs-sources/linux-2.6.17.8'
> make[1]: *** No rule to make target `modules_add'. Stop.
> make[1]: Leaving directory `/lfs-sources/linux-2.6.17.8'
> make: *** [modules_add] Error 2
>
> Can I just copy the module in a specific directory and run depmod?
> Please advice!
You can do that. Just copy it int /lib/.../scsi/qla4xxx/ dir,
run depmod and load it.
Ravi
>
> Thanks!
> Jonathan
>
>
>
ravianand
>
> iscsi-target running on a PC , version 0.4.13.
>
Please forward us /proc/scsi/qla4xxx/<id> and /proc/scsi/scsi output
as well.
Thanx
Ravi Anand
Michael Tokarev
>> On Thu, 31 Aug 2006, Joni wrote:
>> Hi Ravi,
>>
>> Thanks for your reply! I am using kernel 2.6.17.8 and the qla4xxx.ko
>> module has been succesfully built and installed (without messing around
>> with the data_cmnd).
>>
>> May I suggest a modification for extras/build.sh script? If you are
>> compiling the kernel module on a machine running a different kernel
>> than the kernel you are compiling for, you will get some errors.
>> Therefore I edited the extras/build.sh and replaced the setting of the
>> K_VERSION variable to 'K_VERSION=2.6.17.8' instead of 'K_VERSION=`uname
>> -r`'. Also the depmod in the install phase failed, because I needed to
>> run 'depmod 2.6.17.8'. So maybe it would be possible to specify the
>> kernel version as a argument for the build script, which would make it
>> a bit easier.
>
> Thanx for your suggestion. We will incorporate this change in the future
> release.
Please don't ;)
Becase if you do, it will break for everyone who is not running 2.6.17.8.
But noone should run that release, as it's obsolete now... ;)
Instead, i'd say this should be done:
: ${K_VERSION:=`uname -r`}
(instead of current
K_VERSION=`uname -r`
). This way, it'll be possible to do
K_VERSION=1.2.3.4 ./extras/build.sh
if one wants to build for version 1.2.3.4 instead of currently
running version.
/mjt
ravianand
What I meant was we would provide an option to override the kernel version from the cmd line.
Default behaviour would be still the same ie build for running kernel.
Agree ?
Ravi
Joni
>
> What I meant was we would provide an option to override the kernel version from the cmd line.
> Default behaviour would be still the same ie build for running kernel.
>
> Agree ?
Yes, that is what I meant as well. Default should be for active kernel
kernel, but command line for different kernel. Actually that is the
behaviour of depmod.
Thanks for the tips! I will try it tomorrow.
Jonathan
Joni
Thanks for the tips! The iscli tools is now correctly communicating
with the HBA! I'm going to try and connect to the target.
Jonathan
Joni
I managed to establish the connection to our Dell EMC AX-150i on which
I assigned a 200GB virtual disk to the initiator. After setting the
target and a reboot it gives me a /dev/sda device. Have a look at the
dmesg:
QLogic iSCSI HBA Driver (f8882000)
ACPI: PCI Interrupt 0000:0a:01.1[B] -> GSI 33 (level, low) -> IRQ 19
qla4xxx 0000:0a:01.1: Found an ISP4022, irq 19, iobase 0xf880c000
qla4xxx 0000:0a:01.1: Configuring PCI space...
qla4xxx 0000:0a:01.1: Configuring NVRAM ...
qla4xxx 0000:0a:01.1: Initializing firmware..
qla4xxx 0000:0a:01.1: Waiting for Firmware Ready..
qla4xxx 0000:0a:01.1: Firmware Ready..
qla4xxx 0000:0a:01.1: ISNS disabled...
qla4xxx 0000:0a:01.1: Initializing DDBs ...
qla4xxx 0000:0a:01.1: DDB[0] MB0 4000 Tot 1 Next 2 State 0000 ConnErr
00000000 0.0.0.0:3260 ""
qla4xxx 0000:0a:01.1: DDB[0] MB0 4000 Tot 1 Next 2 State 0000 ConnErr
00000000 0.0.0.0:3260 ""
qla4xxx 0000:0a:01.1: DDB[2] MB0 4000 Tot 1 Next 0 State 0004 ConnErr
00000000 10.1.0.1:3260 "iqn.1992-04.com.emc:ax.fcnpr062500215.a0"
qla4xxx 0000:0a:01.1: DDB[2] MB0 4000 Tot 1 Next 0 State 0004 ConnErr
00000000 10.1.0.1:3260 "iqn.1992-04.com.emc:ax.fcnpr062500215.a0"
qla4xxx 0000:0a:01.1: DDB list done..
scsi-qla0-mac=00c0dd0898a6\;
scsi0 : qla4xxx
QLogic iSCSI HBA Driver version: 5.00.07-d2
QLogic ISP4022 @ 0000:0a:01.1 hdma-, host#=0, fw=02.00.00.07
Vendor: DGC Model: RAID 5 Rev: 0220
Type: Direct-Access ANSI SCSI revision: 04
qla4xxx 0000:0a:01.1: scsi(0:0:1:0): Enabled tagged queuing, queue
depth 32.
sda : very big device. try to use READ CAPACITY(16).
SCSI device sda: 4613734400 512-byte hdwr sectors (2362232 MB)
sda: test WP failed, assume Write Enabled
sda: asking for cache data failed
sda: assuming drive cache: write through
sda : very big device. try to use READ CAPACITY(16).
SCSI device sda: 4613734400 512-byte hdwr sectors (2362232 MB)
sda: test WP failed, assume Write Enabled
sda: asking for cache data failed
sda: assuming drive cache: write through
sda:<6>sd 0:0:1:0: Device not ready: <6>: Current: sense key=0x2
ASC=0x4 ASCQ=0x3
end_request: I/O error, dev sda, sector 0
Buffer I/O error on device sda, logical block 0
sd 0:0:1:0: Device not ready: <6>: Current: sense key=0x2
ASC=0x4 ASCQ=0x3
end_request: I/O error, dev sda, sector 0
Buffer I/O error on device sda, logical block 0
unable to read partition table
sd 0:0:1:0: Attached scsi disk sda
QLogic iSCSI IOCTL Module version: v1.00.00
sd 0:0:1:0: Device not ready: <6>: Current: sense key=0x2
ASC=0x4 ASCQ=0x3
end_request: I/O error, dev sda, sector 4613734272
Buffer I/O error on device sda, logical block 576716784
sd 0:0:1:0: Device not ready: <6>: Current: sense key=0x2
ASC=0x4 ASCQ=0x3
end_request: I/O error, dev sda, sector 4613734272
I am still unsure what these Buffer I/O errors are, but maybe it has a
problem with the size of the disk!? I will look a bit further into
that.
However I now have the same problem as Alberto. I also get a kernel
panic when I try to get a list the LUN's (menu 7-7). Did Alberto send
you the error output? Do you want me to send mine?
Thanks!
Jonathan
Joni
> I assigned a 200GB virtual disk to the initiator. After setting the
> target and a reboot it gives me a /dev/sda device. Have a look at the
Uhm, that's a 2200GB virtual disk.
Thanks!
Jonathan
Alberto Cammozzo - mmzz
> Hi Ravi,
>
> I managed to establish the connection to our Dell EMC AX-150i on which
> I assigned a 200GB virtual disk to the initiator. After setting the
> target and a reboot it gives me a /dev/sda device. Have a look at the
> dmesg:
[...]
>
> However I now have the same problem as Alberto. I also get a kernel
> panic when I try to get a list the LUN's (menu 7-7). Did Alberto send
> you the error output? Do you want me to send mine?
I did. This is the relevant part, see if it looks the same.
Will Google groups interface mess up formatting?
BUG: unable to handle kernel NULL pointer dereference at virtual
address 00000018 printing eip:
f9ac5eb3
*pde = 00000000
Oops: 0000 [#1]
SMP
Modules linked in: qisioctl qla4xxx skge via_velocity crc_ccitt b44
8139cp 3c59x sctp configfs hwmon
CPU: 0
EIP: 0060:[<f9ac5eb3>] Not tainted VLI
EFLAGS: 00010082 (2.6.18-rc4 #1)
EIP is at qla4xxx_status_entry+0xf6/0x7a8 [qla4xxx]
eax: 00000000 ebx: f6938000 ecx: f69385a8 edx: f78a1b40
esi: f6f182d8 edi: f6f182d8 ebp: 000000c0 esp: c05d9f00
ds: 007b es: 007b ss: 0068
Process swapper (pid: 0, ti=c05d8000 task=c0517da0 task.ti=c05d8000)
Stack: f7616840 f78a1b40 f6fd0400 f6f182d8 f6f182d8 f6fd0400 00000001
f9ac66e2
f6f182d8 f6fd0400 f6f182d8 00000010 00000046 00000017 f9ac6e5a
f6f182d8
f6f182d8 00000000 f9ac71ae f6f182d8 00000010 00010000 f72ec3e0
00000000
Call Trace:
[<f9ac66e2>] qla4xxx_process_response_queue+0xae/0x310 [qla4xxx]
[<f9ac6e5a>] qla4xxx_interrupt_service_routine+0x32/0x79 [qla4xxx]
[<f9ac71ae>] qla4xxx_intr_handler+0x30d/0x394 [qla4xxx]
[<c01330c1>] handle_IRQ_event+0x21/0x49
[<c0133166>] __do_IRQ+0x7d/0xd7
[<c010480b>] do_IRQ+0x43/0x50
[<c0103146>] common_interrupt+0x1a/0x20
[<c0100cdc>] mwait_idle+0x28/0x38
[<c0100bbd>] cpu_idle+0x64/0x79
[<c05df7c2>] start_kernel+0x1b6/0x1b8
Code: 20 83 3d 50 c9 ad f9 02 75 17 68 e8 c8 ac f9 0f b7 47 42 50 68 40
46 ad f9 e8 6f 23 65 c6 83 c4 0c 8b 54 24 04 8b 42 28 8b 40 08 <8b> 70
18 85 f6 0f 84 41 04 00
00 8b 4c 24 08 8b 44 24 08 0f b6
EIP: [<f9ac5eb3>] qla4xxx_status_entry+0xf6/0x7a8 [qla4xxx] SS:ESP
0068:c05d9f00
<0>Kernel panic - not syncing: Fatal exception in interrupt
BUG: warning at arch/i386/kernel/smp.c:547/smp_call_function()
[<c010af34>] smp_call_function+0x5a/0x11d
[<c0118223>] printk+0xe/0x11
[<c0110131>] bust_spinlocks+0x3d/0x46
[<c010b03d>] smp_send_stop+0x10/0x1b
[<c010aff7>] stop_this_cpu+0x0/0x36
[<c01179f9>] panic+0x3e/0xd8
[<c0103951>] die+0x1b3/0x1ca
[<c011077e>] do_page_fault+0x3ec/0x4c6
[<c0110392>] do_page_fault+0x0/0x4c6
[<c0103265>] error_code+0x39/0x40
[<f9ac5eb3>] qla4xxx_status_entry+0xf6/0x7a8 [qla4xxx]
[<f9ac66e2>] qla4xxx_process_response_queue+0xae/0x310 [qla4xxx]
[<f9ac6e5a>] qla4xxx_interrupt_service_routine+0x32/0x79 [qla4xxx]
[<f9ac71ae>] qla4xxx_intr_handler+0x30d/0x394 [qla4xxx]
[<c01330c1>] handle_IRQ_event+0x21/0x49
[<c0133166>] __do_IRQ+0x7d/0xd7
[<c010480b>] do_IRQ+0x43/0x50
[<c0103146>] common_interrupt+0x1a/0x20
[<c0100cdc>] mwait_idle+0x28/0x38
[<c0100bbd>] cpu_idle+0x64/0x79
[<c05df7c2>] start_kernel+0x1b6/0x1b8
Alberto
ravianand
Its coming back as Not Ready. Make sure
that its connected to the controller where lun
is mapped. Is 150i dual controller where it has
SPA and SPB controller ports ?
Secondly try using a smaller size disk and see
how it goes.
Ravi
ravianand
>
> Hi Ravi,
>
> I managed to establish the connection to our Dell EMC AX-150i on which
> I assigned a 200GB virtual disk to the initiator. After setting the
> target and a reboot it gives me a /dev/sda device. Have a look at the
> dmesg:
>
> panic when I try to get a list the LUN's (menu 7-7). Did Alberto send
> you the error output? Do you want me to send mine?
I recieved the logs from him which I am investigating it right now.
Will keep you guys posted.
Thanx
Ravi
>
>
Ross Anderson
question do have is, how can I get the system to see new luns that are
added? I can open iscli and see the new luns, I've reset the HBA.
However when I cat out /proc/scsi/qla4xxx/<#> It does not show the
changed luns. Only method I've found so far is to rmmod && modprobe
which won't work in a production environment. Under older versions you'd
echo out a command to the proc system. What am I missing?
Thanks,
Ross Anderson
ravianand
Please apply the attached patch as follows :
# cd <driver_src_dir>
# patch < <path_to_patch>/patch_to_address_panic_issue
and re-run the test. Let us know how it goes and send us the log
file/console log to root cause this issue.
Thanx
Ravi
QLogic Corp
ravianand
Try out the following :
# echo "scsi add-single-device X 0 Y Z" >/proc/scsi/scsi
where X -> would be the instance no you see in /proc/scsi/qla4xxx/
for the HBA your tgt is connected to
Y -> is the tgt no
Z -> is the lun no.
Ravi
Alberto Cammozzo - mmzz
I did as told, but I'm not able to log in the target:
qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
00040007 10.10.10.100:3260
"iqn.2001-04.com.example:storage.disk2.sys1.xyz"
qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
00040007 10.10.10.100:3260
"iqn.2001-04.com.example:storage.disk2.sys1.xyz"
Not sure it's the patch's fault...
Even if I didnt change anything on the working setup, I don't have time
now to investigate if it connects again without your patch. Will try
again on monday.
A.
ravianand
>
> Ravi,
> I did as told, but I'm not able to log in the target:
>
> qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
> 00040007 10.10.10.100:3260
> "iqn.2001-04.com.example:storage.disk2.sys1.xyz"
> qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
> 00040007 10.10.10.100:3260
> "iqn.2001-04.com.example:storage.disk2.sys1.xyz"
>
> Not sure it's the patch's fault...
is no active sessions. Can you ping the target to make sure that
there is no conf issue ?
>side there i
> Even if I didnt change anything on the working setup, I don't have time
> now to investigate if it connects again without your patch. Will try
> again on monday.
Sure.
Thanx
Ravi
>
> A.
>
>
>
Alberto Cammozzo - mmzz
ravianand ha scritto:
> >On Sat, 02 Sep 2006, Alberto Cammozzo - mmzz wrote:
> >
> > Ravi,
> > I did as told, but I'm not able to log in the target:
> >
> > qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
> > 00040007 10.10.10.100:3260
> > "iqn.2001-04.com.example:storage.disk2.sys1.xyz"
> > qla4xxx 0000:03:0c.1: DDB[2] MB0 4000 Tot 3 Next 3 State 0006 ConnErr
> > 00040007 10.10.10.100:3260
> > "iqn.2001-04.com.example:storage.disk2.sys1.xyz"
> >
> > Not sure it's the patch's fault...
> I dont think so. Above snippet tells me that the from the HBA side there
> is no active sessions. Can you ping the target to make sure that
> there is no conf issue ?
Yes, it was my fault. I was setting up a sort of wireless testbed ;-)
The patch seems to work; no more kernel panic.
I did several tests, both after reboot and after unloading/reloading
modules.
There is an odd behaviour the first time you list the LUNS: (menu 7/7).
Instead of one LUN it lists a number of them. The second time it lists
only one LUN, correctly.
I'll send you the logs in private to avoid google groups format mess.
bye
Alberto
Joni
Now that this kernel panic is fixed, I assume there will be a patched
driver available? Where would be the best place to keep up to date of
driver updates? Will that be
ftp://ftp.qlogic.com/outgoing/linux/beta/iscsi/5.x/? And where is the
best place to stay in touch with the developers of the qla4xxx drivers?
Would that be here in this group?
Thanks!
Jonathan
Joni
Oh yes and one more thing: is there any effort being made to have the
qla4xxx and qisioctl drivers included in the vanilla kernel? I'm for it
:)
Bye,
Jonathan
ravianand
> @Ravi:
>
> Now that this kernel panic is fixed, I assume there will be a patched
> driver available? Where would be the best place to keep up to date of
> driver updates? Will that be
> ftp://ftp.qlogic.com/outgoing/linux/beta/iscsi/5.x/?
It will be made available at the path you have outlined.
And where is the
> best place to stay in touch with the developers of the qla4xxx drivers?
Please post it in open-iscsi list as you have already been doing.
Myself/David.Somayajulu and couple of other folks are responsible
for maintaining this driver at QLogic. Doug Maxey and Mike.C
has made significant contribution to this driver as well.
> Would that be here in this group?
Yes.
>
> Thanks!
> Jonathan
>
>
>
ravianand
>
>
> Oh yes and one more thing: is there any effort being made to have the
> qla4xxx and qisioctl drivers included in the vanilla kernel? I'm for it
We are working on getting qla4xxx driver included in vanilla kernel, but
its a different driver than qla4xxx/ioctl module combo. qla4xxx/ioctl combo
is an intermediate soultion for people to use our app's until we have
the compelte support for sysfs in upstream driver for our app's to work.
In a nutshell qla4xxx upstream driver is a work in progress and the current driver
for it can be downloaded from the following link :
ftp://ftp.qlogic.com/outgoing/linux/iSCSI/upstream/5.00.05b9-k/
Thanx
Ravi