ICP vortex logo
Headline Logo 
[PrivacyPolicy] [Terms of Usel] [Imprint]
News
FAQs - Linux, BSD, various Distributions
Topics:

Older ICP Controllers and current Linux* distributions

U160 SCSI and older ICP RAID Controller are NOT validated and released for current Linux* distributions. This is especially related to all distributions with kernel 2.4.19 or newer.

Generally older ICP controllers should work with the newer Linux* distributions since the gdth ICP driver typically supports the older controller models. However, since the correct function is extremely dependent on the used hardware / Linux* distribution and the controllers are not tested/validated any longer, there is NO guarantee for proper operation.

The ICP driver and tools, which are required for newer kernel releases, can be downloaded from http://www.icp-vortex.com/english/download/rz_neu/linux/linux_e.htm.

The ICP RAID controllers up to U160 are NOT released for new Linux* distributions, there is NO installation support.



Top of page

RedHat 8.0 - "Segmentation fault" on Systems

Problem:
With a standard Red Hat 8.0 installation "segmentation fault" error messages may be encountered in various situations:
- During system shutdown/reboot
- While starting or using ICP RAID Console (icpcon)
- While accessing the "/proc" file system (grep, cat,...)

Cause:
Seems to be related to "/proc" filesystem accesses. The root cause is unknown.

Solution:
Red Hat offers an update kernel (2.4.18-18.8.0), which solves the problem.
To install the new kernel, please perform the following steps:
- Download update kernel (2.4.18-18.8.0) as an RPM and copy it to the Red Hat 8.0 system
Patch located at: https://rhn.redhat.com/errata/RHSA-2002-262.html
- Use the command "rpm -ivh kernel-*.rpm" to install the new kernel
- Activate the new kernel as the standard boot kernel (e.g. in lilo.conf).

Furthermore it is necessary to reset the text terminal before using the monitoring tool icpcon, otherwise the display is corrupted - simply type "reset" before starting icpcon.



Top of page

SuSE 8.1 - "Segmentation fault" on Systems

Problem:
With a standard SuSE 8.1 installation "segmentation fault" error messages may be encountered in various situations:
- During system shutdown/reboot
- While starting or using ICP RAID Console (icpcon)
- While accessing the "/proc" file system (grep, cat,...)

Cause:
Seems to be related to "/proc" filesystem accesses. The root cause is unknown.

Solution:
SuSE offers an update kernel which solves the problem. This kernel can be downloaded from:
http://www.suse.de/en/private/download/updates/81_i386.html (04 Dec 2002)
and installed with "rpm -Uvh kernel-..." (kernel-... = name of the update kernel).
After the kernel is installed you can activate it as the standard boot kernel (e.g. lilo.conf).



Top of page

SuSE 8.1 - Installation Problems

Problem:
SuSE 8.1 can not be installed - system "freezes" during installation.

Cause:
Depending on the used hardware the ACPI function can not be configured correctly.
Further information see: http://sdb.suse.de/en/sdb/html/81_acpi.html

Solution:
Shortly after booting the SuSE 8.1 Installation CD select "Installation - Safe Settings"
and/or apply the command line switch "acpi=off".



Top of page

"Segmentation fault" on multi processor boards with RedHat 7.3

Problem:
With a standard Red Hat 7.3 installation on multi processor platforms "segmentation fault" error messages may be encountered in various situations:
- During system shutdown
- While starting or using ICP RAID Console (icpcon)
- While accessing the "/proc" file system (grep, cat,...)

Cause:
Seems to be related to "/proc" filesystem accesses. The root cause is unknown.

Solution:
Red Hat offers an updated kernel (2.4.18-5), which solves the problem.
To install the new kernel, perform the following steps:
- Download updated kernel (2.4.18-5) as a RPM update from
http://rhn.redhat.com/errata/RHBA-2002-110.html
- Use the command "rpm -ivh kernel-*.rpm" to install the new kernel
- Include kernel sources into the boot configuration (lilo.conf).

Note:
On the listed website there is no reference to ICP controllers, but the new kernel fixes the problems mentioned above.



Top of page

Red Hat Linux* 7.1 does not detect RAID controllers!
ftp://ftp.redhat.de/pub/rh-addons/driverdisks/ contains a new disk image
with the driver (gdth 1.28) and updated PCI IDs (including the new custom
ID for the ICP controllers after the RZ series).


Top of page

Speed up the boot of the linux kernel

Loading of the kernel can be accelerated using the "compact" switch, in the general section of the file "lilo.conf". The Linux loader then reads several sectors concurrently.

Example: Extract from "lilo.conf"

- / -
boot = /dev/sda
vga =771
read-only
menu-scheme = Wg:kw:Wg:Wg
lba32
prompt
compact
timeout = 80
message = /boot/message
- / -



Top of page

Driver update
Driver update by importing the driver sources
a) Copy the driver sources to /usr/src/linux/drivers/scsi
(using a DOS diskette e.g. using 'mcopy a:gdth.tgz').
b) Unpack the archive using 'tar xvfz gdth.tgz
c) Compile new kernel
(depending on system configuration and Linux Distribution distinctively, if applicable beforehand 'make config', 'make dep', 'make clean' are necessary)


Driver installation/update using a patch
Clean installation of the driver patch:
a) Copy the corresponding patch to the Linux system (best to /usr/src/linux e.g. gdtp2036.gz for Linux 2.0.36)
b) run patch, e.g. for Linux 2.0.36:
cd /usr/src/linux
zcat gdtp2036.gz | patch -p1 2>log_file
Then the file log_file should be checked for any errors
c) Configure kernel / register ICP controller using:
cd /usr/src/linux
make config
d) Determine dependencies /compile kernel:
see LINUX manual (e.g make dep && make zlilo)
e) Reboot

Updating the driver patch:
a) It is best to copy the new patch to /usr/src/linux
b) Undo the old patch using
zcat alter_patch.gz | patch -p1 -R
rm drivers/scsi/gdth*
(The gdth sources with the version less than 1.01 must be manually deleted using rm, because they are not deleted by the -R patch. From driver version 1.01, there are new patch files, which no longer necessitate this deletion stage)
c) Execute patch
zcat neuer_patch.gz | patch -p1 2>log_file
d) Recompile kernel
e) Reboot



Top of page

Installation of ICP controllers under Linux Red Hat 7.1
To install Red Hat Linux 7.1 on an ICP RAID controller, a driver diskette is needed. On it is the current Linux driver for all ICP controllers. This diskette can be downloaded from the following internet address.
ftp://ftp.redhat.de/pub/rh-addons/driverdisks/

Faulty installations diskettes with Red Hat 6.0 and 6.1
Unfortunately the installation diskettes for Red Hat 6.0 are incomplete.
Red Hat will promptly be changing this and accordingly is already providing new diskette images available on its web site.
Or you can also download these images directly from our web site: rh6dd.zip rh6img.zip
(Using the images in rh6dd.img, the three installation diskettes are created under Linux using dd. The archive rh6img.zip
contains the ICP tool iwrite for creating the three diskettes under MSDOS, in addition to the images).
Installation instruction: After starting from the first diskette, the parameter 'Linux supp' must be copied at the Linux prompt.

With Red Hat 6.1 there is also no ICP driver available on the diskettes or on the CD for installation. An additional driver diskette with ICP driver is available from www.redhat.com/support/errata/RHEA1999000-01.html. When installing, boot with "linux dd" to integrate this additional driver.



Top of page

Installation of ICP controllers under Linux Caldera eServer 2.3

To install Caldera eServer 2.3 on an ICP RAID controller at the command prompt the command line

"install insmod=gdth"

must be entered.



Top of page

OpenBSD does not boot with standard ICP controller firmware 2.39.05-R065

Problem:
OpenBSD does not boot with standard ICP controller firmware 2.39.05-R065 with BIOS 7.04H Oct 7, 2003
Affected Controllers: GDT8500RZ, GDT8114RZ, GDT8514RZ, GDT8124RZ, GDT8524RZ, GDT8546RZ, GDT8522RZ

Background:
OpenBSD requires extended memory handling in ICP BIOS.

Solution:
Use the attached Firmware 2.39.05-R069 with BIOS 7.04I Dec 5, 2003

Attention: use this firmware only on above listed controllers.
Note: The driver for OpenBSD was developed by OpenBSD engineering, therefore ICP only can offer limited support for this operating system.



Top of page

gdth driver parameter
Up to driver version 1.07, driver parameters are indicated directly in /usr/src/linux/drivers/scsi/gdth.c. As from driver 1.10, the command append (append = "gdth=…") in /etc/lilo.conf or the LILO boot prompt (gdth=…") can be used to copy parameters to the driver.

Reserving SCSI drives:
Reserving SCSI devices is always necessary whenever so-called 'Direct Access Devices' are to be made available to Linux directly without them having been configured by the ICP controller as host drivers beforehand (e.g. Syquest or Iomega drives). 'Non Direct Access Devices' such as CD-ROM, Streamers, DATs, etc. do not have to be reserved. Warning: Reservation only works with uninitialised SCSI drives. Reservation is supported from gdth driver version 1.05. Up to the ICP driver version 1.07 reservation is carried out directly in driver sources. To do this search in '/usr/src/linux/drivers/scsi/gdth.c' for the entry 'gdth_reserve_str'. A masked example entry can already be found in this section. Modify this entry so that the coordinates point to the SCSI device being reserved (specify adapter, bus, ID and LUN). Save file and recompile kernel. As from ICP driver version 1.10, as already mentioned, the driver parameters can be copied using the LILO append command. Extract from /etc/lilo.conf:

# End LILO global section
# Linux bootable partition config begins
image = /boot/vmlinuz
root = /dev/sda2
label = Linux
append = "gdth=reserve_list:1,0,6,0"
# Linux bootable partition config ends

This gdth driver parameter reserves the SCSI drive connected to ICP controller 1 (second ICP controller), SCSI Bus 0 (channel A), ID 6, LUN 0 for the operating system (raw service). After each change in /etc/lilo.conf, lilo has to be called once. Driver parameters can also be entered at the LILO boot prompt (e.g. gdth=reserve_list:1,0,6,0).

Other driver parameters:
irq1,irq2, etc. Only for GDT3000/GDT3020 ICP EISA controllers and only if their BIOS is disabled (irq1, irq2, etc. corresponding with the IRQs of the aforementioned EISA controller)
disable:Y disables the ICP driver
disable:N enables the ICP driver
reserve_mode:0 does reserves no SCSI devices [*1]
reserve_mode:1 reserves all uninitialised removable disks [*1]
reserve_mode:2 reserves all uninitialised drives [*1]
reserve_list:h,b,t,l,h,b,t,l, reserves drives at the corresponding coordinate where h=host adapter, b=SCSI bus, t=target ID, l=LUN
reverse_scan:Y reverse order of PCI controllers
reverse_scan:N order of the PCI controllers according to the order in the BIOS
max_ids:x x = Number of target IDs per SCSI channel (accelerates boot procedure)
rescan:Y rescan of all channels / IDs
rescan:N only use SCSI devices, which have already been found [*1]

[*1] The driver parameter reserve_mode and rescan:N require ICP controllers with firmware 1.23.00/2.23.00 or higher.

The current default settings for the drivers are as follows:
"gdth=disable:N,reserve_mode:1,reverse_scan:N,max_ids:127,rescan:N"
If the gdth driver is loaded as a module, these parameters can also be copied. When doing this special syntax must however be used. IRQ parameters must be stated with 'IRQ=' (e.g. IRQ=10). Where there are several parameters, instead of using a comma (','), use a space
(' '). All ':' should be replaced with '=', all 'Y' with '1' and all 'N' with '0'.
Default:
"modprobe gdth disable=0 reserve_mode=1 reverse_scan=0 max_ids=127 rescan=0"



Top of page

Notes
/proc-Support:
Read from /proc/scsi/gdth/0, /proc/scsi/gdth/1, .. :
Display for the controller/SCSI bus, firmware version, driver version (e.g. using cat /proc/scsi/gdth/0)
Write to /proc/scsi/gdth/0, .. :
gdth flush for flushing all host drives connected to the controller
gdth flush drive for flushing the host drive with the number of the drive
gdth wbp_off Write-Back switch off permanently [*2]
gdth wbp_on Write-Back switch on permanently [*2]
gdth wb_off Write-Back switch off temporarily (as from FW 1.15/2.15)
gdth wb_on Write-Back switch on temporarily (as from FW 1.15/2.15)

[*2] This Write-Back setting applies for all host drives of the ICP controller and also continues to be effective after a cold start or switching the system on/off !



Top of page

If you have technical questions or problems with ICP Controller, please contact our technical support. Contact us: Monday till Thursday from 8:00am-12:00am and 1:00pm-5:00pm, Friday from 8:00am-12:00am and 1:00pm-3:00pm. (CET)
E-mail: icp_support@adaptec.com
Phone: +49-(0)7132-9620-900


{leer2}
Company
Products
Partners
Events
Solutions
Support
Download
Feedback
Home
german pages
Copyright
blank area