Integrated Management Module (IMM) Firmware Update Change History
IMPORTANT: Prerequisite for updating to IMM firmware version 1.30 (yuooc7e) ========== Updating to IMM firmware version 1.30 (yuooc7e) requires the current IMM firmware version to be 1.10 (yuoo57h), or later. If this prerequisite is not met before updating to version 1.30 (yuooc7e) the system will not power on correctly. If the system is not running a minimum of IMM version 1.10 (yuoo57h), update the IMM before attempting to update to 1.30 (yuooc7e). This prerequisite was a limitation of IMM version 1.30, and it has been fixed in IMM version 1.31(yuooc7f). This prerequisite does not apply to IMM version 1.31 or later.
IMPORTANT: Please DO NOT apply the firmware update (yuoo91e, v1.24) ========= if you are using AEM functions in x3620M3 or x3630M3. The v1.24 firmware will disable the power capping functionality in those two systems. Please remain on the current IMM version or update the firmware to v1.25 or newer. Users not using AEM functions can still update to v1.24.
IMPORTANT: Read before using the MAX5 ========== This Technical Update supplements the documentation that comes with the IBM MAX5 for System x. Keep this document with your other MAX5 documentation. Before you attach the MAX5 to the server and try to use it, you must update the server firmware with the latest level of firmware or code. If you attach and try to use the MAX5 without updating the server firmware, you might get unexpected system behavior or the server might not power on. For special instructions to follow before you attach the MAX5 to the server, go to http://www.ibm.com/support/entry/portal/docdisplay?lndocid=MIGR-5085756.
IBM maintains pages on the World Wide Web where you can get the latest technical information and download device drivers and updates. To access these pages, go to http://www.ibm.com/systems/support/ and follow the instructions.
IMPORTANT: If you are at IMM firmware versions prior to 1.12 (yuoo67c) and ========== updating IMM firmware in Microsoft Windows, when you update IMM to versions between 1.14 (YUOO73M) and 1.30 (YUOOC7E), it will take up to 10 minutes additional wait time; when you update IMM to version 1.30 (YUOOC7E) or later, it will take up to 20 minutes additional wait time. This is due to the iflash tool checking the IMM before flashing. During this time the flashing process may appear hung or unresponsive. The flashing will then complete and there is no functional impact.
IMPORTANT: Make sure the current IMM firmware level is 1.14 or later ========== before you update the uEFI, FPGA, or DSA Preboot firmware. The recommended firmware update sequence is IMM followed by UEFI followed by FPGA followed by DSA Preboot.
IMPORTANT: A firmware change was made in YUOO57H (v1.10) that enforces the ========= requirement for login accounts on the IMM to have passwords that contain both alpha and numeric characters. If you are updating the IMM firmware from YUOO56K (v1.09) or earlier, to YUOO57H (v1.10) or newer, prior to updating the firmware, please ensure that the IMM user account passwords contain both alpha and numeric characters. If the firmware is updated and the passwords do not contain both alpha and numeric characters, you may need to reset the IMM to its default settings using the menus in the UEFI F1 set up display.
IMPORTANT: When updating the IMM firmware from a version earlier than ========= version 1.05 (build YUOO32F) to version 1.05 or newer, the Remote Alert Recipient Email Addresses and the Daylight Saving Time setting will revert to the default setting. Users that utilize these settings will need to set them again after the update.
IMPORTANT: When updating IMM, uEFI, pDSA, FPGA via any method, ========== please follow firmware update best practice document (http://www-947.ibm.com/support/entry/portal/docdisplay?brand= 5000020&lndocid=MIGR-5082923) for suggested update sequence and prerequisite.
IMPORTANT: Due to an Intel errata, if your server is running a 32-bit Linux ========== Operating System, C-states may need to be disabled while updating uEFI, IMM, or DSA Preboot through the Linux command line interface. See RETAIN TIP H195678 for more information. Updates performed through the AMM, Bootable Media Creator, or the IMM web or CLI interfaces do not have this requirement.
IMPORTANT: Updated firmware version numbers might not be visible in SMBIOS ========== tables until after the server is rebooted.
IMPORTANT: After downgrading IMM firmware from version 1.22 ========== to an earlier version, the user needs to reset the IMM to factory default before normal operation.
Version 1.44, YUOOG6C Problem Fixed: - Critical Fix for x3550M3, x3650M3, dx360M3, x3630M3, x3400M3, x3620M3, x3500M3, x3200M3, x3250M3, dx360M2, x3550M2, x3650M2, x3400M2, x3500M2, x3850/x3950X5, x3690X5 -- Updated the firmware with patches to address the OpenSSL vulnerabilities listed in CVE-2014-0224, CVE-2014-0076 Enhancements: N/A
Version 1.43, YUOOG6B Problem Fixed: - Fixed the problem that the LDAP group filter name changed after upgrading IMM. - Fixed the problem on blade servers where the management bus will become unstable if the AMM in slot1 is removed from chassis. - Fixed the problem that the SNMP engine ID is changed after IMM reset. Enhancements: - Add security enhancements.
Version 1.42, YUOOG2C Problem Fixed: - Fixed a NTP sync problem on IMM web - Fixed a time zone setting problem on IMM web
Enhancements: - Added SHA256 and SHA512 support - Add a message in IMM event logs when power on is blocked due to VRD issue on x3850 X5 - Power off the server after N boot attempts without locating a boot device on x3850 X5/x3950 X5/HX5.
Version 1.41, YUOOF7C Problem Fixed: - Fixed the problem on blades where sometimes no throttle event was reported by AMM when CPU temperature is higher than Tcontrol. Enhancements: - Enhanced logging for throttle and Power Management.
Version 1.40, YUOOF7A Problems Fixed: - Fixed the problem on x3690X5 that host OS will be shut down due to wrong overtemp event. - Fixed the problem that "unsupported software version" could be showed in event logs after updating IMM version 1.30 or newer on x3850/x3950 X5, x3690X5 and HX5 - Fixed the problem that SMTP user name field on IMM web page only allows 16 chars. - Fixed the problem on HS22 where IMM reports 0 watt power during initialization
Enhancements: - Enabled IMM to fail over to backup automatically when there is a IMM firmware corruption - Enhanced IMM web GUI response. - Added notification when a corrupted GUID Partition Table (GPT) is encountered by UEFI during POST. Please refer to RETAIN Tip H207759 for details. - Enhanced logging for Power Controls management.
Version 1.39, YUOOF2C Problems Fixed: - Fixed the problem on IBM System x3850X5 where SAP HANA configurations can reset in rare cases. More details in RETAIN Tip H207295.
Version 1.38, YUOOF2B Problems Fixed: - Fixed the problem on x3850 X5/x3950 X5 where the System Fault LED will be lit when only one PSU is installed.
Version 1.37, YUOOE9C Problems Fixed: - Corrected the voltage sensor threshold for HS22 and HS22V system.
Version 1.36, YUOOF2A Enhancements: - Improved IMM web page Accessibility
Version 1.35, YUOOE9B Problems Fixed: - Fixed the thermal issue for Broadcom quad port ethernet (IBM PN 49Y7948., FRU 49Y7949). - Fixed the problem that user authentication for LDAP with SSL will fail when Configured Credentials is enabled. - Updated the thermal profile for NVIDIA M2090 GPU card support in dx360 M3. - Fixed the problem on HS22 systems where pressing the KVM button too soon (right after Power LED starts blinking slowly) would make the KVM owner unable to be changed. - Fixed the problem on blade servers where IMM may report the "Unsupported blade/blower combination" error for 130W Westmere CPUs. - Fixed the problem on x3200 M3 where when power policy is set as "restore" or "always on" in uEFI, enforcing system shutdown during POST will cause automatic system boot-up. - Fixed the problem about SNMP drops intermittently and SNMP data cannot be accessed in a period of time - Fixed the problem on x3850 X5/x3950 X5 Multi-Node systems where "redundancy lost" event is not logged in the Primary Node when one of the power supplies in the Secondary Node is lost. - Optimized CPU power consumption calculation method for x3690 X5. - Fixed the problem on blade servers where IMM does not automatically recover NCSI. - Fixed the problem on blade servers that IMM constantly has "Init Failures" and fails to start up. - Fixed the Readme file problem that some IP addresses of LAN-over-USB instructions are inconsistent. - Fixed the CMOS Voltage Over Threshold problem on HS22V systems. - Fixed the problem that SSH login failed when the setting is configured as "Local first then Ldap" or ldap settings are not valid. - Fixed the problem that the IMM web interface responds too slowly or even fails to respond.
Enhancements: - Enhanced the entropy (randomness) when generating SSL keys. - Added secure SMTP authentication support. - Added I/O board VPD information in IMM web interface - Added the function of letting Location LED remain on after system power-off. - Enhanced event Severity reporting of the QPI Link LED on x3850 X5/x3950 X5 so that IBM Director can interpret it correctly. - Improved IMM version reporting for IBM Director. - Added a function to keep 110V Power Supply Redundancy by throttling the system for x3850 X5/x3950 X5. - Usability enhancement: all events will be checked by default on the alert setting web page. - Enabled ASU to control each sub-category for email alert recipient setting and SNMP alerts setting. - Added new BCH Midplane (that supports 8Gbit Fibre Channel) support. - Added the function to capture all IMM logs with one button click. - Enhanced the DataPower appliance blade support.
Version 1.34, YUOOE3E Enhancement: - Add support of the following fusion-io adapters : 1.2TB High IOPS MLC Adapter for IBM System x 2.4TB High IOPS MLC Adapter for IBM System x 365GB High IOPS MLC PCIe Adapter for IBM System x 785GB High IOPS MLC PCIe Adapter for IBM System x
Version 1.33, YUOOE3C Problems Fixed: - Fixed the problem that a blade's Fault LED is showed as "on" on AMM GUI while there is no active critical event and the LED on the physical blade is off. - Fixed the problem that the "Date" header is missing from the IMM SMTP trap. - Fixed the problem that the "Look Below" LED (CR48) on x3690 X5 Memory Expansion Board may turn on incorrectly when there are no errors detected. - Fixed the problem on x3850 X5/x3950 X5 where the system information in the IMM configuration file may be truncated and the incomplete system information will cause the HTTP server in IMM run into an unstable state. - Fixed the problem that the Scalability LED turns off after the IMM on either node of a 2-Node x3850 X5/x3950 X5 is reset. - Fixed the problem that IMM on x3850 X5/x3950 X5 might fail to start after the IMM firmware is successfully upgraded. - Fixed the problem where "Comm Errors" may occur when the IMM firmware on HX5 is upgraded. - Fixed the problem that dx360 M3 may have PCI errors when a NVIDIA M2090 GPU card is installed. - Fixed the problem that opening the ActiveX remote console closes all open tabs in the same Internet Explorer window. - Fixed the problem that x3550 M3 may occur ServerRAID BBU over temperature and ServerRAID M5016 over heat issues.
Enhancements: - Added Internet Explorer 9 support. - Changed the maximum User's Distinguished Name(DN) for LDAP from 128 bytes to 256 bytes. - Removed the limitation of 255 memberOf groups for an LDAP user in IMM. - Enhanced IMM Web session management. - Added an IMM command line (CLI) command for one time PXE boot option setting. - Added a new unique error code in AMM and changed the text in AMM and IMM for CMOS battery errors. - Updated Russia time zone definition to no more daylight saving time.
Version 1.32, YUOOD4G Problems Fixed: - Fixed a problem where IMM web interface does not lock out a user if the same user exhausts login retries and gets locked out through the SSH interface - Fixed a problem that caused the IMM to report false critical DIMM temperature events - Fixed an intermittent problem where sometimes the Power Supply DC status LED is not illuminated on the x3550 M2/M3, x3650 M2/M3, x3620 M3 and x3630 M3 - Fixed a problem that sometimes causes the IMM web interface to shut down after multiple invalid login attempts Enhancements: - Added Service Advisor support on x3850/x3950 X5 and x3690 X5 - Changed the IMM firmware to allow every service port to be Enabled or Disabled - Changed the IMM firmware to not disable the network interface when there is a conflict IP address on the network - Changed the IMM firmware to permit SSL client certificates to be managed while the SSL client is enabled
Version 1.31, YUOOC7F Problems Fixed: - Fixed the problem that after updating IMM firmware from versions prior to v1.10(YUOO57H) to v1.30(YUOOC7E) the system won't power on correctly
Version 1.30, YUOOC7E Problems Fixed: - Fixed the intermittent Power Supply error on x3550M2/M3 and x3650M2/M3 - Fixed the false alarm message for ArBel Power Supply on x3550M2/M3 and x3650M2/M3 - Fixed the issue that a blade intermittently does not respond to the system management bus on HX5 and HS22/S22V - Fixed the duplicate PECI messages issue on x3850/x3950 X5, x3690X5 and HX5 - Fixed the intermittent problem that a user can not login to IMM web after inventory from IBM Director - Fixed the problem that SMTP test alert doesn't work - Fixed the problem that SNMP remote alert doesn't work - Fixed the intermittent IMM reset problem - Fixed the SSL vulnerability issue Enhancements: - Added 64-bit ActiveX Remote Presence Client support - Added BGE overweight detection on HS22 - Added support of differentiating memory failures happening during UEFI POST and during run-time - Added support of reporting AC restore status on x3850/x3950 X5 - Added the change to include DIMM slot number info in the memory error event message
Version 1.29, YUOOB7F Problem(s) Fixed: - Fixed a problem where false memory configuration error messages are sometimes reported on the x3850/x3950 X5
Version 1.28, YUOOB7C Problem(s) Fixed: - Fixed a problem where the IMM sometimes records duplicate event log entries - Fixed a problem that caused the DIMM Bank number to not be correctly reported - Fixed a problem that prevented the system from powering on when there is a firmware mismatch - Fixed a problem that didn't allow a password to be saved if it contains a & or < character - Fixed a problem where the IMM VPD sometimes is not available on the web page - Fixed a problem where one port on the QLogic 8GB HBA card sometimes is not recognized - Fixed a problem that sometimes caused the DVD drive to disappear in RHEL5 after the IMM is reset - Fixed a false Power Supply failure message - Removed the error message pop up when changing the IMM password to default - Fixed the problem that BBU gets over temp during discharging on x3550 M3 - Corrected a problem where the BOFM address became permanent after BOFM is disabled on the AMM - Corrected a system hang when a Multi-Bit memory error occurs with a mixture of 2GB & 4GB DIMMs installed - Fixed the problem where the reported DC Power consumed on the dx360 M3 is higher than actual AC input power - Fixed the MIB mismatch issue - Fixed the incorrect OID in MIB Enhancements: - Added support for 32-bit ActiveX Remote Presence Clients - Added the ability to change the SMTP port number - Changed the firmware to completely shut down the telnet sevice when the telnet max user count is set to 0 - Added support for LDAP Active Directory Application Mode - Added support for LDAP Authentication Only Mode - Changed the firmware to report a configuration error if the CPU installed on the EXA system is not a Westmere CPU - Changed the firmware to allow the x3690X5 to be powered on sooner - Changed the firmware to decrease the amount of time that it takes for the IMM to become ready after it is reset
Version 1.27, YU00A7C Enhancements: - Added support for the IBM BladeCenter HX5(MT: 1910,7873)
Version 1.26, YU00A6A Enhancements: - Added support for the System x3850/x3950 X5(MT: 7143,7191) and System x3690 X5(MT: 7147,7192)
Version 1.25, YU0091K Problem(s) Fixed: - Fixed a problem that hangs the server when F2 is pressed during POST to start Diagnostics - Fixed a problem in the HS22/HS22v that causes the IMM to report an unexpected throttle condition after the AMM is reset - Fixed a problem that causes the IMM lose its DHCP assigned IP address - Improved the monitoring of hard drive status when the server is powered up - Fixed a problem when using a Japanese keyboard with the remote presence function - Fixed a gradual fan speed reduction due to pressure sensor fluctuation on System x3690 X5, System x3550 M2/3, System x3650 M2/3 - Improved the response to an IO Hub over temperature condition on System x3690 X5, System x3650 M2/M3, and System x3550 M2/M3 - Fixed a problem that IBM Systems Director needs to re-discover IMM in order to receive events from IMM after IMM upgrade Enhancements: - Added support for a 460W power supply in the System x3620 M3 - Added support for a 675W high efficiency power supply in the System x3620 M3 and System x3630 M3
Version 1.24, YU0091E Problem(s) Fixed: - Fixed a problem where the IMM sometimes incorrectly reports a Pwr Rail fault or Pwr 240A_A/Pwr 240A_B fault - Fixed an issue where blade would hang on boot with ABR firmware corruption message - Fixed a problem when downgrading IMM firmware from version 1.22 to earlier levels. The IMM network fails - Fixed a problem that caused the firmware update using IMM Command Line Interface to fail on an IPv6 network - Fixed a problem that caused the IPv6 Link-local address not to be displayed after resetting the IMM to its defaults - Fixed a problem where the IMM fails to set SNMP configuration using ASU - Fixed a problem where IMM cannot set the value 10 for the OSWatchdog value using ASU - Fixed a problem where IMM Remote Media would not re-initialize after network error - Fixed problem if KCS error bit status may not be clear which may cause KCS communication stop - Fixed a problem that IMM web interface hang with Remote Control session opened - Fixed a problem that blade unexpected powered on after power off command Enhancements: - Added IPv6 configuration support via IMM SNMP interface - Added support for BladeCenter PCIe Expansion 4 on BladeCenter HX5 - Added support for BladeCenter GPU Expansion on BladeCenter HS22 - Added support for LDAP sub-tree searching with Novell eDirectory 8.7 - Added support for SNMP trap definitions - Updated fan response and logic for Fusion IO PCIe adapter - Updated ambient temperature sensor calculation
Version 1.22, YUOO84C Problem(s) Fixed: - Fixed a problem where the IMM sometimes incorrectly reports a Processor 2 (CPU 2) internal error - Fixed a problem where after powering down an HS22, the HS22 powered on without user request - Fixed a problem while Log in to IMM might fail if the Host name contains a dash - Fixed a problem where the IMM makes some certificate management settings protected that should not be protected - Fixed a problem where users can't get the full details of an expansion card using the IPMI FRU command - Fixed a problem where IMM would hang after an LDAP login using some German or French characters - Fixed a problem with the IMM SMTP client not conforming to RFC821 - Fixed a problem where the email alert setting won't accept a hyphen - Fixed a problem with the Windows key in the IMM Remote Control feature - Fixed a problem where IMM login via LDAP authentication fails to acquire Supervisor authority via Group membership with an OpenLDAP server. - Fixed a problem where a user with Remote Server Power/Restart Access authority can't change the power off delay - Fixed a problem that prevented a Processor Throttled event from being logged for the x3200 M3 and x3250 M3 models - Fixed a problem that where a user's permission is always set from the LDAP account not the local account profile - Fixed a problem where the CPU2 VCore/CPU2 VDIMM sensors become unreadable after updating the IMM in the x3500 M3 model - Fixed problem for DHCP may get time out if using "Shared NIC" mode but without any LAN cable to be plugged into "Sys Mgmt" port Enhancement: - Added IPv6 support - Added Dynamic DNS support (IPv4 only) - Added support for BladeCenter PCIe Expansion 4 on BladeCenter HS22 - Added BCH-R3 support
Version 1.21, YUOO75V Enhancement: - Added support for BladeCenter Max 5 Memory Expansion Unit (MEU) - Added feature to report SMI Lane failure for BladeCenter HX5
Version 1.20, YUOO75T Enhancement: - Added support for Max 5 Memory Expansion Unit (MEU)for SystemX 3690 X5 - Added support for Max 5 Memory Expansion Unit (MEU)for BladeCenter HX5 - Added feature to show pending firmware level during firmware update process - Added feature to block power on when multi-node complex has firmware mismatch Problem(s) Fixed: - Fixed DASD LED behavior for drives 8-32 - Fixed fan over cooling on SystemX 3690 X5 - Fixed Flash failure due to packet loss - Fixed Power Cycle events not appearing in SystemX 3690 SEL - Fixed DIMM throttle behavior in acoustic mode for BladeCenter HX5 - Fixed CPU type checking
Version 1.19, YUOO75X Enhancement: - Added support for Max 5 Memory Expansion Unit (MEU)for SystemX 3850/3950 X5 - Added support for 130W CPUs in High Performance Blades - Added virtual NIC VPD support - Changed the default value of SerialExitCLIKeySequence to ^[( - Added TPM Physical Presence Sensor support - Added AC metering and capping to System x3850/3950 X5 Problem(s) Fixed: - Fixed server time out authority in web interface - Fixed non-recoverable events that appeared as warning events in web event log - Fixed Server State "unknown" status in web interface - Prevented multiple log entries from being created for the same event - Fixed IPMI multi-session handling when sessions fail - Fixed IMM flash failed at 0 percent update progress error - Fixed Internal Error 255 during FPGA update - Fixed CPU Overtemp messages on FPGA update. - Removed unnecessary events for Power Unit sensors - Fixed failure to log VRD error on voltage fault - Fixed Power Redundancy Sensor handling for 220V configurations - Fixed CPU configuration check for System x3850/3950 X5 - Fixed IMM Network ARP - Fixed flash error "Image transfer timed out - error 15" - Corrected a problem that caused the system to not automatically power on after performing an Automatic BIOS Recovery (ABR)
Version 1.18, YUOO75J Enhancement: - Added support for the System x3630 M3
Version 1.17, YUOO75H Enhancement: - Added support for the System x3690 X5
Version 1.16, YUOO73O Enhancement: - Added support for the BladeCenter System HX5
Version 1.14, YUOO73M Problem(s) Fixed: - Corrected the MIB file to address a compiling problem - Fixed a problem when a second CPU might not been recognized by the IMM, when cycling power on the System x3550 - Fixed a problem where web session may hang after staying on the System Status web page for a while - Disabled the interface to the shared network port when the dedicated systems management network interface is selected. - Fixed a web page pop-up error message that may occur after upgrading the IMM from 48d and 57f - Fixed power/thermal calibration error in the x3620 M3 - Fixed a problem where the IMM may log a CPU Automatically Throttled event or a Redundancy Lost for a cooling zone event in the log after cycling power - Fixed a problem where the IMM doesn't send an ambient temperature alert or IO Hub temperature alert in the x3400 M3 and x3500 M3 - Corrected a remote presence problem that occurs while mounting an ISO image if the remote client does not have any removable devices. Enhancement: - Added a fan control enhancement for performance mode - Added support for additional types of power supplies - Improved the accuracy of power readings on 3550 M3, 3650 M3, and 3620 M3
Version 1.14, YUOO73K (released for multinode System x3850 X5 and x3950 X5 only) Problem(s) Fixed: - Fixed a problem that caused the IMM not to respond to some IPMI commands - Corrected a problem with LDAP authentication when role-based security is enabled - Corrected a problem which could cause a false CPU error to occur - Prevented false automatic boot failure recovery events from being reported Enhancements: - Added multinode support for System x3850 X5 and System x3950 X5 - Added support to activate FPGA updates with a warm boot - Added support for AC power metering and AC power capping on single node System x3850 X5 and System x3950 X5 - Added support to improve permit early power on support - Add support to report the version information for updates that have been flashed but not yet activated - Added support to report an FPGA firmware mismatch
Version 1.12, YUOO67C Problems(s) Fixed: - Adjust the Planar VBAT voltage threshold in HS22 Enhancement: - Added support for the System x3620 M3 - Added support for DC power supplies in the System x3550 M3 and System x3650 M3
Version 1.11, YUOO56O Enhancements: - Improved ABR handling to prevent invalid ABR behavior
Version 1.10, YUOO57J (released for HS22 only) Problem(s) Fixes: - Fixed a problem where the IMM sometimes incorrectly reports a Processor 1 (CPU 1) internal error
Version 1.10, YUOO57H Problem(s) Fixed: - Made changes to improve the robustness of firmware update operations. - Corrected an LDAP authentication problem. - Corrected a problem where SNMP settings could be lost after updating the firmware. - Enable ssh by default. - Permit the Adavnced Setting Utility to enable or disable ssh. - Added support for logging a message when the blade CPU is not compatible with the chassis' blower capacity - Permit the realtime power management controller firmware to be flashed to an earlier level. - Corrected a problem that prevented the remote presence function from working with network address translation - Fixed an issue where blades would throttle/hang when chassis power supply is removed Enhancement: - Added support for the System x3550 M3, System x3650 M3, System x3500 M3, System x3400 M3, System x iDataPlex dx360 M3, and the HS22V blade - Enhanced power measurement and capping capabilities in support of xSmartEnergy control - Increased support for the number of DIMMs from 12 to 18
Version 1.09, YUOO56K Enhancement: - Added support for the System x3850 X5 and the System x3950 X5
Version 1.08, YUOO48F Problem(s) Fixed: - Made changes to improve the robustness of firmware update operations. Limitations: Limitations for versions 1.07 and 1.08: If you are down-leveling the IMM firmware to a version older than 1.07 and the IMM firmware is currently at version 1.07 or 1.08, you must remove AC power from the server before performing a server reboot.
Version 1.07, YUOO48E Prerequisites: IMPORTANT: This firmware update contains changes to permit servers ========== to be powered on sooner after plugging them into AC power. Before installing UEFI with firmware that supports the early power on changes, be sure to update the IMM with this firmware version. UEFI Firmware with early power on support: BladeCenter HS22 UEFI Build P9E131A or newer System x3550 M2 UEFI Build D6E131C or newer System x3650 M2 UEFI Build D6E131C or newer System x3200 M3 UEFI Build GYE129A or newer System x3250 M3 UEFI Build GYE129A or newer System x3400 M2 UEFI Build Y4E131B or newer System x3500 M2 UEFI Build Y4E131B or newer System x iDataplex dx360 M2 UEFI Build TME132A or newer Problem(s) Fixed: - Corrected an Ethernet bug that caused network packets to be dropped - Changed the firmware to improve server performance when the Ethernet over USB interface is enabled - Corrected an intermittent problem where some of the IMM interfaces would not function after flashing the IMM with new firmware - Corrected a problem where LDAP users in nested groups are not permitted access to the IMM - Corrected a problem where the fault LED is not lit on the HS22 server when a hard drive is removed - Fixed a problem where the IMM sometimes incorrectly reports faults when the server is power cycled (for example, a loss of redundancy or a critical fan fault, or a processor is operating in a Degraded State, etc...) - Corrected a problem where the server would sometimes hang at checkpoint 'E8' - Changed the firmware to light the Information LED if the primary UEFI bank becomes corrupted - Corrected a problem that prevented the Remote Disk on Card feature from properly working - Corrected a problem that prevented an HS22 server with a full configuration from powering on - Changed the firmware to send processor information to the Advanced Management Module for display - Corrected a problem where faults for fan numbers 2 and 4 for the System X3200 M3 were swapped - Corrected a problem where the IMM firmware could corrupt itself over time Enhancements: - Changed the firmware to permit servers to be powered on sooner after plugging them into AC power - Improved the performance when displaying pages on the web interface - Added support to permit management traffic over the daughter cards in the HS22 server
Version 1.06, YUOO39B Enhancement: - Added support for Machine Types 4251, 4252, 4261, 7327, 7328
Version 1.05, YUOO32F Limitations: Problem(s) Fixed: - This update package contains updated Management Information Base files for use when managing the server through the IMM's SNMP interface - Fixed a problem where user accounts and settings could become lost - Improved performance when changing settings through the web interface - Corrected a problem where some configuration settings were lost after flashing the IMM with new firmware - Improved the wording of messages in the Chassis Event Log - Restored the ability to the IMM SNMP interface to add or remove user accounts and to set the account security level - Corrected a problem that prevented some configuration settings from being set through the Advanced Setting Utility (ASU) - Fixed a problem that prevented a user from specifying an email alert recipient with a hyphen in the name - Fixed an IMM flash update problem where the IMM erroneously reports a CPU Configuration error and/or illuminates the fault LED on the HS22 Blade Server with no corresponding entry in the Advanced Management Module log - Fixed an IMM flash update issue which can result in blades never completing the Discovery state Enhancement: - Added support for the BladeCenter PCIe Expansion module 3 - Added support for servers running real-time Linux - Enhanced the automatic recovery of communication errors in BladeCenter - Added the ability to configure a hostname for SNMPV3 traps, rather than just an IP address
Version 1.04, YUOO24I Problem(s) Fixed: - This is a critical update for x3550 M2 and x3650 M2. It corrects a possible problem that can occur when adding options such as memory, hard drives, and Central Processing Units (CPUs), and makes the system may become unresponsive. In some cases the server cannot power on again, or network connectivity to the Integrated Management Module is not possible. The problem may also occur when rebooting the IMM while the system is powered on with certain hardware configurations. - Replaced the ibm_rndis_server_os.inf and device.cat in the IMM firmware update package for Windows. The new files are used to workaround a potential performance issue when the IMM's Ethernet over USB is enabled. Enhancement: - Added support for different hard drives - Added support that enables the Advanced Management Module to display CPU information for the HS22 blade.
Version 1.03, YUOO23C Problem(s) Fixed: - Removed the empty virtual drives that were presented to the server by default. Virtual drives are now not presented to the server unless media is present and mounted from the Remote Control client. Enhancement: - Added support for Machine Types 7836, 7837, and 7839
Version 1.02, YUOO19G Problem(s) Fixed: - Corrected a problem where a critical fan fault was erroneously logged when powering off the server
Version 1.01, YUOO19E Enhancement: - Added support for Machine Types 7946, 4198, 7947, 4199, 7321, 6380, 7323
Version 1.00, YUOO18D - Original Release - Supports Machine Types 7870, 1936
(C) Copyright International Business Machines Corporation 2008 - 2013. All rights reserved.
ibm_fw_imm_yuoog6c-1_44_anyos_noarch_chg
=============================================================================== Emulex 16GB FC Firmware Package ===============================================================================
Firmware Version: 1.1.43.3 (Release for IBM System x)
Incremental Interoperability: - VPD-W for Lancer
Problems Fixed: - Emulex FC Adapter does not auto-enable SAN boot mode when boot target configuration is made via IFM
Firmware Version: 1.1.21.9 (Previous release for IBM Rack)
Incremental Interoperability: - IBM Server machine types. Refer to IBM ServerProven for specifics
Problems Fixed: - Fixed FCP_CONF handling bug - No SRR sent after drop write data frame, I/O fails - Adapter does not invoke recovery after dropped read data frame - I/O throughput performance improvements - Link failover when connected to FC Tape - Cosmetic: Cleaned up miscellaneous error handling messages in debug log files - If a new device comes online and attempts to acquire an AL_PA after the link has already come up, the FC link appears to hang until the cable is pulled - Cleaned up link negotiation at 4GB/8GB/16GB speeds with specific switches
- Fixed issue where acceptable indistro device drivers may not pass minium driver check. BIOS Changes Enhancement(s): - Core BIOS and BIOS configuration utility will display "FF" as the PCI slot number when proper slot information is not available. - Increased the disk drive spin-up timeout value to accommodate disk drives having larger spin up times.
Fix(es): - Fixed an issue where the adapter order does not get written to NVM and "Updating Adapter Order" displays on every reboot. - Fixed issue where zero sized volume was showing "4. 0 GiB." - Fixed issue where devices, which are not attached to adapter 0, could not be promoted to positions in the boot order when they were selected in the configuration utility. - Fixed an issue where IOS Banner(POST) not showing SAS Discovery error when more than one HBA is connected to the host system. - Fixed an issue where the SMART status is not correctly displayed for bare drives. - Fixed an issue in SAS topology where a RAID disk is highlighted and attempted to set as preferred boot device or alternate boot device, the configuration utility displays a SAVE/DISCARD prompt. - Fixed an issue where the firmware was showing a large elapsed time when a format operation crosses midnight in the system clock.
Firmware Changes Enhancement(s): - Added support for 4GB DIMM - Added the ability for user the enable/disable of the drive write cache during the BIOS start up.
Fix(es): - Fixed an issue where a request sense command, which is executed after a SATA drive is in active state is moved to idle power condition state, has no 'Additional Sense' information set. - Fixed an issue where protected information error "Logical block reference tag check failed" is received when an IR volume is deleted. - Fixed an issue Self-Test Results log page where values for update parameter code, format and linking, and parameter length are not recorded. - Fixed an issue sending vital product data inquiry with vital product data page set as invalid (0xFF) while diagnostic is running on a SATA drive results in segmentation fault 0x1901. - Fixed an issue where input/output is sent to SATA drives before the drives are ready. - Fixed an issue where fault 8405 was being hit while disabling and enabling the phy to which secondary drive is connected with IOs active. - Fixed an issue where subsequent commands sent to a SATA drive fail after a write same command failure until NCQ error handling is triggered or the device is reset. - Fixed an issue where Report Luns command sent to SATA drive fails if the drive is spun down. - Fixed an issue where hotspare does not resynch with second raid volume when both volumes are in degraded state. - Fixed an issue where incorrect sense data is populated when Security Protocol IN/OUT command is sent to supported drive. - Fixed an issue where allocation length for inquiry SCSI Command is not handled Properly during SATA Drive Format. - Fixed an issue during SATA drive format where vital product data inquiry command is executed for SATA drive with invalid page code, and the command is passed with no data instead for failing with CC and sense data. - Fixed an issue where the host will not indicate the existence of the expander attached drives if the direct attached device re-ordering feature is enabled. - Fixed an issue where protected information errors are observed on a RAID 1 volume its BGI is not completed, and its member physical disks are removed and inserted. - Fixed an issue where the TM response was set to failed if the device is missing due to the DO NOT SEND TASK IU flag being enabled. - Fixed an issue where DMD (Device Missing Delay) timer is not set to the default time when the total Report Device Missing Delay is greater than 0x7F seconds. - Fixed an issue where RAID1 volume fails to resynch when a new compatible bare drive is available in the primary drive slot. - Fixed an issue where error handling for errors, that cause the DMA engines to halt, are not handled properly and multithreaded IOs can hang if the same control block is used too soon. - Fixed an issue where a third hotspare is created even if the number of hotspares currently present in the RAID Configuration is 2. - Fixed an issue where inactive foreign volume information cannot be seen in SASBIOS when a native volume is deleted.
UEFI Changes Enhancement(s) - Additional information such as vendor id and product id added to the HII Create Configuration Screen. - Added ability to check to detect missing hotspares to HII. - Modified to be obtained PCI Slot Number information from the system firmware protected information. - Additional character "M" will be displayed as part of the device display identifier of a multipathed disk. - Ability in HII to detect missing hotspares and provide a facility to delete the missing hotspares. - Implemented Configuration Summary and Simple Virtual Disk Creation Support for Unified Configuration Management. - Modified UEFI driver to wait for a maximum of 10 seconds for the controller to become ready, so controller gets discovered in the OS.
Fix(es): - Fixed an issue where Enable/Disable write cache does not work for SAS drives. - Fixed an issue where HII Configuration utility does not allow the user to create a RAID0 with single disk present in the topology. - Fixed an issue in Create Configuration menu where additional drive information is not getting updated appropriately in HELP area. - Modified to prevent Write Cache settings options from appearing for the IR firmware. - Fixed an issue where SATA volume members and hotspare drives are labeled as 'M' in a multipathed environment. - Fixed an issue to prevent HII from allowing the option to enable/disable write cache for SATA drives. - Fixed an issue where multipathing is disabled in Manufacturing page settings, but volume members and hotspare drives are labeled as "M." - UEFI driver modified to check BIOS Options in order to determine if is should manage the controller.
Fixes: - Fixed issue where HII do not report Multi-LUN devices from external RAID enclosure. - Do not show external RAID LUNs in HII RAID configuration page. - Report NVDATA version consistently in Hex or Decimal. - Fixed issue where update will not install with the 09.101 megaraid_sas device driver.
Note: HII changes may not affect all controllers. Changes for the controller are based on RAID volume support. - HII fix to return EFI_SUCCESS when RETRIEVE Action is given in HII Callback function. - HII fix for Write cache disabling fails for SATA HDD/SSD connected to the IT controller. - HII fix for saved Controller Events Filename. - Behvioral change for the way a volume is created in HII.
Note: The following fixes only effect blade systems that support Bofm. - Fixed boot issues in Windows 2008 where after installation in uEFI the os will not boot due to the BuildDevicePath not allocating enough memory for the DevicePathreturned to caller. - Fixed issue with Bofm configurations where the BOFM supplied addresses for external/switch attach Phys are being assigned to the internal/direct attach Phys. - Fixed a Bofm issue where the SAS WWIDs not reverting to Man Page 5 Default when a Blade is moved back to non-BOFM slot.
Note: HII changes may not affect all controllers. Changes for the controller are based on RAID volume support.
- HII change to Remove Legacy BIOS Enable/Disable option. - Fixed issue where after creating a volume using SATA SSD the Associated Physical Disks always show up as SATA HDD. - Fixed issue where a hotspare could not be created for any volume in the HII browser. - HII fix for the View Physical Disk Properties form where the Serial number of SATA disks were being displaced with a space. - HII fix for the IR help strings displayed when using an IT controller. - Fixed HII issue where a 'Question value mismatch with Option value' error occurs when saying No for creating RAID10/RAID1E volume. - Fixed HII issue where help is not available when having only RAID0 and RAID1 levels in Create Configuration form. - Fixed a problem where the UEFI driver ATA PassThru Protocol is corrupt while trying to access the Mode field. - HII behavior change for 'Manage Foreign Configurations.' - HII change to provide proper extended RAID Levels in help messages based on supported extended RAID Levels.
Version 12.13.0-0179 - Critical ------------------- Enhancement: - Implemented limited band charging for iBBU08. (RETAIN tip H21631) Fixes - Fixed an issue in multipath topology to use an alternate IO path if the current active path is removed. - Fixed an issue where drive states were marked as online before the creation of recontruction volumes completed. - Resolved controller hang condition during system bootup. - Fixed issue causing iBBU07 to always show "Charging" state (RETAIN tip H211025) - Fixed issue causing Virtual Disks to remain in Write-Through state after a relearn (RETAIN tip H211101) - Addressed data exposure issue when using CacheCade
Version 12.12.0-0133 - Suggested ------------------- Fixes: - Addressed an issue where battery over temperature messages could occur during charging. - Fixed issue in the HII broswer where the drive performing copyback will be stated as replacing and not copyback. - Fixed issue in the HII browser where the RPM for SATA drives show "0" - Fixed issue in HII browers where port 4 is shown on both ports 0 and 1 - Fixed HII browser issue where the user will get an error message when a global hot spare is created. - Fix HII browser issue where a RAID volume with greater then 3 spans will fail to create.
Version 12.12.0-0126 - Suggested ------------------- Fixes: - Fixed a issue where the RAID configuration can be lost when updating from firmware package 12.0.1-0097 or below. - Fixed an issue with 4-kilobyte (KB) sector disk support compatibility in NTFS by disabling 512E drive support mode.
Version 12.12.0-0111 - Suggested ------------------- Enhancement: - Implemented limited band charging for iBBU 07. - CacheCade 2.0 read and write caching on SSDs is supported on ServeRAID M5014 and ServeRAID M5015 with the ServeRAID M5000 Series Performance Accelerator Key. Fixes: - Set the compatibility bit to 1 so that patrol reads can be executed on SSD's - Fixed an issue where EEPROM init failures are not handled properly causing the user to see the following message on POST "The native configuration is no longer supported by the current controller and firmware." - Fixed an issue where the controller fails to read the header info from CFI Flash ROM. - Resolved issue switching to WT mode for recently failed batteries.
Version 12.12.0-0098 - Suggested ------------------- Fixes: - Fixed issue where the battery gauge is not accurate for aged batteries due to the discharge profiles of aged batteries not matching that of the new batteries. - Fixed a issue where a reboot would cause a hot spare to revert to 'Unconfigured Good' state when Performance Accelerator Key is attached. (RETAIN Tip H204925)
Version 12.12.0-0085 - Suggested ------------------- Enhancements: - Prevent non IBM firmware from being flashed on IBM ServeRAID branded controllers. - Enabled iBBU08 battery modes. - ibbu battery will no longer be marked as bad when very high capacity is detected. - Dimmer Switch function has been disabled by default.
Fixes: - Fixed a issue where firmware was taking to long to complete the PCI config read, causing the system to generate an NMI. - Fixed issue where enabling security causes the configuration to be missing, and a configured FDE array to be seen as foreign on reboot. - Removed BBU mode select from Web BIOS for all controllers. - Fixed issue where pulling VD 0 from a R60 and rebooting would cause two foreign configurations. - Fixed issue where WebBIOS gives unclear message when enabling FDE on an existing Disk Group. - Fixed issue where iBBU07 batteries display 65K capacity and are marked SOH bad. - Fixed issue where iBBU Absolute Charge is displayed differently - Fixed issue where the logical Unit ID (VPD page 0x83) changed after foreign import. - Fixed issue where the LDBBM entries are incorrect after foreign import. - Fixed issue where the controller does not work with pci set to nomsi in kdump kernel. - Fixed issue where the LED behavior is incorrect, wrong drive will blink activity. - Fixed issue where the controller cannot sustain at least 950 MBps performance with Cache Flush setting one sec. - Fixed issue where the set properties command fails - Fixed issue where the iBBU07 battery reported remaining capacity is much too high (64751 mAh) after a learn cycle. - Fixed issue where extended retries of write verify commands cause a driver timeout. - Fixed issue where the configuration becomes foreign after the firmware fails multiple attempts to access raid key during boot. - Fixed firmware issue where a replacement drive is not seen if the replacement is added during power cycle. - Fixed a issue where an alert should be sent when one phy connector between the SAS controller and SAS expander is degraded. - Enabled puncturing for SATA drives. - Enabled copy back on SMART. - Fixed an intermittent RAID key failure during power-on. - Fixed firmware issue where firmware would cause the controller to (sles11sp1/rhel6) reset when ‘lspci –vvx’ command is executed. - Change for iBBU-08 to support firmware Mode5 WB capacity from 960mAh to 674mAh for low power DDR-2 - Fixed a issue where the iBBU battery will not charge while the server continually resets causing the iBBU battery to over-charge.
Version 12.12.0-0065 - Critical ------------------- Fixes: - Fixed issue where the firmware takes too long to complete a PCI Config Read cycle causing SLES11 to kernel panic. - Fixed issue where MegaCLI may not be able to change drive or controller cache setting resulting in a return code of 0x03. - Added check to return SCSI sense Check condition if PMI is set and LBA is non Zero for SCSI Read Cap 16 - Fixed issue where a blue screen occurs handling RAID 1 with HSP and medium errors. The blue screen occurs because the driver times out while the firmware is retrying sense 0x02. Firmware causes Fatal Error when HSP drive responded with Sense 0x02 for Write/Verify or TUR - Change the boot with pinned cache RAID level support which was only allowed for raid level 0. The check is removed and is allowed for any raid level. - Fixed hang condition or condition where the uEFI driver does not load due to the uEFI Driver issuing a command in poll mode to firmware to early - Remove the following message "Battery cannot initiate transparent learn cycles" when the iBBU08 is under mode 4 or 5 - Fixed issue where BGI will not disable when a user issues megacli -LDBI -Dsbl -L0 -a0
Version 12.12.0-0056 - Critical ------------------- Fixes - Fixed issue where multi-bit ECC's are not reported during POST - Lowered the capacity level at which event notifications are generated for early warning and end of life alerts for iBBU to equal 48 hours
Version 12.12.0-0047 - Critical ------------------- Fixes: - Fixed issue where uEFI WebBIOS version may not display the product name of the second controller - Fixed issue to prevent users from downleveling firmware to older versions - Fixed an issue where firmware reports "Unexpected sense, Invalid field in CDB" for LOG SENSE command sent to Seagate drives - Fixed issue where SMART failures can take longer then 5 minutes to report or in some cases fail to report - Fixed issue where it appears to the user the harddrive cache policy is changed when requested but the change did not occur - Fixed issue where a remote iBBU08 will not be seen by the controller - Fixed data issue When IO is in progress to a RAID 1 array that contains media errors and is under a rebuild - Fixed issue where a drive is not marked as foreign during a drive move between servers causing the import of the configuration to fail - Fixed issue where the uEFI WebBIOS does not accept the + symbol in the key code or pass phrase - Fixed issue for iBBU08
Enhancement: - Enhancement to Reserve/release call for VCS Symantec configurations - Enhancements to iBBU event notifications, including the addition of early warning message. - Raised the capacity level at which event notifications are generated for early warning and end of life alerts for iBBUs
Version 12.12.0-0039 - Critical ------------------- Fixes: - Fixed intermittent initialization failure during POST with the ServeRAID M5025, M5015, and M5014 where the RAID BIOS could fail to load or the user will see one of the following error messages on POST:
RAID Adapter Unrecoverable Error!!! Please check the SDRAM connection.
or
Multi-bit ECC errors were detected on the RAID controller
Version 12.12.0-0037 - Critical ------------------- Enhancement: - Add support for iBBU08 (Battery Backup)
Fixes: - Addressed data issue while reconstructing from RAID1 to RAID0. - Addressed data issue after power lose to ServeRAID controllers. - Addressed data issue in some database applications. - Added fixes from LSI channel products.
Version 12.7.0-0020 - Critical ------------------- Enhancement: - Initial release for ServeRAID M5025 SAS/SATA Controller (46M0830)
Fixes: - Addressed data issue with database applications that use overlapping read write operating. - Addressed data issue which may occur during a reconstruction on some RAID levels that have an associated CacheCade virtual drive enabled. - Fixed issue where the controller may halt if a drive has medium errors that are being used in a check consistency, patrol read, or have delayed writes.
Version 12.0.1-0097 - Critical ------------------- Fixes: - Addresses an issue where data becomes inaccessible when drive is secured after data has been written . Version 12.0.1-0096 - Critical ------------------- Fixes: - Fixed an issue where the BBU may not be seen on POST. - Fixed an issue where the firmware doesn't follow the spinupdelay and maxtargetspinup values set in NVDATA causing spin up of all the drives a at the same time. - Fixed a very intermittent issue where the RAID key (M5000) authentication may not occur on boot leaving users without key options. - Addressed a very rare issue with the expander devices where the firmware megamon could occur if the sgpio configuration cable has errors.
Version 12.0.1-0090 - critical ------------------- Fixes: - Fixed issue where controller properties such as Rebuild, CC, and BGI are not changing when up and down arrow keys are used. - Fixed issue on x3950 M3 servers that can cause a system hang at reboot
Version 12.0.1-0084 - Non-critical ------------------- Enhancement: - Support added for the ServeRAID M5000 Series Advanced Feature Key.
Version 12.0.1-0064 ------------------- - Initial release ibm_fw_sraidmr_m5000-12_13_0-0179-1_linux_32-64_chg
Firmware Update for ServeRAID B5015 SSD Controller
- Fixed issue where the operating system's filesystem goes read only or adapter becomes non-responsive to management tools due to a resource leak in device flush path
- Addressed issue where a device timeouts may occur due to NCQ error handling did not confirm a request had completed successfully (including residual count) after a drive is removed and before using buffer contents. These contents include the failing tag, error status, and, in the case of an ECC error, a failing LBA.
- Fixed issue where the the controller may hangs due to a drive timeout during a hot drive pull.
- Fixed issue where a user may continuously pull, reinsert the drive and initiate rebuilds numerous times resulting in a read command to fail that should of completed.
ServeRAID M and MR10 Series SAS Controller Driver for Linux (also supports MegaRAID 8480 SAS Controller)
Adapters Supported: - ServeRAID M1215 SAS/SATA Controller for IBM System x (46C9114) - ServeRAID M5210e SAS/SATA Controller for IBM System x - ServeRAID M5210 SAS/SATA Controller - ServeRAID M5110 SAS/SATA Controller for IBM System x (81Y4481) - ServeRAID M5110e SAS/SATA Controller for IBM System x - IBM Flex System Storage Expansion Node - ServeRAID M5120 SAS/SATA Controller for IBM System x (81Y4478) - ServeRAID M5016 SAS/SATA Controller for IBM System x (90Y4304) - ServeRAID M5115 SAS/SATA Controller for IBM System x (90Y4390) - ServeRAID M5015 SAS/SATA controller (46M0829) - ServeRAID M5014 SAS/SATA controller (46M0916) - ServeRAID M1115 SAS/SATA Controller for IBM System X (81Y4448) - ServeRAID M1015 SAS/SATA Controller (46M0831) - IBM ServeRAID MR10il SAS controller (44E8767) - IBM ServeRAID MR10is SAS controller (44E8695) - IBM ServeRAID MR10ie (CIOv) SAS controller (46C7167) - IBM ServeRAID MR10i SAS controller (43W4296) - IBM ServeRAID MR10k SAS controller (43W4280) - IBM ServeRAID MR10M SAS controller (43W4339) - MegaRAID 8480 SAS controller (39R8850)
Version 6.602.03.00-1 - Suggested ------------------- Enhancement: - Add support for ServeRAID M1215 SAS/SATA Controller Fixes: - Fixed an issue where Kernel panic occurs due to timeout of MFI/MPT pass-through command, followed by completion through Interrupt context. Note: Device driver version is 6.602.03.00.
Version 6.600.20.00-4 - Suggested ------------------- Fixes: - Fixed reboot issue in SLES 11.3 xen. Note: Device driver version number has not changed. Version 6.600.20.00
Version 6.600.20.00 - Suggested ------------------- Enhancement(s): - Add Support for ServeRAID M5210 and M5210e SAS/SATA Controller
Version 6.506.00.00 - Suggested ------------------- Retain Tip: Tip #28016 Enhancement(s): - Added ability for user to configure the maximum number of MSI-X vectors launched for M51XX and M1XX controllers.
Version 00.00.06.19 - Suggested ------------------- Fixes: - Fixed an issue where a kernel crash may occur in some cases, when megaraid_sas driver is loaded but controller is not present.
Version 00.00.06.15.01 - Suggested ------------------- Fixes: - Corrected version strings in the iso and rpm files. NOTE: No change to the device driver. Packaging fix only.
Version 00.00.06.15 - Suggested ------------------- Enhancements: - Add support for fpRead/WriteCapable & fpRead/WriteAcrossStripe - Add support for SLES 11 SP2.
Fixes: - Mask off MFI_FRAME_IEEE, MFI_FRAME_SGL64, MFI_FRAME_SENSE64 in megasas_mgmt_fw_ioctl() to prevent memory scribble issues from incorrectly built MegaCli frames. - Fixed reglockFlags for degraded RAID 5/6.
Version 00.00.06.12-01 - Suggested ------------------- Fix - Fixed issue where the device driver may fail to install with the following error: "389 SCRIPT INTERNAL INCONISISTENCY! REPORT TO IBM IMMEDIATELY" Note: No change to device driver version.
Version 00.00.06.12 - Suggested ------------------- Enhancement - Add support for ServeRAID M5016
Fixes - LSI Channel fixes
Version 00.00.04.38 - Non-critical ------------------- - Fixed issue where device driver could print message "kernel megasas: Command pool empty!" when the controller and megaraid driver are idle.
Version 00.00.04.37 - Suggested ------------------- - Fixed issue where the ServeRAID M1015 controller will halt during a OCR. - Fixed issue where OCR will not execute when the physical drives are in a power transition mode. - Fix issue where the ServeRAID M1015 controller will halt with a degraded RAID 5 array during IO.
Version 00.00.04.30 - Non-critical ------------------- - Added support for the ServeRAID M5025 SAS/SATA Controller
Fixes: - Fixed issue where during IO firmware faults at line : MonTask: line 1193 in file ../../raid/1078dma.c - Fixed issue where After 64 VD's are created using MSM the 'fdisk -l' command does not work. - Fixed issue where the firmware faults at line 739 in file ../../raid/1078main.c when rebooting after a recon. - Fixed issue where the drive partition is no longer present on the OS after importing a foreign configuration. - Fixed issue where Linux does not show the correct size of VD the second time if the 'fdisk -l' command is executed.
Version 00.00.04.17 - Non-critical ------------------- - Added support for the ServeRAID M1015 SAS/SATA controller. - Added support for the ServeRAID M5000 Series Advanced Feature Key.
Version 00.00.04.10 - Non-critical ------------------- - Added logic to support next-gen SAS2 controllers. - Added support for the ServeRAID M5015 and M5014 SAS/SATA controllers.
Version 00.00.03.23 - Non-critical ------------------- - Added support for the IBM ServeRAID MR10ie (CIOv) SAS controller. - Fixed an issue where, on a Xen kernel with guest OSs running, invoking kdump causes a system reboot. - Fixed an issue that results in a kernel panic when kdump is run on a system with heavy drive I/O.
Version 00.00.03.21 - Suggested ------------------- - Added support for the IBM ServeRAID MR10is SAS controller. - Fixed an issue where a command timeout would occur on a particular SAS hard drive, and would not recover. - Added a shutdown DCMD call to the shutdown routine to ensure a graceful shutdown. - Updated the list of required pre-built binaries. - Fixed an issue the could cause unexpected interrupts. - Fixed an issue where a memory sanity check would sometimes fail for a certain pointer on a 64-bit system, causing some enclosure components to not be seen.
Version 00.00.03.18 - Suggested ------------------- - Added support for the IBM ServeRAID MR10il SAS controller. - Fixed an issue where the OS would not reboot when a rebuild was in progress.
Version 00.00.03.13 - Non-critical ------------------- - Added support the IBM ServeRAID MR10i, MR10k, and MR10M SAS controllers. - Fixed an issue where, on boot to Xen kernel on Red Hat Enterprise Linux (RHEL) 5 64-bit, the user would encounter an “Error 13: Invalid or unsupported executable format” error.
Version 00.00.03.06 - Suggested ------------------- - Added diskdump support for Red Hat Enterprise Linux (RHEL) 3 and RHEL 4. - Set the reply_q_sz value to the maximum number of commands supported by the firmware. - On 2.4 kernels: fixed an issue where, when the system was rebooted, the shutdown routine did not get called; added a notifier for the reboot event and initiated a flush command from the reboot event handler. - Fixed issue where, on RHEL 3 Update 6, the driver would not install on AMD64 systems.
Version 00.00.03.01-1 --------------------- - Initial release ibm_dd_sraidmr_6_602_03_00-1_sles11_32-64_chg
=============================================================================== Emulex FC/FCoE Device Driver for Linux ===============================================================================
Driver Version: 8.3.7.29 (supported on IBM System x, BladeCenter and Flex)
Problems Fixed: - Execution of internal loopback causes a crash - Issue which can cause multipath failover to not work properly
Incremental Interoperability: - IBM Flex System FC5052 2-port 16Gb FC Adapter - IBM Flex System FC5054 4-port 16Gb FC Adapter - IBM Tape unit interoperability - see SSIC for more information - SLES 11.3 on IBM System x and BladeCenter
Driver Vers: 8.3.7.18 (Supported for IBM System x & BladeCenter)
Problems Fixed: - FLOGI sent with Class 2 in-order bit set when Class 2 is set as unsupported class of service - OXID Reused before RA_TOV without RRQ to recover - Intermittently storage paths will not recover after switch reboot
Driver Vers: 8.3.6.4 (Previously supported on IBM Flex)
Problems Fixed: - Fixed debug helper routine failed to dump CQ and EQ entries in non-MSI-X mode - FLOGI sent with Class 2 in-order bit set when Class 2 is set as unsupported class of service - Fixed boot from san failure when SLI4 FC device presented on the same PCI bus - Fix ABTS recovery to be FC spec compliant - Fixed not reporting logical link speed to SCSI midlayer when QoS not enabled - Fix performance problem in pt2pt mode - Fixed SCSI host create showing wrong link speed on SLI3 HBA ports - Update text for misconfigured port messages - Fix discovery bug when swapping cables
Interoperability since previous release: - RHEL6.3 GA Kernel
Driver Vers: 8.3.5.86 (Previously supported on IBM SystemX & BladeCenter)
Problems Fixed: - Kernel panics with FCoE on HS23 in SLES 11.2 x86_64. Also requires a SLES11.2 kernel update - FLOGI sent with Class 2 in-order bit set when Class 2 is set as unsupported class of service
Interoperability since previous release: - RHEL6.3 GA Kernel
Driver Vers: 8.3.5.74 (Previously supported on IBM SystemX, BladeCenter & Flex)
Problems Fixed:
- Fixed system crash due to not providing SCSI error-handling host reset handler - Debug helper utility routines for dumping various SLI4 queues - Fix unsol abts xri lookup - Fixed issues with LPe16000 to LPe16000 discovery - Fix driver crash during back-to-back ramp events - Reregister VPI for SLI3 after cable moved to new Saturn port - Fix log message 2597 displayed when no error is detected - Fixed FCP LOG for easier Finisar trace correlation - Fix kernel panic when going into to sleep state - Fix error message displayed even when not an error - Fix Read Link status data - Fix unable to create vports on FCoE SLI4 adapter - Fix initiator sending flogi after acking flogi from target - Fix the case where an array controller does not apply FCP priority rules - Added support for handling SLI4-port XRI resource-provisioning profile change - Change default DA_ID support from disabled to enabled - Fix bug with driver unload leaving a scsi host for a vport around - Update copyright date for files modified in 2012 - Fix compilation errors on some kernels due to debugfs variable definitions - Used PCI configure space read to flush PCI function reset register write - Fixed system panic when extents enabled with large number of small blocks - Fixed system panic during EEH recovery - Fix resource leak when acc fails for received plogi - Fix SLI4 driver module load and unload test in a loop crashes the system - Fix missing CVL event causing round-robin FCF failover process to stop - Fix deadlock during adapter offline request - Fix same RPI registered multiple times after HBA reset
Driver vers: 8.3.5.66 (Previously supported on IBM SystemX, BladeCenter & Flex)
Problems Fixed:
- Fix SLI4 driver module load and unload test in a loop crashes the system - Fix missing CVL event causing round-robin FCF failover process to stop - Fix deadlock during adapter offline request - Fix same RPI registered multiple times after HBA reset - Fix driver handling of XRI Aborted CQE response - Fixed port and system failure in handling SLI4 FC port function reset - Fix bug with driver not sending a LOGO with vport delete - Make BA_ACC work on a fully qualified exchange - Fix for SLI4 Port delivery for BLS ABORT ACC - Fix KERNEL allocation while lock held - Incorrect usage of bghm for BlockGuard errors - Fix warning on i386 system - Fix ndlp nodelist not empty wait timeout during driver unloading - Fix mailbox and vpi memory leaks - Support character device to take a reference on the driver - Fix for FDISC failures after firmware reset or link bounce - Support debugfs for injecting T10 DIF errors - Fix SLI4 BlockGuard behavior when protection data is generated by HBA - Fix for driver using duplicate RPIs after 16Gb/s adapter port reset - Fix discovery problem when in pt2pt - Fixed failure in handling large CQ/EQ identifiers in an IOV environment - Fix Locking code raises IRQ twice - Fix driver not returning when bad ndlp found in abts error event handling - Support more driver logs in area of SLI4 port error attention and reset recovery - Fix bug with driver returning the wrong ndlp - Sync up with actual upstream lpfc driver code - Support SLI4 16Gb FC Adapter support for T10 DIF / BlockGuard - Merge from upstream: scsi: Fix up files implicitly depending on module.h inclusion - Fix driver behavior when receiving an ADISC - Fixed unbounded firmware revision string from port caused the system panic - Fix bug with driver processing dump command type 4 using 16Gb FC Adapter - Fix driver does not reset port when reset is needed during fw_dump - Fix ELS FDISC failing with local reject / invalid RPI - Changed SLI4 FC port internal loopback from outer serdes to inner internal - Fix REG_RPI fails on SLI4 HBA putting NPort into NPR state - Fix bug with driver processing an els command using 16Gb FC Adapter - Fix NMI seen due to CQE starvation - Fixed SLI4 FC port obtained link type and number dependent on link connection - Fixed SLI4 FC port internal loopback without SFP and external link/loopback plug - Fix driver incorrectly building fcpCdb during scsi command prep - Fixed FCP EQ memory check prevent initialize to single interrupt vector case - Fixed SLI4 16Gbit FC port internal and external diagnostic loopback tests - Fixed system panic during EEH recovery on SLI4 FC port - Fix FLOGI fails with invalid RPI (0x04) after connecting to a arbitrated loop - Fix ELS SCR failing with local reject / invalid RPI - Fix SLI4 port failures after firmware reset or dump - Fix port reset log messages indicate error when no error is seen - Fixed an issue where HBQ buffers are uninitialized in the base driver - Fixed missing shost lock protection on setting/clearing bit in vport->fc_flag - Fixed missing hbalock protection on setting/clearing bit in phba->link_flag - Fixed illegal state transition during driver unload - Fixed system crash on SLI4 FC port looopback diagnostic test with large buffers - Fixed a potential error path that might try to free DMA memory not allocated - Fixed SLI4 FC port Internal and External loopback diagnostic tests - Fix a crash while deleting 256 vports - Fix vport never finished discovery if switch runs out of resources - Fix kernel build warnings - Fix kernel panic when putting board in offline state twice - Fix memory leak when running hba resets to SLI-4 board - Fix driver message 3092 and 3116 to KERN_WARNING - Fix PCI read failure looks like a over temperature event - Fixed driver event reporting to mgmt on error attention due to forced dump - Fix not displaying firmware rev for BE UCNA boards - Fixed ABTS failure logic not rediscovering targets - Fix bug with driver not byte swapping extended mailbox data - Fix premature ndlp cleanup after second target LOGO - Fix Link pull causes I/O failures - Support loop for sli4 fc pass-through - No longer read fcoe parameters if board is not fcoe - Fix bug with driver passing a wrong count value - Support restore state and error log when sysfs board_mode attribute access failed - Fixed a typo in two of the log messages just introduced - Support for SLI4_CONFIG non-embedded COMN_GET_CNTL_ADDL_ATTR pass-through - Support for SLI4 FC port persistent linkdown through config region 23 - Fix for ABTS. Do not free original IOCB whenever ABTS fails - Fixed Linux driver inconsistency in reporting FC host port-state through OCM - Return an error if the mbox sysfs is called - Fix firmware upgrade attempt on unsupported adapter - Fix incomplete message number 2889 - Fix system crash when ASIC fails to initialize - Fix virtual fabrics bit set in FLOGI incorrectly - Fix bug with driver not returning extended mailbox data - Fix unsupported link speed accepted by SLI4 port - Incremented minor management version to reflect additional "protocol" sysfs attribute - Modify WQ handling to use entry_repost - Moved link type definitions to apply to all driver streams - Support a sysfs attribute "protocol" to report SLI4 port link protocol type - Support SLI4 INTF_TYPE and SLI_FAMILY as sub-field to the fwrev sysfs attribute - Support structure to support T10 DIF in SLI4 - Fix performance slowed due to too many register accesses - Fixed a warning when converting from simple_strtoul to strict_strtoul - Support Initial code to inject T10 DIF errors - Fixed a possible driver crash when cfg_fcp_eq_count is zero - Fixed non-embedded COMMON_GET_CNTL_ATTR mbox command failed with status 0x4 - Fixed null pointer dereference after clearing FCP policy rules - Changed the way of SLI4 device detecting physical port name - Fixed FCP policy exchange management rule parsing - Support check to lpfc_decode_firmware_rev to handle releases that have an 'X' - Fixed RPI leaks in ELS protocol handling - Fix kernel crash during boot with SLI4 card installed - Fixed NPIV FDISC failure on SLI4 if-type 2 ports - Fixed driver failed to follow IP reset procedure in SLI4 error attention handler - Fix sysfs lists fabric name for port that has been disconnected from fabric - Support FCP policy exchange management - Changed the timeout to 300 seconds for SLI_CONFIG (0x9B) mailbox command - Fixed fcp underrun reporting - Fixed a driver handling IP reset issue when PCI read return error - Fixed casting problem in calculating tx_words statistics - Support proper error code return for libdfc calling API to make proper decisions - Fix sli4 mailbox status code. Management software has dependencies - Fix compiler warning due to uninitialized local variable - Fix bus reset handler fails with bad failure code - Made proper error code return for libdfc calling API to make proper decisions - Made error log include PCI BAR bitmap from kernel when enable PCI device failed - Fixed sysfs API mismatch for building against SLES11 SP2 beta kernel - Fixed scsi midlayer queuecommand API mismatch for SLES11 SP2 beta kernel - Fixed ctlreg write bug not checking key string and limiting bytes to write - Fix default adapter name for the OCe15100 - Fixed incomplete scsi messages displayed - Fix cable pull failure on interface type 2 SLI-4 adapters - Fix enable_bg config parameter for SLI4 - Fixed not able to perform firmware reset through sysfs board_mode attribute - Fixed not recovering SLI port in handling error attention with RN bit set - Fixed two crash cases when unsolicted ELS ECHO_CMD is received - Fix direct connect does not come up for SLI4 FC ports - Fixed long wait when firmware reset to a SLI port without required privilege - Fix request firmware support for little endian systems - Fix driver firmware to match new firmware image format - Fixed SLI4 device firmware reset with SR-IOV virtual functions - Fix SLI4 CT handling for sequences > 4K - Fixed handling of unsolicited frames for vports - Fixed handling of CVL for vports - Fixed crash when aborting els IOs - Fix up CT and oxid/rxid for unsol rcv frames - Fixed mailbox double free panic - Support fcf priority record selection for fcf failover - Fixed compiler warning for putting large amount of memory on stack - Fixed driver build not building in debugfs even if kernel CONFIG_DEBUG_FS=y is set - Fix issue where the FC/FCoE Async Receive CQE did not scale for 16G FC adapters - FLOGI payload has Multiple N_port_ID set when lpfc_enable_npiv clear - Removed unused argument passed to the lpfc_bsg_diag_mode_enter routine - Fixed new ASIC device-mgmt BSG pass-through failed multi-buffer fw download - Support sysfs logging when writeable parameters are changed - Used PCI config space capability for sysfs get PF supported SR-IOV number of VFs - Support wait for SLI port status register for readyness following firmware reset - iDiag added SLI4 device PCI BAR memory mapped reigster access methods - Consolidated duplicating macro definitions - Fixed failure to enable PCIe AER prevent driver initialize and attach to device - Modified log messages for T10-PI processing - iDiag fixed mailbox capture and dump overlapping bitmap macro definition elx_dd_fc_ibm1212-8_3_7_29-1_sles11_32-64_chg
=============================================== Change History =============================================== Document Name: Change History Document Status: IBM General Avalibility Document Version: 3.2.4 Latest Driver Version: 3.2.4.0 Document Updated: 11/11/2013 Product: Brocade FC HBA / FCoE CNA
=================================================
Please visit www.brocade.com/hba for full documentation, additional drivers, and management applications.
Version 3.2.4.0 New Features and Feature Enhancement: - Support for RHEL5.10 and RHEL6.5
Version 3.2.3.0 New Features and Feature Enhancement: - Support for patching firmware via boot image - Support for Linux PTP - Support for UEFI Health Check Protocol - Support for SLES11SP3
Version 3.2.1.0 New Features and Feature Enhancement: - Support for "IBM Flex System FC5024D 4-port 16Gb FC Adapter" PART_NUMBER "95Y2381" - Support for RHEL5.9 and RHEL6.4
Fixes and Enhancements: - Enhanced Linux driver data path to properly handle IPv6 packets. - Enhanced Linux driver to make sure user configured path tov value is persistent and is not changed due to change in dev_loss_tmo.
Version 3.2.0.0 New Features and Feature Enhancement: - For Brocade 8x5/4x5, I/O interrupt latency values are dynamically updated. Changes are effective for all the ports on the given adapter. - Addressed infrequent PSOD in VMware ESX with iSCSI or VLAN configured traffic in DCB enabled network. - Enhanced BR1860 Tape IO performance. - Improved BR1860 initialization on Xen kernel.
Fixes and Enhancements: - BB Credit Recovery - Forward Error Correction (FEC) - Diagnostic Port (D-Port) - LUN Level IO Latency/Stats - UCM (Unified Configuration Manager) - Windows Network Performance and scalability enhancement
Version 3.1.0.0 Key features and enhancements supported for the software release 3.1.0.0 include the following:
New Features: - D-port support between Brocade 16Gb HBAs and Brocade 16Gb FC switch platforms with FOS v7.0.1 and later releases - FC-AL support for direct attach to storage - Ability to assign minimum bandwidth to each vNIC - Queue Depth adjustment - FC QoS set by percentage - Performance enhancements in Ethernet drivers - Technology Preview support of 1860 HBA on Power PC architecture
Fixes or Enhancements: - Resolved an issue where HCM Agent may not start if BR1860 adapter is configured in NIC mode. - Improved stability with Hyper-V and Teaming configurations. - Resolved I/O errors when running traffic with Brocade HBAs(models 415, 425, 815, 825, and 804), if the server’s PCI max read request size is configured to be 2K or greater with PCI Gen2x8.
Version 3.0.0.0 Fixes and Enhancements: - Added support for Brocade 1860 AnyIO Fabric Adapter (User configurable port mode - FC or NIC or CNA) - Added Fibre Channel 16G Support - Enabled Fabric Assigned Address in Fibre Channel Fabric - Added PF based Virtualization – vHBA - Enhanced CLI, HCM GUI and CIM support in inbox drivers. - Added PF based Virtualization - vNIC - Added iSCSI over DCB Support - Added Multiple Tx Priority Queues Support - Added SNMP Support
Version 2.3.0.3 Fixes and Enhancements: - Enhanced SNIA API to report port speed of 4 and 8Gbps in SNIA format. - Enhanced XIV array interoperability that addresses a data integrity corner case. - Enhanced the transmit path handling in Linux Network Driver to avoid a corner case that could lead to transmit path stall. - Enhanced the Linux Adapter Software installer script to update the initrd with the correct name even when there are multiple copies of the OEM specific initrd(s).
Version 2.3.0.0 Enhancements: - Enabled HP DL120 server to support Boot over SAN with two Brocade HBAs installed. - Improved network driver CPU and performance optimization. - Enabled automated target and LUN discovery of enterprise HDS storage once LUN masking has occurred. - Improved Tivoli interoperability to meet device display requirements.
Version 2.2.0.1 Added capability to handle ADISC frames. Unhandled ADISC use can cause SAN boot failure when booting from IBM storage array.
Version 2.2.0.0 Adds support for bootcode updates to Bladecenter 10G Converged Network Adapters
Version 2.1.0.0 Adds support for bootcode updates to 10G Converged Network Adapters
Version 2.0.0.0 Initial release for 4/8G FC HBAs brcd_dd_fc_bfa-3_2_4_0_sles11_32-64_chg
ServeRAID M1000 Series SAS/SATA Controller Firmware Update
Adapters Supported: ServeRAID M1115 SAS/SATA Controller for IBM System X (81Y4448) ServeRAID M1015 SAS/SATA Controller (46M0831)
Version 20.11.1-0184 - Suggested ------------------- Fixes: - Updated SBR to avoid an errata that could cause correctable CPU errors on some systems. - Fixed an issue where NULL events may show in event log, preventing capture of controller event logs.
Version 20.11.1-0137 - Suggested ------------------- Enhancement(s): - Update recovery mode page (WriteRetryCount and recoveryTimeLimit) during discovery of a drive and during spinup recovery when sense 6/29 is received.
Fixes: - Fixed an issue where certain drives were reporting a temperature of 65 degrees Celsius. - Fixed an issue where systems with SSD HDD will pop up an error message after creating a virtual drive via HII on M1115. - Fixed an issue where the drive state is not modified when a recon virtual drive is created from the reboot path.
Version 20.10.1-0107 - Suggested ------------------- Fixes: - Fixed grammatical errors in WebBIOS. - Fixed an issue in WebBIOS Drive Group Definition screen where enclosures should not be selectable in the Drives column. - Fixed an issue in uEFI-WebBIOS where the white bars do not match up under VD properties page. - Fixed an HII browser issue where a drive performing a copyback displayed the state as "Replacing" but other utilities showed the state as copyback. - Fixed an HII browser display issue where SATA drives showed "0" for the value of RPM under drive properties. - Fixed a display issue where the HDD port number showed incorrectly in an HII browser. - Fixed an HII browser issues where creating a global hotspare gave the user an error message. - Fixed an HII browser issue when security is changed with local key management the changing security report current mode is disabled as external keu management. - Fixed an HII browser issue where the user got the following error message, "The request command has invalid arguments," when changing the security key report. - Fixed an HII browser issue where the user was unable to save the configuration after adding an extra span during array creation. - Fixed an HII browser issue where the return value was not correct when the browser called extract config with a result of null.
Version 20.10.1-0101 - Suggested ------------------- Fixes: - Fixed an issue where a system uefi error will occur when the bootTargetId is set to 0. - Fixed a problem where the firmware will allocate the sata bbl entry for the drive when moved from unsupported to supported. - Fixed a firmware issue where the TargetId for the recon LD should be set to its original LD. - Fixed a HII error message "Question Value mismatch with Option value!" after selecting factory defaults in HII. - Fixed an issue where the firmware fails to create the configuration and utilities fail to report the VD size after the reconstruction.
Version 20.10.1-0091-01 - Non-Critical ------------------- Fix - Fixed issue where update may fail to find all supported controllers in SLES 11 SP 2. Note not change in the firmware; it is the same as package 20.10.1-0091.
Version 20.10.1-0091 - Suggested ------------------- Enhancement: - Add support for ServeRAID M1115
Fixes: - Fixed issue where after starting a reconstruction on a RAID 0 the progress %, does not increase. - Fixed an issue where a new drive, inserted after a drive failure, goes to JBOD and not auto rebuild. - Fixed physical drive count issue that stated more than (16) devices already exists
Version 20.10.1-0069- Suggested ------------------- Enhancement: - ServeRAID M1015 Dimmer Switches are disabled by default.
Fixes: - Fixed issue where enabling security causes the configuration to be missing, and a configured FDE array to be seen as foreign on reboot. - Fixed issue where WebBIOS gives garbled message when enabing FDE on an existing Disk Group - Fixed issue where the logical Unit ID (VPD page 0x83) changed after foreign import. - Fixed issue where the LDBBM entries are incorrect after foreign import. - Fixed a problem where firmware was taking to long to complete the PCI config read. Causing the system to generate an NMI. - Fixed issue where the controller does not work with pci set to nomsi in kdump kernel. - Fixed issue where the LED behavior is incorrect, wrong drive will blink activity. - Fixed issue where the configuration becomes foreign after the firmware fails multiple attempts to access raid key during boot. - Fixed firmware issue where a replacement drive is not seen if the replacement is added during power cycle. - Fixed an intermittent RAID key failure during power-on.
Version 20.10.1-0052- Critical ------------------- Enhancement: - Add support to pass SMART errors for JBOD drives to MSM
Fixes: - Added check to return SCSI sense Check condition if PMI is set and LBA is non zero for SCSI Read Cap 16. - Fixed issue where a bluescreen occurs handling RAID 1 with HSP and medium errors. The bluescreen occurs becausethe driver timesout while fw is retrying sense 0x02. Firmware causes Fatal Error when HSP drive responded with Sense 0x02 for Write/Verify or TUR - Fixed hang condition or condition where the uEFI driver does not load due to the uEFI Driver issuing a command in poll mode to firmware to early
Version 20.10.1-0045 - Critical ------------------- Fix: - Fixed issue where firmware has incomplete writes during power fail and recovery
Version 20.10.1-0036 - Critical ------------------- Fixes: - Fixed issue where uEFI WebBIOS version may not display the product name of the second controller - Fixed issue to prevent users to downlevel firmware to older versions - Fixed an issue where firmware reports "Unexpected sense, Invalid field in CDB" for LOG SENSE command sent to Seagate drives - Fixed issue where SMART failures can take longer then 5 minutes to report or in some cases fail to report - Fixed issue where it appears to the user the harddrive cache policy is changed when requested but the change did not occur
Version 20.10.1-0022 - Critical ------------------- - Fixed issue where Windows 2008 R2 WS08 cannot be installed properly on ServeRAID M1015 SATA JBOD with a uEFI boot. - Added warning message when user tried to change more then 16 drives to JBOD. - Addressed a problem where direct attached drives are given incorrect slot numbers - Added fixes from LSI channel products. - Fixed issue where a single JBOD drive will not boot after a Windows 2008 R2 installation.
Version 20.5.1-0014 - Critical ------------------- - Addressed data issue with database applications that use overlapping read write operating. - Addressed data issue which may occur during a reconstruction on some RAID levels that have an associated CacheCade virtual drive enabled. - Fixed problem where the controller may halt if a drive has medium errors that are being used in a check consistency, patrol read, or have delayed writes.
Version 20.1.1-0059 - Critical ------------------- Fixes: - Fixed an issue where a physical drive that is changed to JBOD from powersave will become unconfigured good again - Addresses an issue where data becomes inaccessible when a drive is secured after data has been written
Version 20.1.1-0049 - Suggested ------------------- Enhancement - Support M1000 key for RAID 5 and SED support
Fixes: - Fixed an issue in M3950M3 where webbios would hang - Fixed an issue where the controller properties in WebBIOS such as Rebuild, CC and BGI are not changing when we use up and down arrow keys. - Fixed an issue where Enclosure ID's are not being shown in MSM
Version 20.1.1-0046 - Suggested ------------------- - Enabled support for the ServeRAID M1000 Series Advanced Feature Key.
Note: All Logical Drives must previously have the encryption enabled before data is written on to the SED drives. Please read Retain TIP H123152 for more details.
Version 20.1.1-0042 ------------------- - Initial release ibm_fw_sraidmr_1000-20_11_1-0184_linux_32-64_chg
 QLogic 10Gb Converged Network Adapter (NIC) for System x and Blades Release Notes October 5, 2013
This software license applies only to QLogic customers. QLogic Corporation. All rights reserved.
Table of Contents
1. Version 2. Changes 3. Fixes 4. Known Issues 5. Additional Information 6. Firmware (Embedded in CNA) 7. Trademarks 8. Notices 9. Contacting Support
1. Version
This is NIC driver release version 1.00.00.32, package version 1.00.00.32.a.
These release notes describe the changes, fixes, known issues, and release details that apply to the QLogic FCoE Adapter driver version for SLES 11 Linux Driver with Kernel 2.6.x.
2. Changes
The following changes have been made to the Converged Network Adapter Networking Driver Update for Linux Kernel 2.6.x:
* Enabled ethtool WoL support * Fixed ethtool loopback test hang * Fixed stat counter
Hardware Support
* QLE8142 PCI-Express to 10Gb FCoE Converged Network Adapter * QMI8142 PCI-Express to 10Gb FCoE Converged Network Adapter
OS Support
These drivers work with the following Linux kernel 2.6.x distributions
* Novell SuSE Linux Enterprise Server 11
Software Components
* Driver 1.00.00.32 for SLES 11
Industry Standards Compliance
N/A
3. Fixes and Enhancements:
Fixed issure where "FC link not coming up after cable pull, only when QLGE module has been loaded."
4. Known Issues
The following known issues may be observed in the Converged Network Adapter Networking Driver Update for Linux Kernel 2.6.x:
Driver Compilation ------------------
When trying to compile the driver by running make install, the following error may occur:
"Linux kernel source not configured - missing version.h"
To resolve this issue, create the version.h file by going to the Linux source tree and entering the following:
make include/linux/version.h
Jumbo Frames System Requirement -------------------------------
Memory allocation failures have been observed on Linux systems that are running Jumbo frames with 64MB of RAM or less. If you are using Jumbo frames, your system may require more than the advertised minimum requirement of 64MB of system memory.
Performance Degradation with Jumbo Frames -----------------------------------------
Degradation in throughput performance may be observed in some Jumbo frames environments. If this is observed, increasing the application's socket buffer size and/or increasing the /proc/sys /net/ipv4/tcp_*mem entry values may help. For more details, see the specific application manual and /usr/src/linux*/Documentation/networking/ip-sysctl.txt.
Multiple Interfaces on Same Ethernet Broadcast Network ------------------------------------------------------
Due to the default ARP behavior on Linux, it is not possible to have one system on two IP networks in the same Ethernet broadcast domain (non-partitioned switch) behave as expected. All Ethernet interfaces will respond to IP traffic for any IP address assigned to the system. This results in unbalanced receive traffic.
If you have multiple interfaces in a server, do either of the following:
* Turn on ARP filtering by entering the following:
echo 1 > /proc/sys/net/ipv4/conf/all/arp_filter
* Install the interfaces in separate broadcast domains: either in different switches or in a switch partitioned to VLANs.
UDP Stress Test Dropped Packet Issue ------------------------------------
Under small packets UDP stress test with 10GbE driver, the Linux system may drop UDP packets due to the fullness of socket buffers. You may want to change the driver's Flow Control variables to the minimum value for controlling packet reception. Or, you can increase the kernel's default buffer sizes for UDP by changing the values in /proc/sys/net/core/rmem_default and rmem_max.
Unplugging Network Cable While ethtool -p is Running ----------------------------------------------------
In kernel versions 2.5.50 and later (including 2.6 kernel), unplugging the network cable while ethtool -p is running will cause the system to become unresponsive to keyboard commands, except for CTRL+ALT+DEL. Restarting the system appears to be the only remedy.
Bonding Configuration ---------------------
While creating the bond, issue the following command:
modprobe -v bonding fail_over_mac=1
Then rebuild the initrd and reboot the system.
5. Additional Information
N/A
6. Firmware
N/A (The Ethernet driver does not have firmware embedded.)
7. Trademarks
QLogic, the QLogic logo, QLA, SANbox, and SANsurfer are registered trademarks of QLogic Corporation.
SANtrack and SAN Express are trademarks of QLogic Corporation.
All other brand and product names are trademarks or registered trademarks of their respective owners.
8. Notices
Information furnished in this document is believed to be accurate and reliable. However, QLogic Corporation assumes no responsibility for its use, nor for any infringements of patents or other rights of third parties which may result from its use. QLogic Corporation reserves the right to change product specifications at any time without notice. Applications described in this document for any of these products are only for illustrative purposes. QLogic Corporation makes no representation nor warranty that such applications are suitable for the specified use without further testing or modification. QLogic Corporation assumes no responsibility for any errors that may appear in this document.
9. Contacting Support
Please feel free to contact your QLogic approved reseller or QLogic Technical Support at any phase of integration for assistance. QLogic Technical Support can be reached by the following methods:
(c) Copyright 2013. All rights reserved worldwide. QLogic, the QLogic logo, and the Powered by QLogic logo are registered trademarks of QLogic Corporation. All other brand and product names are trademarks or registered trademarks of their respective owners. qlgc_dd_nic_qlge-1_00_00_32_a_sles11_32-64_chg
======================================================= Change History for Qlogic Converged Network Adapter STOR FCoE Driver ======================================================= Document Name: Change History Document Status: IBM General Avalibility Document Version: 8.3 Latest Package Version: 8.06.00.10.a Latest Driver Version: 8.06.00.10 Document Updated: 02/15/2014 Product: Qlogic FCoE Linux Device Driver
Please visit www.qlogic.com/hba for full documentation, additional drivers, and management applications.
Version 8.06.00.10 Fixes and Ehancements:
Software/Driver
* Added T10 DiF/DiX support for ISP83xx.
* ISP83xx CNA support * IDC support for three protocols. * Loopback support. * Beacon support.
* Added support for bidirectional commands
* Added support for MCTP dump
* ISP83xx(16G FC) support for Xenserver 6.0 and Oracle UEK.
* SRB Consolidation.
* RSCN handling changes.
* Increase in support of maximum fibre devices.
* Enhancements to support ISP83xx (Limited support). * Driver features supported * 10GB link support only * Port device discovery * LUN discovery * I/O * Perturbations * SG resets of host and SCSI devices.
* Driver features not supported * Beacon support * Loopback * T10 DIX/DIF * SFP read/write * QCC GUI Support * QCC CLI Support * NPIV * Multi-Queue * Boot from SAN * FC-Tape * SLES10.3 driver: * Fix not to MPI reset on ISP83xx is not included.
* Implemeted beacon on/off for ISP82xx.
* Add support for ISP82xx to capture dump (minidump) on failure.
* Provide method for updating I2C attached VPD.
* Resync with latest ISP24xx/25xx firmware (5.08.00).
Industry Standards Compliance ----------------------------- * N/A
3. Bug Fixes
The following fix has been made to the Fibre Channel and FCoE Adapter Inbox Driver Update for Linux Kernel 2.6.x and 3.x:
* 8.04.00.12.xx.y-k * Update the copyright information. * Display the lock owner on lock acquire failure. * Move loopback mode reset after chip reset check. * Extra loopback error handling for ISP83xx. * Do link initialization on get loop id failure. * Do MPI reset only for 81xx.
V2632 (RHEL 6.x, SLES 11.x, Citrix 6.0, UEK): * Revert "qla2xxx: Optimize existing port name server query matching." * Revert "qla2xxx: Avoid losing any fc ports when loop id's are exhausted." * Don't free pool that wasn't allocated. * Unload hangs after issuing BSG commands to vport.
Citrix 6.0: * Create initrd as a part of installation and removal of the driver rpm. * RPM install should install driver under extra instead of driver directory.
RHEL 5.x, SLES 10.x: * MBA_IDC_NOTIFY case should fallthru. * Add fw_dump_size sysfs attribute file. * Add test for validity of virtual host in netlink interface.
SLES 10.x: * Add 8031/2031 cases to request firmware.
* 8.04.00.11.xx.y-k * Move marking fcport online ahead of setting iiDMA speed. * Add 16Gb/s case to get port speed capability. * Get VPD information from common location for CNA. * Dont clear drv active on iospace config failure. * Display that driver is operating in legacy interrupt mode. * Add "16" speed to link speeds array in qla2x00_iidma_fcport() function * Correct race in loop_state assignment during reset handling. * Reject loopback request if one is already in progress. * Update the FTP site references in the driver sources.
V2632 (RHEL 6.x, SLES 11.x Citrix 6.0 and UEK): * Obtain loopback iteration count from bsg request. * Parameterize the link speed string conversion function. * Update ql2xextended_error_logging parameter description with new options. * Fix typo in qla2xxx driver. * Free rsp_data even on error in qla2x00_process_loopback(). * Avoid null pointer dereference in shutdown routine.
RHEL 5.x: * Remove fabric_logout() calls from qla2x00_dev_loss_tmo_callbk() function.
SLES10.x: * Clear local references of rport on device loss timeout notification from FC transport.
* 8.04.00.10.xx.y-k * Honor status value of 2 for report-id acquisition. * Ignore driver ack bit if corresponding presence bit is not set. * Add Gen3 PCIe speed 8GT/s to the log message. * Acquire risc semaphore before resetting ISP.
V2632 (RHEL 6.x, SLES 11.x Citrix 6.0 and UEK): * Ramp down queue depth for attached SCSI devices when driver resources are low. * Determine the number of outstanding commands based on available resources. * Fix for kernel crash when external loopback diagnostic is run through BSG interface. * Fix coccinelle warnings in qla2x00_relogin. * Test and clear FCPORT_UPDATE_NEEDED atomically. * Use correct Request-Q-Out register during bidirectional request processing * Fix typo in qla83xx_fw_dump function. * Move noisy Start scsi failed messages to verbose logging level. * Fix for warnings reported by sparse.
RHEL 5.x: * Inclusion of 4G/8G firmware version 5.08.00. * Log message when qla2xenablemsix module parameter is set to 2( MSI mode) and adpater does not support MSI mode. * Revert "qla2xxx: Use GFF_ID to check FCP-SCSI FC4 type before logging into Nx_Ports" * Revert "qla2xxx: No fcport FC-4 type assignment in GA_NXT response."
Citrix 6.x: * RPM build should hide the directory naming convention from user.
* 8.04.00.09.xx.y-k * No fcport FC-4 type assignment in GA_NXT response. * Fix checkpatch error in qla2xxx_mctp_dump. * Change in setting UNLOADING flag and FC vports logout sequence while unloading qla2xxx driver. * Clear unsupported 'states' during Get-FW-State queries.
V2632 (RHEL 6.x, SLES 11.x Citrix 6.0 and UEK): * Pass correct parameters while creating an initrd. * Remove spurious taking of ha->vport_slock spinlock.
UEK: * Add macro for kernel extended version.
RHEL 5.x: * Use GFF_ID to check FCP-SCSI FC4 type before logging into Nx_Ports. * Pass correct parameters while creating an initrd.
8.04.00.08 Initial release of Qlogic 10Gb FC Driver for RHEL5
qlgc_dd_fc_qla2xxx-8_06_00_10_a_sles11_32-64_chg
ServeRAID m5016 SAS/SATA Controller Firmware Update for Windows
Adapters Supported: ServeRAID M5016 SAS/SATA Controller for IBM System x (90Y4304)
NOTE TO SERVICE - Reference RETAIN #N/A
Version 23.22.0-0024 - Non-Critical ------------------- Enhancement(s): - Display the Native Command Queuing (NCQ) enabled/disabled status of every SATA disk drive in the output of StorCLI and MegaCLI. Fix(es): - Disable PI CRC checking when the drive is not initialized during a CC. - Increased the SuperCap learn time out value for 2GB TMM-B solution. - Fixed an issue for 2GB memory modules where SuperCap does not get connected to the controller and hence cache offload fails. - Fixed an issue where cache policy of virtual drives(VD) does not change to write through(WT) mode even after detecting a bad supercap. - Fixed an issue where a RAID configuration created with physical disks in powersave mode could be losted after rebooting the system. - Fixed an issue where an invalid frame sent during a PIO-mode SATA passthrough command results in resetting the drive. - Fixed an issue where the system is not allocating memory correctly when MR and iMR controllers are in the same system. - Fixed an issue where deep learn does not start immediately when the relative state of charge (RSOC) is just equal to 95%. - Fixed an issue where performing a re-key operation in external key management (EKM) mode without doing a reboot, and then switching to local key management Mode(LKM) results in the Firmware logs showing that a drive security key was created but HII says the key is incorrect. - Fixed an issue where on a local key management (LKM) enabled controller, the user is asked to enter the security key to unlock and import a foreign drive, which is locked and has external key management (EKM) enabled.
Version 23.16.0-0027 - Critical ------------------- Fix(es): - Addressed data exposure issue when using CacheCade - Fixed issue where Supercap shows 0 Capacitance - Fixed iBBU08 and iBBU09 battery issue that caused a VD cache policy to remain in write-through mode after a manual relearn - Fixed iBBU08 and iBBU09 battery issue that caused a VD cache policy to periodically toggle between write-back/write-through mode - The limits for iBBU08 and iBBU09 battery charging were not updating properly in a learn cycle - Ensures a single supercap pack is not supported on 2GB memory module. - Resolves a firmware crash seen while doing continuous sas cable pulls. - Ensures OpROM loads even if segment address is 0xC0000. - Fixed an issue where the adapter is running iMR machine checking after requesting host memory. - Fixed an issue where HII does not allow the user to make use of the remaining free capacity to configure additional VD for R50/R10. - Fixed an issue where an HII: empty warning prompt is given when creating a CacheCade VD with forced writeback option. - Fixed and issue where Windows cannot install without a VGA device on board. - Fixed and issue in EFI WebBIOS were consistency check is still shown in progress after 1 PD is removed from a R1 array.
Version 23.12.0-0011 - Suggested ------------------- Enhancement(s): - Changed BIOS check to determine if running in HAMR, in order to ensure nothing gets broken when user has an iButton which is not programmed. - Changed controller status for safe mode to show controller status as safe mode and disable debug message logging in serial port.
Fix(es): - Addressed an issue where deeply discharged Cache Vault Power Modules would be prematurely be marked as failed. - Fixed an issue where HII hangs when trying to enter configuration management while in safe mode. Changed the safe mode warning message as below: "Controller is running in safe mode. Safe mode is used for diagnostic purposes only. Limited operations such as Configuration Preview, Clear Configuration and View/Change Controller Properties are allowed in this mode. Other operations such as Create Configuration, Import and Clear Foreign Configurations are not allowed. To exit safe mode, correct the fault, and reboot your computer." - Fixed an issue where having more than three adapters results in a calculation error for the offset to store physical disk info for fourth adapter, due to BIOS not clearing upper byte of AX (that is AH which is containing some non zero value). - Fixed an issue where the Wrong spanned drive group size is displayed by WebBIOS (Ctrl-H).
Version 23.7.1-0010 - Suggested ------------------- Fixes: - Fix an issue related to the where the TSOC (EEPROM) is not able to read or write from the firmware which will result in the message "Incompatible secondary iButton present." - Fix a HII broswer issue where Cachecade virtual drive can be selected as a boot device. - Fix a HII broswer issue where the user is unable to clear the controller configuration when there are no drives attached. - Fix an issue where the BIOS doesn't display the RAID level and VD size correctly at BIOS POST.
Version 23.7.1-0009 - Suggested ------------------- Fixes: - Fixed an issue where SoH is bad after a learn cycle but iBBUI shows good status. - Fixed an issue where hi-temp warning was not being displayed.
Version 23.7.1-0004 - Suggested ------------------- Fixes: - Corrected differential signals for Phys 2 and 3 that caused slot numbers in simple swap configurations to be wrong. - Fixed an issue when “maintain PD Failed history” is enabled, a cable is removed and added drives states go to optimal instead of foreign unconfig bad. - Fixed an issue where a capacity change is reported after a rebuild completes on a cache. - Fixed a controller firmware hang when the recovery of a medium error takes to long. - Fixed an issue where the SATA BBM (Bad block manager) list was getting corrupted as the logic was not accounting for a boundary condition of zero. - Fixed an issue where FoD (Feature on Demond) Options are lost because of errors while accessing the TSOC - Fixed an issue where the controller firmware will crash and drives will go offline when a bad block entry is created during a check consistency operation. - Fixed an issue where drives marked as unconfigured bad (foreign) gets marked as unconfigured bad on reboot - Fixed problem where a drive member of a cache cade volume does not get marked as foreign when the drive is pulled with maintain PD history enabled. - Fixed an issue where during a reboot a former hotspare which was in a state of unconfigured good moves to the state hot spare automatically and copy back stops causing the rebuild to start again. - Fixed an issue where a controller fault occurs when a reboot happens while a physical device is in progress of a PD Erase.
Version 23.7.0-0029 - Suggested ------------------- Enhancements: - Added support for Cache Cade 2.1 - Added Support for PI Drives
Fixes: - Fixed an issue where the firmware reports device initialization completion status to be greater than 100% at startup. - Fix an issue where when a secure foreign drive was present, VD creation fails because firmware becomes unresponsive - Added a check for the presence of PD before writing the data arms of a protected LD during BGI. - Fixed Firmware asserts in CacheMemAllocForPrebootApps() when memory was allocated twice by BIOS. - Fixed an issue with USB Cacheoffload restore fails after multiple power cycles on ServeRAID M51xx controllers. - Fixed an issue where querying SMART / PFA Status returns ‘all zeros’ in Sense Data. - Fixed an issue where the import of a config from MR to iMR resulted in a firmware crash. - Fixed an issue where iMR firmware crashed while creating 8/16/32 KB stripe size Lds. - Fixed background initialization watermark conversion to PD LBA. - Fixed write protected R1E data arms during initialization. - Fixed an issue where MR firmware crashed while discarding pinned cache. - Fixed an issue where a dirty LD dependency information of SSD pd has stale entries. - Corrected an issue during system booting from pinned cache, firmware goes to loop message BootMsgDcmd - Fixed an issue where a BGI was aborted and restarted, elapsed time was not reset. - Fixed an issue where the start ref for a shared cache line was not an indicator of the offset into the stripe. - Fixed an issue where the firmware hangs while running CC on a rebuilt R1 VD with SBBM. - Fixed an issue where the firmware was wrongly adding puncture entries. - Fixed an issue where rebuild was ran twice on a target drive with 100% full SBBM table. - Fixed an issue where the firmware was generating powersave change event when DS3 was not enabled, due to this application are not parsing the event. - Fixed an issue where recon from R0 to R1 using 3 TB drive causes firmware to report excessive unexpected sense events. - Fixed an issue, in which the bbuMode would revert to the factory default value of 5, when automatic learn cycle was enabled. - Fixed an issue where cache offload firmware reports unnecessary High Temp Event due to incorrect temp value. - Fixed an issue in which a Patrol Read may not start under certain conditions. - Fixed an issue where the firmware hits montask after discarding the cache when a snapshot VD was involved. - Fixed an issue where the firmware hangs during pin cache destage as dirty lines in that map never become zero. - Fixed an issue where CC fails to correct media error at 2TB location on 2 3TB SATA drives that support WU. - Fixed an issue where a new XML file was added to the nvdata, the buffer to hold the uncompressed nvdata overflowed. - Fixed an issue where an imported drive becomes unconfigured good when secure VD erase was running. - Fixed an issue where CC stops progressing on a R6 VD with media errors. - Fixed an issue where 3TB Hitachi HDD failed in immunity test (Electrostatic Noise test). - Fixed an issue where controller features are lost if RAID Key was not detected - Fixed an issue where performance degraded in RAID 5 Sequential Writes due to command Out-of Ordering. - Fixed an issue where ‘Native configuration is no longer supported’ message in post. - Fixed an issue where (DDR3) Initialization fails with fault 0x91a1 - initial DFIRL selected too low. - Fixed an issue where subsequent drives are dropped when all drives in the stagger spin up drives fails on boot sequence.
Version 23.2.1-0043 - Suggested ------------------- Fixes: - Fixed an issue where "Potential non-optimal configuration due to drive PD commissioned as emergency spare" alarm would trigger erroneously. - Fixed an issue where Copyback does not progress upon inserting drives. - Fixed an issue When allowMixInLD is Disabled, and causes Copyback Fails after Rebuild to Emergency Hot Spare. - Fixed an issue where the Firmware faults while flashing. - Fixed the Querying SMART / PFA Status to Return All Zeros in Sense Data. - Fixed an issue where the Re-inserted Drives becomes Unconfigured_good causing auto rebuilds to fail to start.
Version 23.2.1-0021-01 Non-Critical ------------------- - Packaging fix only to resolve flash failure. Note: Firmware version has not changed.
Version 23.2.1-0021 ------------------- - Initial release ibm_fw_sraidmr_5016-23_22_0-0024_linux_32-64_chg
IBM ServeRAID BR10il SAS/SATA Controller v2 BIOS and Firmware Update BIOS Version 6.30.02.00, Firmware Version 1.30.10.00 and BSD Version 3.16.00.06
Adapters Supported: IBM ServeRAID BR10il SAS/SATA Controller v2 (49Y4731)
NOTE TO SERVICE - Reference RETAIN #N/A
NOTE: This change history is common between several of the IBM and LSI Basic or Integrated RAID SAS controllers for non-BOFM enabled, uEFI systems. Not all levels were released for all controllers.
BIOS Fixes:: - Fixed an issue where the controller was hanging during initialization when an adapter was hung in reset with insufficient information to reset itself. - Fixed an issue where the SAS MPT BIOS CU reports wrong array size.
BDD Fixes - Removed the "SAS/SATA missing" warning message when SAS and SATA arrays are present on the same controller. Firmware Fix - Addressed issue where 6Gb SATA drives will only negotiate at 1.5Gb/s instead of 3Gb/s. - Fixed issue where SAS/SATA mixing errors will occur when the controller is setup to allow mixing of devices. - Fixed issue with large topology of SATA drives where media errors on a single drive will cause one phy to hang after IOs to other drives start to timeout. - Fixed discovery Error 0x2 in maximum topology with SATA devices, - Addressed issue where Credit would be stuck at zero until controller resets resulting in a phy becoming hung and IO failing. Only occurs with very high IO.
Fixes: - Addressed issue with SATA devices where the drive could go missing for several minutes. - Fixed issue where incorrect release of scatter gather list could cause issues with pending IO to missing devices. - Addressed issue where IOs to a host with a bad drive could fail.
Fixes: - Fixed issue where Target mode responds with incorrect WWID after ports are swapped. - Fixed issue with Incorrect/Unexpected discovery errors on RAID storage enclosures. - Fixed issue where device added events are not sent to the host on RAID storage enclosures. - Fixed issue where it is possible for a Incorrect/Duplicate Device handle to get assigned to a target that is a SSP Initiator. - Fixed problem of the boot order selection not being saved
Fixes: - Changed BIOS text to state an IS Volume creation can be 1 to 10 drives. - Fixed issue with 2TB and above drives where SAS MPT BIOS CU reports wrong array size. - Addressed an issue that will maintain Target IDs for attached devices to an enclosure if the enclosure is removed for a short time.
Fixes: - Fixed an issue where, after running heavy I/O to NCQ enabled SATA drives, the controller may log false data underruns for read or write I/Os. The PHY on which the I/Os were reported as underrun may eventually lock up.
Fixes: - Fixed an issue where Solid State Drives (SSDs) were not seen after a link reset. - Fixed an issue where direct attached SATA drives were not reported missing after they were removed. This could result in degraded or failed volumes being reported as healthy.
Fixes: - Fixed an issue introduced by new functionality in package 2.50. The issue would cause a dual attached device to be hidden from both the direct connection and the expander connection. - Changed the behavior during drive quiescing/un-quiescing to prevent the drive metadata from being unnecessarily updated. - Fixed an issue where drives might go offline during a firmware update, due to I/O timeouts caused by the length of time needed to update the firmware. - Fixed an issue where volumes can become inactive due to the metadata on the drives getting out of synch with NVSRAM.
Fixes: - Fixed an issue where, on certain drives, it was not possible to create a RAID 0. - Fixed an issue which could cause drive ID0 to go offline. - Fixed an issue with SSD hot spares being reported as Orphaned. - Fixed an issue where hot spares could not be created when the first physical disk of a roaming IME volume is missing. - Fixed a memory leak which could occur in verification of hot spare. - Fixed several issues involving creating a volume or hot spare immediately after deleting a volume or hot spare. - Fixed various issues with hot spare reporting. - Fixed an issue where, in a multi-path environment under heavy I/O, an open timeout can cause Linux file systems to become read-only. - Fixed an issue where drives may go offline during a firmware flash under Windows. - Fixed an issue that caused a 0x7600 fault while running heavy I/O to expander attached disks. - Fixed an issue where, when a driver at a certain volume ID is roamed, the volume ID would change on the next reboot. - Fixed an issue which caused a 0x266D fault on reinsertion of a hotspare. - Fixed an issue where a volume cannot be deleted due to foreign metadata from a new drive. - Fixed an issue where a 0x8055 fault could occur upon insertion of drives with deleted metadata. - Fixed a rare issue where drives could go missing under a heavy load. - Fixed an issue where a 0x800F fault may be seen at startup on large topologies. - Fixed an issue that could cause a 0x0B83 fault when a SATA drive failed to initialize. - Fixed an issue where the LED of a critical drive would revert to the pattern for an optimal drive before the drive was truly optimal again. - Fixed an issue where a resync would resume, rather than restart, when a primary drive was removed and re-inserted. - Fixed an issue that could result in 0x6001 faults while running I/O. - Removed some extraneous error messages, which were triggered by reboots or firmware flashes, from the firmware log. - Fixed an issue which cause a 0x266D fault during volume creation. - Fixed an issue which could cause the operating system's file system to become read only in a multi-path environment.
Functionality: - Initial release for: IBM 3Gb SAS HBA v2 - Added support for EFI (Extensible Firmware Interface) based systems. - Enhanced support for SSDs (Solid State Drives). - Implemented support for single drive RAID 0. - Added a new volume status 'Sync Pending' for a volume that could be resynching, but is waiting for the first volume to complete resync. - Added ability to remove an orphaned hot spare when its associated, missing, volume is deleted. - Implemented some LED blink pattern changes to account for new hardware. - Added the ability to enable/disable data scrubbing on SSDs. - Enabled functionality to update drive firmware on disks in a degraded volume (the update must be done via the IBM Online SAS/SATA HDD Update Program, which must also contain a fix to enable this function).
Fixes: - Fixed an issue where, with dual solid state disks (SSD) installed, hard drive 0 goes defunct or disappears or a RAID 1 mirror becomes degraded. - Fixed an issue which slightly degraded performance in a Large Drive Count (LDC) configuration. - Fixed an issue where SATA drives may fail when alternate paths are removed and reinserted in a multipath configuration. - Fixed a condition which caused a 0x1904/0x1903 fault when issuing a TARGET RESET to certain SATA drives. - Fixed an issue where a 0x0800 fault occurs when an inactive or non-optimal volume is deleted. - Changed a resource allocation so that topology and other discovery related events aren't affected when other resources are heavily stressed. - Fixed an issue where, if a volume is removed, then reinserted, into a SAS topology, the volume becomes inaccessible. - Fixed an issue where discovery appears to hang due to multiple Expander Removed events sent for a single expander. - Fixed an which causes an OS not to boot if one of the hard disks in the boot volume is removed. An error similar to "Adapter LSI RAID 0/1 SAS 4P (0) was 1: Disk (0) missing after reboot". - Fixed an issue where continuous 'device added' events would be sent during discovery when a controller is attached to a topology with more end devices than are supported. - Fixed an issue that could cause I/O timeouts or 0x0900/0x1600 faults with SATA drives under heavy I/O load during any memory deallocation. - Fixed an issue where multiple WRITE BUFFER commands may cause a 0x4202 fault. - Fixed an issue where a 0x1904/0x1903 fault may be seen when a duplicate I/O results in a double completion. - Prevented task management timeouts which occurred when the SATA engine was involved with multiple error handling. - Fixed an issue that could cause the Initiator Command queue to become corrupt, resulting in a 0x6001 fault. - Fixed an issue where some error information was lost during Record Not Found media errors. - Fixed an issue with the reporting/retrieval of some LDC statistics. - Fixed an issue which may cause firmware initialization failure when many (more than 50) SATA targets are attached. - Fixed an issue which, in rare cases, could cause a 0x266D fault. - Fixed an issue which could cause a 0x1904 fault when issuing task management requests during heavy IO. - Fixed an issue that caused a break of the MPI spec in a scenario where an end device fails SATA initialization or an expander is out of routing table entries. - Fixed an issue where a virtual PHY is counted multiple times, resulting in the same Bus/Target ID being assigned to two different devices. - Fixed a stack overflow issue which could result as a 0x0900 fault or a fault code 11h during system initialization. - Enabled modification of Write Cache settings for RAID 0 volumes. - Fixed an issue which caused an "Insufficient Resources" response for an MPT RAID Passthrough command. - Fixed an issue where SMP requests are sent through the SMP passthrough, a 0x4203 fault occurs. - Fixed an issue where drives with SMART errors are not detected as volume members. - Prevented a false "Device Not Responding" event during discovery under heavy I/O. - Fixed an issue where an Activity LED would erroneously blink when a raw hard drive was added to a previously occupied slot. - Fixed an issue which could result in media errors at 51% rebuild point when a hot spare is present. - Fixed I/O timeouts and/or invalid SAS Expander Page #0 content which could occur if a bad link caused rapid linkup/linkdown states on a drive. - Fixed an issue which could result in a 0xBADn fault (where n is the PHY number) after a HARD RESET primitive is received. - Corrected an issue where the swap out status of a drive is always set to 0, rather than to a meaningful value. - Fixed an issue where, if a card enters fail-safe mode due to unreadable or unreachable settings, a fault 0x0700 was seen and the card would never enter READY state. - Fixed an issue where a hotspare which was created through the BIOS would not be correctly reported until after a reboot. - Corrected an issue where media errors would cause unnecessary device resets. - Fixed an issue where, if a disk is hot swapped with an incompatible disk, the new disk would incorrectly be assigned the volume ID as its target ID. - Prevented multiple paths from being returned unless multipathing is specifically enabled in NVDATA. - Fixed an issue which results in a 0x6014 fault during discovery in complex topologies. - Corrected an issue where pulling a drive in a volume would caused a 0x801F fault. - Fixed an issue where a 0x6002 fault is encountered in LDC topologies. - Ensured the MPI_EVENT_IR2_RC_PD_REMOVED event is properly reported when a foreign disk is removed. - Fixed an issue which could result in a 0x0800 fault under heavy I/O. - Corrected an issue which results in 0x4E01 fault when processing multiple simultaneous NCQ errors. - Fixed an issue where, if four or more drives are configured as any mix of volumes or hot spares, when you delete the volume, the associated hot spare(s) are correctly deleted, but the physical drives don't reappear in BIOS until after a reboot. - Fixed an issue where, if the same LUN is issued two active resets, the second reset could be completed before the first reset. - Fixed an issue which causes SATA initialization failure if a WRITE command is processed concurrently with a RESET command. The volume would then change to a FAILED state. - Ensured that metadata size is automatically detected by the firmware. - Fixed an issue where the BIOS stops responding if a rebuilding hard drive is removed. - Fixed an issue where 0x7600 faults were encountered when running heavy I/O to cascaded x12 expanders. - Fixed an issue where, with a 1078 chip, a 0x266D fault may occur when hot spare is reinserted. - Corrected an issue where SAS or SATA drives may be reported missing after a target reset to an end device. - Fixed an issue where, if the drive at the volume ID slot is roamed, the volume ID might change on the next reboot. - Fixed an issue where there is a very low probability of experiencing data corruption while under very heavy I/O stress. - Reworked the firmware download function into an asynchronous operation to allow I/O to continue during the process. - Fixed an issue where a direct-attached SEP device will not be assigned an ID or handle which will prevent communication with the device. - Fixed an issue where a direct-attached SEP will be incorrectly assigned TargetID 0.
Functionality: - Implemented various changes to the LED blink pattern for SGPIO. - New functionality to ensure that, if a dual-ported device is connected via both direct and expander attached connections, only the direct connection is hidden when the HideNonzeroAttachedPhyIdentifiers bit is set.
Functionality: - Initial release for: IBM ServeRAID BR10i SAS controller ibm_fw_sraidbr_10il-v2-2_75_linux_32-64_chg
=============================================================================== Emulex NIC Device Driver for Linux ===============================================================================
Driver Version: 4.6.267.4 (supported on IBM System x, BladeCenter and Flex)
Problems Fixed: - Under certain conditions, performance issues may be seen with multi-channel in RHEL6.4 KVM configurations - Under certain conditions, performance issues may be seen with multi-channel in RHEL6.4 SR-IOV configurations - SR-IOV configuration section of the Linux NIC Driver Documentation has been modified to improve clarity
Incremental Interoperability: - S3 Resume on IBM Flex - SLES 11.3 on IBM System x and BladeCenter
Known Issues: - SRIOV is supported with the following Linux distributions: RHEL 6.3 and 6.4 64 bit RHEL 6.3 and 6.4 64 bit with KVM (guest support is dependent on the Emulex driver version present in the guest)
SLES11 SP2 and SP3 64 bit SLES11 SP2 and SP3 64 bit with KVM (guest support is dependent on the Emulex driver version present in the guest)
Use of SR-IOV with the XEN hypervisor is not supported at this time Use of SR-IOV with the KVM hypervisor requires the use of macvlan interfaces
Device Driver Version: 4.6.148.0 (Supported on IBM System x & BladeCenter)
Problems Fixed: - Ethernet performance issues in a variety of workloads - VFs are not displayed after driver re-load when is VFs attached to VM - vLAN tagging in Switch Agnostic vNIC mode
Incremental Interoperability: - RHEL 5.9 - RHEL 6.4 - Network performance varies with workload. An ethernet driver tuning script and instructions is now included with the be2net driver. - Switch Independent Mode within hypervisor OS's
Known Issues: - Multi-channel is not supported in RHEL5.x XEN. This is a kernel limitation - Under certain conditions, performance issues may be seen with multi-channel in RHEL6.4 KVM configurations - Under certain conditions, performance issues may be seen with multi-channel in RHEL6.4 SR-IOV configurations - SR-IOV is not supported with RHEL5.x XEN
Driver Version: 4.4.180.0 (Previously supported on IBM Flex) Driver Version: 4.2.389.0 (Previously supported on IBM SystemX & BladeCenter)
Problems Fixed: - Unloading the driver hangs the system after running IPv6 traffic - Link status messages cleaned up when in Switch Agnostic mode - Miscellaneous network performance issues to improve throughput - Miscellaneous problems related to SRIOV to improve usability - Ethtool wol command displays wol enabled even when disabled - Ethtool speed reported as 10000Mb/s for disabled devices in Switch Agnostic mode
Incremental Interoperability: - RHEL6.3 GA Kernel - Increase default B/W for SRIOV vf's to 10Gb from 100Mb
Known Issues: - SR-IOV configuration section of the Linux NIC Driver Documentation will be modified as follows to improve clarity (noted by << >>):
11. Use the Virtual Machine Manager GUI to attach the VF (step 7 on page 34) to the guest operating system by using the add physical device option.
<< Note 1: If user wants to re-configure an already configured setup, it can be done by removing the attached VF from the respective guest operating system using the remove option after selecting the VF attached. And then attaching another VF (step 11).
Note 2: Consult the documentation for your host operating system for help using the Virtual Machine Manager to attach and detach virtual interfaces. >>
12. Start the RHEL 6.x guest operating system. Once the guest operating system is booted, use the lspci command to confirm the visibility of the OneConnect NIC function. The output shows a OneConnect NIC function, for example: 03:00.0 Ethernet controller: Emulex Corp. Emulex OneConnect 10Gb NIC (be3)
13. The OneConnect Ethernet driver automatically loads with the inbox driver, and creates the network interfaces. Use the ifconfig command to verify that the interface is created.
14. After configuring the network interfaces with proper IP addresses, you can send and receive network traffic from the VM. << (Consult the documentation for your host and guest operating systems for network configuration help).>>
Driver Version: 4.1.413.0 (Previously supported on IBM SystemX & BladeCenter)
Problems Fixed:
- WOL changes for Linux NIC driver as mentioned in Bug# 24543 - SR IOV UE when attempt to do multiple (3 times) driver load/unload - Speed is 10000Mb/s for Disabled devices in UMC Mode - Incorrect Duplex value is shown for unconnected port of NC552m in Rhel6.1 32bit - SLES 11 SP2 GMC: Enable support for Lnx NIC Driver - SRIOV: When VGT is configured for a VF, ip link show doesn't display it for the VF configured. - RHEL 5.8 GA: Enable support for Lnx NIC Driver - ip link command with vlan = 4095 has toggling effect - ip link command with vlan =0 should remove guest VGT privilege
Driver Version: 4.1.334.39 (Previously supported on IBM SystemX & BladeCenter)
- SR-IOV: Can't load be2net driver in a VM after the VM has been rebooted - SR-IOV: Unload PF driver with VF driver loaded causes system instability - SR-IOV: VF MAC always gets Emulex OUI - SR-IOV: Partition the limited per port VLAN entries across VFs and PFs - Non Member VLAN packets not received in interface promiscuous mode in BE3 - Avoid CQ full condition in HW - be2net: use RX_FILTER cmd to program multicast addresses - account for skb allocation failures in ethtool stats - Support 60 second delay to Linux NIC driver to allow FAT dump completion on recovery from EEH - Link status is up [ethtool and OCM] when the bandwidth is assign 0 for UMC configuration - Remove msix module parameter - Use V1 version of COMMON_NTWK_LINK_STATUS_QUERY for Lancer. - Emulex: Ethtool failing to update FW - Emulex: Panic in be2net driver in SLES11-SP2-B7 on adapter - TCP ACKs from different connections are getting received in single queue - Multicast filtering feature not working on PPC - Exit of mcast promiscous mode not working - VLAN promiscous mode cannot be set on Lancer - Can't communicate after changing MAC address via VLAN interface - Improper memory accounting in received packets - WOL changes for Linux NIC driver as mentioned in bug - Transparent VLAN tagging feature not working in SR-IOV elx_dd_nic_ibm1212-4_6_267_4-3_sles11_32-64_chg
=============================================== Change History =============================================== Document Name: Change History Document Status: IBM General Avalibility Document Version: 3.2.4 Latest Driver Version: 3.2.4.0 Document Updated: 11/11/2013 Product: Brocade FC HBA / FCoE CNA
=================================================
Please visit www.brocade.com/hba for full documentation, additional drivers, and management applications.
Version 3.2.4.0 New Features and Feature Enhancement: - Support for RHEL5.10 and RHEL6.5
Version 3.2.3.0 New Features and Feature Enhancement: - Support for patching firmware via boot image - Support for Linux PTP - Support for UEFI Health Check Protocol - Support for SLES11SP3
Version 3.2.1.0 New Features and Feature Enhancement: - Support for "IBM Flex System FC5024D 4-port 16Gb FC Adapter" PART_NUMBER "95Y2381" - Support for RHEL5.9 and RHEL6.4
Fixes and Enhancements: - Enhanced Linux driver data path to properly handle IPv6 packets. - Enhanced Linux driver to make sure user configured path tov value is persistent and is not changed due to change in dev_loss_tmo.
Version 3.2.0.0 New Features and Feature Enhancement: - For Brocade 8x5/4x5, I/O interrupt latency values are dynamically updated. Changes are effective for all the ports on the given adapter. - Addressed infrequent PSOD in VMware ESX with iSCSI or VLAN configured traffic in DCB enabled network. - Enhanced BR1860 Tape IO performance. - Improved BR1860 initialization on Xen kernel.
Fixes and Enhancements: - BB Credit Recovery - Forward Error Correction (FEC) - Diagnostic Port (D-Port) - LUN Level IO Latency/Stats - UCM (Unified Configuration Manager) - Windows Network Performance and scalability enhancement
Version 3.1.0.0 Key features and enhancements supported for the software release 3.1.0.0 include the following:
New Features: - D-port support between Brocade 16Gb HBAs and Brocade 16Gb FC switch platforms with FOS v7.0.1 and later releases - FC-AL support for direct attach to storage - Ability to assign minimum bandwidth to each vNIC - Queue Depth adjustment - FC QoS set by percentage - Performance enhancements in Ethernet drivers - Technology Preview support of 1860 HBA on Power PC architecture
Fixes or Enhancements: - Resolved an issue where HCM Agent may not start if BR1860 adapter is configured in NIC mode. - Improved stability with Hyper-V and Teaming configurations. - Resolved I/O errors when running traffic with Brocade HBAs(models 415, 425, 815, 825, and 804), if the server’s PCI max read request size is configured to be 2K or greater with PCI Gen2x8.
Version 3.0.0.0 Fixes and Enhancements: - Added support for Brocade 1860 AnyIO Fabric Adapter (User configurable port mode - FC or NIC or CNA) - Added Fibre Channel 16G Support - Enabled Fabric Assigned Address in Fibre Channel Fabric - Added PF based Virtualization – vHBA - Enhanced CLI, HCM GUI and CIM support in inbox drivers. - Added PF based Virtualization - vNIC - Added iSCSI over DCB Support - Added Multiple Tx Priority Queues Support - Added SNMP Support
Version 2.3.0.3 Fixes and Enhancements: - Enhanced SNIA API to report port speed of 4 and 8Gbps in SNIA format. - Enhanced XIV array interoperability that addresses a data integrity corner case. - Enhanced the transmit path handling in Linux Network Driver to avoid a corner case that could lead to transmit path stall. - Enhanced the Linux Adapter Software installer script to update the initrd with the correct name even when there are multiple copies of the OEM specific initrd(s).
Version 2.3.0.0 Enhancements: - Enabled HP DL120 server to support Boot over SAN with two Brocade HBAs installed. - Improved network driver CPU and performance optimization. - Enabled automated target and LUN discovery of enterprise HDS storage once LUN masking has occurred. - Improved Tivoli interoperability to meet device display requirements.
Version 2.2.0.1 Added capability to handle ADISC frames. Unhandled ADISC use can cause SAN boot failure when booting from IBM storage array.
Version 2.2.0.0 Adds support for bootcode updates to Bladecenter 10G Converged Network Adapters
Version 2.1.0.0 Adds support for bootcode updates to 10G Converged Network Adapters
Version 2.0.0.0 Initial release for 4/8G FC HBAs brcd_dd_nic_bna-3_2_4_0_sles11_32-64_chg
IBM ServeRAID BR10i SAS Controller BIOS and Firmware Update BIOS Version 6.30.02.00, Firmware Version 1.30.10.00 and BSD Version 3.16.00.06
Adapters Supported: IBM ServeRAID BR10i SAS Controller (44E8689)
NOTE TO SERVICE - Reference RETAIN #N/A
NOTE: This change history is common between several of the IBM and LSI Basic or Integrated RAID SAS controllers for non-BOFM enabled, uEFI systems. Not all levels were released for all controllers.
BSD Fixes: - Fixed an issue where the controller was hanging during initialization when an adapter was hung in reset with insufficient information to reset itself. - Fixed an issue where the SAS MPT BIOS CU reports wrong array size.
BSD Fixes - Removed the "SAS/SATA missing" warning message when SAS and SATA arrays are present on the same controller. Firmware Fixes - Addressed issue where 6Gb SATA drives will only negotiate at 1.5Gb/s instead of 3Gb/s. - Fixed issue where SAS/SATA mixing errors will occur when the controller is setup to allow mixing of devices. - Fixed issue with large topology of SATA drives where media errors on a single drive will cause one phy to hang after IOs to other drives start to timeout. - Fixed discovery Error 0x2 in maximum topology with SATA devices, - Addressed issue where Credit would be stuck at zero until controller resets resulting in a phy becoming hung and IO failing. Only occurs with very high IO.
Fixes: - Addressed issue with SATA devices where the drive could go missing for several minutes. - Fixed issue where incorrect release of scatter gather list could cause issues with pending IO to missing devices. - Addressed issue where IOs to a host with a bad drive could fail.
Fixes: - Fixed issue where Target mode responds with incorrect WWID after ports are swapped. - Fixed issue with Incorrect/Unexpected discovery errors on RAID storage enclosures. - Fixed issue where device added events are not sent to the host on RAID storage enclosures. - Fixed issue where it is possible for a Incorrect/Duplicate Device handle to get assigned to a target that is a SSP Initiator. - Fixed problem of the boot order selection not being saved
Fixes: - Changed BIOS text to state an IS Volume creation can be 1 to 10 drives. - Fixed issue with 2TB and above drives where SAS MPT BIOS CU reports wrong array size. - Addressed an issue that will maintain Target IDs for attached devices to an enclosure if the enclosure is removed for a short time.
Fixes: - Fixed an issue where, after running heavy I/O to NCQ enabled SATA drives, the controller may log false data underruns for read or write I/Os. The PHY on which the I/Os were reported as underrun may eventually lock up.
Fixes: - Fixed an issue where Solid State Drives (SSDs) were not seen after a link reset. - Fixed an issue where direct attached SATA drives were not reported missing after they were removed. This could result in degraded or failed volumes being reported as healthy. - Fixed an issue introduced by new functionality in package 2.50. The issue would cause a dual attached device to be hidden from both the direct connection and the expander connection. - Changed the behavior during drive quiescing/un-quiescing to prevent the drive metadata from being unnecessarily updated. - Fixed an issue where drives might go offline during a firmware update, due to I/O timeouts caused by the length of time needed to update the firmware.
Fixes: - Fixed an issue where, on certain drives, it was not possible to create a RAID 0. - Fixed an issue which could cause drive ID0 to go offline. - Fixed an issue with SSD hot spares being reported as Orphaned. - Fixed an issue where hot spares could not be created when the first physical disk of a roaming IME volume is missing. - Fixed a memory leak which could occur in verification of hot spare. - Fixed several issues involving creating a volume or hot spare immediately after deleting a volume or hot spare. - Fixed various issues with hot spare reporting. - Fixed an issue where, in a multi-path environment under heavy I/O, an open timeout can cause Linux file systems to become read-only. - Fixed an issue where drives may go offline during a firmware flash under Windows. - Fixed an issue that caused a 0x7600 fault while running heavy I/O to expander attached disks. - Fixed an issue where, when a driver at a certain volume ID is roamed, the volume ID would change on the next reboot. - Fixed an issue which caused a 0x266D fault on reinsertion of a hotspare. - Fixed an issue where a volume cannot be deleted due to foreign metadata from a new drive. - Fixed an issue where a 0x8055 fault could occur upon insertion of drives with deleted metadata. - Fixed a rare issue where drives could go missing under a heavy load. - Fixed an issue where a 0x800F fault may be seen at startup on large topologies. - Fixed an issue that could cause a 0x0B83 fault when a SATA drive failed to initialize. - Fixed an issue where the LED of a critical drive would revert to the pattern for an optimal drive before the drive was truly optimal again. - Fixed an issue where a resync would resume, rather than restart, when a primary drive was removed and re-inserted. - Fixed an issue that could result in 0x6001 faults while running I/O. - Removed some extraneous error messages, which were triggered by reboots or firmware flashes, from the firmware log. - Fixed an issue which cause a 0x266D fault during volume creation.
Functionality: - Enhanced support for SSDs (Solid State Drives). - Implemented support for single drive RAID 0. - Added a new volume status 'Sync Pending' for a volume that could be resynching, but is waiting for the first volume to complete resync. - Added ability to remove an orphaned hot spare when its associated, missing, volume is deleted. - Implemented some LED blink pattern changes to account for new hardware. - Added the ability to enable/disable data scrubbing on SSDs. - Enabled functionality to update drive firmware on disks in a degraded volume (the update must be done via the IBM Online SAS/SATA HDD Update Program, which must also contain a fix to enable this function).
Fixes: - Fixed an issue which could cause the operating system's file system to become read only in a multi-path environment. - For the 1064 option for System x3200 M2, x3250 M2, or x3350, fixed an issue where the amber LED would not light up on a defunct drive.
Fixes: - Fixed an issue where, with dual solid state disks (SSD) installed, hard drive 0 goes defunct or disappears or a RAID 1 mirror becomes degraded. - Fixed an issue which slightly degraded performance in a Large Drive Count (LDC) configuration. - Fixed an issue where SATA drives may fail when alternate paths are removed and reinserted in a multipath configuration. - Fixed a condition which caused a 0x1904/0x1903 fault when issuing a TARGET RESET to certain SATA drives. - Fixed an issue where a 0x0800 fault occurs when an inactive or non-optimal volume is deleted. - Changed a resource allocation so that topology and other discovery related events aren't affected when other resources are heavily stressed. - Fixed an issue where, if a volume is removed, then reinserted, into a SAS topology, the volume becomes inaccessible. - Fixed an issue where discovery appears to hang due to multiple Expander Removed events sent for a single expander. - Fixed an which causes an OS not to boot if one of the hard disks in the boot volume is removed. An error similar to "Adapter LSI RAID 0/1 SAS 4P (0) was 1: Disk (0) missing after reboot". - Fixed an issue where continuous 'device added' events would be sent during discovery when a controller is attached to a topology with more end devices than are supported. - Fixed an issue that could cause I/O timeouts or 0x0900/0x1600 faults with SATA drives under heavy I/O load during any memory deallocation. - Fixed an issue where multiple WRITE BUFFER commands may cause a 0x4202 fault. - Fixed an issue where a 0x1904/0x1903 fault may be seen when a duplicate I/O results in a double completion. - Prevented task management timeouts which occurred when the SATA engine was involved with multiple error handling. - Fixed an issue that could cause the Initiator Command queue to become corrupt, resulting in a 0x6001 fault. - Fixed an issue where some error information was lost during Record Not Found media errors. - Fixed an issue with the reporting/retrieval of some LDC statistics. - Fixed an issue which may cause firmware initialization failure when many (more than 50) SATA targets are attached. - Fixed an issue which, in rare cases, could cause a 0x266D fault. - Fixed an issue which could cause a 0x1904 fault when issuing task management requests during heavy IO. - Fixed an issue that caused a break of the MPI spec in a scenario where an end device fails SATA initialization or an expander is out of routing table entries. - Fixed an issue where a virtual PHY is counted multiple times, resulting in the same Bus/Target ID being assigned to two different devices. - Fixed a stack overflow issue which could result as a 0x0900 fault or a fault code 11h during system initialization. - Enabled modification of Write Cache settings for RAID 0 volumes. - Fixed an issue which caused an "Insufficient Resources" response for an MPT RAID Passthrough command. - Fixed an issue where SMP requests are sent through the SMP passthrough, a 0x4203 fault occurs. - Fixed an issue where drives with SMART errors are not detected as volume members. - Prevented a false "Device Not Responding" event during discovery under heavy I/O. - Fixed an issue where an Activity LED would erroneously blink when a raw hard drive was added to a previously occupied slot. - Fixed an issue which could result in media errors at 51% rebuild point when a hot spare is present. - Fixed I/O timeouts and/or invalid SAS Expander Page #0 content which could occur if a bad link caused rapid linkup/linkdown states on a drive. - Fixed an issue which could result in a 0xBADn fault (where n is the PHY number) after a HARD RESET primitive is received. - Corrected an issue where the swap out status of a drive is always set to 0, rather than to a meaningful value. - Fixed an issue where, if a card enters fail-safe mode due to unreadable or unreachable settings, a fault 0x0700 was seen and the card would never enter READY state. - Fixed an issue where a hotspare which was created through the BIOS would not be correctly reported until after a reboot. - Corrected an issue where media errors would cause unnecessary device resets. - Fixed an issue where, if a disk is hot swapped with an incompatible disk, the new disk would incorrectly be assigned the volume ID as its target ID. - Prevented multiple paths from being returned unless multipathing is specifically enabled in NVDATA. - Fixed an issue which results in a 0x6014 fault during discovery in complex topologies. - Corrected an issue where pulling a drive in a volume would caused a 0x801F fault. - Fixed an issue where a 0x6002 fault is encountered in LDC topologies. - Ensured the MPI_EVENT_IR2_RC_PD_REMOVED event is properly reported when a foreign disk is removed. - Fixed an issue which could result in a 0x0800 fault under heavy I/O. - Corrected an issue which results in 0x4E01 fault when processing multiple simultaneous NCQ errors. - Fixed an issue where, if four or more drives are configured as any mix of volumes or hot spares, when you delete the volume, the associated hot spare(s) are correctly deleted, but the physical drives don't reappear in BIOS until after a reboot. - Fixed an issue where, if the same LUN is issued two active resets, the second reset could be completed before the first reset. - Fixed an issue which causes SATA initialization failure if a WRITE command is processed concurrently with a RESET command. The volume would then change to a FAILED state. - Ensured that metadata size is automatically detected by the firmware. - Fixed an issue where the BIOS stops responding if a rebuilding hard drive is removed. - Fixed an issue where 0x7600 faults were encountered when running heavy I/O to cascaded x12 expanders. - Fixed an issue where, with a 1078 chip, a 0x266D fault may occur when hot spare is reinserted. - Corrected an issue where SAS or SATA drives may be reported missing after a target reset to an end device. - Fixed an issue where, if the drive at the volume ID slot is roamed, the volume ID might change on the next reboot.
Functionality: - Implemented various changes to the LED blink pattern for SGPIO. - New functionality to ensure that, if a dual-ported device is connected via both direct and expander attached connections, only the direct connection is hidden when the HideNonzeroAttachedPhyIdentifiers bit is set.
Note: - The multi-path read only fix that was added in version 2.45 is not present in 2.50. It is reintroduced in version 2.51.
Fixes: - Fixed an issue where there is a very low probability of experiencing data corruption while under very heavy I/O stress. - Reworked the firmware download function into an asynchronous operation to allow I/O to continue during the process. - Fixed an issue where a direct-attached SEP device will not be assigned an ID or handle which will prevent communication with the device. - Fixed an issue where a direct-attached SEP will be incorrectly assigned TargetID 0.
Fixes: - Fixed an issue where, in a setup with cascaded expanders, a timeout occurred during PortEnable after an adapter initialization. - Increased several preset maximum values to allow all devices to be seen in large topologies. - Fixed an issue where SATA initialization would hang in a multiple initiator environment. - Fixed an issue that could result in a system hang when an active-active MUX that did not support NCQ was detected in the topology. - Fixed several issues which would cause doorbell faults when using SATA drives. Faults include 0x6001, 0x0B85, or 0x6212. - Fixed an issue where "Persistent Table Full" messages were seen even with persistent mapping disabled. - Fixed a SATA initialization error which could occur when I/O was sent to the drive shortly after the DEVICE ADD event. - Fixed a discovery error that was seen when rebooting initiators under heavy I/O in a multi-initiator environment. - Fixed an issue that could cause a 0x7A02 doorbell fault in a 1078-based complex topology environment. - Fixed an issue which caused doorbell fault 0x4203 when an expander is disconnected during discovery. - Fixed a multipathing issue where devices were not seen after a secondary path was removed and readded quickly. - Fixed an issue where, under heavy I/O, the message indicating discovery completion could be lost. - Fixed multiple issues that were specific to multipathing. - Fixed issues where a controller would hang after firmware or BIOS flash. - Fixed an issue that could cause loginfo 0x31123000 during heavy I/O. - Fixed various issues that could cause doorbell faults 0x4104, 0x0B85, 0x266D, 0x2100, 0x4202, or 0x7711. - Fixed an issue where SATA initialization would fail if the drive did not support write caching. - Fixed an issue where, when logs were generated for a passive path, the enclosure handle and slot information for the active path were stored in the logs. - Fixed an issue where all disks become inaccessible if a SATA logical loop is introduced into a topology with SAS drives. - Fixed an issue which caused BIOS timeouts which resulted in the boot device not being found. - Fixed an issue where, during certain MPT BIOS activity, the system BIOS could not read the MPT BIOS, resulting in a "no bootable device" error. - Fixed an issue where, if translated SCSI commands where sent to a SATA drive the I/O would hang. - Fixed an issue where a hard reset sent to one initiator in a multi-initiator environment could cause a cascade of hard resets through the environment. - Fixed an issue where a corrupted sector in a primary volume was not being correctly restored with the correct data from the secondary volume. - Fixed an issue where, if the secondary drive is cold swapped during a volume resync, the resync is not resumed on system start. - Fixed an issue where the driver temperature info is not sent for SCT enabled drives. - Fixed an issue where the second path is not being hidden when a non-volume drive is inserted in a volume ID slot. - Fixed an issue where, during initialization, the SAS BIOS displays 'Adapter Malfunctioning!' Fault 13. - Prevented Device Missing Delay replies when DevMissingTimeout and IoDelayTimeout are disabled. - Fixed and issue where, when inserting a new drive into the topology while I/O is running, I/O to existing drives may experience a delay of ~600msec. - Fixed an issue where the execution of a command in CFGGEN following execution of a command without a ‘noreset’ switch causes controller to hang. - Fixed an issue where firmware update of an online SATA drive would fail. - Fixed several issues where, when detecting whether a device is an HDD or SDD, the device type was detected incorrectly. - Fixed an issue where physical disk firmware download commands were being rejected. - Fixed an IOC hang that could be caused by a SCSI INQUIRY command. - Fixed an issue where, in a multipath environment, a degraded RAID volume with a missing disk is rebooted, it may re-sync to a non-RAID disk. - Fixed a SATA specific issue where a doorbell fault 0x1903 or 0x1904 was seen when hot-plugging an active drive. - Fixed a large topology issue where I/O to the SAS drives times out after a bus reset. - Fixed an issue where, when the last device in a volume is missing, the volume may be calculated as the wrong media type. This will prevent hotplug and failover from working correctly. - Fixed an issue where the volume ID changes after removing the drive at a roamed volume ID. - Fixed an issue where, when a volume's disk is swapped and starts to resync, a controller reset (reboot) may result in either the volume losing a disk that was previously in the volume, and/or the wrong disk reported as the primary. - Fixed an issue where discovery would fail when a 1078 controller is connected to a JBOD with dual-I2C SEP functionality.
Functionality: - Initial release for: IBM BladeCenter LS22 and LS42 onboard LSI 1064e controller - Fixed the ability for the BIOS configuration utility to create hotspares with SSD (Solid State Drive) devices. - Enhanced SATA I/O performance for NCQ drives. - Allowed the host to set the name of a RAID volume. - Added ability to retrieve SATA drive temperature through "Information Exception". - For SGPIO, turned on the fault LED when an IR volume member is detected as missing. - Added ability to specify bus/target ID & slot mapping for direct-attached disks. - Expanded the ability for the backplane to define the sideband interface method (I2C/SGPIO). This was previously enabled for 1078, now it is enabled for 1064 and 1068 as well. - Several changes the blink LED patterns. ibm_fw_sraidbr_10i-2_75_linux_32-64_chg
=============================================================================== Emulex iSCSI Device Driver for Linux ===============================================================================
Driver Version: 4.6.267.4 (supported on IBM System x, BladeCenter and Flex)
Problems Fixed: - RH5.9 kernel panic after rebooting the OS
Device Driver Version: 4.4.175.0 (Previously supported on IBM Flex)
Problems Fixed: - Cleaned up a cosmetic problem with warning messages during driver compile from source - System crashes when we run simultaneous IO to 6 vNIC’s with iSCSI IO - Linux kdump kernel crashes during shutdown after capturing dump, if be2iscsi driver running
Driver Version: 4.1.402.0 (Previously supported on IBM SystemX & BladeCenter)
Problems Fixed:
- LUN is not getting detected for OEM CNA in SLES 11 SP2 64 bit OS - ioctl semaphore deadlock when running load/unload loop with target - SLES 11 SP2 GMC: Enable support for Lnx iSCSI Driver - RHEL 5.8 GA: Enable support for Lnx iSCSI Driver
Driver Version 4.1.334.39 (Previously supported on IBM SystemX & BladeCenter)
Problems Fixed:
- Emulex: Warning messages seen while building Emulex drivers on RHEL platform - Unable to load be2iscsi driver during iSCSI boot OS install - Modprobe be2iscsi hangs when system with max targets login - be2iscsi unload hangs the system - fw rebooted on 1st mbox cmd after driver load > unload > load - Failed to initialize iscsi driver - need to define DMA_32BIT_MASK / DMA_64BIT_MASK in Linux kernel 3.x.x elx_dd_iscsi_ibm1212-4_6_267_4-3_sles11_32-64_chg
Firmware Version: 2.01a11 (w/ 5.12a13 Universal Boot Code) (Supported for IBM System x & BladeCenter)
Problems Fixed: - Emulex FC Adapter does not auto-enable SAN boot mode when boot target configuration is made via IFM (BOFM) - Emulex FC configuration utility is in the wrong F1 setup screen - IBM CIOv adapter connected to Brocade 8 Gig FC BladeCenter switch intermittently Links up at less than 8 Gig
Firmware Version 2.01a3 (w/ 5.12a5 Universal Boot Code)
New Features
- Enhanced timeout notifications for processing abort failures.
Problems Fixed:
- All multi-abort I/Os are returned on the same ring from which the ABORT_MXRI command was received. - Changed the TRCV response in the abort state to CMD RJT to enable the TRSP to be processed. - Corrected an issue with blocked TSEND operations. - Corrected a buffer leak issue with RRQ exchanges. - Corrected an issue where the TRCV IOCB was de-queued improperly causing a XFR_RDY flood. - Corrected an issue where the OXID field was not loaded in the XRI_ABORTED IOCB. - Corrected an issue in recognizing whether an EEPROM is present.
Dependencies:
- The corresponding FC Storport device driver and OneCommand Manager Core-kit (management) will need to be installed for this package to function properly.
1. Fixed an issues booting when more than 8 ports are present in the system and fixed an adapter display issue when more than 8 ports are present in the systems 2. Fixed issues installing and booting O/Ss on ports greater than 8. 3. Fixed some boot time screen display issues.
(2) OpenBoot version 3.10a3
This version of OpenBoot supports FC-AL, Private loop, Public loop, and Fabric Point to Point. Also, it supports multi-initiators and multi-LUNs (0 through 4055)
Problems fixed, features added or other updates:
1. Unrelated to IBM
(3) EFIBoot version 4.12a12
The current EFI (Extensible Firmware Interface) Boot provides 64-bit System boot capability through the use of the EFI Shell. This EFIBoot supports EFI Specification 1.10 and EFI Source Code version 1.10.14.61.
Problems fixed, features added or other updates:
1. Only certain OEM HBAs will locate PlatformToDriverConfiguration- ProtocolGuid. Others just skip and continue processing 2. UEFI2.3 Hii Configuration Utility changes 3. HII Browser Save option does not work
1. On a port connected to a 3PAR "MBOD" cage, the test runs IO to the disk drives and issues a "controlport offline" command to offline the port while IO is running.
2. While doing target side cable pull testing with Block Guard (T10) Enabled we observed that adapter is not sending RSCN to AIX.
Dependencies: - The corresponding FC Storport device driver and OneCommand Manager Core-kit (management) will need to be installed for this package to function properly. elx_fw_fc_ibm1212-2_01a11-11_linux_32-64_chg
Broadcom NetXtreme TG3 Linux Driver
Version 3.136e
Copyright (c) 2010 - 2014 Broadcom Corporation All rights reserved
tg3: Poll cpmu link state on APE + ASF enabled devices
On ASF enabled devices where the mgmt firmware runs on the application processing engine, there is a race between the tg3 driver processing a link change event and the ASF firmware clearing the link changed bit in the EMAC status register. This leads to link notifications to the driver sometimes getting lost.
Poll the CPMU link state as a backup for the normal interrupt path update if ASF is enabled.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Set the MAC clock to the fastest speed during boot code load
On the 5717, 5718 and 5719 devices, the bootcode runs slower when any port doesn't have a link due to clock speed slowing down as part of the link-aware feature. This leads to the driver timing out waiting for the bootcode signature.
This patch overrides the clock policy to the highest frequency just before reset and restores it after the bootcode is up.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
commit e565eec31d9d5b1024311564f6536162df951de4 Author: Michael Chan Date: Fri Jan 3 10:09:12 2014 -0800
tg3: Add unicast filtering support.
Up to 3 additional unicast addresses can be added to the perfect match filter table.
Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
commit f022ae62dd84425192b27100dcf907019f654d02 Author: Michael Chan Date: Fri Jan 3 10:09:11 2014 -0800
tg3: Refactor __tg3_set_mac_addr()
so that individual MAC address filter entries can be set.
Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
tg3: Expand 4g_overflow_test workaround to skb fragments of any size.
The current driver assumes that an skb fragment can only be upto jumbo size. Presumably this was a fast-path optimization. This assumption is no longer true as fragments can be upto 32k.
v2: Remove unnecessary parantheses per Eric Dumazet.
Cc: stable@vger.kernel.org Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Initialize REG_BASE_ADDR at PCI config offset 120 to 0
The new tg3 driver leaves REG_BASE_ADDR (PCI config offset 120) uninitialized. From power on reset this register may have garbage in it. The Register Base Address register defines the device local address of a register. The data pointed to by this location is read or written using the Register Data register (PCI config offset 128). When REG_BASE_ADDR has garbage any read or write of Register Data Register (PCI 128) will cause the PCI bus to lock up. The TCO watchdog will fire and bring down the system.
Signed-off-by: Nat Gurumoorthy Acked-by: Michael Chan Signed-off-by: David S. Miller
tg3: Expand multicast drop counter miscounting fix to 5762
commit 4d95847381228639844c7197deb8b2211274ef22 - "tg3: Workaround rx_discards stat bug", added a workaround for miscounted statistics for multicast packets. This fix needs to be applied to the 5762.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Add flag to disable 1G Half Duplex advertisement
Some link partners have issues if the non-standard 1G half duplex is advertised. This patch adds support for an nvram setting to disable the advertisement.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
rxbds_empty is an informational statistic signifying that a ring full condition was observed. It does not mean an overflow has occurred.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
commit 8496e85c20e7836b3dec97780e40f420a3ae2801 Author: Rafael J. Wysocki Date: Sun Dec 1 02:34:37 2013 +0100
PCI / tg3: Give up chip reset and carrier loss handling if PCI device is not present
Modify tg3_chip_reset() and tg3_close() to check if the PCI network adapter device is accessible at all in order to skip poking it or trying to handle a carrier loss in vain when that's not the case. Introduce a special PCI helper function pci_device_is_present() for this purpose.
Of course, this uncovers the lack of the appropriate RTNL locking in tg3_suspend() and tg3_resume(), so add that locking in there too.
These changes prevent tg3 from burning a CPU at 100% load level for solid several seconds after the Thunderbolt link is disconnected from a Matrox DS1 docking station.
Signed-off-by: Rafael J. Wysocki Acked-by: Michael Chan Signed-off-by: David S. Miller
commit 7260899bde50cbe84bdf0a15b5642e2c3b03db32 Author: Ben Hutchings Date: Mon Nov 18 22:59:43 2013 +0000
tg3: Implement the SIOCGHWTSTAMP ioctl
While we're doing this, fix the error code for SIOCSHWTSTAMP ioctl on non-timestamping hardware.
Compile-tested only.
Signed-off-by: Ben Hutchings
commit 58b187c68d91b54b82f6fd267ee3346ed8770655 Author: Ben Hutchings Date: Thu Nov 14 00:40:56 2013 +0000
tg3: Validate hwtstamp_config completely before applying it
tg3_hwtstamp_ioctl() should validate all fields of hwtstamp_config before making any changes. Currently it sets the TX configuration before validating the rx_filter field.
Compile-tested only.
Signed-off-by: Ben Hutchings Acked-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
commit 85aec73d595b8847f9c4ea571deb127913f0d508 Author: Ivan Vecera Date: Wed Nov 6 14:02:36 2013 +0100
tg3: avoid double-freeing of rx data memory
If build_skb fails the memory associated with the ring buffer is freed but the ri->data member is not zeroed in this case. This causes a double-free of this memory in tg3_free_rings->... path. The patch moves this block after setting ri->data to NULL. It would be nice to fix this bug also in stable >= v3.4 trees.
Cc: Nithin Nayak Sujir Cc: Michael Chan Signed-off-by: Ivan Vecera Acked-by: Michael Chan Signed-off-by: David S. Miller
commit d458cdf712e0c671e8e819abb16ecd6e44f9daec Author: Joe Perches Date: Tue Oct 1 19:04:40 2013 -0700
net:drivers/net: Miscellaneous conversions to ETH_ALEN
Convert the memset/memcpy uses of 6 to ETH_ALEN where appropriate.
Also convert some struct definitions and u8 array declarations of [6] to ETH_ALEN.
Signed-off-by: Joe Perches Acked-by: Arend van Spriel Signed-off-by: David S. Miller
tg3: add support a phy at an address different than 01
When phylib was in use tg3 only searched at address 01 on the mdio bus and did not work with any other address. On the BCM4705 SoCs the switch is connected as a PHY behind the MAC driven by tg3 and it is at PHY address 30 in most cases. This is a preparation patch to allow support for such switches.
phy_addr is set to TG3_PHY_MII_ADDR for all devices, which are using phylib, so this should not change any behavior.
Signed-off-by: Hauke Mehrtens Acked-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
Commit 989038e217e94161862a959e82f9a1ecf8dda152 ("tg3: Don't turn off led on 5719 serdes port 0") added code to skip turning led off on port 0 of the 5719 since it powered down other ports. This workaround needs to be enabled on the 5720 as well.
Cc: stable@vger.kernel.org Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
commit ede23fa8161c1a04aa1b3bf5447812ca14b3fef1 Author: Joe Perches Date: Mon Aug 26 22:45:23 2013 -0700
drivers:net: Convert dma_alloc_coherent(...__GFP_ZERO) to dma_zalloc_coherent
__GFP_ZERO is an uncommon flag and perhaps is better not used. static inline dma_zalloc_coherent exists so convert the uses of dma_alloc_coherent with __GFP_ZERO to the more common kernel style with zalloc.
Remove memset from the static inline dma_zalloc_coherent and add just one use of __GFP_ZERO instead.
Trivially reduces the size of the existing uses of dma_zalloc_coherent.
Realign arguments as appropriate.
Signed-off-by: Joe Perches Acked-by: Neil Horman Acked-by: Jesse Brandeburg Acked-by: Jeff Kirsher Signed-off-by: David S. Miller
While EEH error happens, we might not have network device instance (struct net_device) yet. So we can't access the instance safely and check its link state, which causes kernel crash. The patch fixes it.
EEH: Frozen PE#2 on PHB#3 detected EEH: This PCI device has failed 1 times in the last hour EEH: Notify device drivers to shutdown (NULL net_device): PCI I/O error detected Unable to handle kernel paging request for data at address 0x00000048 Faulting instruction address: 0xd00000001c9387a8 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=1024 NUMA PowerNV : NIP [d00000001c9387a8] .tg3_io_error_detected+0x78/0x2a0 [tg3] LR [d00000001c9387a4] .tg3_io_error_detected+0x74/0x2a0 [tg3] Call Trace: [c000003f93a0f960] [d00000001c9387a4] .tg3_io_error_detected+0x74/0x2a0 [tg3] [c000003f93a0fa30] [c00000000003844c] .eeh_report_error+0xac/0x120 [c000003f93a0fac0] [c0000000000371bc] .eeh_pe_dev_traverse+0x8c/0x150 [c000003f93a0fb60] [c000000000038858] .eeh_handle_normal_event+0x128/0x3d0 [c000003f93a0fbf0] [c000000000038db8] .eeh_handle_event+0x2b8/0x2c0 [c000003f93a0fc90] [c000000000038e80] .eeh_event_handler+0xc0/0x170 [c000003f93a0fd30] [c0000000000cc000] .kthread+0xf0/0x100 [c000003f93a0fe30] [c00000000000a0dc] .ret_from_kernel_thread+0x5c/0x80
Reported-by: Wei Yang Signed-off-by: Gavin Shan Acked-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
Turning off led on port 0 of the 5719 serdes causes all other ports to lose power and stop functioning. Add tg3_phy_led_bug() function to check for this condition. We use a switch() in tg3_phy_led_bug() for consistency with the tg3_phy_power_bug() function.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
commit 6829309926b90c4c32d1f4cafeb600cd34a721e3 Author: Daniel Borkmann Date: Tue Aug 13 11:45:13 2013 -0700
net: tg3: fix NULL pointer dereference in tg3_io_error_detected and tg3_io_slot_reset
Commit d8af4dfd8 ("net/tg3: Fix kernel crash") introduced a possible NULL pointer dereference in tg3 driver when !netdev || !netif_running(netdev) condition is met and netdev is NULL. Then, the jump to the 'done' label calls dev_close() with a netdevice that is NULL. Therefore, only call dev_close() when we have a netdevice, but one that is not running.
[ Add the same checks in tg3_io_slot_reset() per Gavin Shan - by Nithin Nayak Sujir ]
Reported-by: Dave Jones Signed-off-by: Daniel Borkmann Cc: Gavin Shan Cc: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
The 5762 devices sometimes incorrectly treat udp fragments as RMCP packets and route to the APE. This patch sets the RX_MODE_IPV4_FRAG_FIX bit for these devices which enables the proper behaviour.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
The PTP_CAPABLE tg3 devices have a gpio output that is toggled when the free running counter matches a watchdog value. This patch adds support to set the watchdog and enable this feature.
Since the output is controlled via bits in the EAV_REF_CLCK_CTL register, we have to read-modify-write it when we stop/resume.
Cc: Richard Cochran Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Allow NVRAM programming when interface is down
Previously, when the interface was brought down, the driver would set the power state to D3hot. In D3hot, we don't have access to the NVRAM. This patch removes the call to set the power state to PCI_D3hot in close. A following patch will implement the shutdown handler to properly set the D3hot state when the system is going down.
Doing the above means that the TG3_PHYFLG_IS_LOW_POWER should not be checked to validate access to the NVRAM.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
During probe, the driver is incorrectly switching the power to Vaux on the 5717 and later devices. At this point, we are in D0 state and drawing maximum power. We also definitely have Vmain available. It doesn't make sense to switch to Vaux since it has a lesser maximum power draw and we might go over the limit. On a new system, we observe that not all ports are recognized in some of the slots with this call in place.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
When the EEH error is the result of a fenced host bridge, MMIO accesses can be very slow (milliseconds) to timeout and return all 1's, thus causing the driver various timeout loops to take way too long and trigger soft-lockup warnings (in addition to taking minutes to recover).
It might be worthwhile to check if for any of these cases, ffffffff is a valid possible value, and if not, bail early since that means the HW is either gone or isolated. In the meantime, checking that the PCI channel is offline would be workaround of the problem.
Cc: # v3.0+ Signed-off-by: Gavin Shan Signed-off-by: David S. Miller
commit 72bb72b0d98847d22c6fae4e170121f3640f0f60 Author: Michael Chan Date: Mon Jun 17 13:47:25 2013 -0700
tg3: Prevent system hang during repeated EEH errors.
The current tg3 code assumes the pci_error_handlers to be always called in sequence. In particular, during ->error_detected(), NAPI is disabled and the device is shutdown. The device is later reset and NAPI re-enabled in ->slot_reset() and ->resume().
In EEH, if more than 6 errors are detected in a hour, only ->error_detected() will be called. This will leave the driver in an inconsistent state as NAPI is disabled but netif_running state is still true. When the device is later closed, we'll try to disable NAPI again and it will loop forever.
We fix this by closing the device if we encounter any error conditions during the normal sequence of the pci_error_handlers.
v2: Remove the changes in tg3_io_resume() based on Benjamin Poirier's feedback.
Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
Some systems that don't need wake-on-lan may choose to power down the chip on system standby. Upon resume, the power on causes the boot code to startup and initialize the hardware. On one new platform, this is causing the device to go into a bad state due to a race between the driver and boot code, once every several hundred resumes. The same race exists on open since we come up from a power on.
This patch adds a wait for boot code signature at the beginning of tg3_init_hw() which is common to both cases. If there has not been a power-off or the boot code has already completed, the signature will be present and poll_fw() returns immediately. Also return immediately if the device does not have firmware.
Cc: stable@vger.kernel.org Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
Commit 091f0ea30074bc43f9250961b3247af713024bc6 "tg3: Add New 5719 Read DMA workaround" added a workaround for TX DMA stall on the 5719. This workaround needs to be applied to the 5720 as well.
Cc: stable@vger.kernel.org Reported-by: Roland Dreier Tested-by: Roland Dreier Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
The current code calls tg3_set_bdinfo() separately on napi0, followed by a loop that does napi1+. Simplify it by setting bdinfo in the loop for all napi contexts.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Split APE driver state change out of boot reset signature update
Unlike the boot signature that needs to be set before every reset, the ape state only needs to be updated to tell the firmware that the driver is now taking/releasing control of the hardware. Move the calls to tg3_ape_driver_state_change() to better, more appropriate places.
Also, the firmware does not distinguish between SUSPEND and START states anymore. Remove the SUSPEND case in the switch.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
tg3: Ensure boot code has completed initialization before accessing hardware
After resetting the device, the driver waits for a signature to be updated to know that firmware has completed initialization. However, the call to tg3_poll_fw() is being done too late and we're writing to the GRC_MODE register before it has completely initialized, causing contention with firmware. This logic has existed since day one but is causing PCIE link to go down randomly at startup on one platform once every few hundred reboots.
Move the tg3_poll_fw() up to before we write to the GRC_MODE register after reset.
Signed-off-by: Nithin Nayak Sujir Signed-off-by: Michael Chan Signed-off-by: David S. Miller
commit 1cbf9eb85a6601b58f01a71ff10299d2bf5d3365 Author: Nithin Sujir Date: Sat May 18 06:26:55 2013 +0000
tg3: Implement set/get_eee handlers
Reviewed-by: Ben Li Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
commit 5b6c273ad6c3886f30c7c5df7009e489043c59f3 Author: Nithin Sujir Date: Sat May 18 06:26:54 2013 +0000
tg3: Simplify tg3_phy_eee_config_ok() by reusing tg3_eee_pull_config()
eee_config_ok() was checking only for mismatch in advertised settings. This patch expands the scope of eee_config_ok() to check for mismatch in the other eee settings. On mismatch we will require a call to tg3_setup_eee() to push the configured settings to the hardware.
Reviewed-by: Ben Li Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
commit 400dfbaa8d444a29056b051a3d7082dc611e3b12 Author: Nithin Sujir Date: Sat May 18 06:26:53 2013 +0000
tg3: Add tg3_eee_pull_config() function
Add tg3_eee_pull_config() to pull the settings from the hardware and populate the eee structure.
If Link Flap Avoidance is enabled, we pull the eee settings from the hw so as not to cause a phy reset on eee config mismatch later. This requires moving down tg3_setup_eee() below the tg3_pull_config() to not trample existing settings.
Reviewed-by: Ben Li Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
commit 9e2ecbeb250dee67f4bd06a18637c120b48a4865 Author: Nithin Sujir Date: Sat May 18 06:26:52 2013 +0000
tg3: Add ethtool_eee struct and tg3_setup_eee()
Add an eee structure and update it with eee settings. This will be used for set/get_eee operations. Add common function tg3_setup_eee() that will be used in the subsequent patches.
Reviewed-by: Ben Li Signed-off-by: Michael Chan Signed-off-by: Nithin Nayak Sujir Signed-off-by: David S. Miller
The commit f233a976ad15c3b8c54c0157f3c41d23f7514279 (tg3: shows HW time stamping support only if ptp_capable is present) didn't append hardware flags correctly. This patch fixes it.
Signed-off-by: Flavio Leitner Signed-off-by: David S. Miller
net: vlan: add protocol argument to packet tagging functions
Add a protocol argument to the VLAN packet tagging functions. In case of HW tagging, we need that protocol available in the ndo_start_xmit functions, so it is stored in a new field in the skb. The new field fits into a hole (on 64 bit) and doesn't increase the sks's size.
Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller
net: vlan: rename NETIF_F_HW_VLAN_* feature flags to NETIF_F_HW_VLAN_CTAG_*
Rename the hardware VLAN acceleration features to include "CTAG" to indicate that they only support CTAGs. Follow up patches will introduce 802.1ad server provider tagging (STAGs) and require the distinction for hardware not supporting acclerating both.
Signed-off-by: Patrick McHardy Signed-off-by: David S. Miller