mirror of https://github.com/eLvErDe/hwraid.git
574 lines
27 KiB
Plaintext
574 lines
27 KiB
Plaintext
Changes for 1.60:
|
|
|
|
- fix bug in doDeleteVolume2 for setting the volume dev handle in the RAID
|
|
action request.
|
|
- Add option to use default settings on volume creation.
|
|
- Changed default stripe size to 128K
|
|
- fixed a build error that occurred if REGISTER_ACCESS was set to 0 at
|
|
compile time.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.59:
|
|
|
|
- Add code that determines BIOS/FCode compatibility for first MPI 2.0 chips.
|
|
- Add "Dump PCI config space".
|
|
- Modify code that allows changing SAS persistent entries, to handle changing
|
|
a set of mappings from one EnclosureId to another EnclosureId.
|
|
- Enhance display of expander routing tables to accommodate the SAS 2.0
|
|
committee changing the opcode for REPORT EXPANDER ROUTE TABLE LIST from
|
|
17h to 22h.
|
|
- After displaying diagnostic pages, don't ask about making changes except for
|
|
page numbers that are known to be writable.
|
|
- Make modifications to the RAID volume creation code to take drive minimums
|
|
from config pages, rather than hard-coding them. Also, allow selecting a
|
|
stripe size for any volume type except for mirrored (RAID1) volumes.
|
|
- Add ability to send QUERY TASK and ABORT TASK to the "Reset target" routine;
|
|
this is not the ideal place to put them, but it's the best of the current
|
|
alternatives. No allowance is made for specifying the identity of the task
|
|
to be queried or aborted; the intent is that these will be used merely to
|
|
verify that such requests can be issued (not that they'll be successful).
|
|
- Add some missing config pages to "Dump all port state".
|
|
- After resetting the chip (DOS/EFI), wait 20 milliseconds before making any
|
|
chip accesses, per the MPT specification.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.58:
|
|
|
|
- Add option 66, "Show SAS discovery errors".
|
|
- Fix bug introduced in 1.57 that broke option 13 for SAS.
|
|
- Fix bug introduced in 1.56 that made diagnostic buffer actions appear to
|
|
fail for Windows, when in fact they had actually succeeded.
|
|
- Fix options 8, 13, 16, 18, 33, 35, 44, 68, 80, and 81, to accommodate
|
|
differences between MPI 1.x and MPI 2.0 (the new link speed of 6.0 Gbps
|
|
being the most prominent one).
|
|
- Modify expander firmware download to allow boot records (BufferID 10) to
|
|
be headerless. If a header is needed (mode 2), but is not present, add it.
|
|
If a header is not needed (mode 6), but is present, remove it. Also fix
|
|
verification code to be more robust.
|
|
- Fix MPI 2.0 bug that caused duplicate devices to be shown on Windows during
|
|
a device scan.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.57:
|
|
|
|
- Implement RAID actions for MPI 2.0, and implement "Delete hot spare".
|
|
- Add Linux support for MPI 2.0.
|
|
- Modify NVDATA concatenation code for latest MPI 2.0 firmware changes.
|
|
- When downloading expander firmware, if mode is 2, use Write Buffer, but if
|
|
mode is 6 or 7, use SES Download Microcode. Before downloading, verify the
|
|
image for Buffer IDs 0, 2, and 10.
|
|
- For SAS, add "Force full discovery".
|
|
- Add "Identify FLASH device".
|
|
- Fix more endian problems, by adding code that allows the compiler to detect
|
|
them automatically.
|
|
- Skip verification after firmware download when the special bridge firmware
|
|
(1.18.99) is being used.
|
|
- Fix the multi-path detection code (used when creating volumes), which was
|
|
reading enclosure pages incorrectly.
|
|
- Fix the code that handles metadata size, to use ManufacturingPage4 instead
|
|
of IOCPage6 (the former can be modified, while the latter cannot).
|
|
- After reading diagnostic pages, allow the data to be modified and the page
|
|
written back.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.56:
|
|
|
|
- Implement -k flag to suppress the "select a device" mechanism for many
|
|
diagnostic functions, and instead allow directly entering bus and target
|
|
numbers.
|
|
- Fix the -h (help) display, which got broken when paging was implemented.
|
|
- Add some routines to help debug expanders (read/write expander memory,
|
|
read/write expander ISTWI devices, expander firmware download, expander
|
|
reset).
|
|
- Implement all of the diagnostic buffer actions for Solaris (a new driver is
|
|
required). Also, when any diagnostic buffer action fails, display detailed
|
|
information if it's available.
|
|
- Don't try to verify a firmware download for SAS if only a bootloader image
|
|
is active; the required "firmware upload" support is not implemented.
|
|
- Fix endian issues in various places.
|
|
- Fix the code that was trying to detect multi-pathed disks when creating RAID
|
|
volumes; it was not correctly determining whether two enclosures were in fact
|
|
the same enclosure.
|
|
- In the "Log Sense" and "Diagnostic Page" tests, make the buffer larger, and
|
|
check for the device returning more data than requested.
|
|
- When working with RAID physical disks, consistently use the physical disk
|
|
count that is in IOCPage3, vs. the one in IOCPage2.
|
|
- Add path information to the display of RAID physical disks, whenever there
|
|
is more than one path.
|
|
- Enhance "Scan for devices" to also show (hidden) RAID physical disks.
|
|
- Fix the Solaris code that determines device names, to handle SCSI Enclosure
|
|
Services (SES) devices, and to handle SAS disks with long GUID-based names.
|
|
- For "Send SMP request", if the handle entered is 0, prompt for SAS WWID and
|
|
physical port to be entered manually.
|
|
- Fix bug in "Send SCSI CDB" that caused loss of detailed failure information
|
|
(including SCSI status and, if there was a Check Condition, any sense data).
|
|
- More MPI 2.0 work: add a real list of supported devices, fix the expander
|
|
routing table display code, fix the SATA Clear Affiliation code, fix SCSI I/O
|
|
reply handling to correctly obtain sense data.
|
|
- Fix a bug in the EFI implementation of scanf(): %n did not work correctly.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.55:
|
|
|
|
- Extend paged output in menus to cover all output in all circumstances.
|
|
- Make the maximum number of devices shown in menus compile-time configurable,
|
|
with the default remaining at 99.
|
|
- Add the ability to turn off MPT register access at compile time, since this
|
|
causes crashes on some Linux standalone platforms.
|
|
- Accommodate more differences between MPI 1.5 and MPI 2.0 (in SCSI I/O replies
|
|
and event handling). Also, for MPI 2.0, use the BTDH IOCTL to get the driver
|
|
mapping between bus/target and device handle.
|
|
- When the -g flag is used, ask if the fwdiagtool header should be added when
|
|
doing diagnostic buffer reads.
|
|
- Have options 15, 17, and 20 automatically bring the port online under DOS and
|
|
EFI.
|
|
- Expand the PCI information printed to include the segment number.
|
|
- Add verification after firmware downloads (read the image back, and compare
|
|
with what was written).
|
|
- When creating RAID volumes, correctly enforce rules about mixing SAS and
|
|
SATA drives, and SSD and non-SSD drives. Also, use the metadata size that
|
|
is specified in IOC Page 6.
|
|
- Fix code that implements setting the volume name (string length check was
|
|
wrong, and volume addressing was wrong).
|
|
- When deleting a the last volume, delete any hot spares that remain; otherwise
|
|
those physical disks are left stranded.
|
|
- For Solaris, use an IOCTL to get the driver name and version. Use that same
|
|
IOCTL to get PCI information; only if that fails should the "reg" property be
|
|
used to get PCI information.
|
|
- When dumping config pages, dump SAS Expander Page 1 correctly (there is one
|
|
page per expander phy), and fix its name (it was called "SASDevicePage1"!).
|
|
- Implement the new changes for MPI 2.0 NVDATA concatenation. Add IOC Page 8,
|
|
BIOS Page 4, and Driver Mapping Page 0.
|
|
- Add MPI 2.0 support to the DOS/EFI driver.
|
|
- Add "Read Buffer (for firmware upload)" to allow reading back device or
|
|
expander firmware and saving into a file.
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.54:
|
|
|
|
- Fix concatenation bug that caused invalid SAS firmware files to be generated
|
|
- Catch error if no file name is entered for option 15, suboption 8
|
|
- Add ability to issue product-specific SAS IO Unit Control and Diag Data
|
|
Upload requests
|
|
- When modifying SAS IO Unit settings, if "all phys" is selected, the default
|
|
should be to preserve existing settings per phy, rather than overwrite with
|
|
new settings
|
|
- Handle self-configuring expanders when displaying expander route tables
|
|
- When setting a volume name, use the new RAID action
|
|
- Change the display of PCI location information to include the segment number
|
|
- Allow SMP requests and responses to be up to 1024 bytes (per the SAS spec)
|
|
- Fix problem where the maximum target allowed is less than expected, so some
|
|
targets can be missed in a scan
|
|
- Add preliminary parsing of MPI 2.0 NVDATA
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.53:
|
|
|
|
- Add initial support for MPI 2.0
|
|
- Add -q switch, to query files for version information
|
|
- Add -g switch, to guarantee asking of questions that otherwise would be
|
|
suppressed (such as the bus number, when there's only one bus); this makes
|
|
it easier to script lsiutil from Perl
|
|
- Rework the way that ports are found under Solaris, to eliminate what are
|
|
sometimes lengthy delays waiting for prtconf to run -- the new code uses
|
|
the devinfo library to directly walk the device tree
|
|
- Fix memory leaks when using the -j switch (ports not selected were not
|
|
being closed), and when using options 18 or 44
|
|
- Add "Set SAS phy offline/online"
|
|
- Add ability to save persistent mappings to a file, then later restore them
|
|
- Add "Diagnostic Page Test", which reads any SCSI diagnostic pages
|
|
- Add code to enable or disable the read or write caches in SCSI devices
|
|
- Fix incomplete implementation of "Force firmware to fault (with C0FFEE)"
|
|
- Add verification of internal FCode checksum
|
|
- Add verification of PCIR image length
|
|
- Extend "Inject/Repair media error" test so that it works for SATA devices
|
|
- When downloading drive firmware, set default mode to 5 for SATA devices
|
|
- Fix bug in "Show physical disks" that assumed physdisk numbers were dense
|
|
- Detect multi-pathed disks when creating RAID volumes, and suppress their
|
|
inclusion in the device list
|
|
- Add code to allow binary files to be written when dumping chip locations
|
|
(the -g switch must be used for force the question)
|
|
- Fix an endian bug when dumping chip locations below 0x10000000
|
|
- Modified SAS phy/link reset code to ask for a device handle, so that remote
|
|
phys or links (i.e., on expanders) can be reset
|
|
- For DOS/EFI, avoid issuing Port Enable if possible, since that request can
|
|
take a long time to complete
|
|
- Change method of checking for "context reply" after SCSI request passthrough
|
|
under Windows, from looking for a specific MsgContext value, to looking for
|
|
the reply's MsgLength being zero
|
|
- Add retries under Linux if an IOCTL fails with EAGAIN
|
|
- Increase size of DOS/EFI scratch buffer to 1 MB (from 64 KB) to accommodate
|
|
large drive firmware downloads
|
|
- When exiting on DOS/EFI, don't reset a port that was not brought online
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.52:
|
|
|
|
- Make changes required to allow an EBC version for EFI
|
|
- Make changes required to allow native 64-bit IA64 and AMD64 versions
|
|
for Windows
|
|
- Modify Linux workarounds for IOC_INIT (needed to write Manufacturing
|
|
config pages)
|
|
- For Solaris, modify findPorts() to first use "prtconf -D", and then as
|
|
a fallback, use /etc/path_to_inst, so that working ports can be more
|
|
accurately identified
|
|
- Fix several big-endian bugs in the RAID-related options
|
|
- Don't print a stripe size for IM volumes (it's not valid like it is for
|
|
IME and IS volumes)
|
|
- Add support for a few new FC Management toolbox functions
|
|
- Rewrite getOsDeviceName() for Solaris, so that FCA devices are handled
|
|
as well as SCSA devices
|
|
- Fix an off-by-one message when complaining about incorrect VPD end tags
|
|
- Enhance SAS NVDATA parser to handle version 2D
|
|
- Support NPIV for FC by adding a -v switch (to identify which virtual
|
|
port to masquerade as)
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.51:
|
|
|
|
- Add date/time stamp when logging to lsiutil.log
|
|
- Add "SEP (SCSI Enclosure Processor) Test"
|
|
- Add "FC management tools"
|
|
- Add "Force firmware to fault (with C0FFEE)" for SAS
|
|
- Add "Report LUNs Test"
|
|
- Fix dumping of memory regions to work with FC919/FC929
|
|
- Increase chunk size for firmware download/upload, for DOS only
|
|
- Allow Sun's mpt driver to work (for Solaris)
|
|
- Fix endian bugs in firmware download when checking for 1MB/2MB issues
|
|
- Dump enclosure pages when displaying attached devices for SAS
|
|
- Add checks for maximum physical disks in a volume
|
|
- When creating a volume, allow user to specify "no initial sync"
|
|
- Display FC WWNs and SAS WWID as part of -i (aka option 68)
|
|
- Add FC949E A2, 53C1030T A2, SAS1064E B3, SAS1068E B3, and 1078 C2 chips
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.50:
|
|
|
|
- Rearrange some source files to allow easier building on Windows
|
|
and Solaris
|
|
- Add ability to set individual bits, bytes, words, when modifying config
|
|
pages (e.g., when prompted for a value, type "bit3=1", or "bits7:4=2",
|
|
or "byte2=17")
|
|
- Add "Show expander routing tables"
|
|
- Add "Disable/Enable drive firmware update mode" to RAID actions
|
|
- Add "Display Log entries" and "Clear (erase) Log entries"
|
|
- Add "Upload FLASH section"
|
|
- Add "Display version information"
|
|
- Add "Display/Program chip VPD information"
|
|
- Add back deleting of persistent entries for SAS, and add "Change
|
|
Bus/Target of a existing mapping"
|
|
- Add code to firmware download that handles the SAS phase 8 to phase 8
|
|
bridge firmware image (version 1.18.99)
|
|
- When erasing the BIOS, fill with FF rather than 00 to match the FLASH
|
|
erase pattern
|
|
- Add ability to control booting under EFI for "Change FC Port settings"
|
|
- Modify "Change WWID" code to allow for variable-sized config page
|
|
- Under DOS or EFI, after any RAID action, sleep for 5 seconds to allow
|
|
drive metadata to be updated (if necessary)
|
|
- Use the Memory Move toolbox tool to allow "Dump chip memory regions" to
|
|
be able to dump tightly-couple memory for SAS and FC
|
|
- Beef up the display of SAS control settings in "Dump all port state"
|
|
- Display IOCLogInfo value for any RAID action request that fails
|
|
- Enhance concatenation to handle NVDATA versions 25, 28, 29, and 2B
|
|
- Properly return Check Condition Sense Data under DOS and EFI
|
|
- Switch DOS code to use memory space instead of I/O space, whenever
|
|
possible
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.49:
|
|
|
|
- Fix build issues with latest SLES10/RHEL5 drivers
|
|
- Add -j <type> to allow selecting just FC, SCSI or SAS ports
|
|
- Add doWaitForResync (RAID action)
|
|
- Add doRemoveSasDevice
|
|
- Fix -b and -r to obey -p, -0, and -1
|
|
- Fix port and driver naming under Windows StorPort drivers
|
|
- Add 1 MB to 2 MB migration code for SAS (phase 9), DOS and EFI only
|
|
- Fix doCreateVolume to obey volume member limits
|
|
- Remove all tabs from output, replace with spaces
|
|
- Add parsing of NVDATA version 28
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.48:
|
|
|
|
- Add -w, -ww, and -www flags to capture debug output
|
|
- Add -x flag to concatenate SAS firmware and NVDATA files
|
|
- Switch from using PortType as a selector, to ProductIdType
|
|
- For DOS and EFI, make sure port is ready for each new option
|
|
- For ports stuck in reset, don't prompt for a firmware file unless and
|
|
until the port is selected
|
|
- For EFI, do a firmware upload to handle FWDLB cases (no FLASH)
|
|
- Add paged menu mode
|
|
- Move "Beacon on/off" to be FC-specific
|
|
- Add "Erase non-volatile adapter storage"
|
|
- Add "Diagnostic Buffer actions"
|
|
- Add "Program manufacturing information" for FC and SAS
|
|
- Add "Set personal WWNN/WWPN" for FC
|
|
- Add "Send SCSI CDB"
|
|
- Remove most persistence actions for SAS, as they no longer are supported
|
|
by firmware
|
|
- Add "Inject/Repair media error"
|
|
- Add "Set/Clear write protect"
|
|
- When downloading SAS firmware, make sure the attached NVDATA is correct
|
|
for the port's board type
|
|
- Fix SAS firmware validity checking to correctly detect wrong chip type
|
|
and wrong chip revision
|
|
- Allow setting/clearing Native Command Queue Disable and SATA Write Cache
|
|
Disable for SAS
|
|
- Allow "Read/Write/Verify Logical Blocks" to touch individual volume
|
|
members, just like "Inject/Repair media error"
|
|
- Fix incorrect ordering of LBA High/Middle/Low parts in "Send SATA
|
|
request"
|
|
- Add dumping of VPD data (Manufacturing Page 1), if non-zero, to board
|
|
info display
|
|
- Fix incorrect defaulting of FC's HostSCSIID (15 is no longer correct)
|
|
- Fix EFI code to truncate files properly
|
|
- Fix EFI code to use I/O space properly, if required (1030/1030T)
|
|
- Fix DOS code to not enable chip interrupts (caused a hang while exiting
|
|
on some platforms)
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.47:
|
|
|
|
- Add support for EFI
|
|
- Add a workaround for older Linux mptbase drivers (fix IOC_INIT failing)
|
|
- Add doWriteLogicalBlocks and doVerifyLogicalBlocks
|
|
- Add generic SATA passthrough
|
|
- Allow -b (showBoardInfo) in combination with any other arguments
|
|
- Increase range of Initiator Device Timeout value, by using the
|
|
supported scale factor (x16) if necessary
|
|
- Allow setting the Device Missing Report Delay and Device Missing
|
|
I/O Delay values
|
|
- Fix a bug in configuring wide ports
|
|
- Add support for RAID Volumes of size > 2 TB
|
|
- Add missing config pages to doPortStateDump (-d)
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.46:
|
|
|
|
- Add -0 and -1 options, to skip IOC numbers 0 and 1, for certain
|
|
operations (so, lsiutil -1 99 would only reset each chip once)
|
|
- Enhance display to show IOC number (same as PCI function number)
|
|
- Check to see if the firmware is in Operational state before trying the
|
|
first MPT command, to avoid causing a chip reset if the firmware is in
|
|
Fault state (if so, option 99 will perform a chip reset, if desired)
|
|
- Fix code that obtains MPT register addresses for Linux (was broken for
|
|
some systems with 2.6 kernels)
|
|
- Fix code that displays operating system device names for Linux (was
|
|
displaying incorrect names for some SAS and FC devices whose SCSI
|
|
bus/target numbers were not equal to their MPT bus/target numbers)
|
|
- Move code that enables MPT register access inside the loops for options
|
|
51 and 52, so that access continues to work even if the chip is reset
|
|
- Add display of IOCFactsReply and PortFactsReply to port state that is
|
|
dumped by -d, and only display FC Port Page 3 pages that are valid
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.45:
|
|
|
|
- Fixed get2bytes/get3bytes/get4bytes macros to properly truncate
|
|
- Fixed formats for 64-bit integers when printed (Windows is different)
|
|
- Added doDriveFirmwareDownload
|
|
- Added doDisplaySfpPages
|
|
- Added doDefaultPhyRegsSettings
|
|
- Extended event displaying to Linux and Solaris
|
|
- Added some control over port configuration in doSasIoUnitSettings
|
|
- Reworked getOsDeviceName slightly for Linux and Solaris
|
|
- Fixed bug in doPortStateSummary which reversed Yes and No for
|
|
SSP Initiator Enabled and SSP Target Enabled displays
|
|
- Added support for FC949E A1
|
|
- Fixed Linux version of doMptCommand to correctly gather Sense Data
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.44:
|
|
|
|
- Upgrade to 1.05.13 MPI headers
|
|
- Add doReadLogicalBlocks and doDisplayOsDeviceNames
|
|
- Add EEDP support
|
|
- Add an optional duration to the monitor frequency (-m f,d)
|
|
- Add "reset port" option to all submenus too
|
|
- Don't send quiesce/unquiesce RAID actions to SAS IR (not supported)
|
|
- Don't coerce RAID volume size if it's small (less than 1 GB)
|
|
- Add SAS1068E B1 as a known device
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.43:
|
|
|
|
- Add Clear ACA (option 24)
|
|
- Fix a Linux buffer overrun that caused only the first port to be
|
|
discovered for some architectures
|
|
- Fix ManufacturingPage2 checksum offset for FC949E
|
|
- Guard against divide-by-zero when displaying volume resync percent
|
|
- Prevent a target from being chosen twice when creating a volume
|
|
- Add the flag that deletes PhysDisks when deleting a Volume
|
|
- Fix a DOS bug (internal buffer corruption) by rearranging the order
|
|
of operations when starting a new command -- bring the chip online
|
|
first, before loading the scratch buffers
|
|
- Correctly set RequestDataLength to NOT include the CRC bytes, when
|
|
sending hand-built SMP Passthrough commands
|
|
- Add a hardware workaround for the FC949E to the Firmware Download Boot
|
|
code for DOS
|
|
- Fix Adapter Reset for SAS1078 (DOS only)
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.42:
|
|
|
|
- Add Log Sense Test
|
|
- Add SATA SMART Read Test
|
|
- Add Create Hot Spare
|
|
- Add Send SMP Request
|
|
- Add support for the 1078
|
|
- Change number of LUNs probed to 64 for FC and SAS (leave SCSI alone)
|
|
- In doFcPortSettings, preserve original HardALPA setting if user quits
|
|
- Fix endian issue when creating volumes (default settings were incorrect
|
|
for SPARC and PowerPC)
|
|
- Fix typo in doPhysDiskSettings that used the wrong Action value
|
|
- Ignore IOC Page 3 in doShowNonDefaultSettings
|
|
- Ignore LOG pages in doRestoreDefaultSettings
|
|
- Fix DOS MPT driver to not poll the Doorbell while DMA is active
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.41:
|
|
|
|
- Fix build errors on some Linux systems
|
|
- Filter out bad "what" strings (any of the first 4 characters are bogus)
|
|
- Fix missing 50/51/52/53 from main menu on Linux systems
|
|
- Add checking for 1064 A3 SAS firmware images
|
|
- Use current config page data if non-volatile data is invalid
|
|
- Add "Read Default Write NVRAM" to "Test configuration page actions"
|
|
- Resolve SAS firmware hang by not reading LOG page types
|
|
- Remove IO Unit Page 3 from "Restore default settings"
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.40:
|
|
|
|
- Added MPT and chip register access for Linux
|
|
- Added doShowNonDefaultSettings and doRestoreDefaultSettings
|
|
- Fixed commands to consistently check for errors
|
|
- Added check for overflow of volume size when creating a volume
|
|
- Fixed "free" calls in doDisplayCurrentEvents
|
|
- Added retries around Windows IOCTL calls, to handle BUSY status
|
|
- Added warning message if any Linux IOCTL call is rejected
|
|
- Added speed/width negotiation to DOS driver for SCSI
|
|
- Added a watchdog timer routine to DOS driver to catch faults
|
|
- Fixed firmware download boot for 929X in DOS driver
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.39:
|
|
|
|
- Add DOS support
|
|
- Add 949E support
|
|
- Add MPT register access and chip register access
|
|
- Move diag_targets[99] from stack storage to global storage
|
|
- Add Read Block Limits (like Read Capacity but for tapes)
|
|
- Initialize memory after allocation, always
|
|
- When menu options are picked from -a, don't print the menu itself
|
|
- When showing current firmware version, print both hex and decimal
|
|
- Fix bug in displaying phy error counters that nearly always hid errors
|
|
- Allow 2-member IS volumes (previously, 2-member volumes were always IM)
|
|
- Zero errno (the last error that occurred) prior to each syscall
|
|
- Fixed bug that made "lsiutil -e" (or "lsiutil -u") just exit
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.38:
|
|
|
|
- Update to latest header files, 1.05.10
|
|
- Display chip rev as well as chip name
|
|
- Add Display Phy Counters and Clear Phy Counters
|
|
- Make display of firmware version be accurate, always
|
|
- Add Trigger Analyzer with Echo
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.37:
|
|
|
|
- Fixed all mallocs that did not have frees
|
|
- Added Read Capacity and Mode Page tests
|
|
- Added -n (companion to -y) to answer questions automatically
|
|
- In Identify, print current running firmware version as well as FLASH
|
|
versions
|
|
- If mptscsih is not loaded, quit (from Scan for Devices) early
|
|
- Increase metadata size for SAS IR volumes
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.36:
|
|
|
|
- Simplify menus using "expert mode", which is enabled with -e on the
|
|
command line, or by typing "e" at any menu prompt
|
|
- Add display of transfer statistics for FC (-m, for "monitor")
|
|
- Allow -h to print help even if no ports found
|
|
- Fix firmware download and config page reads for old versions of MPT (1.00)
|
|
- Allow easier erasing of BIOS/FCode images
|
|
- Add an explicit error message for Linux if a SCSI I/O fails due to the
|
|
mptscsih driver not being loaded
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.35:
|
|
|
|
- Check for new Windows split-driver names (lsi_fc, lsi_scsi, lsi_sas)
|
|
- Display all "what" strings that are found
|
|
- Add SATA Clear Affiliation test
|
|
- Add ability to monitor current events (Linux only so far)
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.34:
|
|
|
|
- Add -b option, which gets board info (from manufacturing config pages)
|
|
- Add SATA Identify Drive test
|
|
- Fix bug in -p where ranges didn't quite work right
|
|
- Add routine to get the length of a variable config page, and use it
|
|
- Add warning for trying to use an old FC929X firmware image on an FC929XL
|
|
- When setting FC devices persistent, set SCAN_ID and SCAN_LUNS flags too
|
|
- Add display of what phys are connected to what, in Display Attached Devices
|
|
- Remove all knowledge of EEDP for now
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.33:
|
|
|
|
- On Linux, try "modprobe mptctl" and "mknod /dev/mptctl c 10 220" if
|
|
IOCTLs don't work
|
|
- Fix bug when manually setting SAS persistence
|
|
- Add additional checking to distinguish 929X from 929XL
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.32:
|
|
|
|
- Added coercion of disk size when making an IR volume
|
|
|
|
-----------------------------------------------------------------------------
|
|
|
|
Changes for 1.31:
|
|
|
|
- Added -a to allow arguments to be entered on the command line
|
|
- Added Beacon on/off
|
|
- Detected Sun FCode for 1030 (and hopefully for 1064 too)
|
|
- Fixed bug when manually entering more than one SAS persistent address
|
|
- Fixed bug when underrun occurs but some data is transferred (NOT an error!)
|
|
|