Documentation/scsi/ChangeLog.megaraid

   1 Release Date    : Mon Mar 07 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
   2 Current Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module)
   3 Older Version   : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
   4
   5 1.      Added IOCTL backward compatibility.
   6         Convert megaraid_mm driver to new compat_ioctl entry points.
   7         I don't have easy access to hardware, so only compile tested.
   8                 - Signed-off-by:Andi Kleen <ak@muc.de>
   9
  10 2.      megaraid_mbox fix: wrong order of arguments in memset()
  11         That, BTW, shows why cross-builds are useful-the only indication of
  12         problem had been a new warning showing up in sparse output on alpha
  13         build (number of exceeding 256 got truncated).
  14                 - Signed-off-by: Al Viro
  15                 <viro@parcelfarce.linux.theplanet.co.uk>
  16
  17 3.      Convert pci_module_init to pci_register_driver
  18         Convert from pci_module_init to pci_register_driver
  19         (from:http://kerneljanitors.org/TODO)
  20                 - Signed-off-by: Domen Puncer <domen@coderock.org>
  21
  22 4.      Use the pre defined DMA mask constants from dma-mapping.h
  23         Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
  24         pci_set_dma_mask() or pci_set_consistend_dma_mask(). See
  25         http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for more
  26         details.
  27                 Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
  28                 Signed-off-by: Domen Puncer <domen@coderock.org>
  29
  30 5.      Remove SSID checking for Dobson, Lindsay, and Verde based products.
  31         Checking the SSVID/SSID for controllers which have Dobson, Lindsay,
  32         and Verde is unnecessary because device ID has been assigned by LSI
  33         and it is unique value. So, all controllers with these IOPs have to be
  34         supported by the driver regardless SSVID/SSID.
  35
  36 6.      Date Thu, 27 Jan 2005 04:31:09 +0100
  37         From Herbert Poetzl <>
  38         Subject RFC: assert_spin_locked() for 2.6
  39
  40         Greetings!
  41
  42         overcautious programming will kill your kernel ;)
  43         ever thought about checking a spin_lock or even
  44         asserting that it must be held (maybe just for
  45         spinlock debugging?) ...
  46
  47         there are several checks present in the kernel
  48         where somebody does a variation on the following:
  49
  50           BUG_ON(!spin_is_locked(&some_lock));
  51
  52         so what's wrong about that? nothing, unless you
  53         compile the code with CONFIG_DEBUG_SPINLOCK but
  54         without CONFIG_SMP ... in which case the BUG()
  55         will kill your kernel ...
  56
  57         maybe it's not advised to make such assertions,
  58         but here is a solution which works for me ...
  59         (compile tested for sh, x86_64 and x86, boot/run
  60         tested for x86 only)
  61
  62         best,
  63         Herbert
  64
  65                 - Herbert Poetzl <herbert@13thfloor.at>, Thu, 27 Jan 2005
  66
  67 Release Date    : Thu Feb 03 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
  68 Current Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
  69 Older Version   : 2.20.4.4 (scsi module), 2.20.2.4 (cmm module)
  70
  71 1.      Modified name of two attributes in scsi_host_template.
  72         On Wed, 2005-02-02 at 10:56 -0500, Ju, Seokmann wrote:
  73         > +     .sdev_attrs                     = megaraid_device_attrs,
  74         > +     .shost_attrs                    = megaraid_class_device_attrs,
  75
  76         These are, perhaps, slightly confusing names.
  77         The terms device and class_device have well defined meanings in the
  78         generic device model, neither of which is what you mean here.
  79         Why not simply megaraid_sdev_attrs and megaraid_shost_attrs?
  80
  81         Other than this, it looks fine to me too.
  82
  83 Release Date    : Thu Jan 27 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
  84 Current Version : 2.20.4.4 (scsi module), 2.20.2.5 (cmm module)
  85 Older Version   : 2.20.4.3 (scsi module), 2.20.2.4 (cmm module)
  86
  87 1.      Bump up the version of scsi module due to its conflict.
  88
  89 Release Date    : Thu Jan 21 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
  90 Current Version : 2.20.4.3 (scsi module), 2.20.2.5 (cmm module)
  91 Older Version   : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
  92
  93 1.      Remove driver ioctl for logical drive to scsi address translation and
  94         replace with the sysfs attribute. To remove drives and change
  95         capacity, application shall now use the device attribute to get the
  96         logical drive number for a scsi device. For adding newly created
  97         logical drives, class device attribute would be required to uniquely
  98         identify each controller.
  99                 - Atul Mukker <atulm@lsil.com>
 100
 101         "James, I've been thinking about this a little more, and you may be on
 102         to something here. Let each driver add files as such:"
 103
 104                 - Matt Domsch <Matt_Domsch@dell.com>, 12.15.2004
 105                  linux-scsi mailing list
 106
 107
 108         "Then, if you simply publish your LD number as an extra parameter of
 109         the device, you can look through /sys to find it."
 110
 111                 - James Bottomley <James.Bottomley@SteelEye.com>, 01.03.2005
 112                  linux-scsi mailing list
 113
 114
 115         "I don't see why not ... it's your driver, you can publish whatever
 116         extra information you need as scsi_device attributes; that was one of
 117         the designs of the extensible attribute system."
 118
 119                 - James Bottomley <James.Bottomley@SteelEye.com>, 01.06.2005
 120                  linux-scsi mailing list
 121
 122 2.      Add AMI megaraid support - Brian King <brking@charter.net>
 123                 PCI_VENDOR_ID_AMI, PCI_DEVICE_ID_AMI_MEGARAID3,
 124                 PCI_VENDOR_ID_AMI, PCI_SUBSYS_ID_PERC3_DC,
 125
 126 3.      Make some code static - Adrian Bunk <bunk@stusta.de>
 127         Date:   Mon, 15 Nov 2004 03:14:57 +0100
 128
 129         The patch below makes some needlessly global code static.
 130         -wait_queue_head_t wait_q;
 131         +static wait_queue_head_t wait_q;
 132
 133         Signed-off-by: Adrian Bunk <bunk@stusta.de>
 134
 135 4.      Added NEC ROMB support - NEC MegaRAID PCI Express ROMB controller
 136                 PCI_VENDOR_ID_LSI_LOGIC, PCI_DEVICE_ID_MEGARAID_NEC_ROMB_2E,
 137                 PCI_SUBSYS_ID_NEC, PCI_SUBSYS_ID_MEGARAID_NEC_ROMB_2E,
 138
 139 5.      Fixed Tape drive issue : For any Direct CDB command to physical device
 140         including tape, timeout value set by driver was 10 minutes. With this
 141         value, most of command will return within timeout. However, for those
 142         command like ERASE or FORMAT, it takes more than an hour depends on
 143         capacity of the device and the command could be terminated before it
 144         completes.
 145         To address this issue, the 'timeout' field in the DCDB command will
 146         have NO TIMEOUT (i.e., 4) value as its timeout on DCDB command.
 147
 148
 149
 150 Release Date    : Thu Dec  9 19:10:23 EST 2004
 151         - Sreenivas Bagalkote <sreenib@lsil.com>
 152
 153 Current Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
 154 Older Version   : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
 155
 156 i.      Introduced driver ioctl that returns scsi address for a given ld.
 157
 158         "Why can't the existing sysfs interfaces be used to do this?"
 159                 - Brian King (brking@us.ibm.com)
 160
 161         "I've looked into solving this another way, but I cannot see how
 162         to get this driver-private mapping of logical drive number-> HCTL
 163         without putting code something like this into the driver."
 164
 165         "...and by providing a mapping a function to userspace, the driver
 166         is free to change its mapping algorithm in the future if necessary .."
 167                 - Matt Domsch (Matt_Domsch@dell.com)
 168
 169 Release Date    : Thu Dec  9 19:02:14 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
 170
 171 Current Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
 172 Older Version   : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
 173
 174 i.      Fix a bug in kioc's dma buffer deallocation
 175
 176 Release Date    : Thu Nov  4 18:24:56 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
 177
 178 Current Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
 179 Older Version   : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
 180
 181 i.      Handle IOCTL cmd timeouts more properly.
 182
 183 ii.     pci_dma_sync_{sg,single}_for_cpu was introduced into megaraid_mbox
 184         incorrectly (instead of _for_device). Changed to appropriate
 185         pci_dma_sync_{sg,single}_for_device.
 186
 187 Release Date    : Wed Oct 06 11:15:29 EDT 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
 188 Current Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
 189 Older Version   : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
 190
 191 i.      Remove CONFIG_COMPAT around register_ioctl32_conversion
 192
 193 Release Date    : Mon Sep 27 22:15:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
 194 Current Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
 195 Older Version   : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
 196
 197 i.      Fix data corruption. Because of a typo in the driver, the IO packets
 198         were wrongly shared by the ioctl path. This causes a whole IO command
 199         to be replaced by an incoming ioctl command.
 200
 201 Release Date    : Tue Aug 24 09:43:35 EDT 2004 - Atul Mukker <atulm@lsil.com>
 202 Current Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
 203 Older Version   : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
 204
 205 i.      Function reordering so that inline functions are defined before they
 206         are actually used. It is now mandatory for GCC 3.4.1 (current stable)
 207
 208         Declare some heavy-weight functions to be non-inlined,
 209         megaraid_mbox_build_cmd, megaraid_mbox_runpendq,
 210         megaraid_mbox_prepare_pthru, megaraid_mbox_prepare_epthru,
 211         megaraid_busywait_mbox
 212
 213                 - Andrew Morton <akpm@osdl.org>, 08.19.2004
 214                 linux-scsi mailing list
 215
 216         "Something else to clean up after inclusion: every instance of an
 217         inline function is actually rendered as a full function call, because
 218         the function is always used before it is defined.  Atul, please
 219         re-arrange the code to eliminate the need for most (all) of the
 220         function prototypes at the top of each file, and define (not just
 221         declare with a prototype) each inline function before its first use"
 222
 223                 - Matt Domsch <Matt_Domsch@dell.com>, 07.27.2004
 224                 linux-scsi mailing list
 225
 226
 227 ii.     Display elapsed time (countdown) while waiting for FW to boot.
 228
 229 iii.    Module compilation reorder in Makefile so that unresolved symbols do
 230         not occur when driver is compiled non-modular.
 231
 232                 Patrick J. LoPresti <patl@users.sourceforge.net>, 8.22.2004
 233                 linux-scsi mailing list
 234
 235
 236 Release Date    : Thu Aug 19 09:58:33 EDT 2004 - Atul Mukker <atulm@lsil.com>
 237 Current Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
 238 Older Version   : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
 239
 240 i.      When copying the mailbox packets, copy only first 14 bytes (for 32-bit
 241         mailboxes) and only first 22 bytes (for 64-bit mailboxes). This is to
 242         avoid getting the stale values for busy bit. We want to set the busy
 243         bit just before issuing command to the FW.
 244
 245 ii.     In the reset handling, if the reseted command is not owned by the
 246         driver, do not (wrongly) print information for the "attached" driver
 247         packet.
 248
 249 iii.    Have extended wait when issuing command in synchronous mode. This is
 250         required for the cases where the option ROM is disabled and there is
 251         no BIOS to start the controller. The FW starts to boot after receiving
 252         the first command from the driver. The current driver has 1 second
 253         timeout for the synchronous commands, which is far less than what is
 254         actually required. We now wait up to MBOX_RESET_TIME (180 seconds) for
 255         FW boot process.
 256
 257 iv.     In megaraid_mbox_product_info, clear the mailbox contents completely
 258         before preparing the command for inquiry3. This is to ensure that the
 259         FW does not get junk values in the command.
 260
 261 v.      Do away with the redundant LSI_CONFIG_COMPAT redefinition for
 262         CONFIG_COMPAT. Replace <asm/ioctl32.h> with <linux/ioctl32.h>
 263
 264                 - James Bottomley <James.Bottomley@SteelEye.com>, 08.17.2004
 265                  linux-scsi mailing list
 266
 267 vi.     Add support for 64-bit applications. Current drivers assume only
 268         32-bit applications, even on 64-bit platforms. Use the "data" and
 269         "buffer" fields of the mimd_t structure, instead of embedded 32-bit
 270         addresses in application mailbox and passthru structures.
 271
 272 vii.    Move the function declarations for the management module from
 273         megaraid_mm.h to megaraid_mm.c
 274
 275                 - Andrew Morton <akpm@osdl.org>, 08.19.2004
 276                 linux-scsi mailing list
 277
 278 viii.   Change default values for MEGARAID_NEWGEN, MEGARAID_MM, and
 279         MEGARAID_MAILBOX to 'n' in Kconfig.megaraid
 280
 281                 - Andrew Morton <akpm@osdl.org>, 08.19.2004
 282                 linux-scsi mailing list
 283
 284 ix.     replace udelay with msleep
 285
 286 x.      Typos corrected in comments and whitespace adjustments, explicit
 287         grouping of expressions.
 288
 289
 290 Release Date    : Fri Jul 23 15:22:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
 291 Current Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
 292 Older Version   : 2.20.1.0 (scsi module), 2.20.0.0 (cmm module)
 293
 294 i.      Add PCI ids for Acer ROMB 2E solution
 295
 296 ii.     Add PCI ids for I4
 297
 298 iii.    Typo corrected for subsys id for megaraid sata 300-4x
 299
 300 iv.     Remove yield() while mailbox handshake in synchronous commands
 301
 302
 303         "My other main gripe is things like this:
 304
 305         +       // wait for maximum 1 second for status to post
 306         +       for (i = 0; i < 40000; i++) {
 307         +               if (mbox->numstatus != 0xFF) break;
 308         +               udelay(25); yield();
 309         +       }
 310
 311         which litter the driver.  Use of yield() in drivers is deprecated."
 312
 313                 - James Bottomley <James.Bottomley@SteelEye.com>, 07.14.2004
 314                  linux-scsi mailing list
 315
 316 v.      Remove redundant __megaraid_busywait_mbox routine
 317
 318 vi.     Fix bug in the managment module, which causes a system lockup when the
 319         IO module is loaded and then unloaded, followed by executing any
 320         management utility. The current version of management module does not
 321         handle the adapter unregister properly.
 322
 323         Specifically, it still keeps a reference to the unregistered
 324         controllers. To avoid this, the static array adapters has been
 325         replaced by a dynamic list, which gets updated every time an adapter
 326         is added or removed.
 327
 328         Also, during unregistration of the IO module, the resources are
 329         now released in the exact reverse order of the allocation time
 330         sequence.
 331
 332
 333 Release Date    : Fri Jun 25 18:58:43 EDT 2004 - Atul Mukker <atulm@lsil.com>
 334 Current Version : 2.20.1.0
 335 Older Version   : megaraid 2.20.0.1
 336
 337 i.      Stale list pointer in adapter causes kernel panic when module
 338         megaraid_mbox is unloaded
 339
 340
 341 Release Date    : Thu Jun 24 20:37:11 EDT 2004 - Atul Mukker <atulm@lsil.com>
 342 Current Version : 2.20.0.1
 343 Older Version   : megaraid 2.20.0.00
 344
 345 i.      Modules are not 'y' by default, but depend on current definition of
 346         SCSI & PCI.
 347
 348 ii.     Redundant structure mraid_driver_t removed.
 349
 350 iii.    Miscellaneous indentation and goto/label fixes.
 351                 - Christoph Hellwig <hch@infradead.org>, 06.24.2004 linux-scsi
 352
 353 iv.     scsi_host_put(), do just before completing HBA shutdown.
 354
 355
 356
 357 Release Date    : Mon Jun 21 19:53:54 EDT 2004 - Atul Mukker <atulm@lsil.com>
 358 Current Version : 2.20.0.0
 359 Older Version   : megaraid 2.20.0.rc2 and 2.00.3
 360
 361 i.      Independent module to interact with userland applications and
 362         multiplex command to low level RAID module(s).
 363
 364         "Shared code in a third module, a "library module", is an acceptable
 365         solution. modprobe automatically loads dependent modules, so users
 366         running "modprobe driver1" or "modprobe driver2" would automatically
 367         load the shared library module."
 368
 369                 - Jeff Garzik <jgarzik@pobox.com> 02.25.2004 LKML
 370
 371         "As Jeff hinted, if your userspace<->driver API is consistent between
 372         your new MPT-based RAID controllers and your existing megaraid driver,
 373         then perhaps you need a single small helper module (lsiioctl or some
 374         better name), loaded by both mptraid and megaraid automatically, which
 375         handles registering the /dev/megaraid node dynamically. In this case,
 376         both mptraid and megaraid would register with lsiioctl for each
 377         adapter discovered, and lsiioctl would essentially be a switch,
 378         redirecting userspace tool ioctls to the appropriate driver."
 379
 380                 - Matt Domsch <Matt_Domsch@dell.com> 02.25.2004 LKML
 381
 382 ii.     Remove C99 initializations from pci_device id.
 383
 384         "pci_id_table_g would be much more readable when not using C99
 385         initializers.
 386         PCI table doesn't change, there's lots of users that prefer the more
 387         readable variant.  And it's really far less and much easier to grok
 388         lines without C99 initializers."
 389
 390                 - Christoph Hellwig <hch@infradead.org>, 05.28.2004 linux-scsi
 391
 392 iii.    Many fixes as suggested by Christoph Hellwig <hch@infradead.org> on
 393         linux-scsi, 05.28.2004
 394
 395 iv.     We now support up to 32 parallel ioctl commands instead of current 1.
 396         There is a conscious effort to let memory allocation not fail for ioctl
 397         commands.
 398
 399 v.      Do away with internal memory management. Use pci_pool_(create|alloc)
 400         instead.
 401
 402 vi.     Kill tasklet when unloading the driver.
 403
 404 vii.    Do not use "host_lock', driver has fine-grain locks now to protect all
 405         data structures.
 406
 407 viii.   Optimize the build scatter-gather list routine. The callers already
 408         know the data transfer address and length.
 409
 410 ix.     Better implementation of error handling and recovery. Driver now
 411         performs extended errors recovery for instances like scsi cable pull.
 412
 413 x.      Disassociate the management commands with an overlaid scsi command.
 414         Driver now treats the management packets as special packets and has a
 415         dedicated callback routine.