1Release Date : Fri Nov 11 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com> 2Current Version : 2.20.4.7 (scsi module), 2.20.2.6 (cmm module) 3Older Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module) 4 51. Sorted out PCI IDs to remove megaraid support overlaps. 6 Based on the patch from Daniel, sorted out PCI IDs along with 7 charactor node name change from 'megadev' to 'megadev_legacy' to avoid 8 conflict. 9 --- 10 Hopefully we'll be getting the build restriction zapped much sooner, 11 but we should also be thinking about totally removing the hardware 12 support overlap in the megaraid drivers. 13 14 This patch pencils in a date of Feb 06 for this, and performs some 15 printk abuse in hope that existing legacy users might pick up on what's 16 going on. 17 18 Signed-off-by: Daniel Drake <dsd@gentoo.org> 19 --- 20 212. Fixed a issue: megaraid always fails to reset handler. 22 --- 23 I found that the megaraid driver always fails to reset the 24 adapter with the following message: 25 megaraid: resetting the host... 26 megaraid mbox: reset sequence completed successfully 27 megaraid: fast sync command timed out 28 megaraid: reservation reset failed 29 when the "Cluster mode" of the adapter BIOS is enabled. 30 So, whenever the reset occurs, the adapter goes to 31 offline and just become unavailable. 32 33 Jun'ichi Nomura [mailto:jnomura@mtc.biglobe.ne.jp] 34 --- 35 36Release Date : Mon Mar 07 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com> 37Current Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module) 38Older Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module) 39 401. Added IOCTL backward compatibility. 41 Convert megaraid_mm driver to new compat_ioctl entry points. 42 I don't have easy access to hardware, so only compile tested. 43 - Signed-off-by:Andi Kleen <ak@muc.de> 44 452. megaraid_mbox fix: wrong order of arguments in memset() 46 That, BTW, shows why cross-builds are useful-the only indication of 47 problem had been a new warning showing up in sparse output on alpha 48 build (number of exceeding 256 got truncated). 49 - Signed-off-by: Al Viro 50 <viro@parcelfarce.linux.theplanet.co.uk> 51 523. Convert pci_module_init to pci_register_driver 53 Convert from pci_module_init to pci_register_driver 54 (from:http://kerneljanitors.org/TODO) 55 - Signed-off-by: Domen Puncer <domen@coderock.org> 56 574. Use the pre defined DMA mask constants from dma-mapping.h 58 Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling 59 pci_set_dma_mask() or pci_set_consistend_dma_mask(). See 60 http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for more 61 details. 62 Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch> 63 Signed-off-by: Domen Puncer <domen@coderock.org> 64 655. Remove SSID checking for Dobson, Lindsay, and Verde based products. 66 Checking the SSVID/SSID for controllers which have Dobson, Lindsay, 67 and Verde is unnecessary because device ID has been assigned by LSI 68 and it is unique value. So, all controllers with these IOPs have to be 69 supported by the driver regardless SSVID/SSID. 70 716. Date Thu, 27 Jan 2005 04:31:09 +0100 72 From Herbert Poetzl <> 73 Subject RFC: assert_spin_locked() for 2.6 74 75 Greetings! 76 77 overcautious programming will kill your kernel ;) 78 ever thought about checking a spin_lock or even 79 asserting that it must be held (maybe just for 80 spinlock debugging?) ... 81 82 there are several checks present in the kernel 83 where somebody does a variation on the following: 84 85 BUG_ON(!spin_is_locked(&some_lock)); 86 87 so what's wrong about that? nothing, unless you 88 compile the code with CONFIG_DEBUG_SPINLOCK but 89 without CONFIG_SMP ... in which case the BUG() 90 will kill your kernel ... 91 92 maybe it's not advised to make such assertions, 93 but here is a solution which works for me ... 94 (compile tested for sh, x86_64 and x86, boot/run 95 tested for x86 only) 96 97 best, 98 Herbert 99 100 - Herbert Poetzl <herbert@13thfloor.at>, Thu, 27 Jan 2005 101 102Release Date : Thu Feb 03 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com> 103Current Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module) 104Older Version : 2.20.4.4 (scsi module), 2.20.2.4 (cmm module) 105 1061. Modified name of two attributes in scsi_host_template. 107 On Wed, 2005-02-02 at 10:56 -0500, Ju, Seokmann wrote: 108 > + .sdev_attrs = megaraid_device_attrs, 109 > + .shost_attrs = megaraid_class_device_attrs, 110 111 These are, perhaps, slightly confusing names. 112 The terms device and class_device have well defined meanings in the 113 generic device model, neither of which is what you mean here. 114 Why not simply megaraid_sdev_attrs and megaraid_shost_attrs? 115 116 Other than this, it looks fine to me too. 117 118Release Date : Thu Jan 27 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com> 119Current Version : 2.20.4.4 (scsi module), 2.20.2.5 (cmm module) 120Older Version : 2.20.4.3 (scsi module), 2.20.2.4 (cmm module) 121 1221. Bump up the version of scsi module due to its conflict. 123 124Release Date : Thu Jan 21 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com> 125Current Version : 2.20.4.3 (scsi module), 2.20.2.5 (cmm module) 126Older Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module) 127 1281. Remove driver ioctl for logical drive to scsi address translation and 129 replace with the sysfs attribute. To remove drives and change 130 capacity, application shall now use the device attribute to get the 131 logical drive number for a scsi device. For adding newly created 132 logical drives, class device attribute would be required to uniquely 133 identify each controller. 134 - Atul Mukker <atulm@lsil.com> 135 136 "James, I've been thinking about this a little more, and you may be on 137 to something here. Let each driver add files as such:" 138 139 - Matt Domsch <Matt_Domsch@dell.com>, 12.15.2004 140 linux-scsi mailing list 141 142 143 "Then, if you simply publish your LD number as an extra parameter of 144 the device, you can look through /sys to find it." 145 146 - James Bottomley <James.Bottomley@SteelEye.com>, 01.03.2005 147 linux-scsi mailing list 148 149 150 "I don't see why not ... it's your driver, you can publish whatever 151 extra information you need as scsi_device attributes; that was one of 152 the designs of the extensible attribute system." 153 154 - James Bottomley <James.Bottomley@SteelEye.com>, 01.06.2005 155 linux-scsi mailing list 156 1572. Add AMI megaraid support - Brian King <brking@charter.net> 158 PCI_VENDOR_ID_AMI, PCI_DEVICE_ID_AMI_MEGARAID3, 159 PCI_VENDOR_ID_AMI, PCI_SUBSYS_ID_PERC3_DC, 160 1613. Make some code static - Adrian Bunk <bunk@stusta.de> 162 Date: Mon, 15 Nov 2004 03:14:57 +0100 163 164 The patch below makes some needlessly global code static. 165 -wait_queue_head_t wait_q; 166 +static wait_queue_head_t wait_q; 167 168 Signed-off-by: Adrian Bunk <bunk@stusta.de> 169 1704. Added NEC ROMB support - NEC MegaRAID PCI Express ROMB controller 171 PCI_VENDOR_ID_LSI_LOGIC, PCI_DEVICE_ID_MEGARAID_NEC_ROMB_2E, 172 PCI_SUBSYS_ID_NEC, PCI_SUBSYS_ID_MEGARAID_NEC_ROMB_2E, 173 1745. Fixed Tape drive issue : For any Direct CDB command to physical device 175 including tape, timeout value set by driver was 10 minutes. With this 176 value, most of command will return within timeout. However, for those 177 command like ERASE or FORMAT, it takes more than an hour depends on 178 capacity of the device and the command could be terminated before it 179 completes. 180 To address this issue, the 'timeout' field in the DCDB command will 181 have NO TIMEOUT (i.e., 4) value as its timeout on DCDB command. 182 183 184 185Release Date : Thu Dec 9 19:10:23 EST 2004 186 - Sreenivas Bagalkote <sreenib@lsil.com> 187 188Current Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module) 189Older Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module) 190 191i. Introduced driver ioctl that returns scsi address for a given ld. 192 193 "Why can't the existing sysfs interfaces be used to do this?" 194 - Brian King (brking@us.ibm.com) 195 196 "I've looked into solving this another way, but I cannot see how 197 to get this driver-private mapping of logical drive number-> HCTL 198 without putting code something like this into the driver." 199 200 "...and by providing a mapping a function to userspace, the driver 201 is free to change its mapping algorithm in the future if necessary .." 202 - Matt Domsch (Matt_Domsch@dell.com) 203 204Release Date : Thu Dec 9 19:02:14 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com> 205 206Current Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module) 207Older Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module) 208 209i. Fix a bug in kioc's dma buffer deallocation 210 211Release Date : Thu Nov 4 18:24:56 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com> 212 213Current Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module) 214Older Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module) 215 216i. Handle IOCTL cmd timeouts more properly. 217 218ii. pci_dma_sync_{sg,single}_for_cpu was introduced into megaraid_mbox 219 incorrectly (instead of _for_device). Changed to appropriate 220 pci_dma_sync_{sg,single}_for_device. 221 222Release Date : Wed Oct 06 11:15:29 EDT 2004 - Sreenivas Bagalkote <sreenib@lsil.com> 223Current Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module) 224Older Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module) 225 226i. Remove CONFIG_COMPAT around register_ioctl32_conversion 227 228Release Date : Mon Sep 27 22:15:07 EDT 2004 - Atul Mukker <atulm@lsil.com> 229Current Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module) 230Older Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module) 231 232i. Fix data corruption. Because of a typo in the driver, the IO packets 233 were wrongly shared by the ioctl path. This causes a whole IO command 234 to be replaced by an incoming ioctl command. 235 236Release Date : Tue Aug 24 09:43:35 EDT 2004 - Atul Mukker <atulm@lsil.com> 237Current Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module) 238Older Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module) 239 240i. Function reordering so that inline functions are defined before they 241 are actually used. It is now mandatory for GCC 3.4.1 (current stable) 242 243 Declare some heavy-weight functions to be non-inlined, 244 megaraid_mbox_build_cmd, megaraid_mbox_runpendq, 245 megaraid_mbox_prepare_pthru, megaraid_mbox_prepare_epthru, 246 megaraid_busywait_mbox 247 248 - Andrew Morton <akpm@osdl.org>, 08.19.2004 249 linux-scsi mailing list 250 251 "Something else to clean up after inclusion: every instance of an 252 inline function is actually rendered as a full function call, because 253 the function is always used before it is defined. Atul, please 254 re-arrange the code to eliminate the need for most (all) of the 255 function prototypes at the top of each file, and define (not just 256 declare with a prototype) each inline function before its first use" 257 258 - Matt Domsch <Matt_Domsch@dell.com>, 07.27.2004 259 linux-scsi mailing list 260 261 262ii. Display elapsed time (countdown) while waiting for FW to boot. 263 264iii. Module compilation reorder in Makefile so that unresolved symbols do 265 not occur when driver is compiled non-modular. 266 267 Patrick J. LoPresti <patl@users.sourceforge.net>, 8.22.2004 268 linux-scsi mailing list 269 270 271Release Date : Thu Aug 19 09:58:33 EDT 2004 - Atul Mukker <atulm@lsil.com> 272Current Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module) 273Older Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module) 274 275i. When copying the mailbox packets, copy only first 14 bytes (for 32-bit 276 mailboxes) and only first 22 bytes (for 64-bit mailboxes). This is to 277 avoid getting the stale values for busy bit. We want to set the busy 278 bit just before issuing command to the FW. 279 280ii. In the reset handling, if the reseted command is not owned by the 281 driver, do not (wrongly) print information for the "attached" driver 282 packet. 283 284iii. Have extended wait when issuing command in synchronous mode. This is 285 required for the cases where the option ROM is disabled and there is 286 no BIOS to start the controller. The FW starts to boot after receiving 287 the first command from the driver. The current driver has 1 second 288 timeout for the synchronous commands, which is far less than what is 289 actually required. We now wait up to MBOX_RESET_TIME (180 seconds) for 290 FW boot process. 291 292iv. In megaraid_mbox_product_info, clear the mailbox contents completely 293 before preparing the command for inquiry3. This is to ensure that the 294 FW does not get junk values in the command. 295 296v. Do away with the redundant LSI_CONFIG_COMPAT redefinition for 297 CONFIG_COMPAT. Replace <asm/ioctl32.h> with <linux/ioctl32.h> 298 299 - James Bottomley <James.Bottomley@SteelEye.com>, 08.17.2004 300 linux-scsi mailing list 301 302vi. Add support for 64-bit applications. Current drivers assume only 303 32-bit applications, even on 64-bit platforms. Use the "data" and 304 "buffer" fields of the mimd_t structure, instead of embedded 32-bit 305 addresses in application mailbox and passthru structures. 306 307vii. Move the function declarations for the management module from 308 megaraid_mm.h to megaraid_mm.c 309 310 - Andrew Morton <akpm@osdl.org>, 08.19.2004 311 linux-scsi mailing list 312 313viii. Change default values for MEGARAID_NEWGEN, MEGARAID_MM, and 314 MEGARAID_MAILBOX to 'n' in Kconfig.megaraid 315 316 - Andrew Morton <akpm@osdl.org>, 08.19.2004 317 linux-scsi mailing list 318 319ix. replace udelay with msleep 320 321x. Typos corrected in comments and whitespace adjustments, explicit 322 grouping of expressions. 323 324 325Release Date : Fri Jul 23 15:22:07 EDT 2004 - Atul Mukker <atulm@lsil.com> 326Current Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module) 327Older Version : 2.20.1.0 (scsi module), 2.20.0.0 (cmm module) 328 329i. Add PCI ids for Acer ROMB 2E solution 330 331ii. Add PCI ids for I4 332 333iii. Typo corrected for subsys id for megaraid sata 300-4x 334 335iv. Remove yield() while mailbox handshake in synchronous commands 336 337 338 "My other main gripe is things like this: 339 340 + // wait for maximum 1 second for status to post 341 + for (i = 0; i < 40000; i++) { 342 + if (mbox->numstatus != 0xFF) break; 343 + udelay(25); yield(); 344 + } 345 346 which litter the driver. Use of yield() in drivers is deprecated." 347 348 - James Bottomley <James.Bottomley@SteelEye.com>, 07.14.2004 349 linux-scsi mailing list 350 351v. Remove redundant __megaraid_busywait_mbox routine 352 353vi. Fix bug in the managment module, which causes a system lockup when the 354 IO module is loaded and then unloaded, followed by executing any 355 management utility. The current version of management module does not 356 handle the adapter unregister properly. 357 358 Specifically, it still keeps a reference to the unregistered 359 controllers. To avoid this, the static array adapters has been 360 replaced by a dynamic list, which gets updated every time an adapter 361 is added or removed. 362 363 Also, during unregistration of the IO module, the resources are 364 now released in the exact reverse order of the allocation time 365 sequence. 366 367 368Release Date : Fri Jun 25 18:58:43 EDT 2004 - Atul Mukker <atulm@lsil.com> 369Current Version : 2.20.1.0 370Older Version : megaraid 2.20.0.1 371 372i. Stale list pointer in adapter causes kernel panic when module 373 megaraid_mbox is unloaded 374 375 376Release Date : Thu Jun 24 20:37:11 EDT 2004 - Atul Mukker <atulm@lsil.com> 377Current Version : 2.20.0.1 378Older Version : megaraid 2.20.0.00 379 380i. Modules are not 'y' by default, but depend on current definition of 381 SCSI & PCI. 382 383ii. Redundant structure mraid_driver_t removed. 384 385iii. Miscellaneous indentation and goto/label fixes. 386 - Christoph Hellwig <hch@infradead.org>, 06.24.2004 linux-scsi 387 388iv. scsi_host_put(), do just before completing HBA shutdown. 389 390 391 392Release Date : Mon Jun 21 19:53:54 EDT 2004 - Atul Mukker <atulm@lsil.com> 393Current Version : 2.20.0.0 394Older Version : megaraid 2.20.0.rc2 and 2.00.3 395 396i. Independent module to interact with userland applications and 397 multiplex command to low level RAID module(s). 398 399 "Shared code in a third module, a "library module", is an acceptable 400 solution. modprobe automatically loads dependent modules, so users 401 running "modprobe driver1" or "modprobe driver2" would automatically 402 load the shared library module." 403 404 - Jeff Garzik <jgarzik@pobox.com> 02.25.2004 LKML 405 406 "As Jeff hinted, if your userspace<->driver API is consistent between 407 your new MPT-based RAID controllers and your existing megaraid driver, 408 then perhaps you need a single small helper module (lsiioctl or some 409 better name), loaded by both mptraid and megaraid automatically, which 410 handles registering the /dev/megaraid node dynamically. In this case, 411 both mptraid and megaraid would register with lsiioctl for each 412 adapter discovered, and lsiioctl would essentially be a switch, 413 redirecting userspace tool ioctls to the appropriate driver." 414 415 - Matt Domsch <Matt_Domsch@dell.com> 02.25.2004 LKML 416 417ii. Remove C99 initializations from pci_device id. 418 419 "pci_id_table_g would be much more readable when not using C99 420 initializers. 421 PCI table doesn't change, there's lots of users that prefer the more 422 readable variant. And it's really far less and much easier to grok 423 lines without C99 initializers." 424 425 - Christoph Hellwig <hch@infradead.org>, 05.28.2004 linux-scsi 426 427iii. Many fixes as suggested by Christoph Hellwig <hch@infradead.org> on 428 linux-scsi, 05.28.2004 429 430iv. We now support up to 32 parallel ioctl commands instead of current 1. 431 There is a conscious effort to let memory allocation not fail for ioctl 432 commands. 433 434v. Do away with internal memory management. Use pci_pool_(create|alloc) 435 instead. 436 437vi. Kill tasklet when unloading the driver. 438 439vii. Do not use "host_lock', driver has fine-grain locks now to protect all 440 data structures. 441 442viii. Optimize the build scatter-gather list routine. The callers already 443 know the data transfer address and length. 444 445ix. Better implementation of error handling and recovery. Driver now 446 performs extended errors recovery for instances like scsi cable pull. 447 448x. Disassociate the management commands with an overlaid scsi command. 449 Driver now treats the management packets as special packets and has a 450 dedicated callback routine. 451