1 Release Date : Fri May 19 09:31:45 EST 2006 - Seokmann Ju <sju@lsil.com>
2 Current Version : 2.20.4.9 (scsi module), 2.20.2.6 (cmm module)
3 Older Version : 2.20.4.8 (scsi module), 2.20.2.6 (cmm module)
5 1. Fixed a bug in megaraid_init_mbox().
6 Customer reported "garbage in file on x86_64 platform".
7 Root Cause: the driver registered controllers as 64-bit DMA capable
8 for those which are not support it.
9 Fix: Made change in the function inserting identification machanism
10 identifying 64-bit DMA capable controllers.
12 > -----Original Message-----
13 > From: Vasily Averin [mailto:vvs@sw.ru]
14 > Sent: Thursday, May 04, 2006 2:49 PM
15 > To: linux-scsi@vger.kernel.org; Kolli, Neela; Mukker, Atul;
16 > Ju, Seokmann; Bagalkote, Sreenivas;
17 > James.Bottomley@SteelEye.com; devel@openvz.org
18 > Subject: megaraid_mbox: garbage in file
22 > I've investigated customers claim on the unstable work of
23 > their node and found a
24 > strange effect: reading from some files leads to the
25 > "attempt to access beyond end of device" messages.
27 > I've checked filesystem, memory on the node, motherboard BIOS
29 > does not help and issue still has been reproduced by simple
32 > Reproducer is simple:
34 > echo 0xffffffff >/proc/sys/dev/scsi/logging_level ;
35 > cat /vz/private/101/root/etc/ld.so.cache >/tmp/ttt ;
36 > echo 0 >/proc/sys/dev/scsi/logging
38 > It leads to the following messages in dmesg
40 > sd_init_command: disk=sda, block=871769260, count=26
41 > sda : block=871769260
42 > sda : reading 26/26 512 byte blocks.
43 > scsi_add_timer: scmd: f79ed980, time: 7500, (c02b1420)
44 > sd 0:1:0:0: send 0xf79ed980 sd 0:1:0:0:
45 > command: Read (10): 28 00 33 f6 24 ac 00 00 1a 00
46 > buffer = 0xf7cfb540, bufflen = 13312, done = 0xc0366b40,
47 > queuecommand 0xc0344010
48 > leaving scsi_dispatch_cmnd()
49 > scsi_delete_timer: scmd: f79ed980, rtn: 1
50 > sd 0:1:0:0: done 0xf79ed980 SUCCESS 0 sd 0:1:0:0:
51 > command: Read (10): 28 00 33 f6 24 ac 00 00 1a 00
52 > scsi host busy 1 failed 0
53 > sd 0:1:0:0: Notifying upper driver of completion (result 0)
54 > sd_rw_intr: sda: res=0x0
55 > 26 sectors total, 13312 bytes done.
57 > attempt to access beyond end of device
58 > sda6: rw=0, want=1044134458, limit=951401367
59 > Buffer I/O error on device sda6, logical block 522067228
60 > attempt to access beyond end of device
62 2. When INQUIRY with EVPD bit set issued to the MegaRAID controller,
63 system memory gets corrupted.
64 Root Cause: MegaRAID F/W handle the INQUIRY with EVPD bit set
66 Fix: MegaRAID F/W has fixed the problem and being process of release,
67 soon. Meanwhile, driver will filter out the request.
69 Release Date : Mon Apr 11 12:27:22 EST 2006 - Seokmann Ju <sju@lsil.com>
70 Current Version : 2.20.4.8 (scsi module), 2.20.2.6 (cmm module)
71 Older Version : 2.20.4.7 (scsi module), 2.20.2.6 (cmm module)
73 1. Fixed a bug in megaraid_reset_handler().
74 Customer reported "Unable to handle kernel NULL pointer dereference
75 at virtual address 00000000" when system goes to reset condition
76 for some reason. It happened randomly.
77 Root Cause: in the megaraid_reset_handler(), there is possibility not
78 returning pending packets in the pend_list if there are multiple
80 Fix: Made the change in the driver so that it will return all packets
83 2. Added change request.
84 As found in the following URL, rmb() only didn't help the
85 problem. I had to increase the loop counter to 0xFFFFFF. (6 F's)
86 http://marc.theaimsgroup.com/?l=linux-scsi&m=110971060502497&w=2
88 I attached a patch for your reference, too.
89 Could you check and get this fix in your driver?
94 Release Date : Fri Nov 11 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
95 Current Version : 2.20.4.7 (scsi module), 2.20.2.6 (cmm module)
96 Older Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module)
98 1. Sorted out PCI IDs to remove megaraid support overlaps.
99 Based on the patch from Daniel, sorted out PCI IDs along with
100 charactor node name change from 'megadev' to 'megadev_legacy' to avoid
103 Hopefully we'll be getting the build restriction zapped much sooner,
104 but we should also be thinking about totally removing the hardware
105 support overlap in the megaraid drivers.
107 This patch pencils in a date of Feb 06 for this, and performs some
108 printk abuse in hope that existing legacy users might pick up on what's
111 Signed-off-by: Daniel Drake <dsd@gentoo.org>
114 2. Fixed a issue: megaraid always fails to reset handler.
116 I found that the megaraid driver always fails to reset the
117 adapter with the following message:
118 megaraid: resetting the host...
119 megaraid mbox: reset sequence completed successfully
120 megaraid: fast sync command timed out
121 megaraid: reservation reset failed
122 when the "Cluster mode" of the adapter BIOS is enabled.
123 So, whenever the reset occurs, the adapter goes to
124 offline and just become unavailable.
126 Jun'ichi Nomura [mailto:jnomura@mtc.biglobe.ne.jp]
129 Release Date : Mon Mar 07 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
130 Current Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module)
131 Older Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
133 1. Added IOCTL backward compatibility.
134 Convert megaraid_mm driver to new compat_ioctl entry points.
135 I don't have easy access to hardware, so only compile tested.
136 - Signed-off-by:Andi Kleen <ak@muc.de>
138 2. megaraid_mbox fix: wrong order of arguments in memset()
139 That, BTW, shows why cross-builds are useful-the only indication of
140 problem had been a new warning showing up in sparse output on alpha
141 build (number of exceeding 256 got truncated).
142 - Signed-off-by: Al Viro
143 <viro@parcelfarce.linux.theplanet.co.uk>
145 3. Convert pci_module_init to pci_register_driver
146 Convert from pci_module_init to pci_register_driver
147 (from:http://kerneljanitors.org/TODO)
148 - Signed-off-by: Domen Puncer <domen@coderock.org>
150 4. Use the pre defined DMA mask constants from dma-mapping.h
151 Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
152 pci_set_dma_mask() or pci_set_consistend_dma_mask(). See
153 http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for more
155 Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
156 Signed-off-by: Domen Puncer <domen@coderock.org>
158 5. Remove SSID checking for Dobson, Lindsay, and Verde based products.
159 Checking the SSVID/SSID for controllers which have Dobson, Lindsay,
160 and Verde is unnecessary because device ID has been assigned by LSI
161 and it is unique value. So, all controllers with these IOPs have to be
162 supported by the driver regardless SSVID/SSID.
164 6. Date Thu, 27 Jan 2005 04:31:09 +0100
165 From Herbert Poetzl <>
166 Subject RFC: assert_spin_locked() for 2.6
170 overcautious programming will kill your kernel ;)
171 ever thought about checking a spin_lock or even
172 asserting that it must be held (maybe just for
173 spinlock debugging?) ...
175 there are several checks present in the kernel
176 where somebody does a variation on the following:
178 BUG_ON(!spin_is_locked(&some_lock));
180 so what's wrong about that? nothing, unless you
181 compile the code with CONFIG_DEBUG_SPINLOCK but
182 without CONFIG_SMP ... in which case the BUG()
183 will kill your kernel ...
185 maybe it's not advised to make such assertions,
186 but here is a solution which works for me ...
187 (compile tested for sh, x86_64 and x86, boot/run
193 - Herbert Poetzl <herbert@13thfloor.at>, Thu, 27 Jan 2005
195 Release Date : Thu Feb 03 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
196 Current Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
197 Older Version : 2.20.4.4 (scsi module), 2.20.2.4 (cmm module)
199 1. Modified name of two attributes in scsi_host_template.
200 On Wed, 2005-02-02 at 10:56 -0500, Ju, Seokmann wrote:
201 > + .sdev_attrs = megaraid_device_attrs,
202 > + .shost_attrs = megaraid_class_device_attrs,
204 These are, perhaps, slightly confusing names.
205 The terms device and class_device have well defined meanings in the
206 generic device model, neither of which is what you mean here.
207 Why not simply megaraid_sdev_attrs and megaraid_shost_attrs?
209 Other than this, it looks fine to me too.
211 Release Date : Thu Jan 27 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
212 Current Version : 2.20.4.4 (scsi module), 2.20.2.5 (cmm module)
213 Older Version : 2.20.4.3 (scsi module), 2.20.2.4 (cmm module)
215 1. Bump up the version of scsi module due to its conflict.
217 Release Date : Thu Jan 21 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
218 Current Version : 2.20.4.3 (scsi module), 2.20.2.5 (cmm module)
219 Older Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
221 1. Remove driver ioctl for logical drive to scsi address translation and
222 replace with the sysfs attribute. To remove drives and change
223 capacity, application shall now use the device attribute to get the
224 logical drive number for a scsi device. For adding newly created
225 logical drives, class device attribute would be required to uniquely
226 identify each controller.
227 - Atul Mukker <atulm@lsil.com>
229 "James, I've been thinking about this a little more, and you may be on
230 to something here. Let each driver add files as such:"
232 - Matt Domsch <Matt_Domsch@dell.com>, 12.15.2004
233 linux-scsi mailing list
236 "Then, if you simply publish your LD number as an extra parameter of
237 the device, you can look through /sys to find it."
239 - James Bottomley <James.Bottomley@SteelEye.com>, 01.03.2005
240 linux-scsi mailing list
243 "I don't see why not ... it's your driver, you can publish whatever
244 extra information you need as scsi_device attributes; that was one of
245 the designs of the extensible attribute system."
247 - James Bottomley <James.Bottomley@SteelEye.com>, 01.06.2005
248 linux-scsi mailing list
250 2. Add AMI megaraid support - Brian King <brking@charter.net>
251 PCI_VENDOR_ID_AMI, PCI_DEVICE_ID_AMI_MEGARAID3,
252 PCI_VENDOR_ID_AMI, PCI_SUBSYS_ID_PERC3_DC,
254 3. Make some code static - Adrian Bunk <bunk@stusta.de>
255 Date: Mon, 15 Nov 2004 03:14:57 +0100
257 The patch below makes some needlessly global code static.
258 -wait_queue_head_t wait_q;
259 +static wait_queue_head_t wait_q;
261 Signed-off-by: Adrian Bunk <bunk@stusta.de>
263 4. Added NEC ROMB support - NEC MegaRAID PCI Express ROMB controller
264 PCI_VENDOR_ID_LSI_LOGIC, PCI_DEVICE_ID_MEGARAID_NEC_ROMB_2E,
265 PCI_SUBSYS_ID_NEC, PCI_SUBSYS_ID_MEGARAID_NEC_ROMB_2E,
267 5. Fixed Tape drive issue : For any Direct CDB command to physical device
268 including tape, timeout value set by driver was 10 minutes. With this
269 value, most of command will return within timeout. However, for those
270 command like ERASE or FORMAT, it takes more than an hour depends on
271 capacity of the device and the command could be terminated before it
273 To address this issue, the 'timeout' field in the DCDB command will
274 have NO TIMEOUT (i.e., 4) value as its timeout on DCDB command.
278 Release Date : Thu Dec 9 19:10:23 EST 2004
279 - Sreenivas Bagalkote <sreenib@lsil.com>
281 Current Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
282 Older Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
284 i. Introduced driver ioctl that returns scsi address for a given ld.
286 "Why can't the existing sysfs interfaces be used to do this?"
287 - Brian King (brking@us.ibm.com)
289 "I've looked into solving this another way, but I cannot see how
290 to get this driver-private mapping of logical drive number-> HCTL
291 without putting code something like this into the driver."
293 "...and by providing a mapping a function to userspace, the driver
294 is free to change its mapping algorithm in the future if necessary .."
295 - Matt Domsch (Matt_Domsch@dell.com)
297 Release Date : Thu Dec 9 19:02:14 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
299 Current Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
300 Older Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
302 i. Fix a bug in kioc's dma buffer deallocation
304 Release Date : Thu Nov 4 18:24:56 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
306 Current Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
307 Older Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
309 i. Handle IOCTL cmd timeouts more properly.
311 ii. pci_dma_sync_{sg,single}_for_cpu was introduced into megaraid_mbox
312 incorrectly (instead of _for_device). Changed to appropriate
313 pci_dma_sync_{sg,single}_for_device.
315 Release Date : Wed Oct 06 11:15:29 EDT 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
316 Current Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
317 Older Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
319 i. Remove CONFIG_COMPAT around register_ioctl32_conversion
321 Release Date : Mon Sep 27 22:15:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
322 Current Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
323 Older Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
325 i. Fix data corruption. Because of a typo in the driver, the IO packets
326 were wrongly shared by the ioctl path. This causes a whole IO command
327 to be replaced by an incoming ioctl command.
329 Release Date : Tue Aug 24 09:43:35 EDT 2004 - Atul Mukker <atulm@lsil.com>
330 Current Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
331 Older Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
333 i. Function reordering so that inline functions are defined before they
334 are actually used. It is now mandatory for GCC 3.4.1 (current stable)
336 Declare some heavy-weight functions to be non-inlined,
337 megaraid_mbox_build_cmd, megaraid_mbox_runpendq,
338 megaraid_mbox_prepare_pthru, megaraid_mbox_prepare_epthru,
339 megaraid_busywait_mbox
341 - Andrew Morton <akpm@osdl.org>, 08.19.2004
342 linux-scsi mailing list
344 "Something else to clean up after inclusion: every instance of an
345 inline function is actually rendered as a full function call, because
346 the function is always used before it is defined. Atul, please
347 re-arrange the code to eliminate the need for most (all) of the
348 function prototypes at the top of each file, and define (not just
349 declare with a prototype) each inline function before its first use"
351 - Matt Domsch <Matt_Domsch@dell.com>, 07.27.2004
352 linux-scsi mailing list
355 ii. Display elapsed time (countdown) while waiting for FW to boot.
357 iii. Module compilation reorder in Makefile so that unresolved symbols do
358 not occur when driver is compiled non-modular.
360 Patrick J. LoPresti <patl@users.sourceforge.net>, 8.22.2004
361 linux-scsi mailing list
364 Release Date : Thu Aug 19 09:58:33 EDT 2004 - Atul Mukker <atulm@lsil.com>
365 Current Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
366 Older Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
368 i. When copying the mailbox packets, copy only first 14 bytes (for 32-bit
369 mailboxes) and only first 22 bytes (for 64-bit mailboxes). This is to
370 avoid getting the stale values for busy bit. We want to set the busy
371 bit just before issuing command to the FW.
373 ii. In the reset handling, if the reseted command is not owned by the
374 driver, do not (wrongly) print information for the "attached" driver
377 iii. Have extended wait when issuing command in synchronous mode. This is
378 required for the cases where the option ROM is disabled and there is
379 no BIOS to start the controller. The FW starts to boot after receiving
380 the first command from the driver. The current driver has 1 second
381 timeout for the synchronous commands, which is far less than what is
382 actually required. We now wait up to MBOX_RESET_TIME (180 seconds) for
385 iv. In megaraid_mbox_product_info, clear the mailbox contents completely
386 before preparing the command for inquiry3. This is to ensure that the
387 FW does not get junk values in the command.
389 v. Do away with the redundant LSI_CONFIG_COMPAT redefinition for
390 CONFIG_COMPAT. Replace <asm/ioctl32.h> with <linux/ioctl32.h>
392 - James Bottomley <James.Bottomley@SteelEye.com>, 08.17.2004
393 linux-scsi mailing list
395 vi. Add support for 64-bit applications. Current drivers assume only
396 32-bit applications, even on 64-bit platforms. Use the "data" and
397 "buffer" fields of the mimd_t structure, instead of embedded 32-bit
398 addresses in application mailbox and passthru structures.
400 vii. Move the function declarations for the management module from
401 megaraid_mm.h to megaraid_mm.c
403 - Andrew Morton <akpm@osdl.org>, 08.19.2004
404 linux-scsi mailing list
406 viii. Change default values for MEGARAID_NEWGEN, MEGARAID_MM, and
407 MEGARAID_MAILBOX to 'n' in Kconfig.megaraid
409 - Andrew Morton <akpm@osdl.org>, 08.19.2004
410 linux-scsi mailing list
412 ix. replace udelay with msleep
414 x. Typos corrected in comments and whitespace adjustments, explicit
415 grouping of expressions.
418 Release Date : Fri Jul 23 15:22:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
419 Current Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
420 Older Version : 2.20.1.0 (scsi module), 2.20.0.0 (cmm module)
422 i. Add PCI ids for Acer ROMB 2E solution
424 ii. Add PCI ids for I4
426 iii. Typo corrected for subsys id for megaraid sata 300-4x
428 iv. Remove yield() while mailbox handshake in synchronous commands
431 "My other main gripe is things like this:
433 + // wait for maximum 1 second for status to post
434 + for (i = 0; i < 40000; i++) {
435 + if (mbox->numstatus != 0xFF) break;
436 + udelay(25); yield();
439 which litter the driver. Use of yield() in drivers is deprecated."
441 - James Bottomley <James.Bottomley@SteelEye.com>, 07.14.2004
442 linux-scsi mailing list
444 v. Remove redundant __megaraid_busywait_mbox routine
446 vi. Fix bug in the managment module, which causes a system lockup when the
447 IO module is loaded and then unloaded, followed by executing any
448 management utility. The current version of management module does not
449 handle the adapter unregister properly.
451 Specifically, it still keeps a reference to the unregistered
452 controllers. To avoid this, the static array adapters has been
453 replaced by a dynamic list, which gets updated every time an adapter
456 Also, during unregistration of the IO module, the resources are
457 now released in the exact reverse order of the allocation time
461 Release Date : Fri Jun 25 18:58:43 EDT 2004 - Atul Mukker <atulm@lsil.com>
462 Current Version : 2.20.1.0
463 Older Version : megaraid 2.20.0.1
465 i. Stale list pointer in adapter causes kernel panic when module
466 megaraid_mbox is unloaded
469 Release Date : Thu Jun 24 20:37:11 EDT 2004 - Atul Mukker <atulm@lsil.com>
470 Current Version : 2.20.0.1
471 Older Version : megaraid 2.20.0.00
473 i. Modules are not 'y' by default, but depend on current definition of
476 ii. Redundant structure mraid_driver_t removed.
478 iii. Miscellaneous indentation and goto/label fixes.
479 - Christoph Hellwig <hch@infradead.org>, 06.24.2004 linux-scsi
481 iv. scsi_host_put(), do just before completing HBA shutdown.
485 Release Date : Mon Jun 21 19:53:54 EDT 2004 - Atul Mukker <atulm@lsil.com>
486 Current Version : 2.20.0.0
487 Older Version : megaraid 2.20.0.rc2 and 2.00.3
489 i. Independent module to interact with userland applications and
490 multiplex command to low level RAID module(s).
492 "Shared code in a third module, a "library module", is an acceptable
493 solution. modprobe automatically loads dependent modules, so users
494 running "modprobe driver1" or "modprobe driver2" would automatically
495 load the shared library module."
497 - Jeff Garzik <jgarzik@pobox.com> 02.25.2004 LKML
499 "As Jeff hinted, if your userspace<->driver API is consistent between
500 your new MPT-based RAID controllers and your existing megaraid driver,
501 then perhaps you need a single small helper module (lsiioctl or some
502 better name), loaded by both mptraid and megaraid automatically, which
503 handles registering the /dev/megaraid node dynamically. In this case,
504 both mptraid and megaraid would register with lsiioctl for each
505 adapter discovered, and lsiioctl would essentially be a switch,
506 redirecting userspace tool ioctls to the appropriate driver."
508 - Matt Domsch <Matt_Domsch@dell.com> 02.25.2004 LKML
510 ii. Remove C99 initializations from pci_device id.
512 "pci_id_table_g would be much more readable when not using C99
514 PCI table doesn't change, there's lots of users that prefer the more
515 readable variant. And it's really far less and much easier to grok
516 lines without C99 initializers."
518 - Christoph Hellwig <hch@infradead.org>, 05.28.2004 linux-scsi
520 iii. Many fixes as suggested by Christoph Hellwig <hch@infradead.org> on
521 linux-scsi, 05.28.2004
523 iv. We now support up to 32 parallel ioctl commands instead of current 1.
524 There is a conscious effort to let memory allocation not fail for ioctl
527 v. Do away with internal memory management. Use pci_pool_(create|alloc)
530 vi. Kill tasklet when unloading the driver.
532 vii. Do not use "host_lock', driver has fine-grain locks now to protect all
535 viii. Optimize the build scatter-gather list routine. The callers already
536 know the data transfer address and length.
538 ix. Better implementation of error handling and recovery. Driver now
539 performs extended errors recovery for instances like scsi cable pull.
541 x. Disassociate the management commands with an overlaid scsi command.
542 Driver now treats the management packets as special packets and has a
543 dedicated callback routine.