blob: c173806c91face5d6c1046d83268525a42c8de87 [file] [log] [blame]
Ju, Seokmannc005fb42006-04-27 02:33:06 -07001Release Date : Mon Apr 11 12:27:22 EST 2006 - Seokmann Ju <sju@lsil.com>
2Current Version : 2.20.4.8 (scsi module), 2.20.2.6 (cmm module)
3Older Version : 2.20.4.7 (scsi module), 2.20.2.6 (cmm module)
4
51. Fixed a bug in megaraid_reset_handler().
6 Customer reported "Unable to handle kernel NULL pointer dereference
7 at virtual address 00000000" when system goes to reset condition
8 for some reason. It happened randomly.
9 Root Cause: in the megaraid_reset_handler(), there is possibility not
10 returning pending packets in the pend_list if there are multiple
11 pending packets.
12 Fix: Made the change in the driver so that it will return all packets
13 in the pend_list.
14
152. Added change request.
16 As found in the following URL, rmb() only didn't help the
17 problem. I had to increase the loop counter to 0xFFFFFF. (6 F's)
18 http://marc.theaimsgroup.com/?l=linux-scsi&m=110971060502497&w=2
19
20 I attached a patch for your reference, too.
21 Could you check and get this fix in your driver?
22
23 Best Regards,
24 Jun'ichi Nomura
25
Ju, Seokmanned7e8ef2005-11-17 13:17:25 -050026Release Date : Fri Nov 11 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
27Current Version : 2.20.4.7 (scsi module), 2.20.2.6 (cmm module)
28Older Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module)
29
301. Sorted out PCI IDs to remove megaraid support overlaps.
31 Based on the patch from Daniel, sorted out PCI IDs along with
32 charactor node name change from 'megadev' to 'megadev_legacy' to avoid
33 conflict.
34 ---
35 Hopefully we'll be getting the build restriction zapped much sooner,
36 but we should also be thinking about totally removing the hardware
37 support overlap in the megaraid drivers.
38
39 This patch pencils in a date of Feb 06 for this, and performs some
40 printk abuse in hope that existing legacy users might pick up on what's
41 going on.
42
43 Signed-off-by: Daniel Drake <dsd@gentoo.org>
44 ---
45
462. Fixed a issue: megaraid always fails to reset handler.
47 ---
48 I found that the megaraid driver always fails to reset the
49 adapter with the following message:
50 megaraid: resetting the host...
51 megaraid mbox: reset sequence completed successfully
52 megaraid: fast sync command timed out
53 megaraid: reservation reset failed
54 when the "Cluster mode" of the adapter BIOS is enabled.
55 So, whenever the reset occurs, the adapter goes to
56 offline and just become unavailable.
57
58 Jun'ichi Nomura [mailto:jnomura@mtc.biglobe.ne.jp]
59 ---
60
Ju, Seokmann 672b2d32005-05-16 18:32:17 -040061Release Date : Mon Mar 07 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
62Current Version : 2.20.4.6 (scsi module), 2.20.2.6 (cmm module)
63Older Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
64
651. Added IOCTL backward compatibility.
66 Convert megaraid_mm driver to new compat_ioctl entry points.
67 I don't have easy access to hardware, so only compile tested.
68 - Signed-off-by:Andi Kleen <ak@muc.de>
69
702. megaraid_mbox fix: wrong order of arguments in memset()
71 That, BTW, shows why cross-builds are useful-the only indication of
72 problem had been a new warning showing up in sparse output on alpha
73 build (number of exceeding 256 got truncated).
74 - Signed-off-by: Al Viro
75 <viro@parcelfarce.linux.theplanet.co.uk>
76
773. Convert pci_module_init to pci_register_driver
78 Convert from pci_module_init to pci_register_driver
79 (from:http://kerneljanitors.org/TODO)
80 - Signed-off-by: Domen Puncer <domen@coderock.org>
81
824. Use the pre defined DMA mask constants from dma-mapping.h
83 Use the DMA_{64,32}BIT_MASK constants from dma-mapping.h when calling
84 pci_set_dma_mask() or pci_set_consistend_dma_mask(). See
85 http://marc.theaimsgroup.com/?t=108001993000001&r=1&w=2 for more
86 details.
87 Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
88 Signed-off-by: Domen Puncer <domen@coderock.org>
89
905. Remove SSID checking for Dobson, Lindsay, and Verde based products.
91 Checking the SSVID/SSID for controllers which have Dobson, Lindsay,
92 and Verde is unnecessary because device ID has been assigned by LSI
93 and it is unique value. So, all controllers with these IOPs have to be
94 supported by the driver regardless SSVID/SSID.
95
966. Date Thu, 27 Jan 2005 04:31:09 +0100
97 From Herbert Poetzl <>
98 Subject RFC: assert_spin_locked() for 2.6
99
100 Greetings!
101
102 overcautious programming will kill your kernel ;)
103 ever thought about checking a spin_lock or even
104 asserting that it must be held (maybe just for
105 spinlock debugging?) ...
106
107 there are several checks present in the kernel
108 where somebody does a variation on the following:
109
110 BUG_ON(!spin_is_locked(&some_lock));
111
112 so what's wrong about that? nothing, unless you
113 compile the code with CONFIG_DEBUG_SPINLOCK but
114 without CONFIG_SMP ... in which case the BUG()
115 will kill your kernel ...
116
117 maybe it's not advised to make such assertions,
118 but here is a solution which works for me ...
119 (compile tested for sh, x86_64 and x86, boot/run
120 tested for x86 only)
121
122 best,
123 Herbert
124
125 - Herbert Poetzl <herbert@13thfloor.at>, Thu, 27 Jan 2005
126
Linus Torvalds1da177e2005-04-16 15:20:36 -0700127Release Date : Thu Feb 03 12:27:22 EST 2005 - Seokmann Ju <sju@lsil.com>
128Current Version : 2.20.4.5 (scsi module), 2.20.2.5 (cmm module)
129Older Version : 2.20.4.4 (scsi module), 2.20.2.4 (cmm module)
130
1311. Modified name of two attributes in scsi_host_template.
132 On Wed, 2005-02-02 at 10:56 -0500, Ju, Seokmann wrote:
133 > + .sdev_attrs = megaraid_device_attrs,
134 > + .shost_attrs = megaraid_class_device_attrs,
135
136 These are, perhaps, slightly confusing names.
137 The terms device and class_device have well defined meanings in the
138 generic device model, neither of which is what you mean here.
139 Why not simply megaraid_sdev_attrs and megaraid_shost_attrs?
140
141 Other than this, it looks fine to me too.
142
143Release Date : Thu Jan 27 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
144Current Version : 2.20.4.4 (scsi module), 2.20.2.5 (cmm module)
145Older Version : 2.20.4.3 (scsi module), 2.20.2.4 (cmm module)
146
1471. Bump up the version of scsi module due to its conflict.
148
149Release Date : Thu Jan 21 00:01:03 EST 2005 - Atul Mukker <atulm@lsil.com>
150Current Version : 2.20.4.3 (scsi module), 2.20.2.5 (cmm module)
151Older Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
152
1531. Remove driver ioctl for logical drive to scsi address translation and
154 replace with the sysfs attribute. To remove drives and change
155 capacity, application shall now use the device attribute to get the
156 logical drive number for a scsi device. For adding newly created
157 logical drives, class device attribute would be required to uniquely
158 identify each controller.
159 - Atul Mukker <atulm@lsil.com>
160
161 "James, I've been thinking about this a little more, and you may be on
162 to something here. Let each driver add files as such:"
163
164 - Matt Domsch <Matt_Domsch@dell.com>, 12.15.2004
165 linux-scsi mailing list
166
167
168 "Then, if you simply publish your LD number as an extra parameter of
169 the device, you can look through /sys to find it."
170
171 - James Bottomley <James.Bottomley@SteelEye.com>, 01.03.2005
172 linux-scsi mailing list
173
174
175 "I don't see why not ... it's your driver, you can publish whatever
176 extra information you need as scsi_device attributes; that was one of
177 the designs of the extensible attribute system."
178
179 - James Bottomley <James.Bottomley@SteelEye.com>, 01.06.2005
180 linux-scsi mailing list
181
1822. Add AMI megaraid support - Brian King <brking@charter.net>
183 PCI_VENDOR_ID_AMI, PCI_DEVICE_ID_AMI_MEGARAID3,
184 PCI_VENDOR_ID_AMI, PCI_SUBSYS_ID_PERC3_DC,
185
1863. Make some code static - Adrian Bunk <bunk@stusta.de>
187 Date: Mon, 15 Nov 2004 03:14:57 +0100
188
189 The patch below makes some needlessly global code static.
190 -wait_queue_head_t wait_q;
191 +static wait_queue_head_t wait_q;
192
193 Signed-off-by: Adrian Bunk <bunk@stusta.de>
194
1954. Added NEC ROMB support - NEC MegaRAID PCI Express ROMB controller
196 PCI_VENDOR_ID_LSI_LOGIC, PCI_DEVICE_ID_MEGARAID_NEC_ROMB_2E,
197 PCI_SUBSYS_ID_NEC, PCI_SUBSYS_ID_MEGARAID_NEC_ROMB_2E,
198
1995. Fixed Tape drive issue : For any Direct CDB command to physical device
200 including tape, timeout value set by driver was 10 minutes. With this
201 value, most of command will return within timeout. However, for those
202 command like ERASE or FORMAT, it takes more than an hour depends on
203 capacity of the device and the command could be terminated before it
204 completes.
205 To address this issue, the 'timeout' field in the DCDB command will
206 have NO TIMEOUT (i.e., 4) value as its timeout on DCDB command.
207
208
209
210Release Date : Thu Dec 9 19:10:23 EST 2004
211 - Sreenivas Bagalkote <sreenib@lsil.com>
212
213Current Version : 2.20.4.2 (scsi module), 2.20.2.4 (cmm module)
214Older Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
215
216i. Introduced driver ioctl that returns scsi address for a given ld.
217
218 "Why can't the existing sysfs interfaces be used to do this?"
219 - Brian King (brking@us.ibm.com)
220
221 "I've looked into solving this another way, but I cannot see how
222 to get this driver-private mapping of logical drive number-> HCTL
223 without putting code something like this into the driver."
224
225 "...and by providing a mapping a function to userspace, the driver
226 is free to change its mapping algorithm in the future if necessary .."
227 - Matt Domsch (Matt_Domsch@dell.com)
228
229Release Date : Thu Dec 9 19:02:14 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
230
231Current Version : 2.20.4.1 (scsi module), 2.20.2.3 (cmm module)
232Older Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
233
234i. Fix a bug in kioc's dma buffer deallocation
235
236Release Date : Thu Nov 4 18:24:56 EST 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
237
238Current Version : 2.20.4.1 (scsi module), 2.20.2.2 (cmm module)
239Older Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
240
241i. Handle IOCTL cmd timeouts more properly.
242
243ii. pci_dma_sync_{sg,single}_for_cpu was introduced into megaraid_mbox
244 incorrectly (instead of _for_device). Changed to appropriate
245 pci_dma_sync_{sg,single}_for_device.
246
247Release Date : Wed Oct 06 11:15:29 EDT 2004 - Sreenivas Bagalkote <sreenib@lsil.com>
248Current Version : 2.20.4.0 (scsi module), 2.20.2.1 (cmm module)
249Older Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
250
251i. Remove CONFIG_COMPAT around register_ioctl32_conversion
252
253Release Date : Mon Sep 27 22:15:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
254Current Version : 2.20.4.0 (scsi module), 2.20.2.0 (cmm module)
255Older Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
256
257i. Fix data corruption. Because of a typo in the driver, the IO packets
258 were wrongly shared by the ioctl path. This causes a whole IO command
259 to be replaced by an incoming ioctl command.
260
261Release Date : Tue Aug 24 09:43:35 EDT 2004 - Atul Mukker <atulm@lsil.com>
262Current Version : 2.20.3.1 (scsi module), 2.20.2.0 (cmm module)
263Older Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
264
265i. Function reordering so that inline functions are defined before they
266 are actually used. It is now mandatory for GCC 3.4.1 (current stable)
267
268 Declare some heavy-weight functions to be non-inlined,
269 megaraid_mbox_build_cmd, megaraid_mbox_runpendq,
270 megaraid_mbox_prepare_pthru, megaraid_mbox_prepare_epthru,
271 megaraid_busywait_mbox
272
273 - Andrew Morton <akpm@osdl.org>, 08.19.2004
274 linux-scsi mailing list
275
276 "Something else to clean up after inclusion: every instance of an
277 inline function is actually rendered as a full function call, because
278 the function is always used before it is defined. Atul, please
279 re-arrange the code to eliminate the need for most (all) of the
280 function prototypes at the top of each file, and define (not just
281 declare with a prototype) each inline function before its first use"
282
283 - Matt Domsch <Matt_Domsch@dell.com>, 07.27.2004
284 linux-scsi mailing list
285
286
287ii. Display elapsed time (countdown) while waiting for FW to boot.
288
289iii. Module compilation reorder in Makefile so that unresolved symbols do
290 not occur when driver is compiled non-modular.
291
292 Patrick J. LoPresti <patl@users.sourceforge.net>, 8.22.2004
293 linux-scsi mailing list
294
295
296Release Date : Thu Aug 19 09:58:33 EDT 2004 - Atul Mukker <atulm@lsil.com>
297Current Version : 2.20.3.0 (scsi module), 2.20.2.0 (cmm module)
298Older Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
299
300i. When copying the mailbox packets, copy only first 14 bytes (for 32-bit
301 mailboxes) and only first 22 bytes (for 64-bit mailboxes). This is to
302 avoid getting the stale values for busy bit. We want to set the busy
303 bit just before issuing command to the FW.
304
305ii. In the reset handling, if the reseted command is not owned by the
306 driver, do not (wrongly) print information for the "attached" driver
307 packet.
308
309iii. Have extended wait when issuing command in synchronous mode. This is
310 required for the cases where the option ROM is disabled and there is
311 no BIOS to start the controller. The FW starts to boot after receiving
312 the first command from the driver. The current driver has 1 second
313 timeout for the synchronous commands, which is far less than what is
314 actually required. We now wait up to MBOX_RESET_TIME (180 seconds) for
315 FW boot process.
316
317iv. In megaraid_mbox_product_info, clear the mailbox contents completely
318 before preparing the command for inquiry3. This is to ensure that the
319 FW does not get junk values in the command.
320
321v. Do away with the redundant LSI_CONFIG_COMPAT redefinition for
322 CONFIG_COMPAT. Replace <asm/ioctl32.h> with <linux/ioctl32.h>
323
324 - James Bottomley <James.Bottomley@SteelEye.com>, 08.17.2004
325 linux-scsi mailing list
326
327vi. Add support for 64-bit applications. Current drivers assume only
328 32-bit applications, even on 64-bit platforms. Use the "data" and
329 "buffer" fields of the mimd_t structure, instead of embedded 32-bit
330 addresses in application mailbox and passthru structures.
331
332vii. Move the function declarations for the management module from
333 megaraid_mm.h to megaraid_mm.c
334
335 - Andrew Morton <akpm@osdl.org>, 08.19.2004
336 linux-scsi mailing list
337
338viii. Change default values for MEGARAID_NEWGEN, MEGARAID_MM, and
339 MEGARAID_MAILBOX to 'n' in Kconfig.megaraid
340
341 - Andrew Morton <akpm@osdl.org>, 08.19.2004
342 linux-scsi mailing list
343
344ix. replace udelay with msleep
345
346x. Typos corrected in comments and whitespace adjustments, explicit
347 grouping of expressions.
348
349
350Release Date : Fri Jul 23 15:22:07 EDT 2004 - Atul Mukker <atulm@lsil.com>
351Current Version : 2.20.2.0 (scsi module), 2.20.1.0 (cmm module)
352Older Version : 2.20.1.0 (scsi module), 2.20.0.0 (cmm module)
353
354i. Add PCI ids for Acer ROMB 2E solution
355
356ii. Add PCI ids for I4
357
358iii. Typo corrected for subsys id for megaraid sata 300-4x
359
360iv. Remove yield() while mailbox handshake in synchronous commands
361
362
363 "My other main gripe is things like this:
364
365 + // wait for maximum 1 second for status to post
366 + for (i = 0; i < 40000; i++) {
367 + if (mbox->numstatus != 0xFF) break;
368 + udelay(25); yield();
369 + }
370
371 which litter the driver. Use of yield() in drivers is deprecated."
372
373 - James Bottomley <James.Bottomley@SteelEye.com>, 07.14.2004
374 linux-scsi mailing list
375
376v. Remove redundant __megaraid_busywait_mbox routine
377
378vi. Fix bug in the managment module, which causes a system lockup when the
379 IO module is loaded and then unloaded, followed by executing any
380 management utility. The current version of management module does not
381 handle the adapter unregister properly.
382
383 Specifically, it still keeps a reference to the unregistered
384 controllers. To avoid this, the static array adapters has been
385 replaced by a dynamic list, which gets updated every time an adapter
386 is added or removed.
387
388 Also, during unregistration of the IO module, the resources are
389 now released in the exact reverse order of the allocation time
390 sequence.
391
392
393Release Date : Fri Jun 25 18:58:43 EDT 2004 - Atul Mukker <atulm@lsil.com>
394Current Version : 2.20.1.0
395Older Version : megaraid 2.20.0.1
396
397i. Stale list pointer in adapter causes kernel panic when module
398 megaraid_mbox is unloaded
399
400
401Release Date : Thu Jun 24 20:37:11 EDT 2004 - Atul Mukker <atulm@lsil.com>
402Current Version : 2.20.0.1
403Older Version : megaraid 2.20.0.00
404
405i. Modules are not 'y' by default, but depend on current definition of
406 SCSI & PCI.
407
408ii. Redundant structure mraid_driver_t removed.
409
410iii. Miscellaneous indentation and goto/label fixes.
411 - Christoph Hellwig <hch@infradead.org>, 06.24.2004 linux-scsi
412
413iv. scsi_host_put(), do just before completing HBA shutdown.
414
415
416
417Release Date : Mon Jun 21 19:53:54 EDT 2004 - Atul Mukker <atulm@lsil.com>
418Current Version : 2.20.0.0
419Older Version : megaraid 2.20.0.rc2 and 2.00.3
420
421i. Independent module to interact with userland applications and
422 multiplex command to low level RAID module(s).
423
424 "Shared code in a third module, a "library module", is an acceptable
425 solution. modprobe automatically loads dependent modules, so users
426 running "modprobe driver1" or "modprobe driver2" would automatically
427 load the shared library module."
428
429 - Jeff Garzik <jgarzik@pobox.com> 02.25.2004 LKML
430
431 "As Jeff hinted, if your userspace<->driver API is consistent between
432 your new MPT-based RAID controllers and your existing megaraid driver,
433 then perhaps you need a single small helper module (lsiioctl or some
434 better name), loaded by both mptraid and megaraid automatically, which
435 handles registering the /dev/megaraid node dynamically. In this case,
436 both mptraid and megaraid would register with lsiioctl for each
437 adapter discovered, and lsiioctl would essentially be a switch,
438 redirecting userspace tool ioctls to the appropriate driver."
439
440 - Matt Domsch <Matt_Domsch@dell.com> 02.25.2004 LKML
441
442ii. Remove C99 initializations from pci_device id.
443
444 "pci_id_table_g would be much more readable when not using C99
445 initializers.
446 PCI table doesn't change, there's lots of users that prefer the more
447 readable variant. And it's really far less and much easier to grok
448 lines without C99 initializers."
449
450 - Christoph Hellwig <hch@infradead.org>, 05.28.2004 linux-scsi
451
452iii. Many fixes as suggested by Christoph Hellwig <hch@infradead.org> on
453 linux-scsi, 05.28.2004
454
455iv. We now support up to 32 parallel ioctl commands instead of current 1.
456 There is a conscious effort to let memory allocation not fail for ioctl
457 commands.
458
459v. Do away with internal memory management. Use pci_pool_(create|alloc)
460 instead.
461
462vi. Kill tasklet when unloading the driver.
463
464vii. Do not use "host_lock', driver has fine-grain locks now to protect all
465 data structures.
466
467viii. Optimize the build scatter-gather list routine. The callers already
468 know the data transfer address and length.
469
470ix. Better implementation of error handling and recovery. Driver now
471 performs extended errors recovery for instances like scsi cable pull.
472
473x. Disassociate the management commands with an overlaid scsi command.
474 Driver now treats the management packets as special packets and has a
475 dedicated callback routine.