2 Wed Dec 3 15:34:42 EST 2003 - Atul Mukker <atulm@lsil.com>
3 1. pci_dma_sync_sg(), 2nd argument is pointer to scatter-gather list. All
4 previous drivers have a pointer which is incremented beyond the end of
5 the scatter-gather list
6 2. Remove 'ipdev', pci device pointer, to allocate memory for internal
7 commands. pci_alloc_consistent() guaranteed to allocate memory below
8 4GB, which is a requirement for these commands -
9 Tom Coughlan <coughlan@redhat.com>
10 3. Added support for LSI SATA PCI-X controllers
11 4. Advanced Server 3.0 version of the driver is the reference driver now.
12 For all other kernels, we create patches.
13 5. Removed superfluous white spaces.
14 6. Corrected some comments
17 Tue Nov 4 14:33:43 EST 2003 - Atul Mukker <atulm@lsil.com>
18 1. Added vendor ids and device ids for PCI-Express controllers,
19 PERC4E/Si, PERC4E/Di, PERC4E/DC, PERC4E/SC
20 2. Backport some minor changes from 2.6 kernel version of the driver.
23 Fri Oct 3 18:04:29 EDT 2003 - Atul Mukker <atulm@lsil.com>
24 1. Minor changes which are brough in when sync'ing with kernel 2.9-test6.
25 2. Use sizeof(mbox_t) in synchronous routines instead of hard-coded value
27 3. De-couple adapter->host->pci_dev. Replace with adapter->pdev
30 Thu Sep 4 17:49:42 EDT 2003 - Atul Mukker <atulm@lsil.com>
31 i. For extended passthru commands, 64-bit scatter-gather list and 64-bit
32 mailbox address must only be used if the controller supports extended
33 CDBs and 64-bit addressing and the kernel is configured to support
35 With 2.00.8 and previous drivers, if the controllers supports extended
36 CDBs but the kernel does not support high memory - the driver prepares
37 a 32-bit sg list but the mailbox address is chosen to be 64-bit. This
38 causes FW to incorrectly read 32-bit (8 bytes) sg list as 64-bit
39 (12 bytes) sglist. This would cause IO from non-disk devices to fail
42 Wed Aug 27 16:40:05 EDT 2003 - Atul Mukker <atulm@lsil.com>
43 i. Make sure the value of number of statuses, completed command id array,
44 and the mailbox status fields are updated in host memory before we
45 read and interpret them. For this to happen - mailbox numstatus and
46 command id array's first element are invalidated. The ISR busy waits
47 till valid values are obtained in these two fields. Now since the
48 status field is between these two fields it is assumed that status
49 values is sane when write to numstatus and completed id array is
51 ii. while returning from the ioctl handler for the SCSI passthru commands
52 a direct access was made to the user address. Now the user structure
53 is copied in before the required field is accessed.
54 iii. Remove redundant volatile casts from pending command counter pend_cmds.
57 Fri Aug 1 11:01:11 EDT 2003 - Atul Mukker <atulm@lsil.com>
58 i. Adapter lock re-definition so that patch for kernels w/o per host lock
59 is less intrusive - Jens Axboe <axboe@suse.de>
61 ii. While in abort and reset handling, check for non-empty pending list is
62 invalid. The intent is to wait for pending commands in FW to complete,
63 not the pending commands with the driver - Atul Mukker <atulm@lsil.com>
66 Wed Jul 30 11:35:31 EDT 2003 - Sreenivas Bagalkote <sreenib@lsil.com>
67 i. Declare the function prototypes used for "/proc" within the compiler
68 directives #ifdef CONFIG_PROC_FS. Similary, move global/local variable
69 declarations and code fragments related to /proc within the directives.
70 - Mark Haverkamp <markh@osdl.org>
72 ii. Initialize host->lock with adapter->lock.
74 iii. Wait for mailbox status to become valid
75 - Haruo Tomita <haruo.tomita@toshiba.co.jp>
77 iv. Right structure is passed to FW for 4-span read configuration command.
78 - Sreenivas Bagalkote <sreenivas.bagalkote@lsil.com>
81 Thu Apr 24 14:06:55 EDT 2003 - Atul Mukker <atulm@lsil.com>
82 i. Do not use repeated allocations for "pci_dev" for internal
83 allocations. Allocate the handle at load time and set the DMA mask for
84 allocations below 4GB.
86 ii. Remove superfluous definitions, mid-layer /proc entry and synchronous
89 iii. Logical drive numbers were incremented in passthru structure after
90 random-deletion operation! Also, the logical drive deletion command is
91 issued in polled mode because of the racy nature of interrupt based
94 iv. Add support for Intel's subsystem vendor id
97 Thu Apr 17 15:58:58 EDT 2003
98 i. Do not put the completed SCSI commands in a list. Complete them to
99 mid-layer as soon as we have completed them and reclaimed our
102 ii. Break ISR functionality in two portions. The lower half serves as
103 interrupt acknowledgment sequence for the firmware. This half can also
104 be called from other places in the driver, e.g., the abort and reset
107 iii. New abort and reset handling. In these situations, driver allows more
108 time for the firmware to complete outstanding requests instead of
109 failing handlers right-away.
112 Wed Jan 29 09:13:44 EST 200 - Atul Mukker <atulm@lsil.com>
113 i. Change the handshake in ISR while acknowledging interrupts. Write the
114 valid interrupt pattern 0x10001234 as soon as it is read from the
115 outdoor register. In existing driver and on certain platform, invalid
116 command ids were being returned.
118 Also, do not wait on status be become 0xFF, since FW can return this
119 status in certain circumstances.
121 Initialize the numstatus field of mailbox to 0xFF so that we can wait
122 on this wait in next interrupt. Firmware does not change its value
123 unless there are some status to be posted
125 ii. Specify the logical drive number while issuing the RESERVATION_STATUS
127 iii. Reduce the default mailbox busy wait time from 300us to 10us. This is
128 done to avoid a possible deadlock in FW because of longer bust waits.
130 iv. The max outstanding commands are reduced to 126 because that's the
131 safest value on all FW.
133 v. Number of sectors per IO are reduced to 128 (64kb), because FW needs
134 resources in special circumstances like check consistency, rebuilds
137 vi. max_commands is no longer a module parameter because of iv.
140 i. Intermediate release with kernel specific code
144 Wed Dec 4 14:34:51 EST 2002 - Atul Mukker <atulm@lsil.com>
145 i. Making the older IO based controllers to work with this driver
149 Fri Nov 15 10:59:44 EST 2002 - Atul Mukker <atulm@lsil.com>
150 i. Release host lock before issuing internal command to reset
151 reservations in megaraid_reset() and reacquire after internal command