| 1 |
1626 |
jcastillo |
Sun Dec 13 18:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 2 |
|
|
* revision 3.1e
|
| 3 |
|
|
- Same work-around as for the 53c876 rev <= 0x15 for 53c896 rev 1:
|
| 4 |
|
|
Disable overlapped arbitration. This will not make difference
|
| 5 |
|
|
since the chip has on-chip RAM.
|
| 6 |
|
|
|
| 7 |
|
|
Thu Nov 26 22:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 8 |
|
|
* revision 3.1d
|
| 9 |
|
|
- The SISL RAID change requires now remap_pci_mem() stuff to be
|
| 10 |
|
|
compiled for __i386__ when normal IOs are used.
|
| 11 |
|
|
- Minor spelling fixes in doc files.
|
| 12 |
|
|
|
| 13 |
|
|
Sat Nov 21 18:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 14 |
|
|
* revision 3.1c
|
| 15 |
|
|
- Ignore chips that are driven by SISL RAID (DAC 960).
|
| 16 |
|
|
Change sent by Leonard Zubkoff and slightly reworked.
|
| 17 |
|
|
- Still a buglet in the tags initial settings that needed to be fixed.
|
| 18 |
|
|
It was not possible to disable TGQ at system startup for devices
|
| 19 |
|
|
that claim TGQ support. The driver used at least 2 for the queue
|
| 20 |
|
|
depth but did'nt keep track of user settings for tags depth lower
|
| 21 |
|
|
than 2.
|
| 22 |
|
|
|
| 23 |
|
|
Wed Nov 11 10:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 24 |
|
|
* revision 3.1b
|
| 25 |
|
|
- The driver was unhappy when configured with default_tags > MAX_TAGS
|
| 26 |
|
|
Hopefully doubly-fixed.
|
| 27 |
|
|
- Update the Configure.help driver section that speaks of TAGS.
|
| 28 |
|
|
|
| 29 |
|
|
Wed Oct 21 21:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 30 |
|
|
* revision 3.1a
|
| 31 |
|
|
- Changes from Eddie Dost for Sparc and Alpha:
|
| 32 |
|
|
ioremap/iounmap support for Sparc.
|
| 33 |
|
|
pcivtophys changed to bus_dvma_to_phys.
|
| 34 |
|
|
- Add the 53c876 description to the chip table. This is only usefull
|
| 35 |
|
|
for printing the right name of the controller.
|
| 36 |
|
|
- DEL-441 Item 2 work-around for the 53c876 rev <= 5 (0x15).
|
| 37 |
|
|
- Add additionnal checking of INQUIRY data:
|
| 38 |
|
|
Check INQUIRY data received length is at least 7. Byte 7 of
|
| 39 |
|
|
inquiry data contains device features bits and the driver might
|
| 40 |
|
|
be confused by garbage. Also check peripheral qualifier.
|
| 41 |
|
|
- Cleanup of the SCSI tasks management:
|
| 42 |
|
|
Remove the special case for 32 tags. Now the driver only uses the
|
| 43 |
|
|
scheme that allows up to 64 tags per LUN.
|
| 44 |
|
|
Merge some code from the 896 driver.
|
| 45 |
|
|
Use a 1,3,5,...MAXTAGS*2+1 tag numbering. Previous driver could
|
| 46 |
|
|
use any tag number from 1 to 253 and some non conformant devices
|
| 47 |
|
|
might have problems with large tag numbers.
|
| 48 |
|
|
- 'no_sync' changed to 'no_disc' in the README file. This is an old
|
| 49 |
|
|
and trivial mistake that seems to demonstrate the README file is
|
| 50 |
|
|
not often read. :)
|
| 51 |
|
|
|
| 52 |
|
|
Sun Oct 4 14:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 53 |
|
|
* revision 3.0i
|
| 54 |
|
|
- Cosmetic changes for sparc (but not for the driver) that needs
|
| 55 |
|
|
__irq_itoa() to be used for printed IRQ value to be understandable.
|
| 56 |
|
|
- Some problems with the driver that didn't occur using driver 2.5f
|
| 57 |
|
|
were due to a SCSI selection problem triggered by a clearly
|
| 58 |
|
|
documented feature that in fact seems not to work: (53C8XX chips
|
| 59 |
|
|
are claimed by the manuals to be able to execute SCSI scripts just
|
| 60 |
|
|
after abitration while the SCSI core is performing SCSI selection).
|
| 61 |
|
|
This optimization is broken and has been removed.
|
| 62 |
|
|
- Some broken scsi devices are confused when a negotiation is started
|
| 63 |
|
|
on a LUN that does not correspond to a real device. According to
|
| 64 |
|
|
SCSI specs, this is a device firmware bug. This has been worked
|
| 65 |
|
|
around by only starting negotiation if the LUN has previously be
|
| 66 |
|
|
used for at least 1 successful SCSI command.
|
| 67 |
|
|
- The 'last message sent' printed out on M_REJECT message reception
|
| 68 |
|
|
was read from the SFBR i/o register after the previous message had
|
| 69 |
|
|
been sent.
|
| 70 |
|
|
This was not correct and affects all previous driver versions and
|
| 71 |
|
|
the original FreeBSD one as well. The SCSI scripts has been fixed
|
| 72 |
|
|
so that it now provides the right information to the C code.
|
| 73 |
|
|
|
| 74 |
|
|
Sat Jul 18 13:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 75 |
|
|
* revision 3.0g
|
| 76 |
|
|
- Preliminary fixes for Big Endian (sent by Eddie C. Dost).
|
| 77 |
|
|
Big Endian architectures should work again with the driver.
|
| 78 |
|
|
Eddie's patch has been partially applied since current 2.1.109
|
| 79 |
|
|
does not have all the Sparc changes of the vger tree.
|
| 80 |
|
|
- Use of BITS_PER_LONG instead of (~0UL == 0xffffffffUL) has fixed
|
| 81 |
|
|
the problem observed when the driver was compiled using EGCS or
|
| 82 |
|
|
PGCC.
|
| 83 |
|
|
|
| 84 |
|
|
Mon Jul 13 20:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 85 |
|
|
* revision 3.0f
|
| 86 |
|
|
- Some spelling fixes.
|
| 87 |
|
|
- linux/config.h misplaced in ncr53c8xx.h
|
| 88 |
|
|
- MODULE_PARM stuff added for linux 2.1.
|
| 89 |
|
|
- check INQUIRY response data format is exactly 2.
|
| 90 |
|
|
- use BITS_PER_LONG if defined.
|
| 91 |
|
|
|
| 92 |
|
|
Sun Jun 28 12:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 93 |
|
|
* revision 3.0e
|
| 94 |
|
|
- Some cleanup, spelling fixes, version checks, documentations
|
| 95 |
|
|
changes, etc ...
|
| 96 |
|
|
|
| 97 |
|
|
Sat Jun 20 20:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 98 |
|
|
* revision 3.0c
|
| 99 |
|
|
- Add a boot setup option that allows to set up device queue depths
|
| 100 |
|
|
at boot-up. This option is very usefull since Linux does not
|
| 101 |
|
|
allow to change scsi device queue depth once the system has been
|
| 102 |
|
|
booted up.
|
| 103 |
|
|
|
| 104 |
|
|
Sun Jun 15 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 105 |
|
|
* revision 3.0a
|
| 106 |
|
|
- Support for up to 64 TAGS per LUN.
|
| 107 |
|
|
- Rewrite the TARGET vs LUN capabilities management.
|
| 108 |
|
|
CmdQueue is now handled as a LUN capability as it shall be.
|
| 109 |
|
|
This also fixes a bug triggered when disabling tagged command
|
| 110 |
|
|
queuing for a device that had this feature enabled.
|
| 111 |
|
|
- Remove the ncr_opennings() stuff that was useless under Linux
|
| 112 |
|
|
and hard to understand to me.
|
| 113 |
|
|
- Add "setverbose" procfs driver command. It allows to tune
|
| 114 |
|
|
verbose level after boot-up. Setting this level to zero, for
|
| 115 |
|
|
example avoid flooding the syslog file.
|
| 116 |
|
|
- Add KERN_XXX to some printk's.
|
| 117 |
|
|
|
| 118 |
|
|
Tue Jun 10 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 119 |
|
|
* revision 3.0
|
| 120 |
|
|
- Linux config changes for 2.0.34:
|
| 121 |
|
|
Remove NVRAM detection config option. This option is now enabled
|
| 122 |
|
|
by default but can be disabled by editing the driver header file.
|
| 123 |
|
|
Add a PROFILE config option.
|
| 124 |
|
|
- Update Configure.help
|
| 125 |
|
|
- Add calls to new function mdelay() for milli-seconds delay if
|
| 126 |
|
|
kernel version >= 2.1.105.
|
| 127 |
|
|
- Replace all printf(s) by printk(s). After all, the ncr53c8xx is
|
| 128 |
|
|
a driver for Linux.
|
| 129 |
|
|
- Perform auto-sense on COMMAND TERMINATED. Not sure it is usefull.
|
| 130 |
|
|
- Some other minor changes.
|
| 131 |
|
|
|
| 132 |
|
|
Tue Jun 4 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 133 |
|
|
* revision 2.6n
|
| 134 |
|
|
- Code cleanup and simplification:
|
| 135 |
|
|
Remove kernel 1.2.X and 1.3.X support.
|
| 136 |
|
|
Remove the _old_ target capabilities table.
|
| 137 |
|
|
Remove the error recovery code that have'nt been really usefull.
|
| 138 |
|
|
Use a single alignment boundary (CACHE_LINE_SIZE) for data
|
| 139 |
|
|
structures.
|
| 140 |
|
|
- Several aggressive SCRIPTS optimizations and changes:
|
| 141 |
|
|
Reselect SCRIPTS code rewritten.
|
| 142 |
|
|
Support for selection/reselection without ATN.
|
| 143 |
|
|
And some others.
|
| 144 |
|
|
- Miscallaneous changes in the C code:
|
| 145 |
|
|
Count actual number of CCB queued to the controller (future use).
|
| 146 |
|
|
Lots of other minor changes.
|
| 147 |
|
|
|
| 148 |
|
|
Wed May 13 20:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 149 |
|
|
* revision 2.6m
|
| 150 |
|
|
- Problem of missed SCSI bus reset with the 53C895 fixed by
|
| 151 |
|
|
Richard Waltham. The 53C895 needs about 650 us for the bus
|
| 152 |
|
|
mode to settle. Delays used while resetting the controller
|
| 153 |
|
|
and the bus have been adjusted. Thanks Richard!
|
| 154 |
|
|
- Some simplification for 64 bit arch done ccb address testing.
|
| 155 |
|
|
- Add a check of the MSG_OUT phase after Selection with ATN.
|
| 156 |
|
|
- The new tagged queue stuff seems ok, so some informationnal
|
| 157 |
|
|
message have been conditionned by verbose >= 3.
|
| 158 |
|
|
- Donnot reset if a SBMC interrupt reports the same bus mode.
|
| 159 |
|
|
- Print out the whole driver set-up. Some options were missing and
|
| 160 |
|
|
the print statement was misplaced for modules.
|
| 161 |
|
|
- Ignore a SCSI parity interrupt if the chip is not connected to
|
| 162 |
|
|
the SCSI bus.
|
| 163 |
|
|
|
| 164 |
|
|
Sat May 1 16:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 165 |
|
|
* revision 2.6l
|
| 166 |
|
|
- Add CCB done queue support for Alpha and perhaps some other
|
| 167 |
|
|
architectures.
|
| 168 |
|
|
- Add some barriers to enforce memory ordering for x86 and
|
| 169 |
|
|
Alpha architectures.
|
| 170 |
|
|
- Fix something that looks like an old bug in the nego SIR
|
| 171 |
|
|
interrupt code in case of negotiation failure.
|
| 172 |
|
|
|
| 173 |
|
|
Sat Apr 25 21:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 174 |
|
|
* revision 2.6k
|
| 175 |
|
|
- Remove all accesses to the on-chip RAM from the C code:
|
| 176 |
|
|
Use SCRIPTS to load the on-chip RAM.
|
| 177 |
|
|
Use SCRIPTS to repair the start queue on selection timeout.
|
| 178 |
|
|
Use the copy of script in main memory to calculate the chip
|
| 179 |
|
|
context on phase mismatch.
|
| 180 |
|
|
- The above allows now to use the on-chip RAM without requiring
|
| 181 |
|
|
to get access to the on-chip RAM from the C code. This makes
|
| 182 |
|
|
on-chip RAM useable for linux-1.2.13 and for Linux-Alpha for
|
| 183 |
|
|
instance.
|
| 184 |
|
|
- Some simplifications and cleanups in the SCRIPTS and C code.
|
| 185 |
|
|
- Buglet fixed in parity error recovery SCRIPTS (never tested).
|
| 186 |
|
|
- Minor updates in README.ncr53c8xx.
|
| 187 |
|
|
|
| 188 |
|
|
Wed Apr 15 21:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 189 |
|
|
* revision 2.6j
|
| 190 |
|
|
- Incorporate changes from linux-2.1.95 ncr53c8xx driver version.
|
| 191 |
|
|
- Add SMP support for linux-2.1.95 and above.
|
| 192 |
|
|
- Fix a bug when QUEUE FULL is returned and no commands are
|
| 193 |
|
|
disconnected. This happens with Atlas I / L912 and may happen
|
| 194 |
|
|
with Atlas II / LXY4.
|
| 195 |
|
|
- Nail another one on CHECK condition when requeuing the command
|
| 196 |
|
|
for auto-sense.
|
| 197 |
|
|
- Call scsi_done() for all completed commands after interrupt
|
| 198 |
|
|
handling.
|
| 199 |
|
|
- Increase the done queue to 24 entries.
|
| 200 |
|
|
|
| 201 |
|
|
Sat Apr 4 20:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 202 |
|
|
* revision 2.6i
|
| 203 |
|
|
- CTEST0 is used by the 53C885 for Power Management and
|
| 204 |
|
|
priority setting between the 2 functions.
|
| 205 |
|
|
Use SDID instead as actual target number. Just have had to
|
| 206 |
|
|
overwrite it with SSID on reselection.
|
| 207 |
|
|
- Split DATA_IN and DATA_OUT scripts into 2 sub-scripts.
|
| 208 |
|
|
64 segments are moved from on-chip RAM scripts.
|
| 209 |
|
|
If more segments, a script in main memory is used for the
|
| 210 |
|
|
additionnal segments.
|
| 211 |
|
|
- Since the SCRIPTS processor continues SCRIPTS execution after
|
| 212 |
|
|
having won arbitration, do some stuff prior to testing any SCSI
|
| 213 |
|
|
phase on reselection. This should have the vertue to process
|
| 214 |
|
|
scripts in parallel with the SCSI core performing selection.
|
| 215 |
|
|
- Increase the done queue to 12 entries.
|
| 216 |
|
|
|
| 217 |
|
|
Sun Mar 29 12:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 218 |
|
|
* revision 2.6h
|
| 219 |
|
|
- Some fixes.
|
| 220 |
|
|
|
| 221 |
|
|
Tue Mar 26 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 222 |
|
|
* revision 2.6g
|
| 223 |
|
|
- New done queue. 8 entries by default (6 always useable).
|
| 224 |
|
|
Can be increased if needed.
|
| 225 |
|
|
- Resources management using doubly linked queues.
|
| 226 |
|
|
- New auto-sense and QUEUE FULL handling that does not need to
|
| 227 |
|
|
stall the NCR queue any more.
|
| 228 |
|
|
- New CCB starvation avoiding algorithm.
|
| 229 |
|
|
- Prepare CCBs for SCSI commands that cannot be queued, instead of
|
| 230 |
|
|
inserting these commands into the waiting list. The waiting list
|
| 231 |
|
|
is now only used while resetting and when memory for CCBs is not
|
| 232 |
|
|
yet available?
|
| 233 |
|
|
|
| 234 |
|
|
Sun Feb 8 22:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 235 |
|
|
* revision 2.6f
|
| 236 |
|
|
- Some fixes in order to really support the 53C895, at least with
|
| 237 |
|
|
FAST-20 devices.
|
| 238 |
|
|
- Heavy changes in the target/lun resources management to allow
|
| 239 |
|
|
the scripts to jump directly to the CCB on reselection instead
|
| 240 |
|
|
of walking on the lun CCBs list. Up to 32 tags per lun are now
|
| 241 |
|
|
supported without script processor and PCI traffic overhead.
|
| 242 |
|
|
|
| 243 |
|
|
Sun Jan 11 22:00 1998 Gerard Roudier (groudier@club-internet.fr)
|
| 244 |
|
|
* revision 2.6d
|
| 245 |
|
|
- new (different ?) implementation of the start queue:
|
| 246 |
|
|
Use a simple CALL to a launch script in the CCB.
|
| 247 |
|
|
- implement a minimal done queue (1 entry :-) ).
|
| 248 |
|
|
this avoid scanning all CCBs on INT FLY (Only scan all CCBs, on
|
| 249 |
|
|
overflow). Hit ratio is better than 99.9 % on my system, so no
|
| 250 |
|
|
need to have a larger done queue.
|
| 251 |
|
|
- generalization of the restart of CCB on special condition as
|
| 252 |
|
|
Abort, QUEUE FULL, CHECK CONDITION.
|
| 253 |
|
|
This has been called 'silly scheduler'.
|
| 254 |
|
|
- make all the profiling code conditionned by a config option.
|
| 255 |
|
|
This spare some PCI traffic and C code when this feature is not
|
| 256 |
|
|
needed.
|
| 257 |
|
|
- handle more cleanly the situation where direction is unknown.
|
| 258 |
|
|
The pointers patching is now performed by the SCRIPTS processor.
|
| 259 |
|
|
- remove some useless scripts instructions.
|
| 260 |
|
|
|
| 261 |
|
|
Ported from driver 2.5 series:
|
| 262 |
|
|
------------------------------
|
| 263 |
|
|
- Use FAST-5 instead of SLOW for slow scsi devices according to
|
| 264 |
|
|
new SPI-2 draft.
|
| 265 |
|
|
- Make some changes in order to accomodate with 875 rev <= 3
|
| 266 |
|
|
device errata listing 397. Minor consequences are:
|
| 267 |
|
|
. Leave use of PCI Write and Invalidate under user control.
|
| 268 |
|
|
Now, by default the driver does not enable PCI MWI and option
|
| 269 |
|
|
'specf:y' is required in order to enable this feature.
|
| 270 |
|
|
. Memory Read Line is not enabled for 875 and 875-like chips.
|
| 271 |
|
|
. Programmed burst length set to 64 DWORDS (instead of 128).
|
| 272 |
|
|
(Note: SYMBIOS uses 32 DWORDS for the SDMS BIOS)
|
| 273 |
|
|
- Add 'buschk' boot option.
|
| 274 |
|
|
This option enables checking of SCSI BUS data lines after SCSI
|
| 275 |
|
|
RESET (set by default). (Submitted by Richard Waltham).
|
| 276 |
|
|
- Update the README file.
|
| 277 |
|
|
- Dispatch CONDITION MET and RESERVATION CONFLICT scsi status
|
| 278 |
|
|
as OK driver status.
|
| 279 |
|
|
- Update the README file and the Symbios NVRAM format definition
|
| 280 |
|
|
with removable media flags values (available with SDMS 4.09).
|
| 281 |
|
|
- Several PCI configuration registers fix-ups for powerpc.
|
| 282 |
|
|
(Patch sent by Cort).
|
