Root/Documentation/scsi/ChangeLog.sym53c8xx

1Sat May 12 12:00 2001 Gerard Roudier (groudier@club-internet.fr)
2    * version sym53c8xx-1.7.3c
3    - Ensure LEDC bit in GPCNTL is cleared when reading the NVRAM.
4      Fix sent by Stig Telfer <stig@api-networks.com>.
5    - Backport from SYM-2 the work-around that allows to support
6      hardwares that fail PCI parity checking.
7    - Check that we received at least 8 bytes of INQUIRY response
8      for byte 7, that contains device capabilities, to be valid.
9    - Define scsi_set_pci_device() as nil for kernel < 2.4.4.
10    - + A couple of minor changes.
11      
12Sat Apr 7 19:30 2001 Gerard Roudier (groudier@club-internet.fr)
13    * version sym53c8xx-1.7.3b
14    - Fix an unaligned LOAD from scripts (was used as dummy read).
15    - In ncr_soft_reset(), only try to ABORT the current operation
16      for chips that support SRUN bit in ISTAT1 and if SCRIPTS are
17      currently running, as 896 and 1010 manuals suggest.
18    - In the CCB abort path, do not assume that the CCB is currently
19      queued to SCRIPTS. This is not always true, notably after a
20      QUEUE FULL status or when using untagged commands.
21
22Sun Mar 4 18:30 2001 Gerard Roudier (groudier@club-internet.fr)
23    * version sym53c8xx-1.7.3a
24    - Fix an issue in the ncr_int_udc() (unexpected disconnect)
25      handling. If the DSA didn't match a CCB, a bad write to
26      memory could happen.
27
28Mon Feb 12 22:30 2001 Gerard Roudier (groudier@club-internet.fr)
29    * version sym53c8xx-1.7.3
30    - Support for hppa.
31      Tiny patch sent to me by Robert Hirst.
32    - Tiny patch for ia64 sent to me by Pamela Delaney.
33
34Tue Feb 6 13:30 2001 Gerard Roudier (groudier@club-internet.fr)
35    * version sym53c8xx-1.7.3-pre1
36    - Call pci_enable_device() as AC wants this to be done.
37    - Get both the BAR cookies used by CPU and actual PCI BAR
38      values used from SCRIPTS. Recent PCI chips are able to
39      access themselves using internal cycles, but they compare
40      BAR values to destination address to make decision.
41      Earlier chips simply use PCI transactions to access IO
42      registers from SCRIPTS.
43      The bus_dvma_to_mem() interface that reverses the actual
44      PCI BAR value from the BAR cookie is now useless.
45      This point had been discussed at the list and the solution
46      got approved by PCI code maintainer (Martin Mares).
47    - Merge changes for linux-2.4 that declare the host template
48      in the driver object also when the driver is statically
49      linked with the kernel.
50    - Increase SCSI message size up to 12 bytes, given that 8
51      bytes was not enough for the PPR message (fix).
52    - Add field 'maxoffs_st' (max offset for ST data transfers).
53      The C1010 supports offset 62 in DT mode but only 31 in
54      ST mode, to 2 different values for the max SCSI offset
55      are needed. Replace the obviously wrong masking of the
56      offset against 0x1f for ST mode by a lowering to
57      maxoffs_st of the SCSI offset in ST mode.
58    - Refine a work-around for the C1010-66. Revision 1 does
59      not requires extra cycles in DT DATA OUT phase.
60    - Add a missing endian-ization (abrt_tbl.addr).
61    - Minor clean-up in the np structure for fields accessed
62      from SCRIPTS that requires special alignments.
63
64Sun Sep 24 21:30 2000 Gerard Roudier (groudier@club-internet.fr)
65    * version sym53c8xx-1.7.2
66    - Remove the hack for PPC added in previous driver version.
67    - Add FE_DAC feature bit to distinguish between 64 bit PCI
68      addressing (FE_DAC) and 64 bit PCI interface (FE_64BIT).
69    - Get rid of the boot command line "ultra:" argument.
70      This parameter wasn't that clever since we can use "sync:"
71      for Ultra/Ultra2 settings, and for Ultra3 we may want to
72      pass PPR options (for now only DT clocking).
73    - Add FE_VARCLK feature bit that indicates that SCSI clock
74      frequency may vary depending on board design and thus,
75      the driver should try to evaluate the SCSI clock.
76    - Simplify the way the driver determine the SCSI clock:
77      ULTRA3 -> 160 MHz, ULTRA2 -> 80 MHz otherwise 40 MHz.
78      Measure the SCSI clock frequency if FE_VARCLK is set.
79    - Remove FE_CLK80 feature bit that got useless.
80    - Add support for the SYM53C875A (Pamela Delaney).
81
82Wed Jul 26 23:30 2000 Gerard Roudier (groudier@club-internet.fr)
83    * version sym53c8xx-1.7.1
84    - Provide OpenFirmware path through the proc FS on PPC.
85    - Download of on-chip SRAM using memcpy_toio() doesn't work
86      on PPC. Restore previous method (MEMORY MOVE from SCRIPTS).
87    - Remove trailing argument #2 from a couple of #undefs.
88
89Sun Jul 09 16:30 2000 Gerard Roudier (groudier@club-internet.fr)
90    * version sym53c8xx-1.7.0
91    - Remove the PROFILE C and SCRIPTS code.
92      This facility was not this useful and thus was not longer
93      desirable given the increasing complexity of the driver code.
94    - Merges from FreeBSD sym-1.6.2 driver:
95      * Clarify memory barriers needed by the driver for architectures
96        that implement a weak memory ordering.
97      * Simpler handling of illegal phases and data overrun from
98        SCRIPTS. These errors are now immediately reported to
99        the C code by an interrupt.
100      * Sync the residual handling code with sym-1.6.2 and now
101        report `resid' to user for linux version >= 2.3.99
102    - General cleanup:
103      Move definitions for barriers and IO/MMIO operations to the
104      sym53c8xx_defs.h header files. They are now shared by the
105      both drivers.
106      Remove unused options that claimed to optimize for the 896.
107      If fact, they were not this clever. :)
108      Use SCSI_NCR_IOMAPPED instead of NCR_IOMAPPED.
109      Remove a couple of unused fields from data structures.
110
111Thu May 11 12:40 2000 Pam Delaney (pam.delaney@lsil.com)
112    * version sym53c8xx-1.6b
113    - Merged version.
114
115Mon Apr 24 12:00 2000 Gerard Roudier (groudier@club-internet.fr)
116    * version sym53c8xx-1.5m
117    - Return value 1 (instead of 0) from the driver setup routine.
118    - Do not enable PCI DAC cycles. This just broke support for
119      SYM534C896 on sparc64. Problem fixed by David S. Miller.
120
121Fri Apr 14 9:00 2000 Pam Delaney (pam.delaney@lsil.com)
122    * version sym53c8xx-1.6b-9
123    - Added 53C1010_66 support.
124    - Small fix to integrity checking code.
125    - Removed requirement for integrity checking if want to run
126      at ultra 3.
127 
128Sat Apr 1 12:00 2000 Gerard Roudier (groudier@club-internet.fr)
129    * version sym53c8xx-1.5l
130    - Tiny change for __sparc__ appeared in 2.3.99-pre4.1 that
131      applies to cache line size (? Probably from David S Miller).
132    - Make sure no data transfer will happen for Scsi_Cmnd requests
133      that supply SCSI_DATA_NONE direction (this avoids some BUG()
134      statement in the PCI code when a data buffer is also supplied).
135
136Sat Mar 11 12:00 2000 Gerard Roudier (groudier@club-internet.fr)
137    * version sym53c8xx-1.6b-5
138    - Test against expected data transfer direction from SCRIPTS.
139    - Add support for the new dynamic dma mapping kernel interface.
140      Requires Linux-2.3.47 (tested with pre-2.3.47-6).
141      Many thanks to David S. Miller for his preliminary changes
142      that have been useful guidelines.
143    - Get data transfer direction from the scsi command structure
144      (Scsi_Cmnd) with kernels that provide this information.
145
146Mon Mar 6 23:30 2000 Gerard Roudier (groudier@club-internet.fr)
147    * version sym53c8xx-1.5k
148    - Test against expected data transfer direction from SCRIPTS.
149    - Revert the change in 'ncr_flush_done_cmds()' but unmap the
150      scsi dma buffer prior to queueing the command to our done
151      list.
152    - Miscellaneous (minor) fixes in the code added in driver
153      version 1.5j.
154
155Mon Feb 14 4:00 2000 Pam Delaney (pam.delaney@lsil.com)
156     * version sym53c8xx-pre-1.6b-2.
157    - Updated the SCRIPTS error handling of the SWIDE
158      condition - to remove any reads of the sbdl
159      register. Changes needed because the 896 and 1010
160      chips will check parity in some special circumstances.
161      This will cause a parity error interrupt if not in
162      data phase. Changes based on those made in the
163      FreeBSD driver version 1.3.2.
164
165Sun Feb 20 11:00 2000 Gerard Roudier (groudier@club-internet.fr)
166    * version sym53c8xx-1.5j
167    - Add support for the new dynamic dma mapping kernel interface.
168      Requires Linux-2.3.47 (tested with pre-2.3.47-6).
169      Many thanks to David S. Miller for his preliminary changes
170      that have been useful guidelines, for having reviewed the
171      code and having tested this driver version on Ultra-Sparc.
172    - 2 tiny bugs fixed in the PCI wrapper that provides support
173      for early kernels without pci device structure.
174    - Get data transfer direction from the scsi command structure
175      (Scsi_Cmnd) with kernels that provide this information.
176    - Fix an old bug that only affected 896 rev. 1 when driver
177      profile support option was set in kernel configuration.
178
179Fri Jan 14 14:00 2000 Pam Delaney (pam.delaney@lsil.com)
180     * version sym53c8xx-pre-1.6b-1.
181    - Merge parallel driver series 1.61 and 1.5e
182
183Tue Jan 11 14:00 2000 Pam Delaney (pam.delaney@lsil.com)
184    * version sym53c8xx-1.61
185    - Added support for mounting disks on wide-narrow-wide
186      scsi configurations.
187    - Modified offset to be a maximum of 31 in ST mode,
188      62 in DT mode.
189    - Based off of 1.60
190
191Mon Jan 10 10:00 2000 Pam Delaney (pam.delaney@lsil.com)
192    * version sym53c8xx-1.60
193    - Added capability to use the integrity checking code
194      in the kernel (optional).
195    - Added PPR negotiation.
196    - Added support for 53C1010 Ultra 3 part.
197    - Based off of 1.5f
198
199Sat Jan 8 22:00 2000 Gerard Roudier (groudier@club-internet.fr)
200    * version sym53c8xx-1.5h
201    - Add year 2000 copyright.
202    - Display correctly bus signals when bus is detected wrong.
203    - Some fix for Sparc from DSM that went directly to kernel tree.
204
205Mon Dec 6 22:00 1999 Gerard Roudier (groudier@club-internet.fr)
206    * version sym53c8xx-1.5g
207    - Change messages written by the driver at initialisation and
208      through the /proc FS (rather cosmetic changes that consist in
209      printing out the PCI bus number and PCI device/function).
210    - Ensure the SCRIPTS processor is stopped while calibrating the
211      SCSI clock (the initialisation code has been a bit reworked).
212      Change moved to the FreeBSD sym_hipd driver).
213    - Some fixes in the MODIFY_DP/IGN_RESIDUE code and residual
214      calculation (moved from FreeBSD sym_hipd driver).
215    - Add NVRAM support for Tekram boards that use 24C16 EEPROM.
216      Code moved from the FreeBSD sym_hipd driver, since it has
217      been that one that got this feature first.
218    - Definitely disable overlapped PCI arbitration for all dual
219      function chips, since I cannot make sure for what chip revisions
220      it is actually safe.
221    - Add support for the SYM53C1510D (also for ncr53c8xx).
222    - Fix up properly the PCI latency timer when needed or asked for.
223    - Get rid of the old PCI bios interface, but preserve kernel 2.0
224      compatibility from a simple wrapper.
225    - Update the poor Tekram sync factor table.
226    - Fix in a tiny 'printk' bug that may oops in case of extended
227      errors (unrecovered parity error, data overrun, etc ...)
228      (Sent by Pamela Delaney from LSILOGIC)
229    - Remove the compilation condition about having to acquire the
230      io_request_lock since it seems to be a definite feature now.:)
231    - Change get_pages by GetPages since Linux >= 2.3.27 now wants
232      get_pages to ever be used as a kernel symbol (from 2.3.27).
233    - proc_dir structure no longer needed for kernel >= 2.3.27.
234
235Sun Oct 3 19:00 1999 Gerard Roudier (groudier@club-internet.fr)
236    * version sym53c8xx-1.5f
237    - Change the way the driver checks the PCI clock frequency, so
238      that overclocked PCI BUS up to 48 MHz will not be refused.
239      The more the BUS is overclocked, the less the driver will
240      guarantee that its measure of the SCSI clock is correct.
241    - Backport some minor improvements of SCRIPTS from the sym_hipd
242      driver.
243    - Backport the code rewrite of the START QUEUE dequeuing (on
244      bad scsi status received) from the sym_hipd driver.
245
246Sat Sep 11 11:00 1999 Gerard Roudier (groudier@club-internet.fr)
247    * version sym53c8xx-1.5e
248    - New linux-2.3.13 __setup scheme support added.
249    - Cleanup of the extended error status handling:
250      Use 1 bit per error type.
251    - Also save the extended error status prior to auto-sense.
252    - Add the FE_DIFF chip feature bit to indicate support of
253      diff probing from GPIO3 (825/825A/876/875).
254    - Remove the quirk handling that has been useless since day one.
255    - Work-around PCI chips being reported twice on some platforms.
256    - Add some redundant PCI reads in order to deal with common
257      bridge misbehaviour regarding posted write flushing.
258    - Add some other conditionnal code for people who have to deal
259      with really broken bridges (they will have to edit a source
260      file to try these options).
261    - Handle correctly (hopefully) jiffies wrap-around.
262    - Restore the entry used to detect 875 until revision 0xff.
263      (I removed it inadvertently, it seems :) )
264    - Replace __initfunc() which is deprecated stuff by __init which
265      is not yet so. ;-)
266    - Rewrite the MESSAGE IN scripts more generic by using a MOVE
267      table indirect. Extended messages of any size are accepted now.
268      (Size is limited to 8 for now, but a constant is just to be
269      increased if necessary)
270    - Fix some bug in the fully untested MDP handling:) and share
271      some code between MDP handling and residual calculation.
272    - Calculate the data transfer residual as the 2's complement
273      integer (A positive value in returned on data overrun, and
274      a negative one on underrun).
275    - Add support of some 'resource handling' for linux-2.3.13.
276      Basically the BARs have been changed to something more complex
277      in the pci_dev structure.
278    - Remove some deprecated code.
279
280Sat Jun 5 11:00 1999 Gerard Roudier (groudier@club-internet.fr)
281    * version sym53c8xx-1.5c
282    - Do not negotiate on auto-sense if we are currently using 8 bit
283      async transfer for the target.
284    - Only check for SISL/RAID on i386 platforms.
285      (A problem has been reported on PPC with that code).
286    - On MSG REJECT for a negotiation, the driver attempted to restart
287      the SCRIPT processor when this one was already running.
288
289Sat May 29 12:00 1999 Gerard Roudier (groudier@club-internet.fr)
290    * version sym53c8xx-1.5b
291    - Force negotiation prior auto-sense.
292      This ensures that the driver will be able to grab the sense data
293      from a device that has received a BUS DEVICE RESET message from
294      another initiator.
295    - Complete all disconnected CCBs for a logical UNIT if we are told
296      about a UNIT ATTENTION for a RESET condition by this target.
297    - Add the control command 'cleardev' that allows to send a ABORT
298      message to a logical UNIT (for test purpose).
299
300Tue May 25 23:00 1999 Gerard Roudier (groudier@club-internet.fr)
301    * version sym53c8xx-1.5a
302    - Add support for task abort and bus device reset SCSI message
303      and implement proper synchonisation with SCRIPTS to handle
304          correctly task abortion without races.
305    - Send an ABORT message (if untagged) or ABORT TAG message (if tagged)
306      when the driver is told to abort a command that is disconnected and
307      complete the command with appropriate error.
308      If the aborted command is not yet started, remove it from the start
309      queue and complete it with error.
310    - Add the control command 'resetdev' that allows to send a BUS
311      DEVICE RESET message to a target (for test purpose).
312    - Clean-up some unused or useless code.
313
314Fri May 21 23:00 1999 Gerard Roudier (groudier@club-internet.fr)
315    * version sym53c8xx-1.5
316    - Add support for CHMOV with Wide controllers.
317    - Handling of the SWIDE (low byte residue at the end of a CHMOV
318      in DATA IN phase with WIDE transfer when the byte count gets odd).
319    - Handling of the IGNORE WIDE RESIDUE message.
320      Handled from SCRIPTS as possible with some optimizations when both
321      a wide device and the controller are odd at the same time (SWIDE
322      present and IGNORE WIDE RESIDUE message on the BUS at the same time).
323    - Check against data OVERRUN/UNDERRUN condition at the end of a data
324      transfer, whatever a SWIDE is present (OVERRUN in DATA IN phase)
325      or the SODL is full (UNDERRUN in DATA out phase).
326    - Handling of the MODIFY DATA POINTER message.
327      This one cannot be handled from SCRIPTS, but hopefully it will not
328      happen very often. :)
329    - Large rewrite of the SCSI MESSAGE handling.
330
331Sun May 9 11:00 1999 Gerard Roudier (groudier@club-internet.fr)
332    * version sym53c8xx-1.4
333    - Support for IMMEDIATE ARBITRATION.
334      See the README file for detailed information about this feature.
335      Requires both a compile option and a boot option.
336    - Minor SCRIPTS optimization in reselection pattern for LUN 0.
337    - Simpler algorithm to deal with SCSI command starvation.
338      Just use 2 tag counters in flip/flop and switch to the other
339      one every 3 seconds.
340    - Do some work in SCRIPTS after the SELECT instruction and prior
341      to testing for a PHASE. SYMBIOS say this feature is working fine.
342      (Btw, only problems with Toshiba 3401B had been reported).
343    - Measure the PCI clock speed and do not attach controllers if
344      result is greater than 37 MHz. Since the precision of the
345      algorithm (from Stefan Esser) is better than 2%, this should
346      be fine.
347    - Fix the misdetection of SYM53C875E (was detected as a 876).
348    - Fix the misdetection of SYM53C810 not A (was detected as a 810A).
349    - Support for up to 256 TAGS per LUN (CMD_PER_LUN).
350      Currently limited to 255 due to Linux limitation. :)
351    - Support for up to 508 active commands (CAN_QUEUE).
352    - Support for the 53C895A by Pamela Delaney <pam.delaney@lsil.com>
353      The 53C895A contains all of the features of the 896 but has only
354      one channel and has a 32 bit PCI bus. It does 64 bit PCI addressing
355      using dual cycle PCI data transfers.
356    - Miscellaneous minor fixes.
357    - Some additions to the README.ncr53c8xx file.
358
359Tue Apr 15 10:00 1999 Gerard Roudier (groudier@club-internet.fr)
360    * version sym53c8xx-1.3e
361    - Support for any number of LUNs (64) (SPI2-compliant).
362      (Btw, this may only be ever useful under linux-2.2 ;-))
363
364Sun Apr 11 10:00 1999 Gerard Roudier (groudier@club-internet.fr)
365    * version sym53c8xx-1.3d
366    - Add 'hostid:#id' boot option. This option allows to change the
367      default SCSI id the driver uses for controllers.
368    - Make SCRIPTS not use self-mastering for PCI.
369      There were still 2 places the driver used this feature of the
370      53C8XX family.
371    - Move some data structures (nvram layouts and driver set-up) to
372      the sym53c8xx_defs.h file. So, the both drivers will share them.
373    - Set MAX LUNS to 16 (instead of 8).
374
375Sat Mar 20 21:00 1999 Gerard Roudier (groudier@club-internet.fr)
376    * version sym53c8xx-1.3b
377    - Add support for NCR PQS PDS.
378      James Bottomley <James.Bottomley@columbiasc.ncr.com>
379    - Allow value 0 for host ID.
380    - Support more than 8 controllers (> 40 in fact :-) )
381    - Add 'excl=#ioaddr' boot option: exclude controller.
382      (Version 1.3a driver)
383
384Thu Mar 11 23:00 1999 Gerard Roudier (groudier@club-internet.fr)
385    * version sym53c8xx-1.3 (8xx-896 driver bundle)
386    - Equivalent changes as ncr53c8xx-3.2 due to the driver bundle.
387      (See Changelog.ncr53c8xx)
388    - Do a normal soft reset as first chip reset, since aborting current
389      operation may raise an interrupt we are not able to handle since
390      the interrupt handler is not yet established.
391
392Sat Mar 6 11:00 1999 Gerard Roudier (groudier@club-internet.fr)
393    * version sym53c8xx-1.2b
394    - Fix some oooold bug that hangs the bus if a device rejects a
395      negotiation. Btw, the corresponding stuff also needed some cleanup
396      and thus the change is a bit larger than it could have been.
397    - Still some typo that made compilation fail for 64 bit (trivial fix).
398
399Sun Feb 21 20:00 1999 Gerard Roudier (groudier@club-internet.fr)
400    * version sym53c8xx-1.2a
401    - The rewrite of the interrupt handling broke the SBMC interrupt
402      handling due to a 1 bit mask tiny error. Hopefully fixed.
403    - If INQUIRY came from a scatter list, the driver looked into
404      the scatterlist instead of the data.:) Since this should never
405      happen, we just discard the data if use_sg is not zero.
406
407Fri Feb 12 23:00 1999 Gerard Roudier (groudier@club-internet.fr)
408    * version sym53c8xx-1.2
409    - Major rewrite of the interrupt handling and recovery stuff for
410      the support of non compliant SCSI removal, insertion and all
411      kinds of screw-up that may happen on the SCSI BUS.
412      Hopefully, the driver is now unbreakable or may-be, it is just
413      quite brocken. :-)
414      Many thanks to Johnson Russel (Symbios) for having responded to
415      my questions and for his interesting advices and comments about
416      support of SCSI hot-plug.
417    - Add 'recovery' option to driver set-up.
418    - Negotiate SYNC data transfers with CCS devices.
419    - Deal correctly with 64 bit PCI address registers on Linux 2.2.
420      Pointed out by Leonard Zubkoff.
421
422Sun Jan 31 18:00 1999 Gerard Roudier (groudier@club-internet.fr)
423    * version sym53c8xx-1.1a
424    - Some 896 chip revisions (all for now :-)), may hang-up if the
425      soft reset bit is set at the wrong time while SCRIPTS are running.
426      We need to first abort the current SCRIPTS operation prior to
427      resetting the chip. This fix has been sent to me by SYMBIOS/LSI
428      and I just translated it into ncr53c8xx syntax.
429      Must be considered 100 % trustable, unless I did some mistake
430      when translating it. :-)
431
432Sun Jan 24 18:00 1999 Gerard Roudier (groudier@club-internet.fr)
433    * version sym53c8xx-1.1
434    - Major rewrite of the SCSI parity error handling.
435      The informations contained in the data manuals are incomplete about
436      this feature.
437      I asked SYMBIOS about and got in reply the explanations that are
438      _indeed_ missing in the data manuals.
439    - Allow to tune request_irq() flags from the boot command line using
440      ncr53c8xx=irqm:??, as follows:
441      a) If bit 0x10 is set in irqm, SA_SHIRQ flag is not used.
442      b) If bit 0x20 is set in irqm, SA_INTERRUPT flag is not used.
443      By default the driver uses both SA_SHIRQ and SA_INTERRUPT.
444      Option 'ncr53c8xx=irqm:0x20' may be used when an IRQ is shared by
445      a 53C8XX adapter and a network board.
446    - Fix for 64 bit PCI address register calculation. (Lance Robinson)
447    - Fix for big-endian in phase mismatch handling. (Michal Jaegermann)
448
449Fri Jan 1 20:00 1999 Gerard Roudier (groudier@club-internet.fr)
450    * version sym53c8xx-1.0a
451    - Waiting list look-up didn't work for the first command of the list.
452      Hopefully fixed, but tested on paper only. ;)
453    - Remove the most part of PPC specific code for Linux-2.2.
454      Thanks to Cort.
455    - Some other minors changes.
456
457Sat Dec 19 21:00 1998 Gerard Roudier (groudier@club-internet.fr)
458    * version sym53c8xx-1.0
459    - Define some new IO registers for the 896 (istat1, mbox0, mbox1)
460    - Revamp slightly the Symbios NVRAM lay-out based on the excerpt of
461      the header file I received from Symbios.
462    - Check the PCI bus number for the boot order (Using a fast
463      PCI controller behing a PCI-PCI bridge seems sub-optimal).
464    - Disable overlapped PCI arbitration for the 896 revision 1.
465    - Reduce a bit the number of IO register reads for phase mismatch
466      by reading DWORDS at a time instead of BYTES.
467
468Thu Dec 3 24:00 1998 Gerard Roudier (groudier@club-internet.fr)
469    * version pre-sym53c8xx-0.18
470    - I received this afternoon a 896 from SYMBIOS and started testing
471      the driver with this beast. After having fixed 3 buglets, it worked
472      with all features enabled including the phase mismatch handling
473      from SCRIPTS. Since this feature is not yet tested enough, the
474      boot option 'ncr53c8xx=specf:1' is still required to enable the
475      driver to handle PM from SCRIPTS.
476
477Sun Nov 29 18:00 1998 Gerard Roudier (groudier@club-internet.fr)
478    * version pre-sym53c8xx-0.17
479    - The SISL RAID change requires now remap_pci_mem() stuff to be
480      compiled for __i386__ when normal IOs are used.
481    - The PCI memory read from SCRIPTS that should ensure ordering
482      was in fact misplaced. BTW, this may explain why broken PCI
483      device drivers regarding ordering are working so well. ;-)
484    - Rewrite ncr53c8xx_setup (boot command line options) since the
485      binary code was a bit too bloated in my opinion.
486    - Make the code simpler in the wakeup_done routine.
487
488Tue Nov 24 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
489    * version pre-sym53c8xx-0.16
490    - Add SCSI_NCR_OPTIMIZE_896_1 compile option and 'optim' boot option.
491      When set, the driver unconditionnaly assumes that the interrupt
492      handler is called for command completion, then clears INTF, scans
493      the done queue and returns if some completed CCB is found. If no
494      completed CCB are found, interrupt handling will proceed normally.
495      With a 896 that handles MA from SCRIPTS, this can be a great win,
496      since the driver will never performs PCI read transactions, but
497      only PCI write transactions that may be posted.
498      If the driver haven't to also raise the SIGP this would be perfect.
499      Even with this penalty, I think that this will work great.
500      Obviously this optimization makes sense only if the IRQ is not
501      shared with another device.
502    - Still a buglet in the tags initial settings that needed to be fixed.
503      It was not possible to disable TGQ at system startup for devices
504      that claim TGQ support. The driver used at least 2 for the queue
505      depth but did'nt keep track of user settings for tags depth lower
506      than 2.
507
508Thu Nov 19 23:00 1998 Gerard Roudier (groudier@club-internet.fr)
509    * version pre-sym53c8xx-0.15
510    - Add support for hardware LED control of the 896.
511    - Ignore chips that are driven by SISL RAID (DAC 960).
512      Change sent by Leonard Zubkoff and slightly reworked.
513    - Prevent 810A rev 11 and 860 rev 1 from using cache line based
514      transactions since those early chip revisions may use such on
515      LOAD/STORE instructions (work-around).
516    - Remove some useless and bloat code from the pci init stuff.
517    - Do not use the readX()/writeX() kernel functions for __i386__,
518      since they perform useless masking operations in order to deal
519      with broken driver in 2.1.X kernel.
520
521Wed Nov 11 10:00 1998 Gerard Roudier (groudier@club-internet.fr)
522    * version pre-sym53c8xx-0.14
523    - The driver was unhappy when configured with default_tags > MAX_TAGS
524      Hopefully doubly-fixed.
525    - Set PCI_PARITY in PCI_COMMAND register in not set (PCI fix-up).
526    - Print out some message if phase mismatch is handled from SCRIPTS.
527
528Sun Nov 1 14H00 1998 Gerard Roudier (groudier@club-internet.fr)
529    * version pre-sym53c8xx-0.13
530    - Some rewrite of the device detection code. This code had been
531      patched too much and needed to be face-lifted a bit.
532      Remove all platform dependent fix-ups that was not needed or
533      conflicted with some other driver code as work-arounds.
534      Reread the NVRAM before the calling of ncr_attach(). This spares
535      stack space and so allows to handle more boards.
536      Handle 64 bit base addresses under linux-2.0.X.
537      Set MASTER bit in PCI COMMAND register if not set.
538
539Wed Oct 30 22H00 1998 Gerard Roudier (groudier@club-internet.fr)
540    * version pre-sym53c8xx-0.12
541    - Damned! I just broke the driver for Alpha by leaving a stale
542      instruction in the source code. Hopefully fixed.
543    - Do not set PFEN when it is useless. Doing so we are sure that BOF
544      will be active, since the manual appears to be very unclear on what
545      feature is actually used by the chip when both PFEN and BOF are
546      set.
547
548Sat Oct 24 16H00 1998 Gerard Roudier (groudier@club-internet.fr)
549    * version pre-sym53c8xx-0.11
550    - LOAD/STORE instructions were miscompiled for register offsets
551      beyond 0x7f. This broke accesses to 896' new registers.
552    - Disable by default Phase Mismatch handling from SCRIPTS, since
553      current 896 rev.1 seems not to operate safely with the driver
554      when this feature is enabled (and above LOAD/STORE fix applied).
555      I will change the default to 'enabled' when this problem will be
556      solved.
557      Using boot option 'ncr53c8xx=specf:1' enables this feature.
558    - Implement a work-around (DEL 472 - ITEM 5) that should allow the
559      driver to safely enable hardware phase mismatch with 896 rev. 1.
560
561Tue Oct 20 22H00 1998 Gerard Roudier (groudier@club-internet.fr)
562    * version pre-sym53c8xx-0.10
563    - Add the 53c876 description to the chip table. This is only useful
564      for printing the right name of the controller.
565    - Add additional checking of INQUIRY data:
566      Check INQUIRY data received length is at least 7. Byte 7 of
567      inquiry data contains device features bits and the driver might
568      be confused by garbage. Also check peripheral qualifier.
569    - Use a 1,3,5,...MAXTAGS*2+1 tag numbering. Previous driver could
570      use any tag number from 1 to 253 and some non conformant devices
571      might have problems with large tag numbers.
572    - Use NAME53C and NAME53C8XX for chip name prefix chip family name.
573      Just give a try using "sym53c" and "sym53c8xx" instead of "ncr53c"
574      and "ncr53c8xx". :-)
575
576Sun Oct 11 17H00 1998 Gerard Roudier (groudier@club-internet.fr)
577    * version pre-sym53c8xx-0.9
578    - DEL-441 Item 2 work-around for the 53c876 rev <= 5 (0x15).
579    - Break ncr_scatter() into 2 functions in order to guarantee best
580      possible code optimization for the case we get a scatter list.
581    - Add the code intended to support up to 1 tera-byte for 64 bit systems.
582      It is probably too early, but I wanted to complete the thing.
583
584Sat Oct 3 14H00 1998 Gerard Roudier (groudier@club-internet.fr)
585    * version pre-sym53c8xx-0.8
586    - Do some testing with io_mapped and fix what needed to be so.
587    - Wait for SCSI selection to complete or time-out immediately after
588      the chip won arbitration, since executing SCRIPTS while the SCSI
589      core is performing SCSI selection breaks the selection procedure
590      at least for some chip revisions.
591    - Interrupt the SCRIPTS if a device does not go to MSG OUT phase after
592      having been selected with ATN. Such a situation is not recoverable,
593      better to fail when we are stuck.
594

Archive Download this file



interactive