struct asc_dvc_var needs overrun buffer to be placed on an 8 byte
boundary. advansys defines struct asc_dvc_var:

struct asc_dvc_var {
	...
	uchar overrun_buf[ASC_OVERRUN_BSIZE] __aligned(8);

The problem is that struct asc_dvc_var is placed on
shost->hostdata. So if the hostdata is not on an 8 byte boundary, the
advansys crashes. The hostdata is placed on a sizeof(unsigned long)
boundary so the 8 byte boundary is not garanteed with x86_32.

With 2.6.23 and 2.6.24, the hostdata is on an 8 byte boundary by
chance, but with the current git, it's not.

This patch removes overrun_buf static array and use kzalloc.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Signed-off-by: NAdrian Bunk <bunk@kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

This replaces sizeof sense_buffer with SCSI_SENSE_BUFFERSIZE in
several LLDs. It's a preparation for the future changes to remove
sense_buffer array in scsi_cmnd structure.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: NJames Bottomley <James.Bottomley@HansenPartnership.com>

Fix section mismatch warning:

WARNING: vmlinux.o(.exit.text+0x152a): Section mismatch: reference to .init.data:_asc_def_iop_base (between 'advansys_isa_remove' and 'advansys_exit')
Signed-off-by: NRandy Dunlap <randy.dunlap@oracle.com>
Cc: Matthew Wilcox <willy@debian.org>
Cc: James Bottomley <James.Bottomley@steeleye.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Change PortAddr to be an unsigned int instead of an unsigned short (IO
Port address are 24 bit on parisc).  Fix a couple of printk argument
warnings.  Remove the Kconfig marking as 'BROKEN'.

I haven't removed the #warning yet because virt_to_bus/bus_to_virt are
only eliminated for narrow boards.  Wide boards need more work.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Convert the call to virt_to_bus() into a call to dma_map_single().  Some
architectures may require different DMA addresses for different devices,
so allocate one overrun buffer per host rather than one for all cards.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Replace ASC_VADDR_TO_U32 and ASC_U32_TO_VADDR with an auto-expanding
array that maps pointers to 32-bit IDs and back.  One of the uses of
ASC_VADDR_TO_U32 was in error; it should have been using ADV_VADDR_TO_U32.

Also replace the use of virt_to_bus when setting the sense_address with
a call to dma_map_single() followed by dma_cache_sync.  This part cribbed
from the 53c700 driver.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

- remove the unnecessary map_single path.
- convert to use the new accessors for the sg lists and the parameters.
Signed-off-by: NFUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>

- convert the statistics to not distinguish between single and sg xfers
- replace ASC_CEILING with DIV_ROUND_UP
- remove an obsolete comment
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

This rather complex function boiled down to calling virt_to_bus().
Also get rid of some obsolete defines and variables that could never vary.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

board->carrp is a duplicate of asc_dvc->carrier_buf, so cut out the
middle-man and assign directly to carrier_buf.  Move orig_reqp to adv_dvc
too, since it's wide-board specific.  Also eliminate an unnecessary BUG_ON
(we'll never get there with a NULL carrier_buf, and will crash if we do).
The bulk of this patch is rearranging structures so everything's declared
in the right order.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

It's somewhat neater to make this a pointer to one of two tables
than initialising an array in the driver.  Also delete the unused
AscSynIndexToPeriod and rename host_init_sdtr_index to min_sdtr_index
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

It's always a mistake to have your own index of boards; just use the
scsi host number.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The board lock was essentially identical with the host lock.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Replace ASC_DBG{,1,2,3,4,5} with a single variadic macro ASC_DBG.  As
suggested by Jeff Garzik, include DRV_NAME and __FUNCTION__ in the output.
Change all callers to no longer include the function name in the string.

Enabling ADVANSYS_DEBUG to test this feature shows a lot of other problems
that need to be fixed:
 - Reorder asc_prt_* functions now that their prototypes have been removed.
 - There is no longer a struct device in ASC_DVC_CFG/ADV_DVC_CFG, and it
   wasn't necessarily a PCI device to begin with.  Print the bus_id from
   asc_board->dev instead.
 - isr_callback no longer exists.
 - ASC_DBG_PRT_SCSI_CMND isn't being used, so delete asc_prt_scsi_cmnd
   too.
 - A missing semicolon
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

With the ASC and ADV libraries merged into the driver, there really is
no point in reporting their version numbers, or even trying to maintain
them.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

asc_board_t was simply a typedef for struct asc_board.  ASC_BOARDP()
can be replaced by shost_priv() except in the ASC_STATS* macros which
rely on the cast; add an explicit cast there.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

There were two blocks of ASC_IERR definitions; one for narrow and one for
wide boards.  Some of the same names were used (with the same values),
and some of the same values were used with different names.  This could
only lead to confusion, so I unified them in one block of definitions
with no overlapping values.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The interrupt number was being stored in 4-5 different places, each with
its own type, rules and usage.  Fix this by keeping an unsigned int in
the struct asc_board, and filling it in from the bus probe functions
(since it's different for each of the four bus types).  In order to do
this, we have to allocate the Scsi_Host in the bus probe functions too.
Then we can return an error from advansys_board_found, which requires
a little rearranging of code (and removing of the err_code variable).
Move the Wide Board flag setting into the PCI bus probe function.

Split the AscGetChipIRQ function into three functions (one for each bus
type that needs it) and add some commentary to explain what's going on.
Also get rid of the AscSetChipIRQ function as we only ever set the
interrupt number to the same value it already had.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Replace it with !ASC_NARROW_BOARD
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

It was only ever set; never tested, nor cleared.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

 - Don't need to set ASC_HOST_IN_RESET any more
 - Don't need to test scp->device->host for NULL -- if it's NULL, we
   couldn't've been called.
 - Use scmd_printk instead of ASC_PRINT
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The narrow board used two global structures to set up a command;
unfortunately they weren't locked, so with two boards in the machine,
one call to queuecommand could corrupt the data being used by the other
call to queuecommand.

Fix this by allocating asc_scsi_q on the stack (64 bytes) and using kmalloc
for the asc_sg_head (2k)
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The wide and narrow boards share identical handling of the return value,
except for some trivial error messages.  Move the handling to the common
end of the function.  Also move variable declarations to the arms of
the `if' that they're used in and delete some pointless comments.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The driver was saving a scsi_device for each target, but wasn't doing
anything useful with them.  Just delete the array.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Rearrange a lot of the functions in the file to get rid of all the forward
declarations.
Signed-off-by: NMatthew Wilcox <willy@linux.intel.com>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Follow the example of some other drivers by defining DRV_NAME to be
"advansys".  Prevents spelling mistakes.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

 - Delete comments relating to the previous structure of the driver.
   I have no intention of honouring them ;-)
 - Reformat comments > 80 columns
 - Remove now-obsolete comments from advansys_interrupt
 - Change adv_get_sglist() from do {} while (1) to for (;;)
 - Return void from AscInitQLinkVar()
 - Take out a level of indentation in adv_get_sglist()
 - Reduce indentation level of AscAsyncFix()
 - Remove unused macros
 - Refactor AscSendScsiQueue slightly
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

 - Replace ASC_ASSERT() with BUG_ON().
   In a few places, get rid of the assertion altogether -- the ensuing
   crash will tell us all we need to know.  Use BUG() where it fits better
   than BUG_ON().  Also fix a fencepost error in advansys_proc_info().
 - Replace DvcSleepMilliSecond with mdelay.
   Despite its name using 'sleep', the implementation was a delay.
   I've marked some places with XXX where we should probably be using
   msleep instead.  They need to be audited to be sure we can sleep in
   that context.
 - Replace DvcDelayMicroSecond with udelay.
 - Replace DvcDelayNanoSecond with udelay too.
   All callers were multiples of 1000.
 - Remove DvcEnterCritical and DvcLeaveCritical.
   These functions are no-ops, and as the comments said, the spinlock
   protects the critical sections.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

AscExeScsiQueue() has one caller, and it passes the address of a
variable; this cannot ever be NULL.  This is the only place that ever
sets ASCQ_ERR_SCSIQ_NULL_PTR, so delete that error code too, as well as
several other unused ASCQ_ERR codes.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The SCSI midlayer won't send commands greater than ->max_cmd_len.
So the checks on length in asc_build_req and adv_build_req are obsolete
and can be deleted, but also we have to set the max_cmd_len in
advansys_board_found().  Also move the length definitions together, and
write a helpful comment.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Move as much as possible outside the critical section in queuecommand, eg:
 - Set the scsi_done field before acquiring the lock
 - Call asc_scsi_done after dropping the lock

Also remove a comment suggesting we should enable interrupts (now we do)
and do some minor reformatting for readability.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The `active' list was used in the reset handler -- but the midlayer
guarantees that list is empty by the time the reset handler is invoked.
It was also checked in the interrupt routines to be sure that this command
belonged to this board, but we don't need to check that either.  We can
then delete the asc_prt_target_stats() function as it will never print
anything, along with asc_enqueue(), asc_dequeue_list(), asc_rmqueue(),
asc_scsi_done_list(), struct asc_queue, ASC_QUEUE_EMPTY, ASC_TID_ALL,
ASC_FRONT, ASC_BACK, and all the REQ* macros.  Also remove this item from
the todo list.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

 - Move the guts of asc_scsi_done_list() into a new function, asc_scsi_done.
 - Call asc_scsi_done() in asc_isr_callback() and adv_isr_callback().  The
   comment was wrong; scsi_done cannot enable interrupts.
 - All other places which queued an scp on the done list are error paths
   for queuecommand, and so we can just call asc_scsi_done() in queuecommand
   if we receive an error.
 - We no longer need to keep a list of done requests in advansys_interrupt
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

The midlayer guarantees it won't call ->queuecommand for a host which
is handling a reset condition.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

If the adapter is busy, return the request to the midlayer rather than
queueing it in the driver.  asc_execute_queue() and asc_dequeue() become
unused, and we don't need to print out stats on the waiting queue any more.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Move the error reporting into AscInitGetConfig, AdvInitGetConfig and
AscInitSetConfig.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

It's just data, so format it to something that looks more visually
appealing (and saves some lines)
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Split AdvLoadMicrocode out of AdvInitAsc3550Driver, AdvInitAsc38C0800Driver
and AdvInitAsc38C1600Driver.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

Split AdvBuildCarrierFreelist out of AdvInitAsc3550Driver,
AdvInitAsc38C0800Driver and AdvInitAsc38C1600Driver.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>

You can't have two drivers for the same bus type with the same name.
Since ISA and VLB are both isa_drivers, rename the VLB one to advansys_vlb.
Signed-off-by: NMatthew Wilcox <matthew@wil.cx>
Signed-off-by: NJames Bottomley <James.Bottomley@SteelEye.com>