linux.git/drivers/scsi/arm, branch v4.5

ncr5380: Fix soft lockups

2016-01-07T02:43:09+00:00

Because of the rudimentary design of the chip, it is necessary to poll the
SCSI bus signals during PIO and this tends to hog the CPU. The driver will
accept new commands while others execute, and this causes a soft lockup
because the workqueue item will not terminate until the issue queue is
emptied.

When exercising dmx3191d using sequential IO from dd, the driver is sent
512 KiB WRITE commands and 128 KiB READs. For a PIO transfer, the rate is
is only about 300 KiB/s, so these are long-running commands. And although
PDMA may run at several MiB/s, interrupts are disabled for the duration
of the transfer.

Fix the unresponsiveness and soft lockup issues by calling cond_resched()
after each command is completed and by limiting max_sectors for drivers
that don't implement real DMA.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Use standard list data structure

2016-01-07T02:43:07+00:00

The NCR5380 drivers have a home-spun linked list implementation for
scsi_cmnd structs that uses cmd->host_scribble as a 'next' pointer. Adopt
the standard list_head data structure and list operations instead. Remove
the eh_abort_handler rather than convert it. Doing the conversion would
only be churn because the existing EH handlers don't work and get replaced
in a subsequent patch.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Remove command list debug code

2016-01-07T02:43:05+00:00

Some NCR5380 hosts offer a .show_info method to access the contents of
the various command list data structures from a procfs file. When NDEBUG
is set, the same information is sent to the console during EH.

The two core drivers, atari_NCR5380.c and NCR5380.c differ here. Because
it is just for debugging, the easiest way to fix the discrepancy is
simply remove this code.

The only remaining users of NCR5380_show_info() and NCR5380_write_info()
are drivers that define PSEUDO_DMA. The others have no use for the
.show_info method, so don't initialize it.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Cleanup #include directives

2016-01-07T02:43:02+00:00

Remove unused includes (stat.h, signal.h, proc_fs.h) and move includes
needed by the core drivers into the common header (delay.h etc).

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Implement NCR5380_dma_xfer_len and remove LIMIT_TRANSFERSIZE macro

2016-01-07T02:42:56+00:00

Follow the example of the atari_NCR5380.c core driver and adopt the
NCR5380_dma_xfer_len() hook. Implement NCR5380_dma_xfer_len() for dtc.c
and g_NCR5380.c to take care of the limitations of these cards. Keep the
default for drivers using PSEUDO_DMA.

Eliminate the unused macro LIMIT_TRANSFERSIZE.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Introduce unbound workqueue

2016-01-07T02:42:56+00:00

Allocate a work queue that will permit busy waiting and sleeping. This
means NCR5380_init() can potentially fail, so add this error path.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Tested-by: Michael Schmitz 
Signed-off-by: Martin K. Petersen

ncr5380: Cleanup bogus {request,release}_region() calls

2016-01-07T02:42:55+00:00

Commit 8b801ead3d7a ("[ARM] rpc: update Acorn SCSI drivers to modern ecard
interfaces") neglected to remove a request_region() call in cumana_1.c.

Commit eda32612f7b2 ("[PATCH] give all LLDD driver a ->release method") in
history/history.git added some pointless release_region() calls in dtc.c,
pas16.c and t128.c.

Fix these issues.

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Signed-off-by: Martin K. Petersen

ncr5380: Split NCR5380_init() into two functions

2016-01-07T02:42:52+00:00

This patch splits the NCR5380_init() function into two parts, similar
to the scheme used with atari_NCR5380.c. This avoids two problems.

Firstly, NCR5380_init() may perform a bus reset, which would cause the
chip to assert IRQ. The chip is unable to mask its bus reset interrupt.
Drivers can't call request_irq() before calling NCR5380_init(), because
initialization must happen before the interrupt handler executes. If
driver initialization causes an interrupt it may be problematic on some
platforms. To avoid that, first move the bus reset code into
NCR5380_maybe_reset_bus().

Secondly, NCR5380_init() contains some board-specific interrupt setup code
for the NCR53C400 that does not belong in the core driver. In moving this
code, better not re-order interrupt initialization and bus reset. Again,
the solution is to move the bus reset code into NCR5380_maybe_reset_bus().

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Signed-off-by: Martin K. Petersen

ncr5380: Remove NCR5380_local_declare and NCR5380_setup macros

2016-01-07T02:42:52+00:00

The NCR5380_local_declare and NCR5380_setup macros exist to define and
initialize a particular local variable, to provide the address of the
chip registers needed for the driver's implementation of its
NCR5380_read/write register access macros.

In cumana_1 and macscsi, these macros generate pointless code like this,
	struct Scsi_Host *_instance;
	_instance = instance;

In pas16, the use of NCR5380_read/write in pas16_hw_detect() requires that
the io_port local variable has been defined and initialized, but the
NCR5380_local_declare and NCR5380_setup macros can't be used for that
purpose because the Scsi_Host struct has not yet been instantiated.

Moreover, these macros were removed from atari_NCR5380.c long ago and
now they constitute yet another discrepancy between the two core driver
forks.

Remove these "optimizations".

Signed-off-by: Finn Thain 
Reviewed-by: Hannes Reinecke 
Tested-by: Ondrej Zary 
Signed-off-by: Martin K. Petersen

scsi: Do not set cmd_per_lun to 1 in the host template

2015-06-01T01:06:28+00:00

'0' is now used as the default cmd_per_lun value,
so there's no need to explicitly set it to '1' in the
host template.

Signed-off-by: Hannes Reinecke 
Reviewed-by: Christoph Hellwig 
Signed-off-by: James Bottomley