2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* NCR 5380 generic driver routines . These should make it * trivial *
2016-01-03 08:06:10 +03:00
* to implement 5380 SCSI drivers under Linux with a non - trantor
* architecture .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Note that these routines also work with NR53c400 family chips .
2005-04-17 02:20:36 +04:00
*
* Copyright 1993 , Drew Eckhardt
2016-01-03 08:06:10 +03:00
* Visionary Computing
* ( Unix and Linux consulting and custom programming )
* drew @ colorado . edu
* + 1 ( 303 ) 666 - 5836
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* For more information , please consult
2005-04-17 02:20:36 +04:00
*
* NCR 5380 Family
* SCSI Protocol Controller
* Databook
*
* NCR Microelectronics
* 1635 Aeroplaza Drive
* Colorado Springs , CO 80916
* 1 + ( 719 ) 578 - 3400
* 1 + ( 800 ) 334 - 5454
*/
/*
2016-01-03 08:06:08 +03:00
* With contributions from Ray Van Tassle , Ingmar Baumgart ,
* Ronald van Cuijlenborg , Alan Cox and others .
2005-04-17 02:20:36 +04:00
*/
2016-03-23 13:10:20 +03:00
/* Ported to Atari by Roman Hodek and others. */
2016-03-23 13:10:21 +03:00
/* Adapted for the Sun 3 by Sam Creasey. */
2005-04-17 02:20:36 +04:00
/*
* Design
*
2016-01-03 08:06:09 +03:00
* This is a generic 5380 driver . To use it on a different platform ,
2005-04-17 02:20:36 +04:00
* one simply writes appropriate system specific macros ( ie , data
2016-01-03 08:06:09 +03:00
* transfer - some PC ' s will use the I / O bus , 68 K ' s must use
2005-04-17 02:20:36 +04:00
* memory mapped ) and drops this file in their ' C ' wrapper .
*
2016-01-03 08:06:09 +03:00
* As far as command queueing , two queues are maintained for
2005-04-17 02:20:36 +04:00
* each 5380 in the system - commands that haven ' t been issued yet ,
2016-01-03 08:06:09 +03:00
* and commands that are currently executing . This means that an
* unlimited number of commands may be queued , letting
* more commands propagate from the higher driver levels giving higher
* throughput . Note that both I_T_L and I_T_L_Q nexuses are supported ,
* allowing multiple commands to propagate all the way to a SCSI - II device
2005-04-17 02:20:36 +04:00
* while a command is already executing .
*
*
2016-01-03 08:06:09 +03:00
* Issues specific to the NCR5380 :
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* When used in a PIO or pseudo - dma mode , the NCR5380 is a braindead
* piece of hardware that requires you to sit in a loop polling for
* the REQ signal as long as you are connected . Some devices are
* brain dead ( ie , many TEXEL CD ROM drives ) and won ' t disconnect
ncr5380: Rework disconnect versus poll logic
The atari_NCR5380.c and NCR5380.c core drivers differ in their handling of
target disconnection. This is partly because atari_NCR5380.c had all of
the polling and sleeping removed to become entirely interrupt-driven, and
it is partly because of damage done to NCR5380.c after atari_NCR5380.c was
forked. See commit 37cd23b44929 ("Linux 2.1.105") in history/history.git.
The polling changes that were made in v2.1.105 are questionable at best:
if REQ is not already asserted when NCR5380_transfer_pio() is invoked, and
if the expected phase is DATA IN or DATA OUT, the function will schedule
main() to execute after USLEEP_SLEEP jiffies and then return. The problems
here are the expected REQ timing and the sleep interval*. Avoid this issue
by using NCR5380_poll_politely() instead of scheduling main().
The atari_NCR5380.c core driver requires the use of the chip interrupt and
always permits target disconnection. It sets the cmd->device->disconnect
flag when a device disconnects, but never tests this flag.
The NCR5380.c core driver permits disconnection only when
instance->irq != NO_IRQ. It sets the cmd->device->disconnect flag when
a device disconnects and it tests this flag in a couple of places:
1. During NCR5380_information_transfer(), following COMMAND OUT phase,
if !cmd->device->disconnect, the initiator will take a guess as to
whether or not the target will then choose to go to MESSAGE IN phase
and disconnect. If the driver guesses "yes", it will schedule main()
to execute after USLEEP_SLEEP jiffies and then return there.
Unfortunately the driver may guess "yes" even after it has denied
the target the disconnection privilege. When the target does not
disconnect, the sleep can be beneficial, assuming the sleep interval
is appropriate (mostly it is not*).
And even if the driver guesses "yes" correctly, and the target would
then disconnect, the driver still has to go through the MESSAGE IN
phase in order to get to BUS FREE phase. The main loop can do nothing
useful until BUS FREE, and sleeping just delays the phase transition.
2. If !cmd->device->disconnect and REQ is not already asserted when
NCR5380_information_transfer() is invoked, the function polls for REQ
for USLEEP_POLL jiffies. If REQ is not asserted, it then schedules
main() to execute after USLEEP_SLEEP jiffies and returns.
The idea is apparently to yeild the CPU while waiting for REQ.
This is conditional upon !cmd->device->disconnect, but there seems
to be no rhyme or reason for that. For example, the flag may be
unset because disconnection privilege was denied because the driver
has no IRQ. Or the flag may be unset because the device has never
needed to disconnect before. Or if the flag is set, disconnection
may have no relevance to the present bus phase.
Another deficiency of the existing algorithm is as follows. When the
driver has no IRQ, it prevents disconnection, and generally polls and
sleeps more than it would normally. Now, if the driver is going to poll
anyway, why not allow the target to disconnect? That way the driver can do
something useful with the bus instead of polling unproductively!
Avoid this pointless latency, complexity and guesswork by using
NCR5380_poll_politely() instead of scheduling main().
* For g_NCR5380, the time intervals for USLEEP_SLEEP and USLEEP_POLL are
200 ms and 10 ms, respectively. They are 20 ms and 200 ms respectively
for the other NCR5380 drivers. There doesn't seem to be any reason for
this discrepancy. The timing seems to have no relation to the type of
adapter. Bizarrely, the timing in g_NCR5380 seems to relate only to one
particular type of target device. This patch attempts to solve the
problem for all NCR5380 drivers and all target devices.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:26 +03:00
* while doing long seek operations . [ . . . ] These
2005-04-17 02:20:36 +04:00
* broken devices are the exception rather than the rule and I ' d rather
* spend my time optimizing for the normal case .
*
* Architecture :
*
* At the heart of the design is a coroutine , NCR5380_main ,
* which is started from a workqueue for each NCR5380 host in the
* system . It attempts to establish I_T_L or I_T_L_Q nexuses by
* removing the commands from the issue queue and calling
2016-01-03 08:06:09 +03:00
* NCR5380_select ( ) if a nexus is not established .
2005-04-17 02:20:36 +04:00
*
* Once a nexus is established , the NCR5380_information_transfer ( )
* phase goes through the various phases as instructed by the target .
* if the target goes into MSG IN and sends a DISCONNECT message ,
* the command structure is placed into the per instance disconnected
2016-01-03 08:06:09 +03:00
* queue , and NCR5380_main tries to find more work . If the target is
2005-04-17 02:20:36 +04:00
* idle for too long , the system will try to sleep .
*
* If a command has disconnected , eventually an interrupt will trigger ,
* calling NCR5380_intr ( ) which will in turn call NCR5380_reselect
* to reestablish a nexus . This will run main if necessary .
*
2016-01-03 08:06:09 +03:00
* On command termination , the done function will be called as
2005-04-17 02:20:36 +04:00
* appropriate .
*
2016-01-03 08:06:09 +03:00
* SCSI pointers are maintained in the SCp field of SCSI command
2005-04-17 02:20:36 +04:00
* structures , being initialized after the command is connected
* in NCR5380_select , and set as appropriate in NCR5380_information_transfer .
* Note that in violation of the standard , an implicit SAVE POINTERS operation
* is done , since some BROKEN disks fail to issue an explicit SAVE POINTERS .
*/
/*
* Using this file :
* This file a skeleton Linux SCSI driver for the NCR 5380 series
2016-01-03 08:06:09 +03:00
* of chips . To use it , you write an architecture specific functions
2005-04-17 02:20:36 +04:00
* and macros and include this file in your driver .
*
2016-01-03 08:06:09 +03:00
* These macros control options :
* AUTOPROBE_IRQ - if defined , the NCR5380_probe_irq ( ) function will be
2016-01-03 08:06:10 +03:00
* defined .
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* AUTOSENSE - if defined , REQUEST SENSE will be performed automatically
2016-01-03 08:06:10 +03:00
* for commands that return with a CHECK CONDITION status .
2005-04-17 02:20:36 +04:00
*
* DIFFERENTIAL - if defined , NCR53c81 chips will use external differential
2016-01-03 08:06:10 +03:00
* transceivers .
2005-04-17 02:20:36 +04:00
*
* PSEUDO_DMA - if defined , PSEUDO DMA is used during the data transfer phases .
*
2016-03-23 13:10:19 +03:00
* REAL_DMA - if defined , REAL DMA is used during the data transfer phases .
*
2005-04-17 02:20:36 +04:00
* These macros MUST be defined :
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* NCR5380_read ( register ) - read from the specified register
*
2016-01-03 08:06:09 +03:00
* NCR5380_write ( register , value ) - write to the specific register
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* NCR5380_implementation_fields - additional fields needed for this
2016-01-03 08:06:10 +03:00
* specific implementation of the NCR5380
2005-04-17 02:20:36 +04:00
*
* Either real DMA * or * pseudo DMA may be implemented
*
* NCR5380_dma_write_setup ( instance , src , count ) - initialize
* NCR5380_dma_read_setup ( instance , dst , count ) - initialize
* NCR5380_dma_residual ( instance ) ; - residual count
*
* The generic driver is initialized by calling NCR5380_init ( instance ) ,
2016-01-03 08:06:09 +03:00
* after setting the appropriate host specific fields and ID . If the
2005-04-17 02:20:36 +04:00
* driver wishes to autoprobe for an IRQ line , the NCR5380_probe_irq ( instance ,
* possible ) function may be used .
*/
2016-03-23 13:10:16 +03:00
# ifndef NCR5380_io_delay
# define NCR5380_io_delay(x)
# endif
2016-03-23 13:10:20 +03:00
# ifndef NCR5380_acquire_dma_irq
# define NCR5380_acquire_dma_irq(x) (1)
# endif
# ifndef NCR5380_release_dma_irq
# define NCR5380_release_dma_irq(x)
# endif
2016-01-03 08:05:06 +03:00
static int do_abort ( struct Scsi_Host * ) ;
static void do_reset ( struct Scsi_Host * ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:08 +03:00
/**
2016-01-03 08:06:11 +03:00
* initialize_SCp - init the scsi pointer field
2016-01-03 08:06:10 +03:00
* @ cmd : command block to set up
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Set up the internal fields in the SCSI command .
2005-04-17 02:20:36 +04:00
*/
2014-11-12 08:12:02 +03:00
static inline void initialize_SCp ( struct scsi_cmnd * cmd )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:06:09 +03:00
/*
* Initialize the Scsi Pointer field so that all of the commands in the
2005-04-17 02:20:36 +04:00
* various queues are valid .
*/
2007-11-05 12:23:35 +03:00
if ( scsi_bufflen ( cmd ) ) {
cmd - > SCp . buffer = scsi_sglist ( cmd ) ;
cmd - > SCp . buffers_residual = scsi_sg_count ( cmd ) - 1 ;
2007-10-22 23:19:53 +04:00
cmd - > SCp . ptr = sg_virt ( cmd - > SCp . buffer ) ;
2005-04-17 02:20:36 +04:00
cmd - > SCp . this_residual = cmd - > SCp . buffer - > length ;
} else {
cmd - > SCp . buffer = NULL ;
cmd - > SCp . buffers_residual = 0 ;
2007-11-05 12:23:35 +03:00
cmd - > SCp . ptr = NULL ;
cmd - > SCp . this_residual = 0 ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:00 +03:00
cmd - > SCp . Status = 0 ;
cmd - > SCp . Message = 0 ;
2005-04-17 02:20:36 +04:00
}
/**
2016-01-03 08:05:41 +03:00
* NCR5380_poll_politely2 - wait for two chip register values
2016-01-03 08:05:22 +03:00
* @ instance : controller to poll
2016-01-03 08:05:41 +03:00
* @ reg1 : 5380 register to poll
* @ bit1 : Bitmask to check
* @ val1 : Expected value
* @ reg2 : Second 5380 register to poll
* @ bit2 : Second bitmask to check
* @ val2 : Second expected value
2016-01-03 08:05:22 +03:00
* @ wait : Time - out in jiffies
*
* Polls the chip in a reasonably efficient manner waiting for an
* event to occur . After a short quick poll we begin to yield the CPU
* ( if possible ) . In irq contexts the time - out is arbitrarily limited .
* Callers may hold locks as long as they are held in irq mode .
*
2016-01-03 08:05:41 +03:00
* Returns 0 if either or both event ( s ) occurred otherwise - ETIMEDOUT .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:05:22 +03:00
2016-01-03 08:05:41 +03:00
static int NCR5380_poll_politely2 ( struct Scsi_Host * instance ,
int reg1 , int bit1 , int val1 ,
int reg2 , int bit2 , int val2 , int wait )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:22 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
unsigned long deadline = jiffies + wait ;
unsigned long n ;
/* Busy-wait for up to 10 ms */
n = min ( 10000U , jiffies_to_usecs ( wait ) ) ;
n * = hostdata - > accesses_per_ms ;
2016-01-03 08:05:41 +03:00
n / = 2000 ;
2016-01-03 08:05:22 +03:00
do {
2016-01-03 08:05:41 +03:00
if ( ( NCR5380_read ( reg1 ) & bit1 ) = = val1 )
return 0 ;
if ( ( NCR5380_read ( reg2 ) & bit2 ) = = val2 )
2005-04-17 02:20:36 +04:00
return 0 ;
cpu_relax ( ) ;
2016-01-03 08:05:22 +03:00
} while ( n - - ) ;
if ( irqs_disabled ( ) | | in_interrupt ( ) )
return - ETIMEDOUT ;
/* Repeatedly sleep for 1 ms until deadline */
while ( time_is_after_jiffies ( deadline ) ) {
schedule_timeout_uninterruptible ( 1 ) ;
2016-01-03 08:05:41 +03:00
if ( ( NCR5380_read ( reg1 ) & bit1 ) = = val1 )
return 0 ;
if ( ( NCR5380_read ( reg2 ) & bit2 ) = = val2 )
2005-04-17 02:20:36 +04:00
return 0 ;
}
2016-01-03 08:05:22 +03:00
2005-04-17 02:20:36 +04:00
return - ETIMEDOUT ;
}
2005-09-08 02:18:24 +04:00
# if NDEBUG
2005-04-17 02:20:36 +04:00
static struct {
unsigned char mask ;
const char * name ;
2016-01-03 08:06:09 +03:00
} signals [ ] = {
{ SR_DBP , " PARITY " } ,
{ SR_RST , " RST " } ,
{ SR_BSY , " BSY " } ,
{ SR_REQ , " REQ " } ,
{ SR_MSG , " MSG " } ,
{ SR_CD , " CD " } ,
{ SR_IO , " IO " } ,
{ SR_SEL , " SEL " } ,
2005-04-17 02:20:36 +04:00
{ 0 , NULL }
2016-01-03 08:06:09 +03:00
} ,
2005-04-17 02:20:36 +04:00
basrs [ ] = {
2016-03-23 13:10:25 +03:00
{ BASR_END_DMA_TRANSFER , " END OF DMA " } ,
{ BASR_DRQ , " DRQ " } ,
{ BASR_PARITY_ERROR , " PARITY ERROR " } ,
{ BASR_IRQ , " IRQ " } ,
{ BASR_PHASE_MATCH , " PHASE MATCH " } ,
{ BASR_BUSY_ERROR , " BUSY ERROR " } ,
2016-01-03 08:06:09 +03:00
{ BASR_ATN , " ATN " } ,
{ BASR_ACK , " ACK " } ,
2005-04-17 02:20:36 +04:00
{ 0 , NULL }
2016-01-03 08:06:09 +03:00
} ,
icrs [ ] = {
{ ICR_ASSERT_RST , " ASSERT RST " } ,
2016-03-23 13:10:25 +03:00
{ ICR_ARBITRATION_PROGRESS , " ARB. IN PROGRESS " } ,
{ ICR_ARBITRATION_LOST , " LOST ARB. " } ,
2016-01-03 08:06:09 +03:00
{ ICR_ASSERT_ACK , " ASSERT ACK " } ,
{ ICR_ASSERT_BSY , " ASSERT BSY " } ,
{ ICR_ASSERT_SEL , " ASSERT SEL " } ,
{ ICR_ASSERT_ATN , " ASSERT ATN " } ,
{ ICR_ASSERT_DATA , " ASSERT DATA " } ,
2005-04-17 02:20:36 +04:00
{ 0 , NULL }
2016-01-03 08:06:09 +03:00
} ,
mrs [ ] = {
2016-03-23 13:10:25 +03:00
{ MR_BLOCK_DMA_MODE , " BLOCK DMA MODE " } ,
{ MR_TARGET , " TARGET " } ,
{ MR_ENABLE_PAR_CHECK , " PARITY CHECK " } ,
{ MR_ENABLE_PAR_INTR , " PARITY INTR " } ,
{ MR_ENABLE_EOP_INTR , " EOP INTR " } ,
{ MR_MONITOR_BSY , " MONITOR BSY " } ,
{ MR_DMA_MODE , " DMA MODE " } ,
{ MR_ARBITRATE , " ARBITRATE " } ,
2005-04-17 02:20:36 +04:00
{ 0 , NULL }
} ;
/**
2016-01-03 08:06:11 +03:00
* NCR5380_print - print scsi bus signals
* @ instance : adapter state to dump
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Print the SCSI bus signals for debugging purposes
2005-04-17 02:20:36 +04:00
*/
static void NCR5380_print ( struct Scsi_Host * instance )
{
unsigned char status , data , basr , mr , icr , i ;
data = NCR5380_read ( CURRENT_SCSI_DATA_REG ) ;
status = NCR5380_read ( STATUS_REG ) ;
mr = NCR5380_read ( MODE_REG ) ;
icr = NCR5380_read ( INITIATOR_COMMAND_REG ) ;
basr = NCR5380_read ( BUS_AND_STATUS_REG ) ;
2016-03-23 13:10:25 +03:00
printk ( KERN_DEBUG " SR = 0x%02x : " , status ) ;
2005-04-17 02:20:36 +04:00
for ( i = 0 ; signals [ i ] . mask ; + + i )
if ( status & signals [ i ] . mask )
2016-03-23 13:10:25 +03:00
printk ( KERN_CONT " %s, " , signals [ i ] . name ) ;
printk ( KERN_CONT " \n BASR = 0x%02x : " , basr ) ;
2005-04-17 02:20:36 +04:00
for ( i = 0 ; basrs [ i ] . mask ; + + i )
if ( basr & basrs [ i ] . mask )
2016-03-23 13:10:25 +03:00
printk ( KERN_CONT " %s, " , basrs [ i ] . name ) ;
printk ( KERN_CONT " \n ICR = 0x%02x : " , icr ) ;
2005-04-17 02:20:36 +04:00
for ( i = 0 ; icrs [ i ] . mask ; + + i )
if ( icr & icrs [ i ] . mask )
2016-03-23 13:10:25 +03:00
printk ( KERN_CONT " %s, " , icrs [ i ] . name ) ;
printk ( KERN_CONT " \n MR = 0x%02x : " , mr ) ;
2005-04-17 02:20:36 +04:00
for ( i = 0 ; mrs [ i ] . mask ; + + i )
if ( mr & mrs [ i ] . mask )
2016-03-23 13:10:25 +03:00
printk ( KERN_CONT " %s, " , mrs [ i ] . name ) ;
printk ( KERN_CONT " \n " ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:11 +03:00
static struct {
unsigned char value ;
const char * name ;
} phases [ ] = {
{ PHASE_DATAOUT , " DATAOUT " } ,
{ PHASE_DATAIN , " DATAIN " } ,
{ PHASE_CMDOUT , " CMDOUT " } ,
{ PHASE_STATIN , " STATIN " } ,
{ PHASE_MSGOUT , " MSGOUT " } ,
{ PHASE_MSGIN , " MSGIN " } ,
{ PHASE_UNKNOWN , " UNKNOWN " }
} ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:08 +03:00
/**
2016-01-03 08:06:11 +03:00
* NCR5380_print_phase - show SCSI phase
2016-01-03 08:06:10 +03:00
* @ instance : adapter to dump
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Print the current SCSI phase for debugging purposes
2005-04-17 02:20:36 +04:00
*/
static void NCR5380_print_phase ( struct Scsi_Host * instance )
{
unsigned char status ;
int i ;
status = NCR5380_read ( STATUS_REG ) ;
if ( ! ( status & SR_REQ ) )
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_DEBUG , instance , " REQ not asserted, phase unknown. \n " ) ;
2005-04-17 02:20:36 +04:00
else {
2016-01-03 08:06:11 +03:00
for ( i = 0 ; ( phases [ i ] . value ! = PHASE_UNKNOWN ) & &
( phases [ i ] . value ! = ( status & PHASE_MASK ) ) ; + + i )
;
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_DEBUG , instance , " phase %s \n " , phases [ i ] . name ) ;
2005-04-17 02:20:36 +04:00
}
}
# endif
2016-10-11 12:23:23 +03:00
static int probe_irq ;
2005-04-17 02:20:36 +04:00
/**
2016-01-03 08:06:10 +03:00
* probe_intr - helper for IRQ autoprobe
* @ irq : interrupt number
* @ dev_id : unused
* @ regs : unused
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Set a flag to indicate the IRQ in question was received . This is
* used by the IRQ probe code .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:06:09 +03:00
2016-10-11 12:23:23 +03:00
static irqreturn_t probe_intr ( int irq , void * dev_id )
2005-04-17 02:20:36 +04:00
{
probe_irq = irq ;
return IRQ_HANDLED ;
}
/**
2016-01-03 08:06:10 +03:00
* NCR5380_probe_irq - find the IRQ of an NCR5380
* @ instance : NCR5380 controller
* @ possible : bitmask of ISA IRQ lines
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Autoprobe for the IRQ line used by the NCR5380 by triggering an IRQ
* and then looking to see what interrupt actually turned up .
2005-04-17 02:20:36 +04:00
*/
2016-10-11 12:23:23 +03:00
static int __maybe_unused NCR5380_probe_irq ( struct Scsi_Host * instance ,
2007-05-24 01:41:56 +04:00
int possible )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2005-04-17 02:20:36 +04:00
unsigned long timeout ;
int trying_irqs , i , mask ;
2014-11-12 08:11:56 +03:00
for ( trying_irqs = 0 , i = 1 , mask = 2 ; i < 16 ; + + i , mask < < = 1 )
2014-03-05 09:09:41 +04:00
if ( ( mask & possible ) & & ( request_irq ( i , & probe_intr , 0 , " NCR-probe " , NULL ) = = 0 ) )
2005-04-17 02:20:36 +04:00
trying_irqs | = mask ;
2015-02-04 21:30:20 +03:00
timeout = jiffies + msecs_to_jiffies ( 250 ) ;
2014-11-12 08:11:56 +03:00
probe_irq = NO_IRQ ;
2005-04-17 02:20:36 +04:00
/*
* A interrupt is triggered whenever BSY = false , SEL = true
2016-01-03 08:06:09 +03:00
* and a bit set in the SELECT_ENABLE_REG is asserted on the
2005-04-17 02:20:36 +04:00
* SCSI bus .
*
* Note that the bus is only driven when the phase control signals
* ( I / O , C / D , and MSG ) match those in the TCR , so we must reset that
* to zero .
*/
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
NCR5380_write ( OUTPUT_DATA_REG , hostdata - > id_mask ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_DATA | ICR_ASSERT_SEL ) ;
2014-11-12 08:11:56 +03:00
while ( probe_irq = = NO_IRQ & & time_before ( jiffies , timeout ) )
2005-11-07 12:01:20 +03:00
schedule_timeout_uninterruptible ( 1 ) ;
2016-01-03 08:06:09 +03:00
2005-04-17 02:20:36 +04:00
NCR5380_write ( SELECT_ENABLE_REG , 0 ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2014-11-12 08:11:56 +03:00
for ( i = 1 , mask = 2 ; i < 16 ; + + i , mask < < = 1 )
2005-04-17 02:20:36 +04:00
if ( trying_irqs & mask )
free_irq ( i , NULL ) ;
return probe_irq ;
}
/**
2016-01-03 08:06:10 +03:00
* NCR58380_info - report driver and host information
* @ instance : relevant scsi host instance
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* For use as the host template info ( ) handler .
2005-04-17 02:20:36 +04:00
*/
2014-11-12 08:11:58 +03:00
static const char * NCR5380_info ( struct Scsi_Host * instance )
2005-04-17 02:20:36 +04:00
{
2014-11-12 08:11:58 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
return hostdata - > info ;
}
static void prepare_info ( struct Scsi_Host * instance )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
snprintf ( hostdata - > info , sizeof ( hostdata - > info ) ,
" %s, io_port 0x%lx, n_io_port %d, "
" base 0x%lx, irq %d, "
" can_queue %d, cmd_per_lun %d, "
" sg_tablesize %d, this_id %d, "
2016-01-03 08:05:50 +03:00
" flags { %s%s%s}, "
2014-11-12 08:11:58 +03:00
" options { %s} " ,
instance - > hostt - > name , instance - > io_port , instance - > n_io_port ,
instance - > base , instance - > irq ,
instance - > can_queue , instance - > cmd_per_lun ,
instance - > sg_tablesize , instance - > this_id ,
2016-03-23 13:10:14 +03:00
hostdata - > flags & FLAG_DMA_FIXUP ? " DMA_FIXUP " : " " ,
2014-11-12 08:11:58 +03:00
hostdata - > flags & FLAG_NO_PSEUDO_DMA ? " NO_PSEUDO_DMA " : " " ,
2016-01-03 08:05:11 +03:00
hostdata - > flags & FLAG_TOSHIBA_DELAY ? " TOSHIBA_DELAY " : " " ,
2005-04-17 02:20:36 +04:00
# ifdef DIFFERENTIAL
2014-11-12 08:11:58 +03:00
" DIFFERENTIAL "
2005-04-17 02:20:36 +04:00
# endif
# ifdef PARITY
2014-11-12 08:11:58 +03:00
" PARITY "
# endif
" " ) ;
2005-04-17 02:20:36 +04:00
}
/**
2016-01-03 08:06:11 +03:00
* NCR5380_init - initialise an NCR5380
2016-01-03 08:06:10 +03:00
* @ instance : adapter to configure
* @ flags : control flags
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Initializes * instance and corresponding 5380 chip ,
* with flags OR ' d into the initial flags value .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Notes : I assume that the host , hostno , and id bits have been
2016-01-03 08:06:11 +03:00
* set correctly . I don ' t care about the irq and other fields .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Returns 0 for success
2005-04-17 02:20:36 +04:00
*/
2012-12-22 01:08:55 +04:00
static int NCR5380_init ( struct Scsi_Host * instance , int flags )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-01-03 08:05:08 +03:00
int i ;
2016-01-03 08:05:22 +03:00
unsigned long deadline ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:23 +03:00
instance - > max_lun = 7 ;
2016-01-03 08:06:11 +03:00
hostdata - > host = instance ;
2005-04-17 02:20:36 +04:00
hostdata - > id_mask = 1 < < instance - > this_id ;
2016-01-03 08:06:11 +03:00
hostdata - > id_higher_mask = 0 ;
2005-04-17 02:20:36 +04:00
for ( i = hostdata - > id_mask ; i < = 0x80 ; i < < = 1 )
if ( i > hostdata - > id_mask )
hostdata - > id_higher_mask | = i ;
for ( i = 0 ; i < 8 ; + + i )
hostdata - > busy [ i ] = 0 ;
2016-03-23 13:10:12 +03:00
hostdata - > dma_len = 0 ;
2016-01-03 08:05:51 +03:00
spin_lock_init ( & hostdata - > lock ) ;
2005-04-17 02:20:36 +04:00
hostdata - > connected = NULL ;
2016-01-03 08:06:00 +03:00
hostdata - > sensing = NULL ;
INIT_LIST_HEAD ( & hostdata - > autosense ) ;
2016-01-03 08:05:58 +03:00
INIT_LIST_HEAD ( & hostdata - > unissued ) ;
INIT_LIST_HEAD ( & hostdata - > disconnected ) ;
2016-01-03 08:05:42 +03:00
hostdata - > flags = flags ;
2016-01-03 08:06:09 +03:00
2016-01-03 08:05:37 +03:00
INIT_WORK ( & hostdata - > main_task , NCR5380_main ) ;
2016-01-03 08:05:21 +03:00
hostdata - > work_q = alloc_workqueue ( " ncr5380_%d " ,
WQ_UNBOUND | WQ_MEM_RECLAIM ,
1 , instance - > host_no ) ;
if ( ! hostdata - > work_q )
return - ENOMEM ;
2014-11-12 08:11:58 +03:00
prepare_info ( instance ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
NCR5380_write ( MODE_REG , MR_BASE ) ;
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
NCR5380_write ( SELECT_ENABLE_REG , 0 ) ;
2016-01-03 08:05:22 +03:00
/* Calibrate register polling loop */
i = 0 ;
deadline = jiffies + 1 ;
do {
cpu_relax ( ) ;
} while ( time_is_after_jiffies ( deadline ) ) ;
deadline + = msecs_to_jiffies ( 256 ) ;
do {
NCR5380_read ( STATUS_REG ) ;
+ + i ;
cpu_relax ( ) ;
} while ( time_is_after_jiffies ( deadline ) ) ;
hostdata - > accesses_per_ms = i / 256 ;
2016-01-03 08:05:08 +03:00
return 0 ;
}
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:08 +03:00
/**
* NCR5380_maybe_reset_bus - Detect and correct bus wedge problems .
* @ instance : adapter to check
*
* If the system crashed , it may have crashed with a connected target and
* the SCSI bus busy . Check for BUS FREE phase . If not , try to abort the
* currently established nexus , which we know nothing about . Failing that
* do a bus reset .
*
* Note that a bus reset will cause the chip to assert IRQ .
*
* Returns 0 if successful , otherwise - ENXIO .
*/
static int NCR5380_maybe_reset_bus ( struct Scsi_Host * instance )
{
2016-01-03 08:05:11 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-01-03 08:05:08 +03:00
int pass ;
2005-04-17 02:20:36 +04:00
for ( pass = 1 ; ( NCR5380_read ( STATUS_REG ) & SR_BSY ) & & pass < = 6 ; + + pass ) {
switch ( pass ) {
case 1 :
case 3 :
case 5 :
2016-01-03 08:05:10 +03:00
shost_printk ( KERN_ERR , instance , " SCSI bus busy, waiting up to five seconds \n " ) ;
NCR5380_poll_politely ( instance ,
STATUS_REG , SR_BSY , 0 , 5 * HZ ) ;
2005-04-17 02:20:36 +04:00
break ;
case 2 :
2016-01-03 08:05:10 +03:00
shost_printk ( KERN_ERR , instance , " bus busy, attempting abort \n " ) ;
2005-04-17 02:20:36 +04:00
do_abort ( instance ) ;
break ;
case 4 :
2016-01-03 08:05:10 +03:00
shost_printk ( KERN_ERR , instance , " bus busy, attempting reset \n " ) ;
2005-04-17 02:20:36 +04:00
do_reset ( instance ) ;
2016-01-03 08:05:11 +03:00
/* Wait after a reset; the SCSI standard calls for
* 250 ms , we wait 500 ms to be on the safe side .
* But some Toshiba CD - ROMs need ten times that .
*/
if ( hostdata - > flags & FLAG_TOSHIBA_DELAY )
msleep ( 2500 ) ;
else
msleep ( 500 ) ;
2005-04-17 02:20:36 +04:00
break ;
case 6 :
2016-01-03 08:05:10 +03:00
shost_printk ( KERN_ERR , instance , " bus locked solid \n " ) ;
2005-04-17 02:20:36 +04:00
return - ENXIO ;
}
}
return 0 ;
}
/**
2016-01-03 08:06:11 +03:00
* NCR5380_exit - remove an NCR5380
2016-01-03 08:06:10 +03:00
* @ instance : adapter to remove
2016-01-03 08:06:11 +03:00
*
* Assumes that no more work can be queued ( e . g . by NCR5380_intr ) .
2005-04-17 02:20:36 +04:00
*/
2008-01-23 08:39:33 +03:00
static void NCR5380_exit ( struct Scsi_Host * instance )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:37 +03:00
cancel_work_sync ( & hostdata - > main_task ) ;
2016-01-03 08:05:21 +03:00
destroy_workqueue ( hostdata - > work_q ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:59 +03:00
/**
* complete_cmd - finish processing a command and return it to the SCSI ML
* @ instance : the host instance
* @ cmd : command to complete
*/
static void complete_cmd ( struct Scsi_Host * instance ,
struct scsi_cmnd * cmd )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
dsprintk ( NDEBUG_QUEUES , instance , " complete_cmd: cmd %p \n " , cmd ) ;
2016-01-03 08:06:00 +03:00
if ( hostdata - > sensing = = cmd ) {
/* Autosense processing ends here */
if ( ( cmd - > result & 0xff ) ! = SAM_STAT_GOOD ) {
scsi_eh_restore_cmnd ( cmd , & hostdata - > ses ) ;
set_host_byte ( cmd , DID_ERROR ) ;
} else
scsi_eh_restore_cmnd ( cmd , & hostdata - > ses ) ;
hostdata - > sensing = NULL ;
}
2016-01-03 08:05:59 +03:00
hostdata - > busy [ scmd_id ( cmd ) ] & = ~ ( 1 < < cmd - > device - > lun ) ;
cmd - > scsi_done ( cmd ) ;
}
2005-04-17 02:20:36 +04:00
/**
2016-01-03 08:05:29 +03:00
* NCR5380_queue_command - queue a command
* @ instance : the relevant SCSI adapter
* @ cmd : SCSI command
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:05:29 +03:00
* cmd is added to the per - instance issue queue , with minor
* twiddling done to the host specific fields of cmd . If the
* main coroutine is not running , it is restarted .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:05:29 +03:00
static int NCR5380_queue_command ( struct Scsi_Host * instance ,
struct scsi_cmnd * cmd )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:29 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-01-03 08:05:58 +03:00
struct NCR5380_cmd * ncmd = scsi_cmd_priv ( cmd ) ;
2016-01-03 08:05:29 +03:00
unsigned long flags ;
2005-04-17 02:20:36 +04:00
# if (NDEBUG & NDEBUG_NO_WRITE)
switch ( cmd - > cmnd [ 0 ] ) {
case WRITE_6 :
case WRITE_10 :
2016-01-03 08:05:53 +03:00
shost_printk ( KERN_DEBUG , instance , " WRITE attempted with NDEBUG_NO_WRITE set \n " ) ;
2005-04-17 02:20:36 +04:00
cmd - > result = ( DID_ERROR < < 16 ) ;
2016-01-03 08:05:29 +03:00
cmd - > scsi_done ( cmd ) ;
2005-04-17 02:20:36 +04:00
return 0 ;
}
2016-01-03 08:06:11 +03:00
# endif /* (NDEBUG & NDEBUG_NO_WRITE) */
2005-04-17 02:20:36 +04:00
cmd - > result = 0 ;
2016-03-23 13:10:20 +03:00
if ( ! NCR5380_acquire_dma_irq ( instance ) )
return SCSI_MLQUEUE_HOST_BUSY ;
2016-01-03 08:05:51 +03:00
spin_lock_irqsave ( & hostdata - > lock , flags ) ;
2016-01-03 08:05:29 +03:00
2016-01-03 08:06:09 +03:00
/*
* Insert the cmd into the issue queue . Note that REQUEST SENSE
2005-04-17 02:20:36 +04:00
* commands are added to the head of the queue since any command will
2016-01-03 08:06:09 +03:00
* clear the contingent allegiance condition that exists and the
2005-04-17 02:20:36 +04:00
* sense data is only guaranteed to be valid while the condition exists .
*/
2016-01-03 08:05:58 +03:00
if ( cmd - > cmnd [ 0 ] = = REQUEST_SENSE )
list_add ( & ncmd - > list , & hostdata - > unissued ) ;
else
list_add_tail ( & ncmd - > list , & hostdata - > unissued ) ;
2016-01-03 08:05:51 +03:00
spin_unlock_irqrestore ( & hostdata - > lock , flags ) ;
2016-01-03 08:05:29 +03:00
2016-01-03 08:05:53 +03:00
dsprintk ( NDEBUG_QUEUES , instance , " command %p added to %s of queue \n " ,
cmd , ( cmd - > cmnd [ 0 ] = = REQUEST_SENSE ) ? " head " : " tail " ) ;
2005-04-17 02:20:36 +04:00
/* Kick off command processing */
2016-01-03 08:05:37 +03:00
queue_work ( hostdata - > work_q , & hostdata - > main_task ) ;
2005-04-17 02:20:36 +04:00
return 0 ;
}
2016-03-23 13:10:20 +03:00
static inline void maybe_release_dma_irq ( struct Scsi_Host * instance )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
/* Caller does the locking needed to set & test these data atomically */
if ( list_empty ( & hostdata - > disconnected ) & &
list_empty ( & hostdata - > unissued ) & &
list_empty ( & hostdata - > autosense ) & &
! hostdata - > connected & &
! hostdata - > selecting )
NCR5380_release_dma_irq ( instance ) ;
}
2016-01-03 08:06:00 +03:00
/**
* dequeue_next_cmd - dequeue a command for processing
* @ instance : the scsi host instance
*
* Priority is given to commands on the autosense queue . These commands
* need autosense because of a CHECK CONDITION result .
*
* Returns a command pointer if a command is found for a target that is
* not already busy . Otherwise returns NULL .
*/
static struct scsi_cmnd * dequeue_next_cmd ( struct Scsi_Host * instance )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
struct NCR5380_cmd * ncmd ;
struct scsi_cmnd * cmd ;
2016-02-23 02:07:09 +03:00
if ( hostdata - > sensing | | list_empty ( & hostdata - > autosense ) ) {
2016-01-03 08:06:00 +03:00
list_for_each_entry ( ncmd , & hostdata - > unissued , list ) {
cmd = NCR5380_to_scmd ( ncmd ) ;
dsprintk ( NDEBUG_QUEUES , instance , " dequeue: cmd=%p target=%d busy=0x%02x lun=%llu \n " ,
cmd , scmd_id ( cmd ) , hostdata - > busy [ scmd_id ( cmd ) ] , cmd - > device - > lun ) ;
if ( ! ( hostdata - > busy [ scmd_id ( cmd ) ] & ( 1 < < cmd - > device - > lun ) ) ) {
list_del ( & ncmd - > list ) ;
dsprintk ( NDEBUG_QUEUES , instance ,
" dequeue: removed %p from issue queue \n " , cmd ) ;
return cmd ;
}
}
} else {
/* Autosense processing begins here */
ncmd = list_first_entry ( & hostdata - > autosense ,
struct NCR5380_cmd , list ) ;
list_del ( & ncmd - > list ) ;
cmd = NCR5380_to_scmd ( ncmd ) ;
dsprintk ( NDEBUG_QUEUES , instance ,
" dequeue: removed %p from autosense queue \n " , cmd ) ;
scsi_eh_prep_cmnd ( cmd , & hostdata - > ses , NULL , 0 , ~ 0 ) ;
hostdata - > sensing = cmd ;
return cmd ;
}
return NULL ;
}
static void requeue_cmd ( struct Scsi_Host * instance , struct scsi_cmnd * cmd )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
struct NCR5380_cmd * ncmd = scsi_cmd_priv ( cmd ) ;
2016-02-23 02:07:09 +03:00
if ( hostdata - > sensing = = cmd ) {
2016-01-03 08:06:00 +03:00
scsi_eh_restore_cmnd ( cmd , & hostdata - > ses ) ;
list_add ( & ncmd - > list , & hostdata - > autosense ) ;
hostdata - > sensing = NULL ;
} else
list_add ( & ncmd - > list , & hostdata - > unissued ) ;
}
2005-04-17 02:20:36 +04:00
/**
2016-01-03 08:06:11 +03:00
* NCR5380_main - NCR state machines
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* NCR5380_main is a coroutine that runs as long as more work can
* be done on the NCR5380 host adapters in a system . Both
* NCR5380_queue_command ( ) and NCR5380_intr ( ) will try to start it
* in case it is not running .
2005-04-17 02:20:36 +04:00
*/
2006-11-22 17:57:56 +03:00
static void NCR5380_main ( struct work_struct * work )
2005-04-17 02:20:36 +04:00
{
2006-11-22 17:57:56 +03:00
struct NCR5380_hostdata * hostdata =
2016-01-03 08:05:37 +03:00
container_of ( work , struct NCR5380_hostdata , main_task ) ;
2005-04-17 02:20:36 +04:00
struct Scsi_Host * instance = hostdata - > host ;
int done ;
2016-01-03 08:06:09 +03:00
2005-04-17 02:20:36 +04:00
do {
done = 1 ;
2016-01-03 08:05:51 +03:00
2016-01-03 08:06:07 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-02-23 02:07:08 +03:00
while ( ! hostdata - > connected & & ! hostdata - > selecting ) {
struct scsi_cmnd * cmd = dequeue_next_cmd ( instance ) ;
if ( ! cmd )
break ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:00 +03:00
dsprintk ( NDEBUG_MAIN , instance , " main: dequeued %p \n " , cmd ) ;
2014-11-12 08:11:53 +03:00
2016-01-03 08:06:00 +03:00
/*
* Attempt to establish an I_T_L nexus here .
* On success , instance - > hostdata - > connected is set .
* On failure , we must add the command back to the
* issue queue so we can keep trying .
*/
/*
* REQUEST SENSE commands are issued without tagged
* queueing , even on SCSI - II devices because the
* contingent allegiance condition exists for the
* entire unit .
*/
2016-01-03 08:05:51 +03:00
2016-02-23 02:07:08 +03:00
if ( ! NCR5380_select ( instance , cmd ) ) {
2016-01-03 08:06:02 +03:00
dsprintk ( NDEBUG_MAIN , instance , " main: select complete \n " ) ;
2016-03-23 13:10:20 +03:00
maybe_release_dma_irq ( instance ) ;
2016-01-03 08:06:00 +03:00
} else {
dsprintk ( NDEBUG_MAIN | NDEBUG_QUEUES , instance ,
" main: select failed, returning %p to queue \n " , cmd ) ;
requeue_cmd ( instance , cmd ) ;
}
}
2016-03-23 13:10:12 +03:00
if ( hostdata - > connected & & ! hostdata - > dma_len ) {
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_MAIN , instance , " main: performing information transfer \n " ) ;
2005-04-17 02:20:36 +04:00
NCR5380_information_transfer ( instance ) ;
done = 0 ;
2016-01-03 08:05:24 +03:00
}
2016-01-03 08:06:07 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
if ( ! done )
cond_resched ( ) ;
2005-04-17 02:20:36 +04:00
} while ( ! done ) ;
}
2016-03-23 13:10:19 +03:00
/*
* NCR5380_dma_complete - finish DMA transfer
* @ instance : the scsi host instance
*
* Called by the interrupt handler when DMA finishes or a phase
* mismatch occurs ( which would end the DMA transfer ) .
*/
static void NCR5380_dma_complete ( struct Scsi_Host * instance )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
int transferred ;
unsigned char * * data ;
int * count ;
int saved_data = 0 , overrun = 0 ;
unsigned char p ;
if ( hostdata - > read_overruns ) {
p = hostdata - > connected - > SCp . phase ;
if ( p & SR_IO ) {
udelay ( 10 ) ;
if ( ( NCR5380_read ( BUS_AND_STATUS_REG ) &
( BASR_PHASE_MATCH | BASR_ACK ) ) = =
( BASR_PHASE_MATCH | BASR_ACK ) ) {
saved_data = NCR5380_read ( INPUT_DATA_REG ) ;
overrun = 1 ;
dsprintk ( NDEBUG_DMA , instance , " read overrun handled \n " ) ;
}
}
}
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
if ( ( sun3scsi_dma_finish ( rq_data_dir ( hostdata - > connected - > request ) ) ) ) {
pr_err ( " scsi%d: overrun in UDC counter -- not prepared to deal with this! \n " ,
instance - > host_no ) ;
BUG ( ) ;
}
if ( ( NCR5380_read ( BUS_AND_STATUS_REG ) & ( BASR_PHASE_MATCH | BASR_ACK ) ) = =
( BASR_PHASE_MATCH | BASR_ACK ) ) {
pr_err ( " scsi%d: BASR %02x \n " , instance - > host_no ,
NCR5380_read ( BUS_AND_STATUS_REG ) ) ;
pr_err ( " scsi%d: bus stuck in data phase -- probably a single byte overrun! \n " ,
instance - > host_no ) ;
BUG ( ) ;
}
# endif
2016-03-23 13:10:19 +03:00
NCR5380_write ( MODE_REG , MR_BASE ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
NCR5380_read ( RESET_PARITY_INTERRUPT_REG ) ;
transferred = hostdata - > dma_len - NCR5380_dma_residual ( instance ) ;
hostdata - > dma_len = 0 ;
data = ( unsigned char * * ) & hostdata - > connected - > SCp . ptr ;
count = & hostdata - > connected - > SCp . this_residual ;
* data + = transferred ;
* count - = transferred ;
if ( hostdata - > read_overruns ) {
int cnt , toPIO ;
if ( ( NCR5380_read ( STATUS_REG ) & PHASE_MASK ) = = p & & ( p & SR_IO ) ) {
cnt = toPIO = hostdata - > read_overruns ;
if ( overrun ) {
dsprintk ( NDEBUG_DMA , instance ,
" Got an input overrun, using saved byte \n " ) ;
* ( * data ) + + = saved_data ;
( * count ) - - ;
cnt - - ;
toPIO - - ;
}
if ( toPIO > 0 ) {
dsprintk ( NDEBUG_DMA , instance ,
" Doing %d byte PIO to 0x%p \n " , cnt , * data ) ;
NCR5380_transfer_pio ( instance , & p , & cnt , data ) ;
* count - = toPIO - cnt ;
}
}
}
}
2005-04-17 02:20:36 +04:00
/**
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
* NCR5380_intr - generic NCR5380 irq handler
* @ irq : interrupt number
* @ dev_id : device info
*
* Handle interrupts , reestablishing I_T_L or I_T_L_Q nexuses
* from the disconnected queue , and restarting NCR5380_main ( )
* as required .
*
* The chip can assert IRQ in any of six different conditions . The IRQ flag
* is then cleared by reading the Reset Parity / Interrupt Register ( RPIR ) .
* Three of these six conditions are latched in the Bus and Status Register :
* - End of DMA ( cleared by ending DMA Mode )
* - Parity error ( cleared by reading RPIR )
* - Loss of BSY ( cleared by reading RPIR )
* Two conditions have flag bits that are not latched :
* - Bus phase mismatch ( non - maskable in DMA Mode , cleared by ending DMA Mode )
* - Bus reset ( non - maskable )
* The remaining condition has no flag bit at all :
* - Selection / reselection
*
* Hence , establishing the cause ( s ) of any interrupt is partly guesswork .
* In " The DP8490 and DP5380 Comparison Guide " , National Semiconductor
* claimed that " the design of the [DP8490] interrupt logic ensures
* interrupts will not be lost ( they can be on the DP5380 ) . "
* The L5380 / 53 C80 datasheet from LOGIC Devices has more details .
*
* Checking for bus reset by reading RST is futile because of interrupt
* latency , but a bus reset will reset chip logic . Checking for parity error
* is unnecessary because that interrupt is never enabled . A Loss of BSY
* condition will clear DMA Mode . We can tell when this occurs because the
* the Busy Monitor interrupt is enabled together with DMA Mode .
2005-04-17 02:20:36 +04:00
*/
2016-03-23 13:10:27 +03:00
static irqreturn_t __maybe_unused NCR5380_intr ( int irq , void * dev_id )
2005-04-17 02:20:36 +04:00
{
2007-12-14 03:14:14 +03:00
struct Scsi_Host * instance = dev_id ;
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
int handled = 0 ;
2005-04-17 02:20:36 +04:00
unsigned char basr ;
unsigned long flags ;
2016-01-03 08:05:51 +03:00
spin_lock_irqsave ( & hostdata - > lock , flags ) ;
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
basr = NCR5380_read ( BUS_AND_STATUS_REG ) ;
if ( basr & BASR_IRQ ) {
unsigned char mr = NCR5380_read ( MODE_REG ) ;
unsigned char sr = NCR5380_read ( STATUS_REG ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_INTR , instance , " IRQ %d, BASR 0x%02x, SR 0x%02x, MR 0x%02x \n " ,
irq , basr , sr , mr ) ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:19 +03:00
if ( ( mr & MR_DMA_MODE ) | | ( mr & MR_MONITOR_BSY ) ) {
/* Probably End of DMA, Phase Mismatch or Loss of BSY.
* We ack IRQ after clearing Mode Register . Workarounds
* for End of DMA errata need to happen in DMA Mode .
*/
dsprintk ( NDEBUG_INTR , instance , " interrupt in DMA mode \n " ) ;
if ( hostdata - > connected ) {
NCR5380_dma_complete ( instance ) ;
queue_work ( hostdata - > work_q , & hostdata - > main_task ) ;
} else {
NCR5380_write ( MODE_REG , MR_BASE ) ;
NCR5380_read ( RESET_PARITY_INTERRUPT_REG ) ;
}
} else if ( ( NCR5380_read ( CURRENT_SCSI_DATA_REG ) & hostdata - > id_mask ) & &
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
( sr & ( SR_SEL | SR_IO | SR_BSY | SR_RST ) ) = = ( SR_SEL | SR_IO ) ) {
/* Probably reselected */
NCR5380_write ( SELECT_ENABLE_REG , 0 ) ;
NCR5380_read ( RESET_PARITY_INTERRUPT_REG ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_INTR , instance , " interrupt with SEL and IO \n " ) ;
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
if ( ! hostdata - > connected ) {
NCR5380_reselect ( instance ) ;
queue_work ( hostdata - > work_q , & hostdata - > main_task ) ;
2005-04-17 02:20:36 +04:00
}
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
if ( ! hostdata - > connected )
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
} else {
/* Probably Bus Reset */
NCR5380_read ( RESET_PARITY_INTERRUPT_REG ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_INTR , instance , " unknown interrupt \n " ) ;
2016-03-23 13:10:21 +03:00
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_DMA_ENABLE ;
# endif
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
}
handled = 1 ;
} else {
shost_printk ( KERN_NOTICE , instance , " interrupt without IRQ bit \n " ) ;
2016-03-23 13:10:21 +03:00
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_DMA_ENABLE ;
# endif
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
}
2016-01-03 08:05:51 +03:00
spin_unlock_irqrestore ( & hostdata - > lock , flags ) ;
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
return IRQ_RETVAL ( handled ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:09 +03:00
/*
2014-11-12 08:12:02 +03:00
* Function : int NCR5380_select ( struct Scsi_Host * instance ,
2016-01-03 08:06:10 +03:00
* struct scsi_cmnd * cmd )
2005-04-17 02:20:36 +04:00
*
* Purpose : establishes I_T_L or I_T_L_Q nexus for new or existing command ,
2016-01-03 08:06:10 +03:00
* including ARBITRATION , SELECTION , and initial message out for
* IDENTIFY and queue messages .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* Inputs : instance - instantiation of the 5380 driver on which this
2016-01-03 08:06:10 +03:00
* target lives , cmd - SCSI command to execute .
2016-01-03 08:06:09 +03:00
*
2016-01-03 08:06:02 +03:00
* Returns cmd if selection failed but should be retried ,
* NULL if selection failed and should not be retried , or
* NULL if selection succeeded ( hostdata - > connected = = cmd ) .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* Side effects :
2016-01-03 08:06:10 +03:00
* If bus busy , arbitration failed , etc , NCR5380_select ( ) will exit
* with registers as they should have been on entry - ie
* SELECT_ENABLE will be set appropriately , the NCR5380
* will cease to drive any SCSI bus signals .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* If successful : I_T_L or I_T_L_Q nexus will be established ,
* instance - > connected will be set to cmd .
* SELECT interrupt will be disabled .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* If failed ( no target ) : cmd - > scsi_done ( ) will be called , and the
* cmd - > result host byte set to DID_BAD_TARGET .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:06:09 +03:00
2016-01-03 08:06:02 +03:00
static struct scsi_cmnd * NCR5380_select ( struct Scsi_Host * instance ,
struct scsi_cmnd * cmd )
2005-04-17 02:20:36 +04:00
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2005-04-17 02:20:36 +04:00
unsigned char tmp [ 3 ] , phase ;
unsigned char * data ;
int len ;
int err ;
NCR5380_dprint ( NDEBUG_ARBITRATION , instance ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_ARBITRATION , instance , " starting arbitration, id = %d \n " ,
instance - > this_id ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:02 +03:00
/*
* Arbitration and selection phases are slow and involve dropping the
* lock , so we have to watch out for EH . An exception handler may
* change ' selecting ' to NULL . This function will then return NULL
* so that the caller will forget about ' cmd ' . ( During information
* transfer phases , EH may change ' connected ' to NULL . )
*/
hostdata - > selecting = cmd ;
2016-01-03 08:06:09 +03:00
/*
* Set the phase bits to 0 , otherwise the NCR5380 won ' t drive the
2005-04-17 02:20:36 +04:00
* data bus during SELECTION .
*/
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* Start arbitration .
*/
NCR5380_write ( OUTPUT_DATA_REG , hostdata - > id_mask ) ;
NCR5380_write ( MODE_REG , MR_ARBITRATE ) ;
2016-01-03 08:05:35 +03:00
/* The chip now waits for BUS FREE phase. Then after the 800 ns
* Bus Free Delay , arbitration will begin .
*/
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:51 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
2016-01-03 08:05:41 +03:00
err = NCR5380_poll_politely2 ( instance , MODE_REG , MR_ARBITRATE , 0 ,
INITIATOR_COMMAND_REG , ICR_ARBITRATION_PROGRESS ,
ICR_ARBITRATION_PROGRESS , HZ ) ;
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-01-03 08:05:41 +03:00
if ( ! ( NCR5380_read ( MODE_REG ) & MR_ARBITRATE ) ) {
/* Reselection interrupt */
2016-01-03 08:06:02 +03:00
goto out ;
2016-01-03 08:05:41 +03:00
}
2016-02-23 02:07:08 +03:00
if ( ! hostdata - > selecting ) {
/* Command was aborted */
NCR5380_write ( MODE_REG , MR_BASE ) ;
goto out ;
}
2016-01-03 08:05:41 +03:00
if ( err < 0 ) {
NCR5380_write ( MODE_REG , MR_BASE ) ;
shost_printk ( KERN_ERR , instance ,
" select: arbitration timeout \n " ) ;
2016-01-03 08:06:02 +03:00
goto out ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:51 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:35 +03:00
/* The SCSI-2 arbitration delay is 2.4 us */
2005-04-17 02:20:36 +04:00
udelay ( 3 ) ;
/* Check for lost arbitration */
2016-01-03 08:06:11 +03:00
if ( ( NCR5380_read ( INITIATOR_COMMAND_REG ) & ICR_ARBITRATION_LOST ) | |
( NCR5380_read ( CURRENT_SCSI_DATA_REG ) & hostdata - > id_higher_mask ) | |
( NCR5380_read ( INITIATOR_COMMAND_REG ) & ICR_ARBITRATION_LOST ) ) {
2005-04-17 02:20:36 +04:00
NCR5380_write ( MODE_REG , MR_BASE ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_ARBITRATION , instance , " lost arbitration, deasserting MR_ARBITRATE \n " ) ;
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-01-03 08:06:02 +03:00
goto out ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:18 +03:00
/* After/during arbitration, BSY should be asserted.
* IBM DPES - 31080 Version S31Q works now
* Tnx to Thomas_Roesch @ m2 . maus . de for finding this ! ( Roman )
*/
NCR5380_write ( INITIATOR_COMMAND_REG ,
ICR_BASE | ICR_ASSERT_SEL | ICR_ASSERT_BSY ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
* Again , bus clear + bus settle time is 1.2 us , however , this is
2005-04-17 02:20:36 +04:00
* a minimum so we ' ll udelay ceil ( 1.2 )
*/
2016-01-03 08:05:11 +03:00
if ( hostdata - > flags & FLAG_TOSHIBA_DELAY )
udelay ( 15 ) ;
else
udelay ( 2 ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-01-03 08:05:44 +03:00
/* NCR5380_reselect() clears MODE_REG after a reselection interrupt */
if ( ! ( NCR5380_read ( MODE_REG ) & MR_ARBITRATE ) )
2016-01-03 08:06:02 +03:00
goto out ;
if ( ! hostdata - > selecting ) {
NCR5380_write ( MODE_REG , MR_BASE ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
goto out ;
}
2016-01-03 08:05:44 +03:00
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_ARBITRATION , instance , " won arbitration \n " ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
* Now that we have won arbitration , start Selection process , asserting
2005-04-17 02:20:36 +04:00
* the host and target ID ' s on the SCSI bus .
*/
2016-01-03 08:06:13 +03:00
NCR5380_write ( OUTPUT_DATA_REG , hostdata - > id_mask | ( 1 < < scmd_id ( cmd ) ) ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* Raise ATN while SEL is true before BSY goes false from arbitration ,
* since this is the only way to guarantee that we ' ll get a MESSAGE OUT
* phase immediately after selection .
*/
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_BSY |
ICR_ASSERT_DATA | ICR_ASSERT_ATN | ICR_ASSERT_SEL ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( MODE_REG , MR_BASE ) ;
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* Reselect interrupts must be turned off prior to the dropping of BSY ,
* otherwise we will trigger an interrupt .
*/
NCR5380_write ( SELECT_ENABLE_REG , 0 ) ;
2016-01-03 08:05:51 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
2005-04-17 02:20:36 +04:00
/*
2016-01-03 08:06:09 +03:00
* The initiator shall then wait at least two deskew delays and release
2005-04-17 02:20:36 +04:00
* the BSY signal .
*/
2016-01-03 08:06:11 +03:00
udelay ( 1 ) ; /* wingel -- wait two bus deskew delay >2*45ns */
2005-04-17 02:20:36 +04:00
/* Reset BSY */
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_DATA |
ICR_ASSERT_ATN | ICR_ASSERT_SEL ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* Something weird happens when we cease to drive BSY - looks
2016-01-03 08:06:09 +03:00
* like the board / chip is letting us do another read before the
2005-04-17 02:20:36 +04:00
* appropriate propagation delay has expired , and we ' re confusing
* a BSY signal from ourselves as the target ' s response to SELECTION .
*
* A small delay ( the ' C + + ' frontend breaks the pipeline with an
* unnecessary jump , making it work on my 386 - 33 / Trantor T128 , the
2016-01-03 08:06:09 +03:00
* tighter ' C ' code breaks and requires this ) solves the problem -
* the 1 us delay is arbitrary , and only used because this delay will
* be the same on other platforms and since it works here , it should
2005-04-17 02:20:36 +04:00
* work there .
*
* wingel suggests that this could be due to failing to wait
* one deskew delay .
*/
udelay ( 1 ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_SELECTION , instance , " selecting target %d \n " , scmd_id ( cmd ) ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
* The SCSI specification calls for a 250 ms timeout for the actual
2005-04-17 02:20:36 +04:00
* selection .
*/
2016-01-03 08:05:23 +03:00
err = NCR5380_poll_politely ( instance , STATUS_REG , SR_BSY , SR_BSY ,
msecs_to_jiffies ( 250 ) ) ;
2005-04-17 02:20:36 +04:00
if ( ( NCR5380_read ( STATUS_REG ) & ( SR_SEL | SR_IO ) ) = = ( SR_SEL | SR_IO ) ) {
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
NCR5380_reselect ( instance ) ;
ncr5380: Standardize interrupt handling
Because interrupt handling is crucial to the core driver(s), all wrapper
drivers need to agree on this code. This patch removes discrepancies.
NCR5380_intr() in NCR5380.c has the following pointless loop that differs
from the code in atari_NCR5380.c.
done = 1;
do {
/* ... */
} while (!done);
The 'done' flag gets cleared when a reconnected command is to be processed
from the work queue. But in NCR5380.c, the flag is also used to cause the
interrupt conditions to be re-examined. Perhaps this was because
NCR5380_reselect() was expected to cause another interrupt, or perhaps
the remaining present interrupt conditions need to be handled after the
NCR5380_reselect() call?
Actually, both possibilities are bogus, as is the loop itself. It seems
have been overlooked in the hit-and-miss removal of scsi host instance
list iteration many years ago; see history/history.git commit 491447e1fcff
("[PATCH] next NCR5380 updates") and commit 69e1a9482e57 ("[PATCH] fix up
NCR5380 private data"). See also my earlier patch, "Always retry
arbitration and selection".
The datasheet says, "IRQ can be reset simply by reading the Reset
Parity/Interrupt Register". So don't treat the chip IRQ like a
level-triggered interrupt. Of the conditions that set the IRQ flag,
some are level-triggered and some are edge-triggered, which means IRQ
itself must be edge-triggered.
Some interrupt conditions are latched and some are not. Before clearing
the chip IRQ flag, clear all state that may cause it to be raised. That
means clearing the DMA Mode and Busy Monitor bits in the Mode Register
and clearing the host ID in the Select Enable register.
Also clean up some printk's and some comments. Keep atari_NCR5380.c and
NCR5380.c in agreement.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:40 +03:00
if ( ! hostdata - > connected )
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_ERR , instance , " reselection after won arbitration? \n " ) ;
2016-01-03 08:06:02 +03:00
goto out ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:23 +03:00
if ( err < 0 ) {
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-01-03 08:05:23 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
2016-01-03 08:06:02 +03:00
/* Can't touch cmd if it has been reclaimed by the scsi ML */
if ( hostdata - > selecting ) {
cmd - > result = DID_BAD_TARGET < < 16 ;
complete_cmd ( instance , cmd ) ;
dsprintk ( NDEBUG_SELECTION , instance , " target did not respond within 250ms \n " ) ;
cmd = NULL ;
}
goto out ;
2016-01-03 08:05:23 +03:00
}
2016-01-03 08:06:09 +03:00
/*
* No less than two deskew delays after the initiator detects the
* BSY signal is true , it shall release the SEL signal and may
2005-04-17 02:20:36 +04:00
* change the DATA BUS . - wingel
*/
udelay ( 1 ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN ) ;
/*
2016-01-03 08:06:09 +03:00
* Since we followed the SCSI spec , and raised ATN while SEL
2005-04-17 02:20:36 +04:00
* was true but before BSY was false during selection , the information
* transfer phase should be a MESSAGE OUT phase so that we can send the
* IDENTIFY message .
*/
/* Wait for start of REQ/ACK handshake */
err = NCR5380_poll_politely ( instance , STATUS_REG , SR_REQ , SR_REQ , HZ ) ;
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2016-01-03 08:05:32 +03:00
if ( err < 0 ) {
2016-01-03 08:05:35 +03:00
shost_printk ( KERN_ERR , instance , " select: REQ timeout \n " ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
2016-01-03 08:06:02 +03:00
goto out ;
}
if ( ! hostdata - > selecting ) {
do_abort ( instance ) ;
goto out ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_SELECTION , instance , " target %d selected, going into MESSAGE OUT phase. \n " ,
scmd_id ( cmd ) ) ;
2014-11-12 08:11:56 +03:00
tmp [ 0 ] = IDENTIFY ( ( ( instance - > irq = = NO_IRQ ) ? 0 : 1 ) , cmd - > device - > lun ) ;
2005-04-17 02:20:36 +04:00
len = 1 ;
data = tmp ;
phase = PHASE_MSGOUT ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_SELECTION , instance , " nexus established. \n " ) ;
2005-04-17 02:20:36 +04:00
/* XXX need to handle errors here */
2016-01-03 08:05:51 +03:00
2005-04-17 02:20:36 +04:00
hostdata - > connected = cmd ;
2016-01-03 08:06:13 +03:00
hostdata - > busy [ cmd - > device - > id ] | = 1 < < cmd - > device - > lun ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:21 +03:00
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_INTR ;
# endif
2007-09-10 23:37:45 +04:00
initialize_SCp ( cmd ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:02 +03:00
cmd = NULL ;
out :
if ( ! hostdata - > selecting )
return NULL ;
hostdata - > selecting = NULL ;
return cmd ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:09 +03:00
/*
* Function : int NCR5380_transfer_pio ( struct Scsi_Host * instance ,
2016-01-03 08:06:10 +03:00
* unsigned char * phase , int * count , unsigned char * * data )
2005-04-17 02:20:36 +04:00
*
* Purpose : transfers data in given phase using polled I / O
*
2016-01-03 08:06:09 +03:00
* Inputs : instance - instance of driver , * phase - pointer to
2016-01-03 08:06:10 +03:00
* what phase is expected , * count - pointer to number of
* bytes to transfer , * * data - pointer to data pointer .
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* Returns : - 1 when different phase is entered without transferring
2016-01-03 08:06:11 +03:00
* maximum number of bytes , 0 if all bytes are transferred or exit
2016-01-03 08:06:10 +03:00
* is in same phase .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Also , * phase , * count , * data are modified in place .
2005-04-17 02:20:36 +04:00
*
* XXX Note : handling for bus free may be useful .
*/
/*
2016-01-03 08:06:09 +03:00
* Note : this code is not as quick as it could be , however it
2005-04-17 02:20:36 +04:00
* IS 100 % reliable , and for the actual data transfer where speed
* counts , we will always do a pseudo DMA or DMA transfer .
*/
2016-01-03 08:06:11 +03:00
static int NCR5380_transfer_pio ( struct Scsi_Host * instance ,
unsigned char * phase , int * count ,
unsigned char * * data )
{
2005-04-17 02:20:36 +04:00
unsigned char p = * phase , tmp ;
int c = * count ;
unsigned char * d = * data ;
2016-01-03 08:06:09 +03:00
/*
* The NCR5380 chip will only drive the SCSI bus when the
2005-04-17 02:20:36 +04:00
* phase specified in the appropriate bits of the TARGET COMMAND
* REGISTER match the STATUS REGISTER
*/
2016-01-03 08:06:11 +03:00
NCR5380_write ( TARGET_COMMAND_REG , PHASE_SR_TO_TCR ( p ) ) ;
2005-04-17 02:20:36 +04:00
do {
2016-01-03 08:06:09 +03:00
/*
* Wait for assertion of REQ , after which the phase bits will be
* valid
2005-04-17 02:20:36 +04:00
*/
ncr5380: Rework disconnect versus poll logic
The atari_NCR5380.c and NCR5380.c core drivers differ in their handling of
target disconnection. This is partly because atari_NCR5380.c had all of
the polling and sleeping removed to become entirely interrupt-driven, and
it is partly because of damage done to NCR5380.c after atari_NCR5380.c was
forked. See commit 37cd23b44929 ("Linux 2.1.105") in history/history.git.
The polling changes that were made in v2.1.105 are questionable at best:
if REQ is not already asserted when NCR5380_transfer_pio() is invoked, and
if the expected phase is DATA IN or DATA OUT, the function will schedule
main() to execute after USLEEP_SLEEP jiffies and then return. The problems
here are the expected REQ timing and the sleep interval*. Avoid this issue
by using NCR5380_poll_politely() instead of scheduling main().
The atari_NCR5380.c core driver requires the use of the chip interrupt and
always permits target disconnection. It sets the cmd->device->disconnect
flag when a device disconnects, but never tests this flag.
The NCR5380.c core driver permits disconnection only when
instance->irq != NO_IRQ. It sets the cmd->device->disconnect flag when
a device disconnects and it tests this flag in a couple of places:
1. During NCR5380_information_transfer(), following COMMAND OUT phase,
if !cmd->device->disconnect, the initiator will take a guess as to
whether or not the target will then choose to go to MESSAGE IN phase
and disconnect. If the driver guesses "yes", it will schedule main()
to execute after USLEEP_SLEEP jiffies and then return there.
Unfortunately the driver may guess "yes" even after it has denied
the target the disconnection privilege. When the target does not
disconnect, the sleep can be beneficial, assuming the sleep interval
is appropriate (mostly it is not*).
And even if the driver guesses "yes" correctly, and the target would
then disconnect, the driver still has to go through the MESSAGE IN
phase in order to get to BUS FREE phase. The main loop can do nothing
useful until BUS FREE, and sleeping just delays the phase transition.
2. If !cmd->device->disconnect and REQ is not already asserted when
NCR5380_information_transfer() is invoked, the function polls for REQ
for USLEEP_POLL jiffies. If REQ is not asserted, it then schedules
main() to execute after USLEEP_SLEEP jiffies and returns.
The idea is apparently to yeild the CPU while waiting for REQ.
This is conditional upon !cmd->device->disconnect, but there seems
to be no rhyme or reason for that. For example, the flag may be
unset because disconnection privilege was denied because the driver
has no IRQ. Or the flag may be unset because the device has never
needed to disconnect before. Or if the flag is set, disconnection
may have no relevance to the present bus phase.
Another deficiency of the existing algorithm is as follows. When the
driver has no IRQ, it prevents disconnection, and generally polls and
sleeps more than it would normally. Now, if the driver is going to poll
anyway, why not allow the target to disconnect? That way the driver can do
something useful with the bus instead of polling unproductively!
Avoid this pointless latency, complexity and guesswork by using
NCR5380_poll_politely() instead of scheduling main().
* For g_NCR5380, the time intervals for USLEEP_SLEEP and USLEEP_POLL are
200 ms and 10 ms, respectively. They are 20 ms and 200 ms respectively
for the other NCR5380 drivers. There doesn't seem to be any reason for
this discrepancy. The timing seems to have no relation to the type of
adapter. Bizarrely, the timing in g_NCR5380 seems to relate only to one
particular type of target device. This patch attempts to solve the
problem for all NCR5380 drivers and all target devices.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:26 +03:00
if ( NCR5380_poll_politely ( instance , STATUS_REG , SR_REQ , SR_REQ , HZ ) < 0 )
2005-04-17 02:20:36 +04:00
break ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_HANDSHAKE , instance , " REQ asserted \n " ) ;
2005-04-17 02:20:36 +04:00
/* Check for phase mismatch */
ncr5380: Rework disconnect versus poll logic
The atari_NCR5380.c and NCR5380.c core drivers differ in their handling of
target disconnection. This is partly because atari_NCR5380.c had all of
the polling and sleeping removed to become entirely interrupt-driven, and
it is partly because of damage done to NCR5380.c after atari_NCR5380.c was
forked. See commit 37cd23b44929 ("Linux 2.1.105") in history/history.git.
The polling changes that were made in v2.1.105 are questionable at best:
if REQ is not already asserted when NCR5380_transfer_pio() is invoked, and
if the expected phase is DATA IN or DATA OUT, the function will schedule
main() to execute after USLEEP_SLEEP jiffies and then return. The problems
here are the expected REQ timing and the sleep interval*. Avoid this issue
by using NCR5380_poll_politely() instead of scheduling main().
The atari_NCR5380.c core driver requires the use of the chip interrupt and
always permits target disconnection. It sets the cmd->device->disconnect
flag when a device disconnects, but never tests this flag.
The NCR5380.c core driver permits disconnection only when
instance->irq != NO_IRQ. It sets the cmd->device->disconnect flag when
a device disconnects and it tests this flag in a couple of places:
1. During NCR5380_information_transfer(), following COMMAND OUT phase,
if !cmd->device->disconnect, the initiator will take a guess as to
whether or not the target will then choose to go to MESSAGE IN phase
and disconnect. If the driver guesses "yes", it will schedule main()
to execute after USLEEP_SLEEP jiffies and then return there.
Unfortunately the driver may guess "yes" even after it has denied
the target the disconnection privilege. When the target does not
disconnect, the sleep can be beneficial, assuming the sleep interval
is appropriate (mostly it is not*).
And even if the driver guesses "yes" correctly, and the target would
then disconnect, the driver still has to go through the MESSAGE IN
phase in order to get to BUS FREE phase. The main loop can do nothing
useful until BUS FREE, and sleeping just delays the phase transition.
2. If !cmd->device->disconnect and REQ is not already asserted when
NCR5380_information_transfer() is invoked, the function polls for REQ
for USLEEP_POLL jiffies. If REQ is not asserted, it then schedules
main() to execute after USLEEP_SLEEP jiffies and returns.
The idea is apparently to yeild the CPU while waiting for REQ.
This is conditional upon !cmd->device->disconnect, but there seems
to be no rhyme or reason for that. For example, the flag may be
unset because disconnection privilege was denied because the driver
has no IRQ. Or the flag may be unset because the device has never
needed to disconnect before. Or if the flag is set, disconnection
may have no relevance to the present bus phase.
Another deficiency of the existing algorithm is as follows. When the
driver has no IRQ, it prevents disconnection, and generally polls and
sleeps more than it would normally. Now, if the driver is going to poll
anyway, why not allow the target to disconnect? That way the driver can do
something useful with the bus instead of polling unproductively!
Avoid this pointless latency, complexity and guesswork by using
NCR5380_poll_politely() instead of scheduling main().
* For g_NCR5380, the time intervals for USLEEP_SLEEP and USLEEP_POLL are
200 ms and 10 ms, respectively. They are 20 ms and 200 ms respectively
for the other NCR5380 drivers. There doesn't seem to be any reason for
this discrepancy. The timing seems to have no relation to the type of
adapter. Bizarrely, the timing in g_NCR5380 seems to relate only to one
particular type of target device. This patch attempts to solve the
problem for all NCR5380 drivers and all target devices.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:26 +03:00
if ( ( NCR5380_read ( STATUS_REG ) & PHASE_MASK ) ! = p ) {
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_PIO , instance , " phase mismatch \n " ) ;
NCR5380_dprint_phase ( NDEBUG_PIO , instance ) ;
2005-04-17 02:20:36 +04:00
break ;
}
2016-01-03 08:06:11 +03:00
2005-04-17 02:20:36 +04:00
/* Do actual transfer from SCSI bus to / from memory */
if ( ! ( p & SR_IO ) )
NCR5380_write ( OUTPUT_DATA_REG , * d ) ;
else
* d = NCR5380_read ( CURRENT_SCSI_DATA_REG ) ;
+ + d ;
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* The SCSI standard suggests that in MSGOUT phase , the initiator
* should drop ATN on the last byte of the message phase
* after REQ has been asserted for the handshake but before
* the initiator raises ACK .
*/
if ( ! ( p & SR_IO ) ) {
if ( ! ( ( p & SR_MSG ) & & c > 1 ) ) {
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_DATA ) ;
NCR5380_dprint ( NDEBUG_PIO , instance ) ;
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE |
ICR_ASSERT_DATA | ICR_ASSERT_ACK ) ;
2005-04-17 02:20:36 +04:00
} else {
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE |
ICR_ASSERT_DATA | ICR_ASSERT_ATN ) ;
2005-04-17 02:20:36 +04:00
NCR5380_dprint ( NDEBUG_PIO , instance ) ;
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE |
ICR_ASSERT_DATA | ICR_ASSERT_ATN | ICR_ASSERT_ACK ) ;
2005-04-17 02:20:36 +04:00
}
} else {
NCR5380_dprint ( NDEBUG_PIO , instance ) ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ACK ) ;
}
2016-01-03 08:05:27 +03:00
if ( NCR5380_poll_politely ( instance ,
STATUS_REG , SR_REQ , 0 , 5 * HZ ) < 0 )
break ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_HANDSHAKE , instance , " REQ negated, handshake complete \n " ) ;
2005-04-17 02:20:36 +04:00
/*
2016-01-03 08:06:09 +03:00
* We have several special cases to consider during REQ / ACK handshaking :
* 1. We were in MSGOUT phase , and we are on the last byte of the
2016-01-03 08:06:10 +03:00
* message . ATN must be dropped as ACK is dropped .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* 2. We are in a MSGIN phase , and we are on the last byte of the
2016-01-03 08:06:10 +03:00
* message . We must exit with ACK asserted , so that the calling
* code may raise ATN before dropping ACK to reject the message .
2005-04-17 02:20:36 +04:00
*
* 3. ACK and ATN are clear and the target may proceed as normal .
*/
if ( ! ( p = = PHASE_MSGIN & & c = = 1 ) ) {
if ( p = = PHASE_MSGOUT & & c > 1 )
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN ) ;
else
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
}
} while ( - - c ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_PIO , instance , " residual %d \n " , c ) ;
2005-04-17 02:20:36 +04:00
* count = c ;
* data = d ;
tmp = NCR5380_read ( STATUS_REG ) ;
2016-01-03 08:05:27 +03:00
/* The phase read from the bus is valid if either REQ is (already)
* asserted or if ACK hasn ' t been released yet . The latter applies if
* we ' re in MSG IN , DATA IN or STATUS and all bytes have been received .
*/
if ( ( tmp & SR_REQ ) | | ( ( tmp & SR_IO ) & & c = = 0 ) )
2005-04-17 02:20:36 +04:00
* phase = tmp & PHASE_MASK ;
else
* phase = PHASE_UNKNOWN ;
if ( ! c | | ( * phase = = p ) )
return 0 ;
else
return - 1 ;
}
/**
2016-01-03 08:05:10 +03:00
* do_reset - issue a reset command
* @ instance : adapter to reset
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:05:10 +03:00
* Issue a reset sequence to the NCR5380 and try and get the bus
* back into sane shape .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:05:10 +03:00
* This clears the reset interrupt flag because there may be no handler for
* it . When the driver is initialized , the NCR5380_intr ( ) handler has not yet
* been installed . And when in EH we may have released the ST DMA interrupt .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:06:09 +03:00
2016-01-03 08:05:06 +03:00
static void do_reset ( struct Scsi_Host * instance )
{
2016-01-03 08:05:10 +03:00
unsigned long flags ;
local_irq_save ( flags ) ;
NCR5380_write ( TARGET_COMMAND_REG ,
PHASE_SR_TO_TCR ( NCR5380_read ( STATUS_REG ) & PHASE_MASK ) ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_RST ) ;
2016-01-03 08:05:10 +03:00
udelay ( 50 ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2016-01-03 08:05:10 +03:00
( void ) NCR5380_read ( RESET_PARITY_INTERRUPT_REG ) ;
local_irq_restore ( flags ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:34 +03:00
/**
* do_abort - abort the currently established nexus by going to
* MESSAGE OUT phase and sending an ABORT message .
* @ instance : relevant scsi host instance
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:05:34 +03:00
* Returns 0 on success , - 1 on failure .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:05:06 +03:00
static int do_abort ( struct Scsi_Host * instance )
{
2005-04-17 02:20:36 +04:00
unsigned char * msgptr , phase , tmp ;
int len ;
int rc ;
/* Request message out phase */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN ) ;
2016-01-03 08:06:09 +03:00
/*
* Wait for the target to indicate a valid phase by asserting
* REQ . Once this happens , we ' ll have either a MSGOUT phase
* and can immediately send the ABORT message , or we ' ll have some
2005-04-17 02:20:36 +04:00
* other phase and will have to source / sink data .
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* We really don ' t care what value was on the bus or what value
* the target sees , so we just handshake .
*/
2016-01-03 08:05:34 +03:00
rc = NCR5380_poll_politely ( instance , STATUS_REG , SR_REQ , SR_REQ , 10 * HZ ) ;
2016-01-03 08:05:32 +03:00
if ( rc < 0 )
2016-01-03 08:05:34 +03:00
goto timeout ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:33 +03:00
tmp = NCR5380_read ( STATUS_REG ) & PHASE_MASK ;
2016-01-03 08:06:09 +03:00
2005-04-17 02:20:36 +04:00
NCR5380_write ( TARGET_COMMAND_REG , PHASE_SR_TO_TCR ( tmp ) ) ;
2016-01-03 08:05:33 +03:00
if ( tmp ! = PHASE_MSGOUT ) {
2016-01-03 08:06:11 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG ,
ICR_BASE | ICR_ASSERT_ATN | ICR_ASSERT_ACK ) ;
2016-01-03 08:05:06 +03:00
rc = NCR5380_poll_politely ( instance , STATUS_REG , SR_REQ , 0 , 3 * HZ ) ;
2016-01-03 08:05:32 +03:00
if ( rc < 0 )
2016-01-03 08:05:34 +03:00
goto timeout ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:11 +03:00
2005-04-17 02:20:36 +04:00
tmp = ABORT ;
msgptr = & tmp ;
len = 1 ;
phase = PHASE_MSGOUT ;
2016-01-03 08:05:06 +03:00
NCR5380_transfer_pio ( instance , & phase , & len , & msgptr ) ;
2005-04-17 02:20:36 +04:00
/*
* If we got here , and the command completed successfully ,
* we ' re about to go into bus free state .
*/
return len ? - 1 : 0 ;
2016-01-03 08:05:34 +03:00
timeout :
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
return - 1 ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:09 +03:00
/*
* Function : int NCR5380_transfer_dma ( struct Scsi_Host * instance ,
2016-01-03 08:06:10 +03:00
* unsigned char * phase , int * count , unsigned char * * data )
2005-04-17 02:20:36 +04:00
*
* Purpose : transfers data in given phase using either real
2016-01-03 08:06:10 +03:00
* or pseudo DMA .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* Inputs : instance - instance of driver , * phase - pointer to
2016-01-03 08:06:10 +03:00
* what phase is expected , * count - pointer to number of
* bytes to transfer , * * data - pointer to data pointer .
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* Returns : - 1 when different phase is entered without transferring
2016-01-03 08:06:10 +03:00
* maximum number of bytes , 0 if all bytes or transferred or exit
* is in same phase .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:10 +03:00
* Also , * phase , * count , * data are modified in place .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:06:11 +03:00
static int NCR5380_transfer_dma ( struct Scsi_Host * instance ,
unsigned char * phase , int * count ,
unsigned char * * data )
{
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-03-23 13:10:26 +03:00
int c = * count ;
unsigned char p = * phase ;
unsigned char * d = * data ;
2005-04-17 02:20:36 +04:00
unsigned char tmp ;
2016-03-23 13:10:19 +03:00
int result = 0 ;
2005-04-17 02:20:36 +04:00
if ( ( tmp = ( NCR5380_read ( STATUS_REG ) & PHASE_MASK ) ) ! = p ) {
* phase = tmp ;
return - 1 ;
}
2016-03-23 13:10:19 +03:00
hostdata - > connected - > SCp . phase = p ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:19 +03:00
if ( p & SR_IO ) {
if ( hostdata - > read_overruns )
c - = hostdata - > read_overruns ;
else if ( hostdata - > flags & FLAG_DMA_FIXUP )
- - c ;
}
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:19 +03:00
dsprintk ( NDEBUG_DMA , instance , " initializing DMA %s: length %d, address %p \n " ,
( p & SR_IO ) ? " receive " : " send " , c , d ) ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
/* send start chain */
sun3scsi_dma_start ( c , * data ) ;
# endif
2016-03-23 13:10:19 +03:00
NCR5380_write ( TARGET_COMMAND_REG , PHASE_SR_TO_TCR ( p ) ) ;
NCR5380_write ( MODE_REG , MR_BASE | MR_DMA_MODE | MR_MONITOR_BSY |
MR_ENABLE_EOP_INTR ) ;
if ( ! ( hostdata - > flags & FLAG_LATE_DMA_SETUP ) ) {
/* On the Medusa, it is a must to initialize the DMA before
* starting the NCR . This is also the cleaner way for the TT .
*/
if ( p & SR_IO )
result = NCR5380_dma_recv_setup ( instance , d , c ) ;
else
result = NCR5380_dma_send_setup ( instance , d , c ) ;
}
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
2016-01-03 08:06:10 +03:00
* On the PAS16 at least I / O recovery delays are not needed here .
* Everyone else seems to want them .
2005-04-17 02:20:36 +04:00
*/
if ( p & SR_IO ) {
2016-03-23 13:10:21 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2016-03-23 13:10:16 +03:00
NCR5380_io_delay ( 1 ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( START_DMA_INITIATOR_RECEIVE_REG , 0 ) ;
} else {
2016-03-23 13:10:16 +03:00
NCR5380_io_delay ( 1 ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_DATA ) ;
2016-03-23 13:10:16 +03:00
NCR5380_io_delay ( 1 ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( START_DMA_SEND_REG , 0 ) ;
2016-03-23 13:10:16 +03:00
NCR5380_io_delay ( 1 ) ;
2005-04-17 02:20:36 +04:00
}
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_DMA_ENABLE ;
# endif
sun3_dma_active = 1 ;
# endif
2016-03-23 13:10:19 +03:00
if ( hostdata - > flags & FLAG_LATE_DMA_SETUP ) {
/* On the Falcon, the DMA setup must be done after the last
* NCR access , else the DMA setup gets trashed !
*/
if ( p & SR_IO )
result = NCR5380_dma_recv_setup ( instance , d , c ) ;
else
result = NCR5380_dma_send_setup ( instance , d , c ) ;
}
/* On failure, NCR5380_dma_xxxx_setup() returns a negative int. */
if ( result < 0 )
return result ;
/* For real DMA, result is the byte count. DMA interrupt is expected. */
if ( result > 0 ) {
hostdata - > dma_len = result ;
return 0 ;
}
/* The result is zero iff pseudo DMA send/receive was completed. */
hostdata - > dma_len = c ;
2005-04-17 02:20:36 +04:00
/*
2016-03-23 13:10:12 +03:00
* A note regarding the DMA errata workarounds for early NMOS silicon .
2016-01-03 08:06:08 +03:00
*
* For DMA sends , we want to wait until the last byte has been
* transferred out over the bus before we turn off DMA mode . Alas , there
* seems to be no terribly good way of doing this on a 5380 under all
* conditions . For non - scatter - gather operations , we can wait until REQ
* and ACK both go false , or until a phase mismatch occurs . Gather - sends
* are nastier , since the device will be expecting more data than we
* are prepared to send it , and REQ will remain asserted . On a 53 C8 [ 01 ] we
* could test Last Byte Sent to assure transfer ( I imagine this is precisely
* why this signal was added to the newer chips ) but on the older 538 [ 01 ]
* this signal does not exist . The workaround for this lack is a watchdog ;
* we bail out of the wait - loop after a modest amount of wait - time if
* the usual exit conditions are not met . Not a terribly clean or
* correct solution : - %
*
* DMA receive is equally tricky due to a nasty characteristic of the NCR5380 .
* If the chip is in DMA receive mode , it will respond to a target ' s
* REQ by latching the SCSI data into the INPUT DATA register and asserting
* ACK , even if it has _already_ been notified by the DMA controller that
* the current DMA transfer has completed ! If the NCR5380 is then taken
* out of DMA mode , this already - acknowledged byte is lost . This is
* not a problem for " one DMA transfer per READ command " , because
* the situation will never arise . . . either all of the data is DMA ' ed
* properly , or the target switches to MESSAGE IN phase to signal a
* disconnection ( either operation bringing the DMA to a clean halt ) .
* However , in order to handle scatter - receive , we must work around the
2016-03-23 13:10:12 +03:00
* problem . The chosen fix is to DMA fewer bytes , then check for the
2016-01-03 08:06:08 +03:00
* condition before taking the NCR5380 out of DMA mode . One or two extra
* bytes are transferred via PIO as necessary to fill out the original
* request .
2005-04-17 02:20:36 +04:00
*/
2016-03-23 13:10:19 +03:00
if ( hostdata - > flags & FLAG_DMA_FIXUP ) {
if ( p & SR_IO ) {
2005-04-17 02:20:36 +04:00
/*
2016-03-23 13:10:12 +03:00
* The workaround was to transfer fewer bytes than we
2016-01-03 08:06:09 +03:00
* intended to with the pseudo - DMA read function , wait for
2005-04-17 02:20:36 +04:00
* the chip to latch the last byte , read it , and then disable
* pseudo - DMA mode .
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* After REQ is asserted , the NCR5380 asserts DRQ and ACK .
* REQ is deasserted when ACK is asserted , and not reasserted
* until ACK goes false . Since the NCR5380 won ' t lower ACK
* until DACK is asserted , which won ' t happen unless we twiddle
2016-01-03 08:06:09 +03:00
* the DMA port or we take the NCR5380 out of DMA mode , we
* can guarantee that we won ' t handshake another extra
2005-04-17 02:20:36 +04:00
* byte .
*/
2016-01-03 08:05:42 +03:00
if ( NCR5380_poll_politely ( instance , BUS_AND_STATUS_REG ,
BASR_DRQ , BASR_DRQ , HZ ) < 0 ) {
2016-03-23 13:10:18 +03:00
result = - 1 ;
2016-01-03 08:05:42 +03:00
shost_printk ( KERN_ERR , instance , " PDMA read: DRQ timeout \n " ) ;
}
if ( NCR5380_poll_politely ( instance , STATUS_REG ,
SR_REQ , 0 , HZ ) < 0 ) {
2016-03-23 13:10:18 +03:00
result = - 1 ;
2016-01-03 08:05:42 +03:00
shost_printk ( KERN_ERR , instance , " PDMA read: !REQ timeout \n " ) ;
2005-04-17 02:20:36 +04:00
}
2016-03-23 13:10:19 +03:00
d [ * count - 1 ] = NCR5380_read ( INPUT_DATA_REG ) ;
} else {
2005-04-17 02:20:36 +04:00
/*
2016-01-03 08:06:09 +03:00
* Wait for the last byte to be sent . If REQ is being asserted for
* the byte we ' re interested , we ' ll ACK it and it will go false .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:05:42 +03:00
if ( NCR5380_poll_politely2 ( instance ,
BUS_AND_STATUS_REG , BASR_DRQ , BASR_DRQ ,
BUS_AND_STATUS_REG , BASR_PHASE_MATCH , 0 , HZ ) < 0 ) {
2016-03-23 13:10:18 +03:00
result = - 1 ;
2016-01-03 08:05:42 +03:00
shost_printk ( KERN_ERR , instance , " PDMA write: DRQ and phase timeout \n " ) ;
2005-04-17 02:20:36 +04:00
}
}
}
2016-03-23 13:10:19 +03:00
NCR5380_dma_complete ( instance ) ;
2016-03-23 13:10:18 +03:00
return result ;
2005-04-17 02:20:36 +04:00
}
/*
* Function : NCR5380_information_transfer ( struct Scsi_Host * instance )
*
2016-01-03 08:06:09 +03:00
* Purpose : run through the various SCSI phases and do as the target
2016-01-03 08:06:10 +03:00
* directs us to . Operates on the currently connected command ,
* instance - > connected .
2005-04-17 02:20:36 +04:00
*
* Inputs : instance , instance for which we are doing commands
*
2016-01-03 08:06:09 +03:00
* Side effects : SCSI things happen , the disconnected queue will be
2016-01-03 08:06:10 +03:00
* modified if a command disconnects , * instance - > connected will
* change .
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:06:09 +03:00
* XXX Note : we need to watch for bus free or a reset condition here
2016-01-03 08:06:10 +03:00
* to recover from an unexpected bus free condition .
2005-04-17 02:20:36 +04:00
*/
2016-01-03 08:06:11 +03:00
static void NCR5380_information_transfer ( struct Scsi_Host * instance )
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2005-04-17 02:20:36 +04:00
unsigned char msgout = NOP ;
int sink = 0 ;
int len ;
int transfersize ;
unsigned char * data ;
unsigned char phase , tmp , extended_msg [ 10 ] , old_phase = 0xff ;
2016-01-03 08:05:51 +03:00
struct scsi_cmnd * cmd ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:21 +03:00
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_INTR ;
# endif
2016-01-03 08:05:51 +03:00
while ( ( cmd = hostdata - > connected ) ) {
2016-01-03 08:05:58 +03:00
struct NCR5380_cmd * ncmd = scsi_cmd_priv ( cmd ) ;
2005-04-17 02:20:36 +04:00
tmp = NCR5380_read ( STATUS_REG ) ;
/* We only have a valid SCSI phase when REQ is asserted */
if ( tmp & SR_REQ ) {
phase = ( tmp & PHASE_MASK ) ;
if ( phase ! = old_phase ) {
old_phase = phase ;
NCR5380_dprint_phase ( NDEBUG_INFORMATION , instance ) ;
}
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
if ( phase = = PHASE_CMDOUT ) {
void * d ;
unsigned long count ;
if ( ! cmd - > SCp . this_residual & & cmd - > SCp . buffers_residual ) {
count = cmd - > SCp . buffer - > length ;
d = sg_virt ( cmd - > SCp . buffer ) ;
} else {
count = cmd - > SCp . this_residual ;
d = cmd - > SCp . ptr ;
}
if ( sun3_dma_setup_done ! = cmd & &
sun3scsi_dma_xfer_len ( count , cmd ) > 0 ) {
sun3scsi_dma_setup ( instance , d , count ,
rq_data_dir ( cmd - > request ) ) ;
sun3_dma_setup_done = cmd ;
}
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_INTR ;
# endif
}
# endif /* CONFIG_SUN3 */
2005-04-17 02:20:36 +04:00
if ( sink & & ( phase ! = PHASE_MSGOUT ) ) {
NCR5380_write ( TARGET_COMMAND_REG , PHASE_SR_TO_TCR ( tmp ) ) ;
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN |
ICR_ASSERT_ACK ) ;
2016-01-03 08:06:11 +03:00
while ( NCR5380_read ( STATUS_REG ) & SR_REQ )
;
2016-01-03 08:06:13 +03:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE |
ICR_ASSERT_ATN ) ;
2005-04-17 02:20:36 +04:00
sink = 0 ;
continue ;
}
2016-01-03 08:06:11 +03:00
2005-04-17 02:20:36 +04:00
switch ( phase ) {
case PHASE_DATAOUT :
# if (NDEBUG & NDEBUG_NO_DATAOUT)
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_DEBUG , instance , " NDEBUG_NO_DATAOUT set, attempted DATAOUT aborted \n " ) ;
2005-04-17 02:20:36 +04:00
sink = 1 ;
do_abort ( instance ) ;
cmd - > result = DID_ERROR < < 16 ;
2016-01-03 08:05:59 +03:00
complete_cmd ( instance , cmd ) ;
2016-02-23 02:07:07 +03:00
hostdata - > connected = NULL ;
2005-04-17 02:20:36 +04:00
return ;
# endif
2016-01-03 08:05:47 +03:00
case PHASE_DATAIN :
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* If there is no room left in the current buffer in the
* scatter - gather list , move onto the next one .
*/
if ( ! cmd - > SCp . this_residual & & cmd - > SCp . buffers_residual ) {
+ + cmd - > SCp . buffer ;
- - cmd - > SCp . buffers_residual ;
cmd - > SCp . this_residual = cmd - > SCp . buffer - > length ;
2007-10-22 23:19:53 +04:00
cmd - > SCp . ptr = sg_virt ( cmd - > SCp . buffer ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_INFORMATION , instance , " %d bytes and %d buffers left \n " ,
cmd - > SCp . this_residual ,
cmd - > SCp . buffers_residual ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:11 +03:00
2005-04-17 02:20:36 +04:00
/*
2016-01-03 08:06:09 +03:00
* The preferred transfer method is going to be
2005-04-17 02:20:36 +04:00
* PSEUDO - DMA for systems that are strictly PIO ,
* since we can let the hardware do the handshaking .
*
* For this to work , we need to know the transfersize
* ahead of time , since the pseudo - DMA code will sit
* in an unconditional loop .
*/
2016-01-03 08:05:25 +03:00
transfersize = 0 ;
2016-03-23 13:10:11 +03:00
if ( ! cmd - > device - > borken )
2016-01-03 08:05:25 +03:00
transfersize = NCR5380_dma_xfer_len ( instance , cmd , phase ) ;
2016-03-23 13:10:18 +03:00
if ( transfersize > 0 ) {
2005-04-17 02:20:36 +04:00
len = transfersize ;
2016-01-03 08:06:11 +03:00
if ( NCR5380_transfer_dma ( instance , & phase ,
& len , ( unsigned char * * ) & cmd - > SCp . ptr ) ) {
2005-04-17 02:20:36 +04:00
/*
2016-01-03 08:06:11 +03:00
* If the watchdog timer fires , all future
* accesses to this device will use the
* polled - IO .
2005-04-17 02:20:36 +04:00
*/
2005-10-25 02:04:36 +04:00
scmd_printk ( KERN_INFO , cmd ,
2016-01-03 08:06:11 +03:00
" switching to slow handshake \n " ) ;
2005-04-17 02:20:36 +04:00
cmd - > device - > borken = 1 ;
sink = 1 ;
do_abort ( instance ) ;
cmd - > result = DID_ERROR < < 16 ;
/* XXX - need to source or sink data here, as appropriate */
2016-03-23 13:10:19 +03:00
}
2016-03-23 13:10:15 +03:00
} else {
2016-08-31 07:44:56 +03:00
/* Transfer a small chunk so that the
* irq mode lock is not held too long .
2016-02-23 02:07:05 +03:00
*/
2016-08-31 07:44:56 +03:00
transfersize = min ( cmd - > SCp . this_residual ,
NCR5380_PIO_CHUNK_SIZE ) ;
2016-02-23 02:07:05 +03:00
len = transfersize ;
NCR5380_transfer_pio ( instance , & phase , & len ,
2016-01-03 08:06:13 +03:00
( unsigned char * * ) & cmd - > SCp . ptr ) ;
2016-02-23 02:07:05 +03:00
cmd - > SCp . this_residual - = transfersize - len ;
2016-01-03 08:05:51 +03:00
}
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
if ( sun3_dma_setup_done = = cmd )
sun3_dma_setup_done = NULL ;
# endif
2016-02-23 02:07:05 +03:00
return ;
2005-04-17 02:20:36 +04:00
case PHASE_MSGIN :
len = 1 ;
data = & tmp ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
cmd - > SCp . Message = tmp ;
switch ( tmp ) {
case ABORT :
case COMMAND_COMPLETE :
/* Accept message by clearing ACK */
sink = 1 ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2016-01-03 08:05:55 +03:00
dsprintk ( NDEBUG_QUEUES , instance ,
" COMMAND COMPLETE %p target %d lun %llu \n " ,
cmd , scmd_id ( cmd ) , cmd - > device - > lun ) ;
2005-04-17 02:20:36 +04:00
hostdata - > connected = NULL ;
2016-01-03 08:06:00 +03:00
cmd - > result & = ~ 0xffff ;
cmd - > result | = cmd - > SCp . Status ;
cmd - > result | = cmd - > SCp . Message < < 8 ;
2007-09-10 23:37:45 +04:00
2016-01-03 08:06:00 +03:00
if ( cmd - > cmnd [ 0 ] = = REQUEST_SENSE )
2016-01-03 08:05:59 +03:00
complete_cmd ( instance , cmd ) ;
2016-01-03 08:06:00 +03:00
else {
if ( cmd - > SCp . Status = = SAM_STAT_CHECK_CONDITION | |
cmd - > SCp . Status = = SAM_STAT_COMMAND_TERMINATED ) {
dsprintk ( NDEBUG_QUEUES , instance , " autosense: adding cmd %p to tail of autosense queue \n " ,
cmd ) ;
list_add_tail ( & ncmd - > list ,
& hostdata - > autosense ) ;
} else
complete_cmd ( instance , cmd ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:09 +03:00
/*
* Restore phase bits to 0 so an interrupted selection ,
2005-04-17 02:20:36 +04:00
* arbitration can resume .
*/
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
2016-01-03 08:05:44 +03:00
/* Enable reselect interrupts */
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
2016-03-23 13:10:20 +03:00
maybe_release_dma_irq ( instance ) ;
2005-04-17 02:20:36 +04:00
return ;
case MESSAGE_REJECT :
/* Accept message by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
switch ( hostdata - > last_message ) {
case HEAD_OF_QUEUE_TAG :
case ORDERED_QUEUE_TAG :
case SIMPLE_QUEUE_TAG :
cmd - > device - > simple_tags = 0 ;
2014-06-25 17:27:36 +04:00
hostdata - > busy [ cmd - > device - > id ] | = ( 1 < < ( cmd - > device - > lun & 0xFF ) ) ;
2005-04-17 02:20:36 +04:00
break ;
default :
break ;
}
2016-01-03 08:05:31 +03:00
break ;
2016-01-03 08:06:11 +03:00
case DISCONNECT :
/* Accept message by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
hostdata - > connected = NULL ;
list_add ( & ncmd - > list , & hostdata - > disconnected ) ;
dsprintk ( NDEBUG_INFORMATION | NDEBUG_QUEUES ,
instance , " connected command %p for target %d lun %llu moved to disconnected queue \n " ,
cmd , scmd_id ( cmd ) , cmd - > device - > lun ) ;
2016-01-03 08:05:55 +03:00
2016-01-03 08:06:11 +03:00
/*
* Restore phase bits to 0 so an interrupted selection ,
* arbitration can resume .
*/
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:11 +03:00
/* Enable reselect interrupts */
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
2016-03-23 13:10:21 +03:00
# ifdef SUN3_SCSI_VME
dregs - > csr | = CSR_DMA_ENABLE ;
# endif
2016-01-03 08:06:11 +03:00
return ;
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* The SCSI data pointer is * IMPLICITLY * saved on a disconnect
2016-01-03 08:06:09 +03:00
* operation , in violation of the SCSI spec so we can safely
2005-04-17 02:20:36 +04:00
* ignore SAVE / RESTORE pointers calls .
*
2016-01-03 08:06:09 +03:00
* Unfortunately , some disks violate the SCSI spec and
2005-04-17 02:20:36 +04:00
* don ' t issue the required SAVE_POINTERS message before
2016-01-03 08:06:09 +03:00
* disconnecting , and we have to break spec to remain
2005-04-17 02:20:36 +04:00
* compatible .
*/
case SAVE_POINTERS :
case RESTORE_POINTERS :
/* Accept message by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
break ;
case EXTENDED_MESSAGE :
2016-01-03 08:06:08 +03:00
/*
* Start the message buffer with the EXTENDED_MESSAGE
* byte , since spi_print_msg ( ) wants the whole thing .
*/
2005-04-17 02:20:36 +04:00
extended_msg [ 0 ] = EXTENDED_MESSAGE ;
/* Accept first byte by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
2016-01-03 08:05:51 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_EXTENDED , instance , " receiving extended message \n " ) ;
2005-04-17 02:20:36 +04:00
len = 2 ;
data = extended_msg + 1 ;
phase = PHASE_MSGIN ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_EXTENDED , instance , " length %d, code 0x%02x \n " ,
( int ) extended_msg [ 1 ] ,
( int ) extended_msg [ 2 ] ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:45 +03:00
if ( ! len & & extended_msg [ 1 ] > 0 & &
extended_msg [ 1 ] < = sizeof ( extended_msg ) - 2 ) {
2005-04-17 02:20:36 +04:00
/* Accept third byte by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
len = extended_msg [ 1 ] - 1 ;
data = extended_msg + 3 ;
phase = PHASE_MSGIN ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_EXTENDED , instance , " message received, residual %d \n " ,
len ) ;
2005-04-17 02:20:36 +04:00
switch ( extended_msg [ 2 ] ) {
case EXTENDED_SDTR :
case EXTENDED_WDTR :
case EXTENDED_MODIFY_DATA_POINTER :
case EXTENDED_EXTENDED_IDENTIFY :
tmp = 0 ;
}
} else if ( len ) {
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_ERR , instance , " error receiving extended message \n " ) ;
2005-04-17 02:20:36 +04:00
tmp = 0 ;
} else {
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_NOTICE , instance , " extended message code %02x length %d is too long \n " ,
extended_msg [ 2 ] , extended_msg [ 1 ] ) ;
2005-04-17 02:20:36 +04:00
tmp = 0 ;
}
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
if ( ! hostdata - > connected )
return ;
2005-04-17 02:20:36 +04:00
/* Fall through to reject message */
2016-01-03 08:06:09 +03:00
/*
* If we get something weird that we aren ' t expecting ,
2005-04-17 02:20:36 +04:00
* reject it .
*/
default :
if ( ! tmp ) {
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_ERR , instance , " rejecting message " ) ;
2005-12-16 00:22:01 +03:00
spi_print_msg ( extended_msg ) ;
2005-04-17 02:20:36 +04:00
printk ( " \n " ) ;
} else if ( tmp ! = EXTENDED_MESSAGE )
2005-10-25 02:04:36 +04:00
scmd_printk ( KERN_INFO , cmd ,
2016-01-03 08:06:11 +03:00
" rejecting unknown message %02x \n " ,
tmp ) ;
2005-04-17 02:20:36 +04:00
else
2005-10-25 02:04:36 +04:00
scmd_printk ( KERN_INFO , cmd ,
2016-01-03 08:06:11 +03:00
" rejecting unknown extended message code %02x, length %d \n " ,
extended_msg [ 1 ] , extended_msg [ 0 ] ) ;
2005-04-17 02:20:36 +04:00
msgout = MESSAGE_REJECT ;
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ATN ) ;
break ;
2016-01-03 08:06:11 +03:00
} /* switch (tmp) */
2005-04-17 02:20:36 +04:00
break ;
case PHASE_MSGOUT :
len = 1 ;
data = & msgout ;
hostdata - > last_message = msgout ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
if ( msgout = = ABORT ) {
hostdata - > connected = NULL ;
cmd - > result = DID_ERROR < < 16 ;
2016-01-03 08:05:59 +03:00
complete_cmd ( instance , cmd ) ;
2016-03-23 13:10:20 +03:00
maybe_release_dma_irq ( instance ) ;
2005-04-17 02:20:36 +04:00
NCR5380_write ( SELECT_ENABLE_REG , hostdata - > id_mask ) ;
return ;
}
msgout = NOP ;
break ;
case PHASE_CMDOUT :
len = cmd - > cmd_len ;
data = cmd - > cmnd ;
2016-01-03 08:06:09 +03:00
/*
* XXX for performance reasons , on machines with a
* PSEUDO - DMA architecture we should probably
* use the dma transfer function .
2005-04-17 02:20:36 +04:00
*/
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
break ;
case PHASE_STATIN :
len = 1 ;
data = & tmp ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
cmd - > SCp . Status = tmp ;
break ;
default :
2016-01-03 08:06:04 +03:00
shost_printk ( KERN_ERR , instance , " unknown phase \n " ) ;
2014-03-18 04:42:17 +04:00
NCR5380_dprint ( NDEBUG_ANY , instance ) ;
2016-01-03 08:06:11 +03:00
} /* switch(phase) */
ncr5380: Rework disconnect versus poll logic
The atari_NCR5380.c and NCR5380.c core drivers differ in their handling of
target disconnection. This is partly because atari_NCR5380.c had all of
the polling and sleeping removed to become entirely interrupt-driven, and
it is partly because of damage done to NCR5380.c after atari_NCR5380.c was
forked. See commit 37cd23b44929 ("Linux 2.1.105") in history/history.git.
The polling changes that were made in v2.1.105 are questionable at best:
if REQ is not already asserted when NCR5380_transfer_pio() is invoked, and
if the expected phase is DATA IN or DATA OUT, the function will schedule
main() to execute after USLEEP_SLEEP jiffies and then return. The problems
here are the expected REQ timing and the sleep interval*. Avoid this issue
by using NCR5380_poll_politely() instead of scheduling main().
The atari_NCR5380.c core driver requires the use of the chip interrupt and
always permits target disconnection. It sets the cmd->device->disconnect
flag when a device disconnects, but never tests this flag.
The NCR5380.c core driver permits disconnection only when
instance->irq != NO_IRQ. It sets the cmd->device->disconnect flag when
a device disconnects and it tests this flag in a couple of places:
1. During NCR5380_information_transfer(), following COMMAND OUT phase,
if !cmd->device->disconnect, the initiator will take a guess as to
whether or not the target will then choose to go to MESSAGE IN phase
and disconnect. If the driver guesses "yes", it will schedule main()
to execute after USLEEP_SLEEP jiffies and then return there.
Unfortunately the driver may guess "yes" even after it has denied
the target the disconnection privilege. When the target does not
disconnect, the sleep can be beneficial, assuming the sleep interval
is appropriate (mostly it is not*).
And even if the driver guesses "yes" correctly, and the target would
then disconnect, the driver still has to go through the MESSAGE IN
phase in order to get to BUS FREE phase. The main loop can do nothing
useful until BUS FREE, and sleeping just delays the phase transition.
2. If !cmd->device->disconnect and REQ is not already asserted when
NCR5380_information_transfer() is invoked, the function polls for REQ
for USLEEP_POLL jiffies. If REQ is not asserted, it then schedules
main() to execute after USLEEP_SLEEP jiffies and returns.
The idea is apparently to yeild the CPU while waiting for REQ.
This is conditional upon !cmd->device->disconnect, but there seems
to be no rhyme or reason for that. For example, the flag may be
unset because disconnection privilege was denied because the driver
has no IRQ. Or the flag may be unset because the device has never
needed to disconnect before. Or if the flag is set, disconnection
may have no relevance to the present bus phase.
Another deficiency of the existing algorithm is as follows. When the
driver has no IRQ, it prevents disconnection, and generally polls and
sleeps more than it would normally. Now, if the driver is going to poll
anyway, why not allow the target to disconnect? That way the driver can do
something useful with the bus instead of polling unproductively!
Avoid this pointless latency, complexity and guesswork by using
NCR5380_poll_politely() instead of scheduling main().
* For g_NCR5380, the time intervals for USLEEP_SLEEP and USLEEP_POLL are
200 ms and 10 ms, respectively. They are 20 ms and 200 ms respectively
for the other NCR5380 drivers. There doesn't seem to be any reason for
this discrepancy. The timing seems to have no relation to the type of
adapter. Bizarrely, the timing in g_NCR5380 seems to relate only to one
particular type of target device. This patch attempts to solve the
problem for all NCR5380 drivers and all target devices.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:26 +03:00
} else {
2016-01-03 08:05:51 +03:00
spin_unlock_irq ( & hostdata - > lock ) ;
ncr5380: Rework disconnect versus poll logic
The atari_NCR5380.c and NCR5380.c core drivers differ in their handling of
target disconnection. This is partly because atari_NCR5380.c had all of
the polling and sleeping removed to become entirely interrupt-driven, and
it is partly because of damage done to NCR5380.c after atari_NCR5380.c was
forked. See commit 37cd23b44929 ("Linux 2.1.105") in history/history.git.
The polling changes that were made in v2.1.105 are questionable at best:
if REQ is not already asserted when NCR5380_transfer_pio() is invoked, and
if the expected phase is DATA IN or DATA OUT, the function will schedule
main() to execute after USLEEP_SLEEP jiffies and then return. The problems
here are the expected REQ timing and the sleep interval*. Avoid this issue
by using NCR5380_poll_politely() instead of scheduling main().
The atari_NCR5380.c core driver requires the use of the chip interrupt and
always permits target disconnection. It sets the cmd->device->disconnect
flag when a device disconnects, but never tests this flag.
The NCR5380.c core driver permits disconnection only when
instance->irq != NO_IRQ. It sets the cmd->device->disconnect flag when
a device disconnects and it tests this flag in a couple of places:
1. During NCR5380_information_transfer(), following COMMAND OUT phase,
if !cmd->device->disconnect, the initiator will take a guess as to
whether or not the target will then choose to go to MESSAGE IN phase
and disconnect. If the driver guesses "yes", it will schedule main()
to execute after USLEEP_SLEEP jiffies and then return there.
Unfortunately the driver may guess "yes" even after it has denied
the target the disconnection privilege. When the target does not
disconnect, the sleep can be beneficial, assuming the sleep interval
is appropriate (mostly it is not*).
And even if the driver guesses "yes" correctly, and the target would
then disconnect, the driver still has to go through the MESSAGE IN
phase in order to get to BUS FREE phase. The main loop can do nothing
useful until BUS FREE, and sleeping just delays the phase transition.
2. If !cmd->device->disconnect and REQ is not already asserted when
NCR5380_information_transfer() is invoked, the function polls for REQ
for USLEEP_POLL jiffies. If REQ is not asserted, it then schedules
main() to execute after USLEEP_SLEEP jiffies and returns.
The idea is apparently to yeild the CPU while waiting for REQ.
This is conditional upon !cmd->device->disconnect, but there seems
to be no rhyme or reason for that. For example, the flag may be
unset because disconnection privilege was denied because the driver
has no IRQ. Or the flag may be unset because the device has never
needed to disconnect before. Or if the flag is set, disconnection
may have no relevance to the present bus phase.
Another deficiency of the existing algorithm is as follows. When the
driver has no IRQ, it prevents disconnection, and generally polls and
sleeps more than it would normally. Now, if the driver is going to poll
anyway, why not allow the target to disconnect? That way the driver can do
something useful with the bus instead of polling unproductively!
Avoid this pointless latency, complexity and guesswork by using
NCR5380_poll_politely() instead of scheduling main().
* For g_NCR5380, the time intervals for USLEEP_SLEEP and USLEEP_POLL are
200 ms and 10 ms, respectively. They are 20 ms and 200 ms respectively
for the other NCR5380 drivers. There doesn't seem to be any reason for
this discrepancy. The timing seems to have no relation to the type of
adapter. Bizarrely, the timing in g_NCR5380 seems to relate only to one
particular type of target device. This patch attempts to solve the
problem for all NCR5380 drivers and all target devices.
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ondrej Zary <linux@rainbow-software.org>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2016-01-03 08:05:26 +03:00
NCR5380_poll_politely ( instance , STATUS_REG , SR_REQ , SR_REQ , HZ ) ;
2016-01-03 08:05:51 +03:00
spin_lock_irq ( & hostdata - > lock ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:51 +03:00
}
2005-04-17 02:20:36 +04:00
}
/*
* Function : void NCR5380_reselect ( struct Scsi_Host * instance )
*
2016-01-03 08:06:09 +03:00
* Purpose : does reselection , initializing the instance - > connected
2016-01-03 08:06:10 +03:00
* field to point to the scsi_cmnd for which the I_T_L or I_T_L_Q
* nexus has been reestablished ,
2016-01-03 08:06:09 +03:00
*
2005-04-17 02:20:36 +04:00
* Inputs : instance - this instance of the NCR5380 .
*/
2016-01-03 08:06:11 +03:00
static void NCR5380_reselect ( struct Scsi_Host * instance )
{
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2005-04-17 02:20:36 +04:00
unsigned char target_mask ;
2016-03-23 13:10:21 +03:00
unsigned char lun ;
2005-04-17 02:20:36 +04:00
unsigned char msg [ 3 ] ;
2016-01-03 08:05:58 +03:00
struct NCR5380_cmd * ncmd ;
struct scsi_cmnd * tmp ;
2005-04-17 02:20:36 +04:00
/*
* Disable arbitration , etc . since the host adapter obviously
* lost , and tell an interrupted NCR5380_select ( ) to restart .
*/
NCR5380_write ( MODE_REG , MR_BASE ) ;
target_mask = NCR5380_read ( CURRENT_SCSI_DATA_REG ) & ~ ( hostdata - > id_mask ) ;
2016-01-03 08:06:05 +03:00
dsprintk ( NDEBUG_RESELECTION , instance , " reselect \n " ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:09 +03:00
/*
2005-04-17 02:20:36 +04:00
* At this point , we have detected that our SCSI ID is on the bus ,
* SEL is true and BSY was false for at least one bus settle delay
* ( 400 ns ) .
*
* We must assert BSY ourselves , until the target drops the SEL
* signal .
*/
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_BSY ) ;
2016-01-03 08:05:44 +03:00
if ( NCR5380_poll_politely ( instance ,
STATUS_REG , SR_SEL , 0 , 2 * HZ ) < 0 ) {
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
return ;
}
2005-04-17 02:20:36 +04:00
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
/*
* Wait for target to go into MSGIN .
*/
2016-01-03 08:05:32 +03:00
if ( NCR5380_poll_politely ( instance ,
2016-01-03 08:05:44 +03:00
STATUS_REG , SR_REQ , SR_REQ , 2 * HZ ) < 0 ) {
do_abort ( instance ) ;
return ;
}
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
/* acknowledge toggle to MSGIN */
NCR5380_write ( TARGET_COMMAND_REG , PHASE_SR_TO_TCR ( PHASE_MSGIN ) ) ;
2005-04-17 02:20:36 +04:00
2016-03-23 13:10:21 +03:00
/* peek at the byte without really hitting the bus */
msg [ 0 ] = NCR5380_read ( CURRENT_SCSI_DATA_REG ) ;
# else
{
int len = 1 ;
unsigned char * data = msg ;
unsigned char phase = PHASE_MSGIN ;
NCR5380_transfer_pio ( instance , & phase , & len , & data ) ;
if ( len ) {
do_abort ( instance ) ;
return ;
}
2016-01-03 08:05:44 +03:00
}
2016-03-23 13:10:21 +03:00
# endif /* CONFIG_SUN3 */
2016-01-03 08:05:44 +03:00
2005-04-17 02:20:36 +04:00
if ( ! ( msg [ 0 ] & 0x80 ) ) {
2016-01-03 08:05:44 +03:00
shost_printk ( KERN_ERR , instance , " expecting IDENTIFY message, got " ) ;
2005-12-16 00:22:01 +03:00
spi_print_msg ( msg ) ;
2016-01-03 08:05:44 +03:00
printk ( " \n " ) ;
do_abort ( instance ) ;
return ;
}
lun = msg [ 0 ] & 0x07 ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:44 +03:00
/*
* We need to add code for SCSI - II to track which devices have
* I_T_L_Q nexuses established , and which have simple I_T_L
* nexuses so we can chose to do additional data transfer .
*/
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:44 +03:00
/*
* Find the command corresponding to the I_T_L or I_T_L_Q nexus we
* just reestablished , and remove it from the disconnected queue .
*/
2005-04-17 02:20:36 +04:00
2016-01-03 08:05:58 +03:00
tmp = NULL ;
list_for_each_entry ( ncmd , & hostdata - > disconnected , list ) {
struct scsi_cmnd * cmd = NCR5380_to_scmd ( ncmd ) ;
if ( target_mask = = ( 1 < < scmd_id ( cmd ) ) & &
lun = = ( u8 ) cmd - > device - > lun ) {
list_del ( & ncmd - > list ) ;
tmp = cmd ;
2016-01-03 08:05:44 +03:00
break ;
2005-04-17 02:20:36 +04:00
}
}
2016-01-03 08:05:55 +03:00
if ( tmp ) {
dsprintk ( NDEBUG_RESELECTION | NDEBUG_QUEUES , instance ,
" reselect: removed %p from disconnected queue \n " , tmp ) ;
} else {
2016-01-03 08:05:44 +03:00
shost_printk ( KERN_ERR , instance , " target bitmask 0x%02x lun %d not in disconnected queue. \n " ,
target_mask , lun ) ;
/*
2016-01-03 08:06:11 +03:00
* Since we have an established nexus that we can ' t do anything
* with , we must abort it .
2016-01-03 08:05:44 +03:00
*/
2005-04-17 02:20:36 +04:00
do_abort ( instance ) ;
2016-01-03 08:05:44 +03:00
return ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:44 +03:00
2016-03-23 13:10:21 +03:00
# ifdef CONFIG_SUN3
{
void * d ;
unsigned long count ;
if ( ! tmp - > SCp . this_residual & & tmp - > SCp . buffers_residual ) {
count = tmp - > SCp . buffer - > length ;
d = sg_virt ( tmp - > SCp . buffer ) ;
} else {
count = tmp - > SCp . this_residual ;
d = tmp - > SCp . ptr ;
}
if ( sun3_dma_setup_done ! = tmp & &
sun3scsi_dma_xfer_len ( count , tmp ) > 0 ) {
sun3scsi_dma_setup ( instance , d , count ,
rq_data_dir ( tmp - > request ) ) ;
sun3_dma_setup_done = tmp ;
}
}
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE | ICR_ASSERT_ACK ) ;
# endif /* CONFIG_SUN3 */
2016-01-03 08:05:44 +03:00
/* Accept message by clearing ACK */
NCR5380_write ( INITIATOR_COMMAND_REG , ICR_BASE ) ;
hostdata - > connected = tmp ;
2016-03-23 13:10:22 +03:00
dsprintk ( NDEBUG_RESELECTION , instance , " nexus established, target %d, lun %llu \n " ,
scmd_id ( tmp ) , tmp - > device - > lun ) ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:06:01 +03:00
/**
* list_find_cmd - test for presence of a command in a linked list
* @ haystack : list of commands
* @ needle : command to search for
*/
static bool list_find_cmd ( struct list_head * haystack ,
struct scsi_cmnd * needle )
{
struct NCR5380_cmd * ncmd ;
list_for_each_entry ( ncmd , haystack , list )
if ( NCR5380_to_scmd ( ncmd ) = = needle )
return true ;
return false ;
}
/**
* list_remove_cmd - remove a command from linked list
* @ haystack : list of commands
* @ needle : command to remove
*/
static bool list_del_cmd ( struct list_head * haystack ,
struct scsi_cmnd * needle )
{
if ( list_find_cmd ( haystack , needle ) ) {
struct NCR5380_cmd * ncmd = scsi_cmd_priv ( needle ) ;
list_del ( & ncmd - > list ) ;
return true ;
}
return false ;
}
/**
* NCR5380_abort - scsi host eh_abort_handler ( ) method
* @ cmd : the command to be aborted
*
* Try to abort a given command by removing it from queues and / or sending
* the target an abort message . This may not succeed in causing a target
* to abort the command . Nonetheless , the low - level driver must forget about
* the command because the mid - layer reclaims it and it may be re - issued .
*
* The normal path taken by a command is as follows . For EH we trace this
* same path to locate and abort the command .
*
* unissued - > selecting - > [ unissued - > selecting - > ] . . . connected - >
* [ disconnected - > connected - > ] . . .
* [ autosense - > connected - > ] done
*
* If cmd was not found at all then presumably it has already been completed ,
* in which case return SUCCESS to try to avoid further EH measures .
2016-02-23 02:07:07 +03:00
*
2016-01-03 08:06:01 +03:00
* If the command has not completed yet , we must not fail to find it .
2016-02-23 02:07:07 +03:00
* We have no option but to forget the aborted command ( even if it still
* lacks sense data ) . The mid - layer may re - issue a command that is in error
* recovery ( see scsi_send_eh_cmnd ) , but the logic and data structures in
* this driver are such that a command can appear on one queue only .
2016-02-23 02:07:06 +03:00
*
* The lock protects driver data structures , but EH handlers also use it
* to serialize their own execution and prevent their own re - entry .
2005-04-17 02:20:36 +04:00
*/
2014-11-12 08:12:02 +03:00
static int NCR5380_abort ( struct scsi_cmnd * cmd )
{
2005-04-17 02:20:36 +04:00
struct Scsi_Host * instance = cmd - > device - > host ;
2016-01-03 08:05:54 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-01-03 08:05:51 +03:00
unsigned long flags ;
2016-01-03 08:06:01 +03:00
int result = SUCCESS ;
2014-10-24 16:26:58 +04:00
2016-01-03 08:05:51 +03:00
spin_lock_irqsave ( & hostdata - > lock , flags ) ;
2016-01-03 08:05:58 +03:00
# if (NDEBUG & NDEBUG_ANY)
2016-01-03 08:06:01 +03:00
scmd_printk ( KERN_INFO , cmd , __func__ ) ;
2016-01-03 08:05:58 +03:00
# endif
2016-01-03 08:05:52 +03:00
NCR5380_dprint ( NDEBUG_ANY , instance ) ;
NCR5380_dprint_phase ( NDEBUG_ANY , instance ) ;
2005-04-17 02:20:36 +04:00
2016-01-03 08:06:01 +03:00
if ( list_del_cmd ( & hostdata - > unissued , cmd ) ) {
dsprintk ( NDEBUG_ABORT , instance ,
" abort: removed %p from issue queue \n " , cmd ) ;
cmd - > result = DID_ABORT < < 16 ;
cmd - > scsi_done ( cmd ) ; /* No tag or busy flag to worry about */
2016-02-23 02:07:07 +03:00
goto out ;
2016-01-03 08:06:01 +03:00
}
2016-01-03 08:06:02 +03:00
if ( hostdata - > selecting = = cmd ) {
dsprintk ( NDEBUG_ABORT , instance ,
" abort: cmd %p == selecting \n " , cmd ) ;
hostdata - > selecting = NULL ;
cmd - > result = DID_ABORT < < 16 ;
complete_cmd ( instance , cmd ) ;
goto out ;
}
2016-01-03 08:06:01 +03:00
if ( list_del_cmd ( & hostdata - > disconnected , cmd ) ) {
dsprintk ( NDEBUG_ABORT , instance ,
" abort: removed %p from disconnected list \n " , cmd ) ;
2016-02-23 02:07:06 +03:00
/* Can't call NCR5380_select() and send ABORT because that
* means releasing the lock . Need a bus reset .
*/
2016-02-23 02:07:07 +03:00
set_host_byte ( cmd , DID_ERROR ) ;
complete_cmd ( instance , cmd ) ;
2016-02-23 02:07:06 +03:00
result = FAILED ;
goto out ;
2016-01-03 08:06:01 +03:00
}
if ( hostdata - > connected = = cmd ) {
dsprintk ( NDEBUG_ABORT , instance , " abort: cmd %p is connected \n " , cmd ) ;
hostdata - > connected = NULL ;
hostdata - > dma_len = 0 ;
if ( do_abort ( instance ) ) {
set_host_byte ( cmd , DID_ERROR ) ;
complete_cmd ( instance , cmd ) ;
result = FAILED ;
goto out ;
}
set_host_byte ( cmd , DID_ABORT ) ;
2016-02-23 02:07:07 +03:00
complete_cmd ( instance , cmd ) ;
goto out ;
}
if ( list_del_cmd ( & hostdata - > autosense , cmd ) ) {
dsprintk ( NDEBUG_ABORT , instance ,
" abort: removed %p from sense queue \n " , cmd ) ;
set_host_byte ( cmd , DID_ERROR ) ;
2016-01-03 08:06:01 +03:00
complete_cmd ( instance , cmd ) ;
}
out :
if ( result = = FAILED )
dsprintk ( NDEBUG_ABORT , instance , " abort: failed to abort %p \n " , cmd ) ;
else
dsprintk ( NDEBUG_ABORT , instance , " abort: successfully aborted %p \n " , cmd ) ;
queue_work ( hostdata - > work_q , & hostdata - > main_task ) ;
2016-03-23 13:10:20 +03:00
maybe_release_dma_irq ( instance ) ;
2016-01-03 08:05:51 +03:00
spin_unlock_irqrestore ( & hostdata - > lock , flags ) ;
2016-01-03 08:05:58 +03:00
2016-01-03 08:06:01 +03:00
return result ;
2005-04-17 02:20:36 +04:00
}
2016-01-03 08:05:12 +03:00
/**
* NCR5380_bus_reset - reset the SCSI bus
* @ cmd : SCSI command undergoing EH
2005-04-17 02:20:36 +04:00
*
2016-01-03 08:05:12 +03:00
* Returns SUCCESS
2005-04-17 02:20:36 +04:00
*/
2014-11-12 08:12:02 +03:00
static int NCR5380_bus_reset ( struct scsi_cmnd * cmd )
2005-05-28 15:56:31 +04:00
{
struct Scsi_Host * instance = cmd - > device - > host ;
2016-01-03 08:05:51 +03:00
struct NCR5380_hostdata * hostdata = shost_priv ( instance ) ;
2016-01-03 08:06:03 +03:00
int i ;
2016-01-03 08:05:51 +03:00
unsigned long flags ;
2016-01-03 08:06:03 +03:00
struct NCR5380_cmd * ncmd ;
2005-05-28 15:56:31 +04:00
2016-01-03 08:05:51 +03:00
spin_lock_irqsave ( & hostdata - > lock , flags ) ;
2016-01-03 08:05:12 +03:00
# if (NDEBUG & NDEBUG_ANY)
2016-01-03 08:06:03 +03:00
scmd_printk ( KERN_INFO , cmd , __func__ ) ;
2016-01-03 08:05:12 +03:00
# endif
2016-01-03 08:05:52 +03:00
NCR5380_dprint ( NDEBUG_ANY , instance ) ;
NCR5380_dprint_phase ( NDEBUG_ANY , instance ) ;
2005-05-28 15:56:31 +04:00
do_reset ( instance ) ;
2016-01-03 08:05:12 +03:00
2016-01-03 08:06:03 +03:00
/* reset NCR registers */
NCR5380_write ( MODE_REG , MR_BASE ) ;
NCR5380_write ( TARGET_COMMAND_REG , 0 ) ;
NCR5380_write ( SELECT_ENABLE_REG , 0 ) ;
/* After the reset, there are no more connected or disconnected commands
* and no busy units ; so clear the low - level status here to avoid
* conflicts when the mid - level code tries to wake up the affected
* commands !
*/
2016-02-23 02:07:04 +03:00
if ( list_del_cmd ( & hostdata - > unissued , cmd ) ) {
cmd - > result = DID_RESET < < 16 ;
cmd - > scsi_done ( cmd ) ;
}
if ( hostdata - > selecting ) {
hostdata - > selecting - > result = DID_RESET < < 16 ;
complete_cmd ( instance , hostdata - > selecting ) ;
hostdata - > selecting = NULL ;
}
2016-01-03 08:06:03 +03:00
list_for_each_entry ( ncmd , & hostdata - > disconnected , list ) {
struct scsi_cmnd * cmd = NCR5380_to_scmd ( ncmd ) ;
set_host_byte ( cmd , DID_RESET ) ;
2016-03-23 13:10:32 +03:00
complete_cmd ( instance , cmd ) ;
2016-01-03 08:06:03 +03:00
}
2016-02-23 02:07:04 +03:00
INIT_LIST_HEAD ( & hostdata - > disconnected ) ;
2016-01-03 08:06:03 +03:00
list_for_each_entry ( ncmd , & hostdata - > autosense , list ) {
struct scsi_cmnd * cmd = NCR5380_to_scmd ( ncmd ) ;
set_host_byte ( cmd , DID_RESET ) ;
cmd - > scsi_done ( cmd ) ;
}
2016-02-23 02:07:04 +03:00
INIT_LIST_HEAD ( & hostdata - > autosense ) ;
2016-01-03 08:06:03 +03:00
if ( hostdata - > connected ) {
set_host_byte ( hostdata - > connected , DID_RESET ) ;
complete_cmd ( instance , hostdata - > connected ) ;
hostdata - > connected = NULL ;
}
for ( i = 0 ; i < 8 ; + + i )
hostdata - > busy [ i ] = 0 ;
hostdata - > dma_len = 0 ;
queue_work ( hostdata - > work_q , & hostdata - > main_task ) ;
2016-03-23 13:10:20 +03:00
maybe_release_dma_irq ( instance ) ;
2016-01-03 08:05:51 +03:00
spin_unlock_irqrestore ( & hostdata - > lock , flags ) ;
2005-04-17 02:20:36 +04:00
return SUCCESS ;
}