RepoMirrors
/
openwrt
mirror of git://git.openwrt.org/openwrt/openwrt.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

oxnas: sata_oxnas: refactoring phase 2 

- replaced // comments by /* comments */
 - added line-breaks where needed
 - fixed white-space according to kernel style
 - fixed some obvious spelling mistakes in comments and printks
 - removed some unneeded left-overs imported from vendor code-base
 - replaced printk(...) by libata macros where possible

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

SVN-Revision: 43767
This commit is contained in:
Felix Fietkau 2014-12-22 20:47:01 +00:00
parent feda47451f
commit 74a9e1c1b0
1 changed files with 174 additions and 111 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

285
target/linux/oxnas/files/drivers/ata/sata_oxnas.c
View File

@ -391,7 +391,7 @@ static int sata_oxnas_acquire_hw(struct ata_port *ap, int may_sleep,
int timeout_jiffies);
static void sata_oxnas_release_hw(struct ata_port *ap);
static const void *HW_LOCKER_UID = (void*)0xdeadbeef;
static const void *HW_LOCKER_UID = (void *)0xdeadbeef;
/***************************************************************************
* ASIC access
@ -561,7 +561,7 @@ static unsigned int sata_oxnas_qc_issue(struct ata_queued_cmd *qc)
void __iomem *port_base = pd->port_base;
void __iomem *core_base = pd->core_base;
int port_no = qc->ap->port_no;
int no_microcode = ( hd->current_ucode == UNKNOWN_MODE );
int no_microcode = (hd->current_ucode == UNKNOWN_MODE);
u32 reg;
/* check the core is idle */
@ -584,7 +584,7 @@ static unsigned int sata_oxnas_qc_issue(struct ata_queued_cmd *qc)
/* enable passing of error signals to DMA sub-core by clearing the
* appropriate bit */
reg = ioread32(core_base + DATA_PLANE_CTRL);
if(no_microcode)
if (no_microcode)
reg |= (DPC_ERROR_MASK_BIT | (DPC_ERROR_MASK_BIT << 1));
reg &= ~(DPC_ERROR_MASK_BIT << port_no);
iowrite32(reg, core_base + DATA_PLANE_CTRL);
@ -669,60 +669,84 @@ static int __acquire_sata_core(
spin_lock_irqsave(&hd->core_lock, flags);
DPRINTK("Entered uid %p, port %d, h/w count %d, d count %d, callback %p, "
"hw_access %d, core_locked %d, reentrant_port_no %d, isr_callback %p\n",
uid, port_no, hd->hw_lock_count, hd->direct_lock_count, callback, hw_access,
hd->core_locked, hd->reentrant_port_no, hd->isr_callback);
DPRINTK("Entered uid %p, port %d, h/w count %d, d count %d, "
"callback %p, hw_access %d, core_locked %d, "
"reentrant_port_no %d, isr_callback %p\n",
uid, port_no, hd->hw_lock_count, hd->direct_lock_count,
callback, hw_access, hd->core_locked, hd->reentrant_port_no,
hd->isr_callback);
while (!timed_out) {
if (hd->core_locked || (!hw_access && hd->scsi_nonblocking_attempts)) {
if (hd->core_locked ||
(!hw_access && hd->scsi_nonblocking_attempts)) {
/* Can only allow access if from SCSI/SATA stack and if
reentrant access is allowed and this access is to the same
port for which the lock is current held */
* reentrant access is allowed and this access is to the
* same port for which the lock is current held
*/
if (hw_access && (port_no == hd->reentrant_port_no)) {
BUG_ON(!hd->hw_lock_count);
++(hd->hw_lock_count);
DPRINTK("Allow SCSI/SATA re-entrant access to uid %p port %d\n", uid, port_no);
DPRINTK("Allow SCSI/SATA re-entrant access to "
"uid %p port %d\n", uid, port_no);
acquired = 1;
break;
} else if (!hw_access) {
if ((locker_type == SATA_READER) && (hd->current_locker_type == SATA_READER)) {
if ((locker_type == SATA_READER) &&
(hd->current_locker_type == SATA_READER)) {
WARN(1,
"Already locked by reader, uid %p, locker_uid %p, port %d, "
"h/w count %d, d count %d, hw_access %d\n", uid, hd->locker_uid,
port_no, hd->hw_lock_count, hd->direct_lock_count, hw_access);
"Already locked by reader, "
"uid %p, locker_uid %p, "
"port %d, h/w count %d, "
"d count %d, hw_access %d\n",
uid, hd->locker_uid, port_no,
hd->hw_lock_count,
hd->direct_lock_count,
hw_access);
goto check_uid;
}
if ((locker_type != SATA_READER) && (locker_type != SATA_WRITER)) {
if ((locker_type != SATA_READER) &&
(locker_type != SATA_WRITER)) {
goto wait_for_lock;
}
check_uid:
WARN(uid == hd->locker_uid, "Attempt to lock by locker type %d "
"uid %p, already locked by locker type %d with "
"locker_uid %p, port %d, h/w count %d, d count %d, "
"hw_access %d\n", locker_type, uid, hd->current_locker_type,
hd->locker_uid, port_no, hd->hw_lock_count, hd->direct_lock_count, hw_access);
WARN(uid == hd->locker_uid, "Attempt to lock "
"by locker type %d uid %p, already "
"locked by locker type %d with "
"locker_uid %p, port %d, "
"h/w count %d, d count %d, "
"hw_access %d\n", locker_type, uid,
hd->current_locker_type,
hd->locker_uid, port_no,
hd->hw_lock_count,
hd->direct_lock_count, hw_access);
}
} else {
WARN(hd->hw_lock_count || hd->direct_lock_count, "Core unlocked but counts "
"non-zero: uid %p, locker_uid %p, port %d, h/w count %d, "
"d count %d, hw_access %d\n", uid, hd->locker_uid, port_no,
hd->hw_lock_count, hd->direct_lock_count, hw_access);
WARN(hd->hw_lock_count || hd->direct_lock_count,
"Core unlocked but counts non-zero: uid %p, "
"locker_uid %p, port %d, h/w count %d, "
"d count %d, hw_access %d\n", uid,
hd->locker_uid, port_no, hd->hw_lock_count,
hd->direct_lock_count, hw_access);
BUG_ON(hd->current_locker_type != SATA_UNLOCKED);
WARN(hd->locker_uid, "Attempt to lock uid %p when locker_uid %p is "
"non-zero, port %d, h/w count %d, d count %d, hw_access %d\n",
uid, hd->locker_uid, port_no, hd->hw_lock_count, hd->direct_lock_count,
hw_access);
WARN(hd->locker_uid, "Attempt to lock uid %p when "
"locker_uid %p is non-zero, port %d, "
"h/w count %d, d count %d, hw_access %d\n",
uid, hd->locker_uid, port_no, hd->hw_lock_count,
hd->direct_lock_count, hw_access);
if (!hw_access) {
/* Direct access attempting to acquire non-contented lock */
BUG_ON(!callback); // Must have callback for direct access
BUG_ON(hd->reentrant_port_no != -1); // Sanity check lock state
/* Direct access attempting to acquire
* non-contented lock
*/
/* Must have callback for direct access */
BUG_ON(!callback);
/* Sanity check lock state */
BUG_ON(hd->reentrant_port_no != -1);
hd->isr_callback = callback;
hd->isr_arg = arg;
@ -730,12 +754,17 @@ check_uid:
hd->current_locker_type = locker_type;
} else {
/* SCSI/SATA attempting to acquire non-contented lock */
BUG_ON(callback); // No callbacks for SCSI/SATA access
BUG_ON(arg); // No callback args for SCSI/SATA access
/* SCSI/SATA attempting to acquire
* non-contented lock
*/
/* No callbacks for SCSI/SATA access */
BUG_ON(callback);
/* No callback args for SCSI/SATA access */
BUG_ON(arg);
BUG_ON(hd->isr_callback); // Sanity check lock state
BUG_ON(hd->isr_arg); // Sanity check lock state
/* Sanity check lock state */
BUG_ON(hd->isr_callback);
BUG_ON(hd->isr_arg);
++(hd->hw_lock_count);
hd->reentrant_port_no = port_no;
@ -751,66 +780,80 @@ check_uid:
wait_for_lock:
if (!may_sleep) {
DPRINTK("Denying for uid %p locker_type %d, hw_access %d, port %d, "
"current_locker_type %d as cannot sleep\n", uid, locker_type,
hw_access, port_no, hd->current_locker_type);
DPRINTK("Denying for uid %p locker_type %d, "
"hw_access %d, port %d, current_locker_type %d as "
"cannot sleep\n", uid, locker_type, hw_access, port_no,
hd->current_locker_type);
if (hw_access) {
if (hw_access)
++(hd->scsi_nonblocking_attempts);
}
break;
}
// Core is locked and we're allowed to sleep, so wait to be awoken when
// the core is unlocked
/* Core is locked and we're allowed to sleep, so wait to be
* awoken when the core is unlocked
*/
for (;;) {
prepare_to_wait(hw_access ? &hd->scsi_wait_queue : &hd->fast_wait_queue,
&wait, TASK_UNINTERRUPTIBLE);
if (!hd->core_locked && !(!hw_access && hd->scsi_nonblocking_attempts)) {
// We're going to use variables that will have been changed by
// the waker prior to clearing core_locked so we need to ensure
// we see changes to all those variables
prepare_to_wait(hw_access ? &hd->scsi_wait_queue :
&hd->fast_wait_queue,
&wait, TASK_UNINTERRUPTIBLE);
if (!hd->core_locked &&
!(!hw_access && hd->scsi_nonblocking_attempts)) {
/* We're going to use variables that will have
* been changed by the waker prior to clearing
* core_locked so we need to ensure we see
* changes to all those variables
*/
smp_rmb();
break;
}
if (time_after(jiffies, end)) {
printk("__acquire_sata_core() uid %p failing for port %d timed out, "
"locker_uid %p, h/w count %d, d count %d, callback %p, hw_access %d, "
"core_locked %d, reentrant_port_no %d, isr_callback %p, "
"isr_arg %p\n", uid, port_no, hd->locker_uid,
hd->hw_lock_count, hd->direct_lock_count, callback, hw_access,
hd->core_locked, hd->reentrant_port_no, hd->isr_callback,
hd->isr_arg);
printk(KERN_WARNING "__acquire_sata_core() "
"uid %p failing for port %d timed out, "
"locker_uid %p, h/w count %d, "
"d count %d, callback %p, hw_access %d, "
"core_locked %d, reentrant_port_no %d, "
"isr_callback %p, isr_arg %p\n", uid,
port_no, hd->locker_uid,
hd->hw_lock_count,
hd->direct_lock_count, callback,
hw_access, hd->core_locked,
hd->reentrant_port_no, hd->isr_callback,
hd->isr_arg);
timed_out = 1;
break;
}
spin_unlock_irqrestore(&hd->core_lock, flags);
if (!schedule_timeout(4*HZ)) {
printk(KERN_INFO "__acquire_sata_core() uid %p, locker_uid %p, "
"timed-out of schedule(), checking overall timeout\n",
printk(KERN_INFO "__acquire_sata_core() uid %p, "
"locker_uid %p, timed-out of "
"schedule(), checking overall timeout\n",
uid, hd->locker_uid);
}
spin_lock_irqsave(&hd->core_lock, flags);
}
finish_wait(hw_access ? &hd->scsi_wait_queue : &hd->fast_wait_queue, &wait);
finish_wait(hw_access ? &hd->scsi_wait_queue :
&hd->fast_wait_queue, &wait);
}
if (hw_access && acquired) {
if (hd->scsi_nonblocking_attempts) {
if (hd->scsi_nonblocking_attempts)
hd->scsi_nonblocking_attempts = 0;
}
// Wake any other SCSI/SATA waiters so they can get reentrant access to
// the same port if appropriate. This is because if the SATA core is
// locked by fast access, or SCSI/SATA access to other port, then can
// have >1 SCSI/SATA waiters on the wait list so want to give reentrant
// accessors a chance to get access ASAP
if (!list_empty(&hd->scsi_wait_queue.task_list)) {
/* Wake any other SCSI/SATA waiters so they can get reentrant
* access to the same port if appropriate. This is because if
* the SATA core is locked by fast access, or SCSI/SATA access
* to other port, then can have >1 SCSI/SATA waiters on the wait
* list so want to give reentrant accessors a chance to get
* access ASAP
*/
if (!list_empty(&hd->scsi_wait_queue.task_list))
wake_up(&hd->scsi_wait_queue);
}
}
DPRINTK("Leaving uid %p with acquired = %d, port %d, callback %p\n", uid, acquired, port_no, callback);
DPRINTK("Leaving uid %p with acquired = %d, port %d, callback %p\n",
uid, acquired, port_no, callback);
spin_unlock_irqrestore(&hd->core_lock, flags);
@ -838,7 +881,8 @@ int sata_core_has_scsi_waiters(struct ata_host *ah)
struct sata_oxnas_host_priv *hd = ah->private_data;
spin_lock_irqsave(&hd->core_lock, flags);
has_waiters = hd->scsi_nonblocking_attempts || !list_empty(&hd->scsi_wait_queue.task_list);
has_waiters = hd->scsi_nonblocking_attempts ||
!list_empty(&hd->scsi_wait_queue.task_list);
spin_unlock_irqrestore(&hd->core_lock, flags);
return has_waiters;
@ -856,7 +900,9 @@ static int sata_oxnas_acquire_hw(
int may_sleep,
int timeout_jiffies)
{
return __acquire_sata_core(ap->host, ap->port_no, NULL, 0, may_sleep, timeout_jiffies, 1, (void*)HW_LOCKER_UID, SATA_SCSI_STACK);
return __acquire_sata_core(ap->host, ap->port_no, NULL, 0, may_sleep,
timeout_jiffies, 1, (void *)HW_LOCKER_UID,
SATA_SCSI_STACK);
}
/*
@ -870,17 +916,20 @@ static void sata_oxnas_release_hw(struct ata_port *ap)
spin_lock_irqsave(&hd->core_lock, flags);
DPRINTK("Entered port_no = %d, h/w count %d, d count %d, core locked = %d, "
"reentrant_port_no = %d, isr_callback %p\n", ap->port_no,
hd->hw_lock_count, hd->direct_lock_count, hd->core_locked, hd->reentrant_port_no, hd->isr_callback);
DPRINTK("Entered port_no = %d, h/w count %d, d count %d, "
"core locked = %d, reentrant_port_no = %d, isr_callback %p\n",
ap->port_no, hd->hw_lock_count, hd->direct_lock_count,
hd->core_locked, hd->reentrant_port_no, hd->isr_callback);
if (!hd->core_locked) {
/* Nobody holds the SATA lock */
printk(KERN_WARNING "Nobody holds SATA lock, port_no %d\n", ap->port_no);
printk(KERN_WARNING "Nobody holds SATA lock, port_no %d\n",
ap->port_no);
released = 1;
} else if (!hd->hw_lock_count) {
/* SCSI/SATA has released without holding the lock */
printk(KERN_WARNING "SCSI/SATA does not hold SATA lock, port_no %d\n", ap->port_no);
printk(KERN_WARNING "SCSI/SATA does not hold SATA lock, "
"port_no %d\n", ap->port_no);
} else {
/* Trap incorrect usage */
BUG_ON(hd->reentrant_port_no == -1);
@ -888,10 +937,11 @@ static void sata_oxnas_release_hw(struct ata_port *ap)
BUG_ON(hd->direct_lock_count);
BUG_ON(hd->current_locker_type != SATA_SCSI_STACK);
WARN(!hd->locker_uid || (hd->locker_uid != HW_LOCKER_UID), "Invalid locker "
"uid %p, h/w count %d, d count %d, reentrant_port_no %d, "
"core_locked %d, isr_callback %p\n", hd->locker_uid,
hd->hw_lock_count, hd->direct_lock_count, hd->reentrant_port_no,
WARN(!hd->locker_uid || (hd->locker_uid != HW_LOCKER_UID),
"Invalid locker uid %p, h/w count %d, d count %d, "
"reentrant_port_no %d, core_locked %d, "
"isr_callback %p\n", hd->locker_uid, hd->hw_lock_count,
hd->direct_lock_count, hd->reentrant_port_no,
hd->core_locked, hd->isr_callback);
if (--(hd->hw_lock_count)) {
@ -904,11 +954,14 @@ static void sata_oxnas_release_hw(struct ata_port *ap)
hd->locker_uid = 0;
hd->core_locked = 0;
released = 1;
wake_up(!list_empty(&hd->scsi_wait_queue.task_list) ? &hd->scsi_wait_queue : &hd->fast_wait_queue);
wake_up(!list_empty(&hd->scsi_wait_queue.task_list) ?
&hd->scsi_wait_queue :
&hd->fast_wait_queue);
}
}
DPRINTK("Leaving, port_no %d, count %d\n", ap->port_no, hd->hw_lock_count);
DPRINTK("Leaving, port_no %d, count %d\n", ap->port_no,
hd->hw_lock_count);
spin_unlock_irqrestore(&hd->core_lock, flags);
@ -933,7 +986,8 @@ static inline u32 sata_oxnas_hostportbusy(struct ata_port *ap)
return (ioread32(hd->port_base + SATA_COMMAND) & CMD_CORE_BUSY) ||
(hd->n_ports > 1 &&
(ioread32(hd->port_base + PORT_SIZE + SATA_COMMAND) & CMD_CORE_BUSY));
(ioread32(hd->port_base + PORT_SIZE + SATA_COMMAND) &
CMD_CORE_BUSY));
}
static inline u32 sata_oxnas_hostdmabusy(struct ata_port *ap)
@ -970,10 +1024,10 @@ static void sata_oxnas_reset_core(struct ata_host *ah)
udelay(50);
workaround5458(ah);
/* tune for sata compatability */
/* tune for sata compatibility */
sata_oxnas_link_write(ah->ports[0], 0x60, 0x2988);
for (n=0; n < host_priv->n_ports; n++) {
for (n = 0; n < host_priv->n_ports; n++) {
/* each port in turn */
sata_oxnas_link_write(ah->ports[n], 0x70, 0x55629);
}
@ -1387,6 +1441,7 @@ static inline void sata_oxnas_clear_CS_error(struct ata_port *ap)
static inline void sata_oxnas_reset_sgdma(struct ata_port *ap)
{
struct sata_oxnas_port_priv *pd = ap->private_data;
iowrite32(SGDMA_RESETS_CTRL, pd->sgdma_base + SGDMA_RESETS);
}
@ -1418,12 +1473,13 @@ static inline void sata_oxnas_clear_reg_access_error(struct ata_port *ap)
DPRINTK("ENTER\n");
if (reg & INT_REG_ACCESS_ERR) {
printk(KERN_INFO "clearing register access error on port %d\n", ap->port_no);
DPRINTK("clearing register access error on port %d\n",
ap->port_no);
iowrite32(INT_REG_ACCESS_ERR, base + INT_STATUS);
}
reg = ioread32(base + INT_STATUS);
if (reg & INT_REG_ACCESS_ERR)
printk(KERN_INFO "register access error didn't clear\n");
DPRINTK("register access error didn't clear\n");
}
static inline void sata_oxnas_clear_sctl_error(struct ata_port *ap)
@ -1451,7 +1507,8 @@ static int sata_oxnas_cleanup(struct ata_host *ah)
struct sata_oxnas_host_priv *hd = ah->private_data;
int actions_required = 0;
int n;
printk(KERN_INFO "sata_oxnas: reseting SATA core\n");
printk(KERN_INFO "sata_oxnas: resetting SATA core\n");
/* core not recovering, reset it */
mdelay(5);
sata_oxnas_reset_core(ah);
@ -1460,7 +1517,7 @@ static int sata_oxnas_cleanup(struct ata_host *ah)
/* Perform any SATA core re-initialisation after reset post reset init
* needs to be called for both ports as there's one reset for both
* ports */
for (n=0; n < hd->n_ports; n++)
for (n = 0; n < hd->n_ports; n++)
sata_oxnas_post_reset_init(ah->ports[n]);
@ -1477,6 +1534,7 @@ static int sata_oxnas_cleanup(struct ata_host *ah)
static int sata_oxnas_qc_new(struct ata_port *ap)
{
struct sata_oxnas_host_priv *hd = ap->host->private_data;
DPRINTK("port %d\n", ap->port_no);
smp_rmb();
if (hd->port_frozen || hd->port_in_eh)
@ -1494,17 +1552,19 @@ static void sata_oxnas_qc_free(struct ata_queued_cmd *qc)
sata_oxnas_release_hw(qc->ap);
}
static void sata_oxnas_freeze(struct ata_port* ap)
static void sata_oxnas_freeze(struct ata_port *ap)
{
struct sata_oxnas_host_priv *hd = ap->host->private_data;
DPRINTK("\n");
hd->port_frozen |= BIT(ap->port_no);
smp_wmb();
}
static void sata_oxnas_thaw(struct ata_port* ap)
static void sata_oxnas_thaw(struct ata_port *ap)
{
struct sata_oxnas_host_priv *hd = ap->host->private_data;
DPRINTK("\n");
hd->port_frozen &= ~BIT(ap->port_no);
smp_wmb();
@ -1688,13 +1748,15 @@ static u8 sata_oxnas_check_status(struct ata_port *ap)
return status;
}
static inline void sata_oxnas_reset_ucode(struct ata_host *ah, int force, int no_microcode)
static inline void sata_oxnas_reset_ucode(struct ata_host *ah, int force,
int no_microcode)
{
struct sata_oxnas_host_priv *hd = ah->private_data;
DPRINTK("ENTER\n");
if (no_microcode) {
u32 reg;
sata_oxnas_set_mode(ah, UNKNOWN_MODE, force);
reg = ioread32(hd->core_base + DEVICE_CONTROL);
reg |= DEVICE_CONTROL_ATA_ERR_OVERRIDE;
@ -1719,7 +1781,8 @@ static void sata_oxnas_qc_prep(struct ata_queued_cmd *qc)
/* if the port's not connected, complete now with an error */
if (!sata_oxnas_check_link(qc->ap)) {
printk(KERN_ERR "port %d not connected completing with error\n",
ata_port_err(qc->ap,
"port %d not connected completing with error\n",
port_no);
qc->err_mask |= AC_ERR_ATA_BUS;
ata_qc_complete(qc);
@ -2006,12 +2069,12 @@ static int sata_oxnas_bug_6320_workaround(struct ata_port *ap)
bug_present = 1;
} else if (sector_quads_remaining) {
if (is_read) {
printk(KERN_WARNING "SATA read fixup cannot deal with" \
" %d quads remaining\n",
ata_port_warn(ap, "SATA read fixup cannot deal with "
"%d quads remaining\n",
sector_quads_remaining);
} else {
printk(KERN_WARNING "SATA write fixup of %d quads" \
" remaining not supported\n",
ata_port_warn(ap, "SATA write fixup of %d quads "
"remaining not supported\n",
sector_quads_remaining);
}
}
@ -2029,15 +2092,16 @@ static void sata_oxnas_port_irq(struct ata_port *ap, int force_error)
u32 int_status;
unsigned long flags = 0;
DPRINTK("ENTER port %d irqstatus %x\n", ap->port_no, ioread32(port_base + INT_STATUS));
DPRINTK("ENTER port %d irqstatus %x\n", ap->port_no,
ioread32(port_base + INT_STATUS));
if (ap->qc_active & (1 << ATA_TAG_INTERNAL)) {
qc = ata_qc_from_tag(ap, ATA_TAG_INTERNAL);
DPRINTK("completing non-ncq cmd\n");
if (qc) {
if (qc)
ata_qc_complete(qc);
}
return;
}
@ -2064,7 +2128,7 @@ static void sata_oxnas_port_irq(struct ata_port *ap, int force_error)
local_irq_restore(flags);
ata_qc_complete(qc);
} else {
VPRINTK("Ignoring interrupt, can't find the command tag=" \
VPRINTK("Ignoring interrupt, can't find the command tag="
"%d %08x\n", ap->link.active_tag, ap->qc_active);
}
@ -2098,14 +2162,14 @@ static irqreturn_t sata_oxnas_interrupt(int irq, void *dev_instance)
/* loop until there are no more interrupts */
while ((int_status = (ioread32(core_base + CORE_INT_STATUS)) &
(COREINT_END | (COREINT_END << 1)) )) {
(COREINT_END | (COREINT_END << 1)))) {
/* clear any interrupt */
iowrite32(int_status, core_base + CORE_INT_CLEAR);
/* Only need workaround_bug_6320 for single disk systems as dual
* disk will use uCode which prevents this read underrun problem
* from occuring.
* from occurring.
* All single disk systems will use port 0 */
for (port_no = 0; port_no < hd->n_ports; ++port_no) {
/* check the raw end of command interrupt to see if the
@ -2114,9 +2178,10 @@ static irqreturn_t sata_oxnas_interrupt(int irq, void *dev_instance)
if (int_status & mask) {
/* this port had an interrupt, clear it */
iowrite32(mask, core_base + CORE_INT_CLEAR);
bug_present = ( hd->current_ucode == UNKNOWN_MODE ) ?
sata_oxnas_bug_6320_workaround(
ah->ports[port_no]) : 0;
bug_present =
(hd->current_ucode == UNKNOWN_MODE) ?
sata_oxnas_bug_6320_workaround(
ah->ports[port_no]):0;
sata_oxnas_port_irq(ah->ports[port_no],
bug_present);
ret = IRQ_HANDLED;
@ -2193,7 +2258,6 @@ static int sata_oxnas_probe(struct platform_device *ofdev)
struct clk *clk = NULL;
const struct ata_port_info *ppi[] = { &sata_oxnas_port_info, NULL };
const struct ata_port_info *dppi[] = { &sata_oxnas_port_info, &sata_oxnas_port_info, NULL };
of_property_read_u32(ofdev->dev.of_node, "nr-ports", &n_ports);
if (n_ports < 1 || n_ports > SATA_OXNAS_MAX_PORTS)
@ -2275,8 +2339,7 @@ static int sata_oxnas_probe(struct platform_device *ofdev)
host_priv->rst_phy = rstc;
/* allocate host structure */
host = ata_host_alloc_pinfo(&ofdev->dev, n_ports>1 ? dppi : ppi,
n_ports);
host = ata_host_alloc_pinfo(&ofdev->dev, ppi, n_ports);
if (!host) {
retval = -ENOMEM;
@ -2410,4 +2473,4 @@ module_platform_driver(oxnas_sata_driver);
MODULE_LICENSE("GPL");
MODULE_VERSION("1.0");
MODULE_AUTHOR("Oxford Semiconductor Ltd.");
MODULE_DESCRIPTION("934 SATA core controler");
MODULE_DESCRIPTION("low-level driver for Oxford 934 SATA core");