*/
#include <linux/kernel.h>
+#include <linux/pci.h>
#include <scsi/scsi.h>
#include <scsi/scsi_host.h>
#include <scsi/scsi_eh.h>
#include "libata.h"
enum {
+ /* speed down verdicts */
ATA_EH_SPDN_NCQ_OFF = (1 << 0),
ATA_EH_SPDN_SPEED_DOWN = (1 << 1),
ATA_EH_SPDN_FALLBACK_TO_PIO = (1 << 2),
-};
-
-/* Waiting in ->prereset can never be reliable. It's sometimes nice
- * to wait there but it can't be depended upon; otherwise, we wouldn't
- * be resetting. Just give it enough time for most drives to spin up.
- */
-enum {
- ATA_EH_PRERESET_TIMEOUT = 10 * HZ,
- ATA_EH_FASTDRAIN_INTERVAL = 3 * HZ,
+ ATA_EH_SPDN_KEEP_ERRORS = (1 << 3),
+
+ /* error flags */
+ ATA_EFLAG_IS_IO = (1 << 0),
+ ATA_EFLAG_DUBIOUS_XFER = (1 << 1),
+
+ /* error categories */
+ ATA_ECAT_NONE = 0,
+ ATA_ECAT_ATA_BUS = 1,
+ ATA_ECAT_TOUT_HSM = 2,
+ ATA_ECAT_UNK_DEV = 3,
+ ATA_ECAT_DUBIOUS_NONE = 4,
+ ATA_ECAT_DUBIOUS_ATA_BUS = 5,
+ ATA_ECAT_DUBIOUS_TOUT_HSM = 6,
+ ATA_ECAT_DUBIOUS_UNK_DEV = 7,
+ ATA_ECAT_NR = 8,
+
+ ATA_EH_CMD_DFL_TIMEOUT = 5000,
+
+ /* always put at least this amount of time between resets */
+ ATA_EH_RESET_COOL_DOWN = 5000,
+
+ /* Waiting in ->prereset can never be reliable. It's
+ * sometimes nice to wait there but it can't be depended upon;
+ * otherwise, we wouldn't be resetting. Just give it enough
+ * time for most drives to spin up.
+ */
+ ATA_EH_PRERESET_TIMEOUT = 10000,
+ ATA_EH_FASTDRAIN_INTERVAL = 3000,
};
/* The following table determines how we sequence resets. Each entry
* are mostly for error handling, hotplug and retarded devices.
*/
static const unsigned long ata_eh_reset_timeouts[] = {
- 10 * HZ, /* most drives spin up by 10sec */
- 10 * HZ, /* > 99% working drives spin up before 20sec */
- 35 * HZ, /* give > 30 secs of idleness for retarded devices */
- 5 * HZ, /* and sweet one last chance */
- /* > 1 min has elapsed, give up */
+ 10000, /* most drives spin up by 10sec */
+ 10000, /* > 99% working drives spin up before 20sec */
+ 35000, /* give > 30 secs of idleness for retarded devices */
+ 5000, /* and sweet one last chance */
+ ULONG_MAX, /* > 1 min has elapsed, give up */
+};
+
+static const unsigned long ata_eh_identify_timeouts[] = {
+ 5000, /* covers > 99% of successes and not too boring on failures */
+ 10000, /* combined time till here is enough even for media access */
+ 30000, /* for true idiots */
+ ULONG_MAX,
+};
+
+static const unsigned long ata_eh_other_timeouts[] = {
+ 5000, /* same rationale as identify timeout */
+ 10000, /* ditto */
+ /* but no merciful 30sec for other commands, it just isn't worth it */
+ ULONG_MAX,
};
+struct ata_eh_cmd_timeout_ent {
+ const u8 *commands;
+ const unsigned long *timeouts;
+};
+
+/* The following table determines timeouts to use for EH internal
+ * commands. Each table entry is a command class and matches the
+ * commands the entry applies to and the timeout table to use.
+ *
+ * On the retry after a command timed out, the next timeout value from
+ * the table is used. If the table doesn't contain further entries,
+ * the last value is used.
+ *
+ * ehc->cmd_timeout_idx keeps track of which timeout to use per
+ * command class, so if SET_FEATURES times out on the first try, the
+ * next try will use the second timeout value only for that class.
+ */
+#define CMDS(cmds...) (const u8 []){ cmds, 0 }
+static const struct ata_eh_cmd_timeout_ent
+ata_eh_cmd_timeout_table[ATA_EH_CMD_TIMEOUT_TABLE_SIZE] = {
+ { .commands = CMDS(ATA_CMD_ID_ATA, ATA_CMD_ID_ATAPI),
+ .timeouts = ata_eh_identify_timeouts, },
+ { .commands = CMDS(ATA_CMD_READ_NATIVE_MAX, ATA_CMD_READ_NATIVE_MAX_EXT),
+ .timeouts = ata_eh_other_timeouts, },
+ { .commands = CMDS(ATA_CMD_SET_MAX, ATA_CMD_SET_MAX_EXT),
+ .timeouts = ata_eh_other_timeouts, },
+ { .commands = CMDS(ATA_CMD_SET_FEATURES),
+ .timeouts = ata_eh_other_timeouts, },
+ { .commands = CMDS(ATA_CMD_INIT_DEV_PARAMS),
+ .timeouts = ata_eh_other_timeouts, },
+};
+#undef CMDS
+
static void __ata_port_freeze(struct ata_port *ap);
#ifdef CONFIG_PM
static void ata_eh_handle_port_suspend(struct ata_port *ap);
if (offset < 0)
ata_port_desc(ap, "%s %s%llu@0x%llx", name, type, len, start);
else
- ata_port_desc(ap, "%s 0x%llx", name, start + offset);
+ ata_port_desc(ap, "%s 0x%llx", name,
+ start + (unsigned long long)offset);
}
#endif /* CONFIG_PCI */
-static void ata_ering_record(struct ata_ering *ering, int is_io,
+static int ata_lookup_timeout_table(u8 cmd)
+{
+ int i;
+
+ for (i = 0; i < ATA_EH_CMD_TIMEOUT_TABLE_SIZE; i++) {
+ const u8 *cur;
+
+ for (cur = ata_eh_cmd_timeout_table[i].commands; *cur; cur++)
+ if (*cur == cmd)
+ return i;
+ }
+
+ return -1;
+}
+
+/**
+ * ata_internal_cmd_timeout - determine timeout for an internal command
+ * @dev: target device
+ * @cmd: internal command to be issued
+ *
+ * Determine timeout for internal command @cmd for @dev.
+ *
+ * LOCKING:
+ * EH context.
+ *
+ * RETURNS:
+ * Determined timeout.
+ */
+unsigned long ata_internal_cmd_timeout(struct ata_device *dev, u8 cmd)
+{
+ struct ata_eh_context *ehc = &dev->link->eh_context;
+ int ent = ata_lookup_timeout_table(cmd);
+ int idx;
+
+ if (ent < 0)
+ return ATA_EH_CMD_DFL_TIMEOUT;
+
+ idx = ehc->cmd_timeout_idx[dev->devno][ent];
+ return ata_eh_cmd_timeout_table[ent].timeouts[idx];
+}
+
+/**
+ * ata_internal_cmd_timed_out - notification for internal command timeout
+ * @dev: target device
+ * @cmd: internal command which timed out
+ *
+ * Notify EH that internal command @cmd for @dev timed out. This
+ * function should be called only for commands whose timeouts are
+ * determined using ata_internal_cmd_timeout().
+ *
+ * LOCKING:
+ * EH context.
+ */
+void ata_internal_cmd_timed_out(struct ata_device *dev, u8 cmd)
+{
+ struct ata_eh_context *ehc = &dev->link->eh_context;
+ int ent = ata_lookup_timeout_table(cmd);
+ int idx;
+
+ if (ent < 0)
+ return;
+
+ idx = ehc->cmd_timeout_idx[dev->devno][ent];
+ if (ata_eh_cmd_timeout_table[ent].timeouts[idx + 1] != ULONG_MAX)
+ ehc->cmd_timeout_idx[dev->devno][ent]++;
+}
+
+static void ata_ering_record(struct ata_ering *ering, unsigned int eflags,
unsigned int err_mask)
{
struct ata_ering_entry *ent;
ering->cursor %= ATA_ERING_SIZE;
ent = &ering->ring[ering->cursor];
- ent->is_io = is_io;
+ ent->eflags = eflags;
ent->err_mask = err_mask;
ent->timestamp = get_jiffies_64();
}
+static struct ata_ering_entry *ata_ering_top(struct ata_ering *ering)
+{
+ struct ata_ering_entry *ent = &ering->ring[ering->cursor];
+
+ if (ent->err_mask)
+ return ent;
+ return NULL;
+}
+
static void ata_ering_clear(struct ata_ering *ering)
{
memset(ering, 0, sizeof(*ering));
spin_lock_irqsave(ap->lock, flags);
__ata_port_for_each_link(link, ap) {
+ struct ata_eh_context *ehc = &link->eh_context;
+ struct ata_device *dev;
+
memset(&link->eh_context, 0, sizeof(link->eh_context));
link->eh_context.i = link->eh_info;
memset(&link->eh_info, 0, sizeof(link->eh_info));
+
+ ata_link_for_each_dev(dev, link) {
+ int devno = dev->devno;
+
+ ehc->saved_xfer_mode[devno] = dev->xfer_mode;
+ if (ata_ncq_enabled(dev))
+ ehc->saved_ncq_enabled |= 1 << devno;
+ }
+
+ /* set last reset timestamp to some time in the past */
+ ehc->last_reset = jiffies - 60 * HZ;
}
ap->pflags |= ATA_PFLAG_EH_IN_PROGRESS;
}
}
-/**
- * ata_qc_timeout - Handle timeout of queued command
- * @qc: Command that timed out
- *
- * Some part of the kernel (currently, only the SCSI layer)
- * has noticed that the active command on port @ap has not
- * completed after a specified length of time. Handle this
- * condition by disabling DMA (if necessary) and completing
- * transactions, with error if necessary.
- *
- * This also handles the case of the "lost interrupt", where
- * for some reason (possibly hardware bug, possibly driver bug)
- * an interrupt was not delivered to the driver, even though the
- * transaction completed successfully.
- *
- * TODO: kill this function once old EH is gone.
- *
- * LOCKING:
- * Inherited from SCSI layer (none, can sleep)
- */
-static void ata_qc_timeout(struct ata_queued_cmd *qc)
-{
- struct ata_port *ap = qc->ap;
- u8 host_stat = 0, drv_stat;
- unsigned long flags;
-
- DPRINTK("ENTER\n");
-
- ap->hsm_task_state = HSM_ST_IDLE;
-
- spin_lock_irqsave(ap->lock, flags);
-
- switch (qc->tf.protocol) {
-
- case ATA_PROT_DMA:
- case ATA_PROT_ATAPI_DMA:
- host_stat = ap->ops->bmdma_status(ap);
-
- /* before we do anything else, clear DMA-Start bit */
- ap->ops->bmdma_stop(qc);
-
- /* fall through */
-
- default:
- ata_altstatus(ap);
- drv_stat = ata_chk_status(ap);
-
- /* ack bmdma irq events */
- ap->ops->irq_clear(ap);
-
- ata_dev_printk(qc->dev, KERN_ERR, "command 0x%x timeout, "
- "stat 0x%x host_stat 0x%x\n",
- qc->tf.command, drv_stat, host_stat);
-
- /* complete taskfile transaction */
- qc->err_mask |= AC_ERR_TIMEOUT;
- break;
- }
-
- spin_unlock_irqrestore(ap->lock, flags);
-
- ata_eh_qc_complete(qc);
-
- DPRINTK("EXIT\n");
-}
-
-/**
- * ata_eng_timeout - Handle timeout of queued command
- * @ap: Port on which timed-out command is active
- *
- * Some part of the kernel (currently, only the SCSI layer)
- * has noticed that the active command on port @ap has not
- * completed after a specified length of time. Handle this
- * condition by disabling DMA (if necessary) and completing
- * transactions, with error if necessary.
- *
- * This also handles the case of the "lost interrupt", where
- * for some reason (possibly hardware bug, possibly driver bug)
- * an interrupt was not delivered to the driver, even though the
- * transaction completed successfully.
- *
- * TODO: kill this function once old EH is gone.
- *
- * LOCKING:
- * Inherited from SCSI layer (none, can sleep)
- */
-void ata_eng_timeout(struct ata_port *ap)
-{
- DPRINTK("ENTER\n");
-
- ata_qc_timeout(ata_qc_from_tag(ap, ap->link.active_tag));
-
- DPRINTK("EXIT\n");
-}
-
static int ata_eh_nr_in_flight(struct ata_port *ap)
{
unsigned int tag;
/* some qcs have finished, give it another chance */
ap->fastdrain_cnt = cnt;
ap->fastdrain_timer.expires =
- jiffies + ATA_EH_FASTDRAIN_INTERVAL;
+ ata_deadline(jiffies, ATA_EH_FASTDRAIN_INTERVAL);
add_timer(&ap->fastdrain_timer);
}
/* activate fast drain */
ap->fastdrain_cnt = cnt;
- ap->fastdrain_timer.expires = jiffies + ATA_EH_FASTDRAIN_INTERVAL;
+ ap->fastdrain_timer.expires =
+ ata_deadline(jiffies, ATA_EH_FASTDRAIN_INTERVAL);
add_timer(&ap->fastdrain_timer);
}
if (rc == 0)
sata_scr_write(&ap->link, SCR_NOTIFICATION, sntf);
- if (!ap->nr_pmp_links || rc) {
+ if (!sata_pmp_attached(ap) || rc) {
/* PMP is not attached or SNTF is not available */
- if (!ap->nr_pmp_links) {
+ if (!sata_pmp_attached(ap)) {
/* PMP is not attached. Check whether ATAPI
* AN is configured. If so, notify media
* change.
spin_lock_irqsave(ap->lock, flags);
- /* Reset is represented by combination of actions and EHI
- * flags. Suck in all related bits before clearing eh_info to
- * avoid losing requested action.
- */
- if (action & ATA_EH_RESET_MASK) {
- ehc->i.action |= ehi->action & ATA_EH_RESET_MASK;
- ehc->i.flags |= ehi->flags & ATA_EHI_RESET_MODIFIER_MASK;
-
- /* make sure all reset actions are cleared & clear EHI flags */
- action |= ATA_EH_RESET_MASK;
- ehi->flags &= ~ATA_EHI_RESET_MODIFIER_MASK;
- }
-
ata_eh_clear_action(link, dev, ehi, action);
if (!(ehc->i.flags & ATA_EHI_QUIET))
{
struct ata_eh_context *ehc = &link->eh_context;
- /* if reset is complete, clear all reset actions & reset modifier */
- if (action & ATA_EH_RESET_MASK) {
- action |= ATA_EH_RESET_MASK;
- ehc->i.flags &= ~ATA_EHI_RESET_MODIFIER_MASK;
- }
-
ata_eh_clear_action(link, dev, &ehc->i, action);
}
* RETURNS:
* Descriptive string for @err_mask
*/
-static const char * ata_err_string(unsigned int err_mask)
+static const char *ata_err_string(unsigned int err_mask)
{
if (err_mask & AC_ERR_HOST_BUS)
return "host bus error";
* atapi_eh_request_sense - perform ATAPI REQUEST_SENSE
* @dev: device to perform REQUEST_SENSE to
* @sense_buf: result sense data buffer (SCSI_SENSE_BUFFERSIZE bytes long)
+ * @dfl_sense_key: default sense key to use
*
* Perform ATAPI REQUEST_SENSE after the device reported CHECK
* SENSE. This function is EH helper.
* RETURNS:
* 0 on success, AC_ERR_* mask on failure
*/
-static unsigned int atapi_eh_request_sense(struct ata_queued_cmd *qc)
+static unsigned int atapi_eh_request_sense(struct ata_device *dev,
+ u8 *sense_buf, u8 dfl_sense_key)
{
- struct ata_device *dev = qc->dev;
- unsigned char *sense_buf = qc->scsicmd->sense_buffer;
+ u8 cdb[ATAPI_CDB_LEN] =
+ { REQUEST_SENSE, 0, 0, 0, SCSI_SENSE_BUFFERSIZE, 0 };
struct ata_port *ap = dev->link->ap;
struct ata_taskfile tf;
- u8 cdb[ATAPI_CDB_LEN];
DPRINTK("ATAPI request sense\n");
* for the case where they are -not- overwritten
*/
sense_buf[0] = 0x70;
- sense_buf[2] = qc->result_tf.feature >> 4;
+ sense_buf[2] = dfl_sense_key;
/* some devices time out if garbage left in tf */
ata_tf_init(dev, &tf);
- memset(cdb, 0, ATAPI_CDB_LEN);
- cdb[0] = REQUEST_SENSE;
- cdb[4] = SCSI_SENSE_BUFFERSIZE;
-
tf.flags |= ATA_TFLAG_ISADDR | ATA_TFLAG_DEVICE;
tf.command = ATA_CMD_PACKET;
/* is it pointless to prefer PIO for "safety reasons"? */
if (ap->flags & ATA_FLAG_PIO_DMA) {
- tf.protocol = ATA_PROT_ATAPI_DMA;
+ tf.protocol = ATAPI_PROT_DMA;
tf.feature |= ATAPI_PKT_DMA;
} else {
- tf.protocol = ATA_PROT_ATAPI;
- tf.lbam = (8 * 1024) & 0xff;
- tf.lbah = (8 * 1024) >> 8;
+ tf.protocol = ATAPI_PROT_PIO;
+ tf.lbam = SCSI_SENSE_BUFFERSIZE;
+ tf.lbah = 0;
}
return ata_exec_internal(dev, &tf, cdb, DMA_FROM_DEVICE,
unsigned int err_mask = 0, action = 0;
u32 hotplug_mask;
- if (serror & SERR_PERSISTENT) {
- err_mask |= AC_ERR_ATA_BUS;
- action |= ATA_EH_HARDRESET;
- }
- if (serror &
- (SERR_DATA_RECOVERED | SERR_COMM_RECOVERED | SERR_DATA)) {
+ if (serror & (SERR_PERSISTENT | SERR_DATA)) {
err_mask |= AC_ERR_ATA_BUS;
- action |= ATA_EH_SOFTRESET;
+ action |= ATA_EH_RESET;
}
if (serror & SERR_PROTOCOL) {
err_mask |= AC_ERR_HSM;
- action |= ATA_EH_SOFTRESET;
+ action |= ATA_EH_RESET;
}
if (serror & SERR_INTERNAL) {
err_mask |= AC_ERR_SYSTEM;
- action |= ATA_EH_HARDRESET;
+ action |= ATA_EH_RESET;
}
/* Determine whether a hotplug event has occurred. Both
* LOCKING:
* Kernel thread context (may sleep).
*/
-static void ata_eh_analyze_ncq_error(struct ata_link *link)
+void ata_eh_analyze_ncq_error(struct ata_link *link)
{
struct ata_port *ap = link->ap;
struct ata_eh_context *ehc = &link->eh_context;
/* we've got the perpetrator, condemn it */
qc = __ata_qc_from_tag(ap, tag);
memcpy(&qc->result_tf, &tf, sizeof(tf));
+ qc->result_tf.flags = ATA_TFLAG_ISADDR | ATA_TFLAG_LBA | ATA_TFLAG_LBA48;
qc->err_mask |= AC_ERR_DEV | AC_ERR_NCQ;
ehc->i.err_mask &= ~AC_ERR_DEV;
}
if ((stat & (ATA_BUSY | ATA_DRQ | ATA_DRDY)) != ATA_DRDY) {
qc->err_mask |= AC_ERR_HSM;
- return ATA_EH_SOFTRESET;
+ return ATA_EH_RESET;
}
if (stat & (ATA_ERR | ATA_DF))
case ATA_DEV_ATAPI:
if (!(qc->ap->pflags & ATA_PFLAG_FROZEN)) {
- tmp = atapi_eh_request_sense(qc);
+ tmp = atapi_eh_request_sense(qc->dev,
+ qc->scsicmd->sense_buffer,
+ qc->result_tf.feature >> 4);
if (!tmp) {
/* ATA_QCFLAG_SENSE_VALID is used to
* tell atapi_qc_complete() that sense
}
if (qc->err_mask & (AC_ERR_HSM | AC_ERR_TIMEOUT | AC_ERR_ATA_BUS))
- action |= ATA_EH_SOFTRESET;
+ action |= ATA_EH_RESET;
return action;
}
-static int ata_eh_categorize_error(int is_io, unsigned int err_mask)
+static int ata_eh_categorize_error(unsigned int eflags, unsigned int err_mask,
+ int *xfer_ok)
{
+ int base = 0;
+
+ if (!(eflags & ATA_EFLAG_DUBIOUS_XFER))
+ *xfer_ok = 1;
+
+ if (!*xfer_ok)
+ base = ATA_ECAT_DUBIOUS_NONE;
+
if (err_mask & AC_ERR_ATA_BUS)
- return 1;
+ return base + ATA_ECAT_ATA_BUS;
if (err_mask & AC_ERR_TIMEOUT)
- return 2;
+ return base + ATA_ECAT_TOUT_HSM;
- if (is_io) {
+ if (eflags & ATA_EFLAG_IS_IO) {
if (err_mask & AC_ERR_HSM)
- return 2;
+ return base + ATA_ECAT_TOUT_HSM;
if ((err_mask &
(AC_ERR_DEV|AC_ERR_MEDIA|AC_ERR_INVALID)) == AC_ERR_DEV)
- return 3;
+ return base + ATA_ECAT_UNK_DEV;
}
return 0;
struct speed_down_verdict_arg {
u64 since;
- int nr_errors[4];
+ int xfer_ok;
+ int nr_errors[ATA_ECAT_NR];
};
static int speed_down_verdict_cb(struct ata_ering_entry *ent, void *void_arg)
{
struct speed_down_verdict_arg *arg = void_arg;
- int cat = ata_eh_categorize_error(ent->is_io, ent->err_mask);
+ int cat;
if (ent->timestamp < arg->since)
return -1;
+ cat = ata_eh_categorize_error(ent->eflags, ent->err_mask,
+ &arg->xfer_ok);
arg->nr_errors[cat]++;
+
return 0;
}
* whether NCQ needs to be turned off, transfer speed should be
* stepped down, or falling back to PIO is necessary.
*
- * Cat-1 is ATA_BUS error for any command.
+ * ECAT_ATA_BUS : ATA_BUS error for any command
+ *
+ * ECAT_TOUT_HSM : TIMEOUT for any command or HSM violation for
+ * IO commands
*
- * Cat-2 is TIMEOUT for any command or HSM violation for known
- * supported commands.
+ * ECAT_UNK_DEV : Unknown DEV error for IO commands
*
- * Cat-3 is is unclassified DEV error for known supported
- * command.
+ * ECAT_DUBIOUS_* : Identical to above three but occurred while
+ * data transfer hasn't been verified.
*
- * NCQ needs to be turned off if there have been more than 3
- * Cat-2 + Cat-3 errors during last 10 minutes.
+ * Verdicts are
*
- * Speed down is necessary if there have been more than 3 Cat-1 +
- * Cat-2 errors or 10 Cat-3 errors during last 10 minutes.
+ * NCQ_OFF : Turn off NCQ.
*
- * Falling back to PIO mode is necessary if there have been more
- * than 10 Cat-1 + Cat-2 + Cat-3 errors during last 5 minutes.
+ * SPEED_DOWN : Speed down transfer speed but don't fall back
+ * to PIO.
+ *
+ * FALLBACK_TO_PIO : Fall back to PIO.
+ *
+ * Even if multiple verdicts are returned, only one action is
+ * taken per error. An action triggered by non-DUBIOUS errors
+ * clears ering, while one triggered by DUBIOUS_* errors doesn't.
+ * This is to expedite speed down decisions right after device is
+ * initially configured.
+ *
+ * The followings are speed down rules. #1 and #2 deal with
+ * DUBIOUS errors.
+ *
+ * 1. If more than one DUBIOUS_ATA_BUS or DUBIOUS_TOUT_HSM errors
+ * occurred during last 5 mins, SPEED_DOWN and FALLBACK_TO_PIO.
+ *
+ * 2. If more than one DUBIOUS_TOUT_HSM or DUBIOUS_UNK_DEV errors
+ * occurred during last 5 mins, NCQ_OFF.
+ *
+ * 3. If more than 8 ATA_BUS, TOUT_HSM or UNK_DEV errors
+ * ocurred during last 5 mins, FALLBACK_TO_PIO
+ *
+ * 4. If more than 3 TOUT_HSM or UNK_DEV errors occurred
+ * during last 10 mins, NCQ_OFF.
+ *
+ * 5. If more than 3 ATA_BUS or TOUT_HSM errors, or more than 6
+ * UNK_DEV errors occurred during last 10 mins, SPEED_DOWN.
*
* LOCKING:
* Inherited from caller.
struct speed_down_verdict_arg arg;
unsigned int verdict = 0;
- /* scan past 10 mins of error history */
+ /* scan past 5 mins of error history */
memset(&arg, 0, sizeof(arg));
- arg.since = j64 - min(j64, j10mins);
+ arg.since = j64 - min(j64, j5mins);
ata_ering_map(&dev->ering, speed_down_verdict_cb, &arg);
- if (arg.nr_errors[2] + arg.nr_errors[3] > 3)
- verdict |= ATA_EH_SPDN_NCQ_OFF;
- if (arg.nr_errors[1] + arg.nr_errors[2] > 3 || arg.nr_errors[3] > 10)
- verdict |= ATA_EH_SPDN_SPEED_DOWN;
+ if (arg.nr_errors[ATA_ECAT_DUBIOUS_ATA_BUS] +
+ arg.nr_errors[ATA_ECAT_DUBIOUS_TOUT_HSM] > 1)
+ verdict |= ATA_EH_SPDN_SPEED_DOWN |
+ ATA_EH_SPDN_FALLBACK_TO_PIO | ATA_EH_SPDN_KEEP_ERRORS;
+
+ if (arg.nr_errors[ATA_ECAT_DUBIOUS_TOUT_HSM] +
+ arg.nr_errors[ATA_ECAT_DUBIOUS_UNK_DEV] > 1)
+ verdict |= ATA_EH_SPDN_NCQ_OFF | ATA_EH_SPDN_KEEP_ERRORS;
- /* scan past 3 mins of error history */
+ if (arg.nr_errors[ATA_ECAT_ATA_BUS] +
+ arg.nr_errors[ATA_ECAT_TOUT_HSM] +
+ arg.nr_errors[ATA_ECAT_UNK_DEV] > 6)
+ verdict |= ATA_EH_SPDN_FALLBACK_TO_PIO;
+
+ /* scan past 10 mins of error history */
memset(&arg, 0, sizeof(arg));
- arg.since = j64 - min(j64, j5mins);
+ arg.since = j64 - min(j64, j10mins);
ata_ering_map(&dev->ering, speed_down_verdict_cb, &arg);
- if (arg.nr_errors[1] + arg.nr_errors[2] + arg.nr_errors[3] > 10)
- verdict |= ATA_EH_SPDN_FALLBACK_TO_PIO;
+ if (arg.nr_errors[ATA_ECAT_TOUT_HSM] +
+ arg.nr_errors[ATA_ECAT_UNK_DEV] > 3)
+ verdict |= ATA_EH_SPDN_NCQ_OFF;
+
+ if (arg.nr_errors[ATA_ECAT_ATA_BUS] +
+ arg.nr_errors[ATA_ECAT_TOUT_HSM] > 3 ||
+ arg.nr_errors[ATA_ECAT_UNK_DEV] > 6)
+ verdict |= ATA_EH_SPDN_SPEED_DOWN;
return verdict;
}
/**
* ata_eh_speed_down - record error and speed down if necessary
* @dev: Failed device
- * @is_io: Did the device fail during normal IO?
+ * @eflags: mask of ATA_EFLAG_* flags
* @err_mask: err_mask of the error
*
* Record error and examine error history to determine whether
* RETURNS:
* Determined recovery action.
*/
-static unsigned int ata_eh_speed_down(struct ata_device *dev, int is_io,
- unsigned int err_mask)
+static unsigned int ata_eh_speed_down(struct ata_device *dev,
+ unsigned int eflags, unsigned int err_mask)
{
+ struct ata_link *link = dev->link;
+ int xfer_ok = 0;
unsigned int verdict;
unsigned int action = 0;
/* don't bother if Cat-0 error */
- if (ata_eh_categorize_error(is_io, err_mask) == 0)
+ if (ata_eh_categorize_error(eflags, err_mask, &xfer_ok) == 0)
return 0;
/* record error and determine whether speed down is necessary */
- ata_ering_record(&dev->ering, is_io, err_mask);
+ ata_ering_record(&dev->ering, eflags, err_mask);
verdict = ata_eh_speed_down_verdict(dev);
/* turn off NCQ? */
/* speed down? */
if (verdict & ATA_EH_SPDN_SPEED_DOWN) {
/* speed down SATA link speed if possible */
- if (sata_down_spd_limit(dev->link) == 0) {
- action |= ATA_EH_HARDRESET;
+ if (sata_down_spd_limit(link) == 0) {
+ action |= ATA_EH_RESET;
goto done;
}
dev->spdn_cnt++;
if (ata_down_xfermask_limit(dev, sel) == 0) {
- action |= ATA_EH_SOFTRESET;
+ action |= ATA_EH_RESET;
goto done;
}
}
}
/* Fall back to PIO? Slowing down to PIO is meaningless for
- * SATA. Consider it only for PATA.
+ * SATA ATA devices. Consider it only for PATA and SATAPI.
*/
if ((verdict & ATA_EH_SPDN_FALLBACK_TO_PIO) && (dev->spdn_cnt >= 2) &&
- (dev->link->ap->cbl != ATA_CBL_SATA) &&
+ (link->ap->cbl != ATA_CBL_SATA || dev->class == ATA_DEV_ATAPI) &&
(dev->xfer_shift != ATA_SHIFT_PIO)) {
if (ata_down_xfermask_limit(dev, ATA_DNXFER_FORCE_PIO) == 0) {
dev->spdn_cnt = 0;
- action |= ATA_EH_SOFTRESET;
+ action |= ATA_EH_RESET;
goto done;
}
}
return 0;
done:
/* device has been slowed down, blow error history */
- ata_ering_clear(&dev->ering);
+ if (!(verdict & ATA_EH_SPDN_KEEP_ERRORS))
+ ata_ering_clear(&dev->ering);
return action;
}
{
struct ata_port *ap = link->ap;
struct ata_eh_context *ehc = &link->eh_context;
- unsigned int all_err_mask = 0;
- int tag, is_io = 0;
+ struct ata_device *dev;
+ unsigned int all_err_mask = 0, eflags = 0;
+ int tag;
u32 serror;
int rc;
ehc->i.serror |= serror;
ata_eh_analyze_serror(link);
} else if (rc != -EOPNOTSUPP) {
- /* SError read failed, force hardreset and probing */
- ata_ehi_schedule_probe(&ehc->i);
- ehc->i.action |= ATA_EH_HARDRESET;
+ /* SError read failed, force reset and probing */
+ ehc->i.probe_mask |= ATA_ALL_DEVICES;
+ ehc->i.action |= ATA_EH_RESET;
ehc->i.err_mask |= AC_ERR_OTHER;
}
qc->err_mask &= ~AC_ERR_OTHER;
/* SENSE_VALID trumps dev/unknown error and revalidation */
- if (qc->flags & ATA_QCFLAG_SENSE_VALID) {
+ if (qc->flags & ATA_QCFLAG_SENSE_VALID)
qc->err_mask &= ~(AC_ERR_DEV | AC_ERR_OTHER);
- ehc->i.action &= ~ATA_EH_REVALIDATE;
- }
+
+ /* determine whether the command is worth retrying */
+ if (!(qc->err_mask & AC_ERR_INVALID) &&
+ ((qc->flags & ATA_QCFLAG_IO) || qc->err_mask != AC_ERR_DEV))
+ qc->flags |= ATA_QCFLAG_RETRY;
/* accumulate error info */
ehc->i.dev = qc->dev;
all_err_mask |= qc->err_mask;
if (qc->flags & ATA_QCFLAG_IO)
- is_io = 1;
+ eflags |= ATA_EFLAG_IS_IO;
}
/* enforce default EH actions */
if (ap->pflags & ATA_PFLAG_FROZEN ||
all_err_mask & (AC_ERR_HSM | AC_ERR_TIMEOUT))
- ehc->i.action |= ATA_EH_SOFTRESET;
- else if (all_err_mask)
+ ehc->i.action |= ATA_EH_RESET;
+ else if (((eflags & ATA_EFLAG_IS_IO) && all_err_mask) ||
+ (!(eflags & ATA_EFLAG_IS_IO) && (all_err_mask & ~AC_ERR_DEV)))
ehc->i.action |= ATA_EH_REVALIDATE;
- /* if we have offending qcs and the associated failed device */
+ /* If we have offending qcs and the associated failed device,
+ * perform per-dev EH action only on the offending device.
+ */
if (ehc->i.dev) {
- /* speed down */
- ehc->i.action |= ata_eh_speed_down(ehc->i.dev, is_io,
- all_err_mask);
-
- /* perform per-dev EH action only on the offending device */
ehc->i.dev_action[ehc->i.dev->devno] |=
ehc->i.action & ATA_EH_PERDEV_MASK;
ehc->i.action &= ~ATA_EH_PERDEV_MASK;
}
+ /* propagate timeout to host link */
+ if ((all_err_mask & AC_ERR_TIMEOUT) && !ata_is_host_link(link))
+ ap->link.eh_context.i.err_mask |= AC_ERR_TIMEOUT;
+
+ /* record error and consider speeding down */
+ dev = ehc->i.dev;
+ if (!dev && ((ata_link_max_devices(link) == 1 &&
+ ata_dev_enabled(link->device))))
+ dev = link->device;
+
+ if (dev) {
+ if (dev->flags & ATA_DFLAG_DUBIOUS_XFER)
+ eflags |= ATA_EFLAG_DUBIOUS_XFER;
+ ehc->i.action |= ata_eh_speed_down(dev, eflags, all_err_mask);
+ }
+
DPRINTK("EXIT\n");
}
{
struct ata_link *link;
- __ata_port_for_each_link(link, ap)
+ ata_port_for_each_link(link, ap)
ata_eh_link_autopsy(link);
+
+ /* Autopsy of fanout ports can affect host link autopsy.
+ * Perform host link autopsy last.
+ */
+ if (sata_pmp_attached(ap))
+ ata_eh_link_autopsy(&ap->link);
}
/**
for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
- if (!(qc->flags & ATA_QCFLAG_FAILED) || qc->dev->link != link)
+ if (!(qc->flags & ATA_QCFLAG_FAILED) || qc->dev->link != link ||
+ ((qc->flags & ATA_QCFLAG_QUIET) &&
+ qc->err_mask == AC_ERR_DEV))
continue;
if (qc->flags & ATA_QCFLAG_SENSE_VALID && !qc->err_mask)
continue;
}
if (ehc->i.serror)
- ata_port_printk(ap, KERN_ERR,
+ ata_link_printk(link, KERN_ERR,
"SError: { %s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s%s}\n",
ehc->i.serror & SERR_DATA_RECOVERED ? "RecovData " : "",
ehc->i.serror & SERR_COMM_RECOVERED ? "RecovComm " : "",
ehc->i.serror & SERR_LINK_SEQ_ERR ? "LinkSeq " : "",
ehc->i.serror & SERR_TRANS_ST_ERROR ? "TrStaTrns " : "",
ehc->i.serror & SERR_UNRECOG_FIS ? "UnrecFIS " : "",
- ehc->i.serror & SERR_DEV_XCHG ? "DevExch " : "" );
+ ehc->i.serror & SERR_DEV_XCHG ? "DevExch " : "");
for (tag = 0; tag < ATA_MAX_QUEUE; tag++) {
- static const char *dma_str[] = {
- [DMA_BIDIRECTIONAL] = "bidi",
- [DMA_TO_DEVICE] = "out",
- [DMA_FROM_DEVICE] = "in",
- [DMA_NONE] = "",
- };
struct ata_queued_cmd *qc = __ata_qc_from_tag(ap, tag);
struct ata_taskfile *cmd = &qc->tf, *res = &qc->result_tf;
+ const u8 *cdb = qc->cdb;
+ char data_buf[20] = "";
+ char cdb_buf[70] = "";
if (!(qc->flags & ATA_QCFLAG_FAILED) ||
qc->dev->link != link || !qc->err_mask)
continue;
+ if (qc->dma_dir != DMA_NONE) {
+ static const char *dma_str[] = {
+ [DMA_BIDIRECTIONAL] = "bidi",
+ [DMA_TO_DEVICE] = "out",
+ [DMA_FROM_DEVICE] = "in",
+ };
+ static const char *prot_str[] = {
+ [ATA_PROT_PIO] = "pio",
+ [ATA_PROT_DMA] = "dma",
+ [ATA_PROT_NCQ] = "ncq",
+ [ATAPI_PROT_PIO] = "pio",
+ [ATAPI_PROT_DMA] = "dma",
+ };
+
+ snprintf(data_buf, sizeof(data_buf), " %s %u %s",
+ prot_str[qc->tf.protocol], qc->nbytes,
+ dma_str[qc->dma_dir]);
+ }
+
+ if (ata_is_atapi(qc->tf.protocol))
+ snprintf(cdb_buf, sizeof(cdb_buf),
+ "cdb %02x %02x %02x %02x %02x %02x %02x %02x "
+ "%02x %02x %02x %02x %02x %02x %02x %02x\n ",
+ cdb[0], cdb[1], cdb[2], cdb[3],
+ cdb[4], cdb[5], cdb[6], cdb[7],
+ cdb[8], cdb[9], cdb[10], cdb[11],
+ cdb[12], cdb[13], cdb[14], cdb[15]);
+
ata_dev_printk(qc->dev, KERN_ERR,
"cmd %02x/%02x:%02x:%02x:%02x:%02x/%02x:%02x:%02x:%02x:%02x/%02x "
- "tag %d cdb 0x%x data %u %s\n "
+ "tag %d%s\n %s"
"res %02x/%02x:%02x:%02x:%02x:%02x/%02x:%02x:%02x:%02x:%02x/%02x "
"Emask 0x%x (%s)%s\n",
cmd->command, cmd->feature, cmd->nsect,
cmd->lbal, cmd->lbam, cmd->lbah,
cmd->hob_feature, cmd->hob_nsect,
cmd->hob_lbal, cmd->hob_lbam, cmd->hob_lbah,
- cmd->device, qc->tag, qc->cdb[0], qc->nbytes,
- dma_str[qc->dma_dir],
+ cmd->device, qc->tag, data_buf, cdb_buf,
res->command, res->feature, res->nsect,
res->lbal, res->lbam, res->lbah,
res->hob_feature, res->hob_nsect,
qc->err_mask & AC_ERR_NCQ ? " <F>" : "");
if (res->command & (ATA_BUSY | ATA_DRDY | ATA_DF | ATA_DRQ |
- ATA_ERR) ) {
+ ATA_ERR)) {
if (res->command & ATA_BUSY)
ata_dev_printk(qc->dev, KERN_ERR,
- "status: { Busy }\n" );
+ "status: { Busy }\n");
else
ata_dev_printk(qc->dev, KERN_ERR,
"status: { %s%s%s%s}\n",
res->command & ATA_DRDY ? "DRDY " : "",
res->command & ATA_DF ? "DF " : "",
res->command & ATA_DRQ ? "DRQ " : "",
- res->command & ATA_ERR ? "ERR " : "" );
+ res->command & ATA_ERR ? "ERR " : "");
}
if (cmd->command != ATA_CMD_PACKET &&
res->feature & ATA_ICRC ? "ICRC " : "",
res->feature & ATA_UNC ? "UNC " : "",
res->feature & ATA_IDNF ? "IDNF " : "",
- res->feature & ATA_ABORTED ? "ABRT " : "" );
+ res->feature & ATA_ABORTED ? "ABRT " : "");
}
}
unsigned int *classes, unsigned long deadline)
{
struct ata_device *dev;
- int rc;
ata_link_for_each_dev(dev, link)
classes[dev->devno] = ATA_DEV_UNKNOWN;
- rc = reset(link, classes, deadline);
- if (rc)
- return rc;
-
- /* If any class isn't ATA_DEV_UNKNOWN, consider classification
- * is complete and convert all ATA_DEV_UNKNOWN to
- * ATA_DEV_NONE.
- */
- ata_link_for_each_dev(dev, link)
- if (classes[dev->devno] != ATA_DEV_UNKNOWN)
- break;
-
- if (dev) {
- ata_link_for_each_dev(dev, link) {
- if (classes[dev->devno] == ATA_DEV_UNKNOWN)
- classes[dev->devno] = ATA_DEV_NONE;
- }
- }
-
- return 0;
+ return reset(link, classes, deadline);
}
static int ata_eh_followup_srst_needed(struct ata_link *link,
- int rc, int classify,
- const unsigned int *classes)
+ int rc, const unsigned int *classes)
{
- if (link->flags & ATA_LFLAG_NO_SRST)
+ if ((link->flags & ATA_LFLAG_NO_SRST) || ata_link_offline(link))
return 0;
if (rc == -EAGAIN)
return 1;
- if (rc != 0)
- return 0;
- if ((link->ap->flags & ATA_FLAG_PMP) && ata_is_host_link(link))
- return 1;
- if (classify && !(link->flags & ATA_LFLAG_ASSUME_CLASS) &&
- classes[0] == ATA_DEV_UNKNOWN)
+ if (sata_pmp_supported(link->ap) && ata_is_host_link(link))
return 1;
return 0;
}
struct ata_port *ap = link->ap;
struct ata_eh_context *ehc = &link->eh_context;
unsigned int *classes = ehc->classes;
+ unsigned int lflags = link->flags;
int verbose = !(ehc->i.flags & ATA_EHI_QUIET);
- int try = 0;
+ int max_tries = 0, try = 0;
struct ata_device *dev;
- unsigned long deadline;
- unsigned int action;
+ unsigned long deadline, now;
ata_reset_fn_t reset;
unsigned long flags;
- int rc;
+ u32 sstatus;
+ int nr_known, rc;
+
+ /*
+ * Prepare to reset
+ */
+ while (ata_eh_reset_timeouts[max_tries] != ULONG_MAX)
+ max_tries++;
+ if (link->flags & ATA_LFLAG_NO_HRST)
+ hardreset = NULL;
+ if (link->flags & ATA_LFLAG_NO_SRST)
+ softreset = NULL;
+
+ now = jiffies;
+ deadline = ata_deadline(ehc->last_reset, ATA_EH_RESET_COOL_DOWN);
+ if (time_before(now, deadline))
+ schedule_timeout_uninterruptible(deadline - now);
- /* about to reset */
spin_lock_irqsave(ap->lock, flags);
ap->pflags |= ATA_PFLAG_RESETTING;
spin_unlock_irqrestore(ap->lock, flags);
- ata_eh_about_to_do(link, NULL, ehc->i.action & ATA_EH_RESET_MASK);
+ ata_eh_about_to_do(link, NULL, ATA_EH_RESET);
+ ehc->last_reset = jiffies;
- /* Determine which reset to use and record in ehc->i.action.
- * prereset() may examine and modify it.
- */
- action = ehc->i.action;
- ehc->i.action &= ~ATA_EH_RESET_MASK;
- if (softreset && (!hardreset || (!(link->flags & ATA_LFLAG_NO_SRST) &&
- !sata_set_spd_needed(link) &&
- !(action & ATA_EH_HARDRESET))))
- ehc->i.action |= ATA_EH_SOFTRESET;
- else
+ ata_link_for_each_dev(dev, link) {
+ /* If we issue an SRST then an ATA drive (not ATAPI)
+ * may change configuration and be in PIO0 timing. If
+ * we do a hard reset (or are coming from power on)
+ * this is true for ATA or ATAPI. Until we've set a
+ * suitable controller mode we should not touch the
+ * bus as we may be talking too fast.
+ */
+ dev->pio_mode = XFER_PIO_0;
+
+ /* If the controller has a pio mode setup function
+ * then use it to set the chipset to rights. Don't
+ * touch the DMA setup as that will be dealt with when
+ * configuring devices.
+ */
+ if (ap->ops->set_piomode)
+ ap->ops->set_piomode(ap, dev);
+ }
+
+ /* prefer hardreset */
+ reset = NULL;
+ ehc->i.action &= ~ATA_EH_RESET;
+ if (hardreset) {
+ reset = hardreset;
ehc->i.action |= ATA_EH_HARDRESET;
+ } else if (softreset) {
+ reset = softreset;
+ ehc->i.action |= ATA_EH_SOFTRESET;
+ }
if (prereset) {
- rc = prereset(link, jiffies + ATA_EH_PRERESET_TIMEOUT);
+ rc = prereset(link,
+ ata_deadline(jiffies, ATA_EH_PRERESET_TIMEOUT));
if (rc) {
if (rc == -ENOENT) {
ata_link_printk(link, KERN_DEBUG,
"port disabled. ignoring.\n");
- ehc->i.action &= ~ATA_EH_RESET_MASK;
+ ehc->i.action &= ~ATA_EH_RESET;
ata_link_for_each_dev(dev, link)
classes[dev->devno] = ATA_DEV_NONE;
"prereset failed (errno=%d)\n", rc);
goto out;
}
- }
-
- /* prereset() might have modified ehc->i.action */
- if (ehc->i.action & ATA_EH_HARDRESET)
- reset = hardreset;
- else if (ehc->i.action & ATA_EH_SOFTRESET)
- reset = softreset;
- else {
- /* prereset told us not to reset, bang classes and return */
- ata_link_for_each_dev(dev, link)
- classes[dev->devno] = ATA_DEV_NONE;
- rc = 0;
- goto out;
- }
- /* did prereset() screw up? if so, fix up to avoid oopsing */
- if (!reset) {
- if (softreset)
- reset = softreset;
- else
- reset = hardreset;
+ /* prereset() might have cleared ATA_EH_RESET. If so,
+ * bang classes and return.
+ */
+ if (reset && !(ehc->i.action & ATA_EH_RESET)) {
+ ata_link_for_each_dev(dev, link)
+ classes[dev->devno] = ATA_DEV_NONE;
+ rc = 0;
+ goto out;
+ }
}
retry:
- deadline = jiffies + ata_eh_reset_timeouts[try++];
+ /*
+ * Perform reset
+ */
+ ehc->last_reset = jiffies;
+ if (ata_is_host_link(link))
+ ata_eh_freeze_port(ap);
- /* shut up during boot probing */
- if (verbose)
- ata_link_printk(link, KERN_INFO, "%s resetting link\n",
- reset == softreset ? "soft" : "hard");
+ deadline = ata_deadline(jiffies, ata_eh_reset_timeouts[try++]);
- /* mark that this EH session started with reset */
- if (reset == hardreset)
- ehc->i.flags |= ATA_EHI_DID_HARDRESET;
- else
- ehc->i.flags |= ATA_EHI_DID_SOFTRESET;
+ if (reset) {
+ if (verbose)
+ ata_link_printk(link, KERN_INFO, "%s resetting link\n",
+ reset == softreset ? "soft" : "hard");
- rc = ata_do_reset(link, reset, classes, deadline);
+ /* mark that this EH session started with reset */
+ if (reset == hardreset)
+ ehc->i.flags |= ATA_EHI_DID_HARDRESET;
+ else
+ ehc->i.flags |= ATA_EHI_DID_SOFTRESET;
- if (reset == hardreset &&
- ata_eh_followup_srst_needed(link, rc, classify, classes)) {
- /* okay, let's do follow-up softreset */
- reset = softreset;
+ rc = ata_do_reset(link, reset, classes, deadline);
+ if (rc && rc != -EAGAIN)
+ goto fail;
- if (!reset) {
- ata_link_printk(link, KERN_ERR,
- "follow-up softreset required "
- "but no softreset avaliable\n");
- rc = -EINVAL;
- goto out;
- }
+ if (reset == hardreset &&
+ ata_eh_followup_srst_needed(link, rc, classes)) {
+ /* okay, let's do follow-up softreset */
+ reset = softreset;
- ata_eh_about_to_do(link, NULL, ATA_EH_RESET_MASK);
- rc = ata_do_reset(link, reset, classes, deadline);
+ if (!reset) {
+ ata_link_printk(link, KERN_ERR,
+ "follow-up softreset required "
+ "but no softreset avaliable\n");
+ rc = -EINVAL;
+ goto fail;
+ }
- if (rc == 0 && classify && classes[0] == ATA_DEV_UNKNOWN &&
- !(link->flags & ATA_LFLAG_ASSUME_CLASS)) {
- ata_link_printk(link, KERN_ERR,
- "classification failed\n");
- rc = -EINVAL;
- goto out;
+ ata_eh_about_to_do(link, NULL, ATA_EH_RESET);
+ rc = ata_do_reset(link, reset, classes, deadline);
}
+ } else {
+ if (verbose)
+ ata_link_printk(link, KERN_INFO, "no reset method "
+ "available, skipping reset\n");
+ if (!(lflags & ATA_LFLAG_ASSUME_CLASS))
+ lflags |= ATA_LFLAG_ASSUME_ATA;
}
- /* if we skipped follow-up srst, clear rc */
- if (rc == -EAGAIN)
- rc = 0;
-
- if (rc && rc != -ERESTART && try < ARRAY_SIZE(ata_eh_reset_timeouts)) {
- unsigned long now = jiffies;
-
- if (time_before(now, deadline)) {
- unsigned long delta = deadline - jiffies;
-
- ata_link_printk(link, KERN_WARNING, "reset failed "
- "(errno=%d), retrying in %u secs\n",
- rc, (jiffies_to_msecs(delta) + 999) / 1000);
+ /*
+ * Post-reset processing
+ */
+ ata_link_for_each_dev(dev, link) {
+ /* After the reset, the device state is PIO 0 and the
+ * controller state is undefined. Reset also wakes up
+ * drives from sleeping mode.
+ */
+ dev->pio_mode = XFER_PIO_0;
+ dev->flags &= ~ATA_DFLAG_SLEEPING;
- while (delta)
- delta = schedule_timeout_uninterruptible(delta);
- }
+ if (ata_link_offline(link))
+ continue;
- if (rc == -EPIPE ||
- try == ARRAY_SIZE(ata_eh_reset_timeouts) - 1)
- sata_down_spd_limit(link);
- if (hardreset)
- reset = hardreset;
- goto retry;
+ /* apply class override */
+ if (lflags & ATA_LFLAG_ASSUME_ATA)
+ classes[dev->devno] = ATA_DEV_ATA;
+ else if (lflags & ATA_LFLAG_ASSUME_SEMB)
+ classes[dev->devno] = ATA_DEV_SEMB_UNSUP; /* not yet */
}
- if (rc == 0) {
- u32 sstatus;
+ /* record current link speed */
+ if (sata_scr_read(link, SCR_STATUS, &sstatus) == 0)
+ link->sata_spd = (sstatus >> 4) & 0xf;
- ata_link_for_each_dev(dev, link) {
- /* After the reset, the device state is PIO 0
- * and the controller state is undefined.
- * Record the mode.
- */
- dev->pio_mode = XFER_PIO_0;
+ /* thaw the port */
+ if (ata_is_host_link(link))
+ ata_eh_thaw_port(ap);
- if (ata_link_offline(link))
- continue;
+ /* postreset() should clear hardware SError. Although SError
+ * is cleared during link resume, clearing SError here is
+ * necessary as some PHYs raise hotplug events after SRST.
+ * This introduces race condition where hotplug occurs between
+ * reset and here. This race is mediated by cross checking
+ * link onlineness and classification result later.
+ */
+ if (postreset)
+ postreset(link, classes);
+
+ /* clear cached SError */
+ spin_lock_irqsave(link->ap->lock, flags);
+ link->eh_info.serror = 0;
+ spin_unlock_irqrestore(link->ap->lock, flags);
+
+ /* Make sure onlineness and classification result correspond.
+ * Hotplug could have happened during reset and some
+ * controllers fail to wait while a drive is spinning up after
+ * being hotplugged causing misdetection. By cross checking
+ * link onlineness and classification result, those conditions
+ * can be reliably detected and retried.
+ */
+ nr_known = 0;
+ ata_link_for_each_dev(dev, link) {
+ /* convert all ATA_DEV_UNKNOWN to ATA_DEV_NONE */
+ if (classes[dev->devno] == ATA_DEV_UNKNOWN)
+ classes[dev->devno] = ATA_DEV_NONE;
+ else
+ nr_known++;
+ }
- /* apply class override and convert UNKNOWN to NONE */
- if (link->flags & ATA_LFLAG_ASSUME_ATA)
- classes[dev->devno] = ATA_DEV_ATA;
- else if (link->flags & ATA_LFLAG_ASSUME_SEMB)
- classes[dev->devno] = ATA_DEV_SEMB_UNSUP; /* not yet */
- else if (classes[dev->devno] == ATA_DEV_UNKNOWN)
- classes[dev->devno] = ATA_DEV_NONE;
+ if (classify && !nr_known && ata_link_online(link)) {
+ if (try < max_tries) {
+ ata_link_printk(link, KERN_WARNING, "link online but "
+ "device misclassified, retrying\n");
+ rc = -EAGAIN;
+ goto fail;
}
+ ata_link_printk(link, KERN_WARNING,
+ "link online but device misclassified, "
+ "device detection might fail\n");
+ }
- /* record current link speed */
- if (sata_scr_read(link, SCR_STATUS, &sstatus) == 0)
- link->sata_spd = (sstatus >> 4) & 0xf;
-
- if (postreset)
- postreset(link, classes);
+ /* reset successful, schedule revalidation */
+ ata_eh_done(link, NULL, ATA_EH_RESET);
+ ehc->last_reset = jiffies;
+ ehc->i.action |= ATA_EH_REVALIDATE;
- /* reset successful, schedule revalidation */
- ata_eh_done(link, NULL, ehc->i.action & ATA_EH_RESET_MASK);
- ehc->i.action |= ATA_EH_REVALIDATE;
- }
+ rc = 0;
out:
/* clear hotplug flag */
ehc->i.flags &= ~ATA_EHI_HOTPLUGGED;
spin_unlock_irqrestore(ap->lock, flags);
return rc;
+
+ fail:
+ /* if SCR isn't accessible on a fan-out port, PMP needs to be reset */
+ if (!ata_is_host_link(link) &&
+ sata_scr_read(link, SCR_STATUS, &sstatus))
+ rc = -ERESTART;
+
+ if (rc == -ERESTART || try >= max_tries)
+ goto out;
+
+ now = jiffies;
+ if (time_before(now, deadline)) {
+ unsigned long delta = deadline - now;
+
+ ata_link_printk(link, KERN_WARNING,
+ "reset failed (errno=%d), retrying in %u secs\n",
+ rc, DIV_ROUND_UP(jiffies_to_msecs(delta), 1000));
+
+ while (delta)
+ delta = schedule_timeout_uninterruptible(delta);
+ }
+
+ if (rc == -EPIPE || try == max_tries - 1)
+ sata_down_spd_limit(link);
+ if (hardreset)
+ reset = hardreset;
+ goto retry;
}
static int ata_eh_revalidate_and_attach(struct ata_link *link,
}
/* PDIAG- should have been released, ask cable type if post-reset */
- if (ata_is_host_link(link) && ap->ops->cable_detect &&
- (ehc->i.flags & ATA_EHI_DID_RESET))
- ap->cbl = ap->ops->cable_detect(ap);
+ if ((ehc->i.flags & ATA_EHI_DID_RESET) && ata_is_host_link(link)) {
+ if (ap->ops->cable_detect)
+ ap->cbl = ap->ops->cable_detect(ap);
+ ata_force_cbl(ap);
+ }
/* Configure new devices forward such that user doesn't see
* device detection messages backwards.
return rc;
}
+/**
+ * ata_set_mode - Program timings and issue SET FEATURES - XFER
+ * @link: link on which timings will be programmed
+ * @r_failed_dev: out paramter for failed device
+ *
+ * Set ATA device disk transfer mode (PIO3, UDMA6, etc.). If
+ * ata_set_mode() fails, pointer to the failing device is
+ * returned in @r_failed_dev.
+ *
+ * LOCKING:
+ * PCI/etc. bus probe sem.
+ *
+ * RETURNS:
+ * 0 on success, negative errno otherwise
+ */
+int ata_set_mode(struct ata_link *link, struct ata_device **r_failed_dev)
+{
+ struct ata_port *ap = link->ap;
+ struct ata_device *dev;
+ int rc;
+
+ /* if data transfer is verified, clear DUBIOUS_XFER on ering top */
+ ata_link_for_each_dev(dev, link) {
+ if (!(dev->flags & ATA_DFLAG_DUBIOUS_XFER)) {
+ struct ata_ering_entry *ent;
+
+ ent = ata_ering_top(&dev->ering);
+ if (ent)
+ ent->eflags &= ~ATA_EFLAG_DUBIOUS_XFER;
+ }
+ }
+
+ /* has private set_mode? */
+ if (ap->ops->set_mode)
+ rc = ap->ops->set_mode(link, r_failed_dev);
+ else
+ rc = ata_do_set_mode(link, r_failed_dev);
+
+ /* if transfer mode has changed, set DUBIOUS_XFER on device */
+ ata_link_for_each_dev(dev, link) {
+ struct ata_eh_context *ehc = &link->eh_context;
+ u8 saved_xfer_mode = ehc->saved_xfer_mode[dev->devno];
+ u8 saved_ncq = !!(ehc->saved_ncq_enabled & (1 << dev->devno));
+
+ if (dev->xfer_mode != saved_xfer_mode ||
+ ata_ncq_enabled(dev) != saved_ncq)
+ dev->flags |= ATA_DFLAG_DUBIOUS_XFER;
+ }
+
+ return rc;
+}
+
static int ata_link_nr_enabled(struct ata_link *link)
{
struct ata_device *dev;
static int ata_eh_skip_recovery(struct ata_link *link)
{
+ struct ata_port *ap = link->ap;
struct ata_eh_context *ehc = &link->eh_context;
struct ata_device *dev;
if (link->flags & ATA_LFLAG_DISABLED)
return 1;
- /* thaw frozen port, resume link and recover failed devices */
- if ((link->ap->pflags & ATA_PFLAG_FROZEN) ||
- (ehc->i.flags & ATA_EHI_RESUME_LINK) || ata_link_nr_enabled(link))
+ /* thaw frozen port and recover failed devices */
+ if ((ap->pflags & ATA_PFLAG_FROZEN) || ata_link_nr_enabled(link))
+ return 0;
+
+ /* reset at least once if reset is requested */
+ if ((ehc->i.action & ATA_EH_RESET) &&
+ !(ehc->i.flags & ATA_EHI_DID_RESET))
return 0;
/* skip if class codes for all vacant slots are ATA_DEV_NONE */
return 1;
}
+static int ata_eh_schedule_probe(struct ata_device *dev)
+{
+ struct ata_eh_context *ehc = &dev->link->eh_context;
+
+ if (!(ehc->i.probe_mask & (1 << dev->devno)) ||
+ (ehc->did_probe_mask & (1 << dev->devno)))
+ return 0;
+
+ ata_eh_detach_dev(dev);
+ ata_dev_init(dev);
+ ehc->did_probe_mask |= (1 << dev->devno);
+ ehc->i.action |= ATA_EH_RESET;
+ ehc->saved_xfer_mode[dev->devno] = 0;
+ ehc->saved_ncq_enabled &= ~(1 << dev->devno);
+
+ return 1;
+}
+
static int ata_eh_handle_dev_fail(struct ata_device *dev, int err)
{
struct ata_eh_context *ehc = &dev->link->eh_context;
/* give it just one more chance */
ehc->tries[dev->devno] = min(ehc->tries[dev->devno], 1);
case -EIO:
- if (ehc->tries[dev->devno] == 1) {
+ if (ehc->tries[dev->devno] == 1 && dev->pio_mode > XFER_PIO_0) {
/* This is the last chance, better to slow
* down than lose it.
*/
if (ata_link_offline(dev->link))
ata_eh_detach_dev(dev);
- /* probe if requested */
- if ((ehc->i.probe_mask & (1 << dev->devno)) &&
- !(ehc->did_probe_mask & (1 << dev->devno))) {
- ata_eh_detach_dev(dev);
- ata_dev_init(dev);
-
+ /* schedule probe if necessary */
+ if (ata_eh_schedule_probe(dev)) {
ehc->tries[dev->devno] = ATA_EH_DEV_TRIES;
- ehc->did_probe_mask |= (1 << dev->devno);
- ehc->i.action |= ATA_EH_SOFTRESET;
+ memset(ehc->cmd_timeout_idx[dev->devno], 0,
+ sizeof(ehc->cmd_timeout_idx[dev->devno]));
}
return 1;
} else {
- /* soft didn't work? be haaaaard */
- if (ehc->i.flags & ATA_EHI_DID_RESET)
- ehc->i.action |= ATA_EH_HARDRESET;
- else
- ehc->i.action |= ATA_EH_SOFTRESET;
-
+ ehc->i.action |= ATA_EH_RESET;
return 0;
}
}
{
struct ata_link *link;
struct ata_device *dev;
- int nr_failed_devs, nr_disabled_devs;
- int reset, rc;
+ int nr_failed_devs;
+ int rc;
unsigned long flags;
DPRINTK("ENTER\n");
if (dev->flags & ATA_DFLAG_DETACH)
ata_eh_detach_dev(dev);
- if (!ata_dev_enabled(dev) &&
- ((ehc->i.probe_mask & (1 << dev->devno)) &&
- !(ehc->did_probe_mask & (1 << dev->devno)))) {
- ata_eh_detach_dev(dev);
- ata_dev_init(dev);
- ehc->did_probe_mask |= (1 << dev->devno);
- ehc->i.action |= ATA_EH_SOFTRESET;
- }
+ /* schedule probe if necessary */
+ if (!ata_dev_enabled(dev))
+ ata_eh_schedule_probe(dev);
}
}
retry:
rc = 0;
nr_failed_devs = 0;
- nr_disabled_devs = 0;
- reset = 0;
/* if UNLOADING, finish immediately */
if (ap->pflags & ATA_PFLAG_UNLOADING)
if (ata_eh_skip_recovery(link))
ehc->i.action = 0;
- /* do we need to reset? */
- if (ehc->i.action & ATA_EH_RESET_MASK)
- reset = 1;
-
ata_link_for_each_dev(dev, link)
ehc->classes[dev->devno] = ATA_DEV_UNKNOWN;
}
/* reset */
- if (reset) {
- ata_eh_freeze_port(ap);
-
- ata_port_for_each_link(link, ap) {
- struct ata_eh_context *ehc = &link->eh_context;
+ ata_port_for_each_link(link, ap) {
+ struct ata_eh_context *ehc = &link->eh_context;
- if (!(ehc->i.action & ATA_EH_RESET_MASK))
- continue;
+ if (!(ehc->i.action & ATA_EH_RESET))
+ continue;
- rc = ata_eh_reset(link, ata_link_nr_vacant(link),
- prereset, softreset, hardreset,
- postreset);
- if (rc) {
- ata_link_printk(link, KERN_ERR,
- "reset failed, giving up\n");
- goto out;
- }
+ rc = ata_eh_reset(link, ata_link_nr_vacant(link),
+ prereset, softreset, hardreset, postreset);
+ if (rc) {
+ ata_link_printk(link, KERN_ERR,
+ "reset failed, giving up\n");
+ goto out;
}
-
- ata_eh_thaw_port(ap);
}
/* the rest */
ehc->i.flags &= ~ATA_EHI_SETMODE;
}
+ if (ehc->i.action & ATA_EH_LPM)
+ ata_link_for_each_dev(dev, link)
+ ata_dev_enable_pm(dev, ap->pm_policy);
+
/* this link is okay now */
ehc->i.flags = 0;
continue;
- dev_fail:
+dev_fail:
nr_failed_devs++;
- if (ata_eh_handle_dev_fail(dev, rc))
- nr_disabled_devs++;
+ ata_eh_handle_dev_fail(dev, rc);
- if (ap->pflags & ATA_PFLAG_FROZEN)
+ if (ap->pflags & ATA_PFLAG_FROZEN) {
+ /* PMP reset requires working host port.
+ * Can't retry if it's frozen.
+ */
+ if (sata_pmp_attached(ap))
+ goto out;
break;
- }
-
- if (nr_failed_devs) {
- if (nr_failed_devs != nr_disabled_devs) {
- ata_port_printk(ap, KERN_WARNING, "failed to recover "
- "some devices, retrying in 5 secs\n");
- ssleep(5);
- } else {
- /* no device left to recover, repeat fast */
- msleep(500);
}
+ }
+ if (nr_failed_devs)
goto retry;
- }
out:
if (rc && r_failed_link)
* generate sense data in this function,
* considering both err_mask and tf.
*/
- if (qc->err_mask & AC_ERR_INVALID)
- ata_eh_qc_complete(qc);
- else
+ if (qc->flags & ATA_QCFLAG_RETRY)
ata_eh_qc_retry(qc);
+ else
+ ata_eh_qc_complete(qc);
} else {
if (qc->flags & ATA_QCFLAG_SENSE_VALID) {
ata_eh_qc_complete(qc);
/**
* ata_do_eh - do standard error handling
* @ap: host port to handle error for
+ *
* @prereset: prereset method (can be NULL)
* @softreset: softreset method (can be NULL)
* @hardreset: hardreset method (can be NULL)
ata_eh_finish(ap);
}
+/**
+ * ata_std_error_handler - standard error handler
+ * @ap: host port to handle error for
+ *
+ * Standard error handler
+ *
+ * LOCKING:
+ * Kernel thread context (may sleep).
+ */
+void ata_std_error_handler(struct ata_port *ap)
+{
+ struct ata_port_operations *ops = ap->ops;
+ ata_reset_fn_t hardreset = ops->hardreset;
+
+ /* ignore built-in hardreset if SCR access is not available */
+ if (ata_is_builtin_hardreset(hardreset) && !sata_scr_valid(&ap->link))
+ hardreset = NULL;
+
+ ata_do_eh(ap, ops->prereset, ops->softreset, hardreset, ops->postreset);
+}
+
#ifdef CONFIG_PM
/**
* ata_eh_handle_port_suspend - perform port suspend operation
if (ap->ops->port_suspend)
rc = ap->ops->port_suspend(ap, ap->pm_mesg);
+ ata_acpi_set_state(ap, PMSG_SUSPEND);
out:
/* report result */
spin_lock_irqsave(ap->lock, flags);
WARN_ON(!(ap->pflags & ATA_PFLAG_SUSPENDED));
+ ata_acpi_set_state(ap, PMSG_ON);
+
if (ap->ops->port_resume)
rc = ap->ops->port_resume(ap);