Gitweb:     
http://git.kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=14e660e677ddd3574247495aae4ef63eb8899072
Commit:     14e660e677ddd3574247495aae4ef63eb8899072
Parent:     b7cc176c9eb3aa6989ac099efd8bdd6d0eaa784a
Author:     Seokmann Ju <[EMAIL PROTECTED]>
AuthorDate: Thu Sep 20 14:07:36 2007 -0700
Committer:  James Bottomley <[EMAIL PROTECTED]>
CommitDate: Fri Oct 12 14:49:47 2007 -0400

    [SCSI] qla2xxx: Add PCI error recovery support.
    
    Additional cleanups and
    Signed-off-by: Andrew Vasquez <[EMAIL PROTECTED]>
    Signed-off-by: James Bottomley <[EMAIL PROTECTED]>
---
 drivers/scsi/qla2xxx/qla_def.h |    1 +
 drivers/scsi/qla2xxx/qla_isr.c |   28 +++++++++-
 drivers/scsi/qla2xxx/qla_os.c  |  114 ++++++++++++++++++++++++++++++++++++++++
 3 files changed, 142 insertions(+), 1 deletions(-)

diff --git a/drivers/scsi/qla2xxx/qla_def.h b/drivers/scsi/qla2xxx/qla_def.h
index e1e3428..75ab898 100644
--- a/drivers/scsi/qla2xxx/qla_def.h
+++ b/drivers/scsi/qla2xxx/qla_def.h
@@ -23,6 +23,7 @@
 #include <linux/interrupt.h>
 #include <linux/workqueue.h>
 #include <linux/firmware.h>
+#include <linux/aer.h>
 #include <asm/semaphore.h>
 
 #include <scsi/scsi.h>
diff --git a/drivers/scsi/qla2xxx/qla_isr.c b/drivers/scsi/qla2xxx/qla_isr.c
index eecae99..dcfb24b 100644
--- a/drivers/scsi/qla2xxx/qla_isr.c
+++ b/drivers/scsi/qla2xxx/qla_isr.c
@@ -34,6 +34,7 @@ qla2100_intr_handler(int irq, void *dev_id)
        int             status;
        unsigned long   flags;
        unsigned long   iter;
+       uint16_t        hccr;
        uint16_t        mb[4];
 
        ha = (scsi_qla_host_t *) dev_id;
@@ -48,7 +49,23 @@ qla2100_intr_handler(int irq, void *dev_id)
 
        spin_lock_irqsave(&ha->hardware_lock, flags);
        for (iter = 50; iter--; ) {
-               if ((RD_REG_WORD(&reg->istatus) & ISR_RISC_INT) == 0)
+               hccr = RD_REG_WORD(&reg->hccr);
+               if (hccr & HCCR_RISC_PAUSE) {
+                       if (pci_channel_offline(ha->pdev))
+                               break;
+
+                       /*
+                        * Issue a "HARD" reset in order for the RISC interrupt
+                        * bit to be cleared.  Schedule a big hammmer to get
+                        * out of the RISC PAUSED state.
+                        */
+                       WRT_REG_WORD(&reg->hccr, HCCR_RESET_RISC);
+                       RD_REG_WORD(&reg->hccr);
+
+                       ha->isp_ops->fw_dump(ha, 1);
+                       set_bit(ISP_ABORT_NEEDED, &ha->dpc_flags);
+                       break;
+               } else if ((RD_REG_WORD(&reg->istatus) & ISR_RISC_INT) == 0)
                        break;
 
                if (RD_REG_WORD(&reg->semaphore) & BIT_0) {
@@ -127,6 +144,9 @@ qla2300_intr_handler(int irq, void *dev_id)
        for (iter = 50; iter--; ) {
                stat = RD_REG_DWORD(&reg->u.isp2300.host_status);
                if (stat & HSR_RISC_PAUSED) {
+                       if (pci_channel_offline(ha->pdev))
+                               break;
+
                        hccr = RD_REG_WORD(&reg->hccr);
                        if (hccr & (BIT_15 | BIT_13 | BIT_11 | BIT_8))
                                qla_printk(KERN_INFO, ha, "Parity error -- "
@@ -1499,6 +1519,9 @@ qla24xx_intr_handler(int irq, void *dev_id)
        for (iter = 50; iter--; ) {
                stat = RD_REG_DWORD(&reg->host_status);
                if (stat & HSRX_RISC_PAUSED) {
+                       if (pci_channel_offline(ha->pdev))
+                               break;
+
                        hccr = RD_REG_DWORD(&reg->hccr);
 
                        qla_printk(KERN_INFO, ha, "RISC paused -- HCCR=%x, "
@@ -1633,6 +1656,9 @@ qla24xx_msix_default(int irq, void *dev_id)
        for (iter = 50; iter--; ) {
                stat = RD_REG_DWORD(&reg->host_status);
                if (stat & HSRX_RISC_PAUSED) {
+                       if (pci_channel_offline(ha->pdev))
+                               break;
+
                        hccr = RD_REG_DWORD(&reg->hccr);
 
                        qla_printk(KERN_INFO, ha, "RISC paused -- HCCR=%x, "
diff --git a/drivers/scsi/qla2xxx/qla_os.c b/drivers/scsi/qla2xxx/qla_os.c
index 2a03400..a8ab2d3 100644
--- a/drivers/scsi/qla2xxx/qla_os.c
+++ b/drivers/scsi/qla2xxx/qla_os.c
@@ -385,6 +385,11 @@ qla2x00_queuecommand(struct scsi_cmnd *cmd, void 
(*done)(struct scsi_cmnd *))
        srb_t *sp;
        int rval;
 
+       if (unlikely(pci_channel_offline(ha->pdev))) {
+               cmd->result = DID_REQUEUE << 16;
+               goto qc_fail_command;
+       }
+
        rval = fc_remote_port_chkready(rport);
        if (rval) {
                cmd->result = rval;
@@ -447,6 +452,11 @@ qla24xx_queuecommand(struct scsi_cmnd *cmd, void 
(*done)(struct scsi_cmnd *))
        int rval;
        scsi_qla_host_t *pha = to_qla_parent(ha);
 
+       if (unlikely(pci_channel_offline(ha->pdev))) {
+               cmd->result = DID_REQUEUE << 16;
+               goto qc24_fail_command;
+       }
+
        rval = fc_remote_port_chkready(rport);
        if (rval) {
                cmd->result = rval;
@@ -1571,6 +1581,10 @@ qla2x00_probe_one(struct pci_dev *pdev, const struct 
pci_device_id *id)
        if (pci_enable_device(pdev))
                goto probe_out;
 
+       if (pci_find_aer_capability(pdev))
+               if (pci_enable_pcie_error_reporting(pdev))
+                       goto probe_out;
+
        sht = &qla2x00_driver_template;
        if (pdev->device == PCI_DEVICE_ID_QLOGIC_ISP2422 ||
            pdev->device == PCI_DEVICE_ID_QLOGIC_ISP2432 ||
@@ -2814,6 +2828,105 @@ qla2x00_release_firmware(void)
        up(&qla_fw_lock);
 }
 
+static pci_ers_result_t
+qla2xxx_pci_error_detected(struct pci_dev *pdev, pci_channel_state_t state)
+{
+       switch (state) {
+       case pci_channel_io_normal:
+               return PCI_ERS_RESULT_CAN_RECOVER;
+       case pci_channel_io_frozen:
+               pci_disable_device(pdev);
+               return PCI_ERS_RESULT_NEED_RESET;
+       case pci_channel_io_perm_failure:
+               qla2x00_remove_one(pdev);
+               return PCI_ERS_RESULT_DISCONNECT;
+       }
+       return PCI_ERS_RESULT_NEED_RESET;
+}
+
+static pci_ers_result_t
+qla2xxx_pci_mmio_enabled(struct pci_dev *pdev)
+{
+       int risc_paused = 0;
+       uint32_t stat;
+       unsigned long flags;
+       scsi_qla_host_t *ha = pci_get_drvdata(pdev);
+       struct device_reg_2xxx __iomem *reg = &ha->iobase->isp;
+       struct device_reg_24xx __iomem *reg24 = &ha->iobase->isp24;
+
+       spin_lock_irqsave(&ha->hardware_lock, flags);
+       if (IS_QLA2100(ha) || IS_QLA2200(ha)){
+               stat = RD_REG_DWORD(&reg->hccr);
+               if (stat & HCCR_RISC_PAUSE)
+                       risc_paused = 1;
+       } else if (IS_QLA23XX(ha)) {
+               stat = RD_REG_DWORD(&reg->u.isp2300.host_status);
+               if (stat & HSR_RISC_PAUSED)
+                       risc_paused = 1;
+       } else if (IS_FWI2_CAPABLE(ha)) {
+               stat = RD_REG_DWORD(&reg24->host_status);
+               if (stat & HSRX_RISC_PAUSED)
+                       risc_paused = 1;
+       }
+       spin_unlock_irqrestore(&ha->hardware_lock, flags);
+
+       if (risc_paused) {
+               qla_printk(KERN_INFO, ha, "RISC paused -- mmio_enabled, "
+                   "Dumping firmware!\n");
+               ha->isp_ops->fw_dump(ha, 0);
+
+               return PCI_ERS_RESULT_NEED_RESET;
+       } else
+               return PCI_ERS_RESULT_RECOVERED;
+}
+
+static pci_ers_result_t
+qla2xxx_pci_slot_reset(struct pci_dev *pdev)
+{
+       pci_ers_result_t ret = PCI_ERS_RESULT_DISCONNECT;
+       scsi_qla_host_t *ha = pci_get_drvdata(pdev);
+
+       if (pci_enable_device(pdev)) {
+               qla_printk(KERN_WARNING, ha,
+                   "Can't re-enable PCI device after reset.\n");
+
+               return ret;
+       }
+       pci_set_master(pdev);
+
+       if (ha->isp_ops->pci_config(ha))
+               return ret;
+
+       set_bit(ABORT_ISP_ACTIVE, &ha->dpc_flags);
+       if (qla2x00_abort_isp(ha)== QLA_SUCCESS)
+               ret =  PCI_ERS_RESULT_RECOVERED;
+       clear_bit(ABORT_ISP_ACTIVE, &ha->dpc_flags);
+
+       return ret;
+}
+
+static void
+qla2xxx_pci_resume(struct pci_dev *pdev)
+{
+       scsi_qla_host_t *ha = pci_get_drvdata(pdev);
+       int ret;
+
+       ret = qla2x00_wait_for_hba_online(ha);
+       if (ret != QLA_SUCCESS) {
+               qla_printk(KERN_ERR, ha,
+                   "the device failed to resume I/O "
+                   "from slot/link_reset");
+       }
+       pci_cleanup_aer_uncorrect_error_status(pdev);
+}
+
+static struct pci_error_handlers qla2xxx_err_handler = {
+       .error_detected = qla2xxx_pci_error_detected,
+       .mmio_enabled = qla2xxx_pci_mmio_enabled,
+       .slot_reset = qla2xxx_pci_slot_reset,
+       .resume = qla2xxx_pci_resume,
+};
+
 static struct pci_device_id qla2xxx_pci_tbl[] = {
        { PCI_DEVICE(PCI_VENDOR_ID_QLOGIC, PCI_DEVICE_ID_QLOGIC_ISP2100) },
        { PCI_DEVICE(PCI_VENDOR_ID_QLOGIC, PCI_DEVICE_ID_QLOGIC_ISP2200) },
@@ -2839,6 +2952,7 @@ static struct pci_driver qla2xxx_pci_driver = {
        .id_table       = qla2xxx_pci_tbl,
        .probe          = qla2x00_probe_one,
        .remove         = __devexit_p(qla2x00_remove_one),
+       .err_handler    = &qla2xxx_err_handler,
 };
 
 /**
-
To unsubscribe from this list: send the line "unsubscribe git-commits-head" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to