From ef7c80c1f18f2c5eea2dabd214f12e0c93ac29cf Mon Sep 17 00:00:00 2001 From: "Kashyap, Desai" Date: Mon, 5 Apr 2010 14:20:07 +0530 Subject: [PATCH] [SCSI] mpt2sas: Added support for PCIe Advanced Error Recovery. Added support in the driver to support EEH and PCIe Advanced Error Recovery. This involves adding new pci_error_handler interface for recovering the controller from PCI Bus errors, such as SERR and PERR. Some tools are available for simulating PCI errors in order to validate this interface: http://www.kernel.org/pub/linux/utils/pci/aer-inject Signed-off-by: Kashyap Desai Signed-off-by: James Bottomley --- drivers/scsi/mpt2sas/mpt2sas_base.c | 6 ++ drivers/scsi/mpt2sas/mpt2sas_scsih.c | 118 +++++++++++++++++++++++++++++++++++ 2 files changed, 124 insertions(+) diff --git a/drivers/scsi/mpt2sas/mpt2sas_base.c b/drivers/scsi/mpt2sas/mpt2sas_base.c index f980b88..b04ccad 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_base.c +++ b/drivers/scsi/mpt2sas/mpt2sas_base.c @@ -58,6 +58,7 @@ #include #include #include +#include #include "mpt2sas_base.h" @@ -1256,6 +1257,9 @@ mpt2sas_base_map_resources(struct MPT2SAS_ADAPTER *ioc) goto out_fail; } + /* AER (Advanced Error Reporting) hooks */ + pci_enable_pcie_error_reporting(pdev); + pci_set_master(pdev); if (_base_config_dma_addressing(ioc, pdev) != 0) { @@ -1311,6 +1315,7 @@ mpt2sas_base_map_resources(struct MPT2SAS_ADAPTER *ioc) ioc->chip_phys = 0; ioc->pci_irq = -1; pci_release_selected_regions(ioc->pdev, ioc->bars); + pci_disable_pcie_error_reporting(pdev); pci_disable_device(pdev); return r; } @@ -3547,6 +3552,7 @@ mpt2sas_base_free_resources(struct MPT2SAS_ADAPTER *ioc) ioc->pci_irq = -1; ioc->chip_phys = 0; pci_release_selected_regions(ioc->pdev, ioc->bars); + pci_disable_pcie_error_reporting(pdev); pci_disable_device(pdev); return; } diff --git a/drivers/scsi/mpt2sas/mpt2sas_scsih.c b/drivers/scsi/mpt2sas/mpt2sas_scsih.c index 1f0466a..bb5659c 100644 --- a/drivers/scsi/mpt2sas/mpt2sas_scsih.c +++ b/drivers/scsi/mpt2sas/mpt2sas_scsih.c @@ -52,6 +52,7 @@ #include #include #include +#include #include #include @@ -6664,6 +6665,122 @@ _scsih_resume(struct pci_dev *pdev) } #endif /* CONFIG_PM */ +/** + * _scsih_pci_error_detected - Called when a PCI error is detected. + * @pdev: PCI device struct + * @state: PCI channel state + * + * Description: Called when a PCI error is detected. + * + * Return value: + * PCI_ERS_RESULT_NEED_RESET or PCI_ERS_RESULT_DISCONNECT + */ +static pci_ers_result_t +_scsih_pci_error_detected(struct pci_dev *pdev, pci_channel_state_t state) +{ + struct Scsi_Host *shost = pci_get_drvdata(pdev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(shost); + + printk(MPT2SAS_INFO_FMT "PCI error: detected callback, state(%d)!!\n", + ioc->name, state); + + switch (state) { + case pci_channel_io_normal: + return PCI_ERS_RESULT_CAN_RECOVER; + case pci_channel_io_frozen: + scsi_block_requests(ioc->shost); + mpt2sas_base_stop_watchdog(ioc); + mpt2sas_base_free_resources(ioc); + return PCI_ERS_RESULT_NEED_RESET; + case pci_channel_io_perm_failure: + _scsih_remove(pdev); + return PCI_ERS_RESULT_DISCONNECT; + } + return PCI_ERS_RESULT_NEED_RESET; +} + +/** + * _scsih_pci_slot_reset - Called when PCI slot has been reset. + * @pdev: PCI device struct + * + * Description: This routine is called by the pci error recovery + * code after the PCI slot has been reset, just before we + * should resume normal operations. + */ +static pci_ers_result_t +_scsih_pci_slot_reset(struct pci_dev *pdev) +{ + struct Scsi_Host *shost = pci_get_drvdata(pdev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(shost); + int rc; + + printk(MPT2SAS_INFO_FMT "PCI error: slot reset callback!!\n", + ioc->name); + + ioc->pdev = pdev; + rc = mpt2sas_base_map_resources(ioc); + if (rc) + return PCI_ERS_RESULT_DISCONNECT; + + + rc = mpt2sas_base_hard_reset_handler(ioc, CAN_SLEEP, + FORCE_BIG_HAMMER); + + printk(MPT2SAS_WARN_FMT "hard reset: %s\n", ioc->name, + (rc == 0) ? "success" : "failed"); + + if (!rc) + return PCI_ERS_RESULT_RECOVERED; + else + return PCI_ERS_RESULT_DISCONNECT; +} + +/** + * _scsih_pci_resume() - resume normal ops after PCI reset + * @pdev: pointer to PCI device + * + * Called when the error recovery driver tells us that its + * OK to resume normal operation. Use completion to allow + * halted scsi ops to resume. + */ +static void +_scsih_pci_resume(struct pci_dev *pdev) +{ + struct Scsi_Host *shost = pci_get_drvdata(pdev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(shost); + + printk(MPT2SAS_INFO_FMT "PCI error: resume callback!!\n", ioc->name); + + pci_cleanup_aer_uncorrect_error_status(pdev); + mpt2sas_base_start_watchdog(ioc); + scsi_unblock_requests(ioc->shost); +} + +/** + * _scsih_pci_mmio_enabled - Enable MMIO and dump debug registers + * @pdev: pointer to PCI device + */ +static pci_ers_result_t +_scsih_pci_mmio_enabled(struct pci_dev *pdev) +{ + struct Scsi_Host *shost = pci_get_drvdata(pdev); + struct MPT2SAS_ADAPTER *ioc = shost_priv(shost); + + printk(MPT2SAS_INFO_FMT "PCI error: mmio enabled callback!!\n", + ioc->name); + + /* TODO - dump whatever for debugging purposes */ + + /* Request a slot reset. */ + return PCI_ERS_RESULT_NEED_RESET; +} + +static struct pci_error_handlers _scsih_err_handler = { + .error_detected = _scsih_pci_error_detected, + .mmio_enabled = _scsih_pci_mmio_enabled, + .slot_reset = _scsih_pci_slot_reset, + .resume = _scsih_pci_resume, +}; static struct pci_driver scsih_driver = { .name = MPT2SAS_DRIVER_NAME, @@ -6671,6 +6788,7 @@ static struct pci_driver scsih_driver = { .probe = _scsih_probe, .remove = __devexit_p(_scsih_remove), .shutdown = _scsih_shutdown, + .err_handler = &_scsih_err_handler, #ifdef CONFIG_PM .suspend = _scsih_suspend, .resume = _scsih_resume, -- 2.7.4