Hi Greg, hi Sasha,
Please consider to apply both fixes to stable 4.19 tree.
Both apply cleanly to 4.19 stable queue.
Deepak Ukey (1): scsi: pm80xx: Fixed kernel panic during error recovery for SATA drive
Viswas G (1): scsi: pm80xx: panic on ncq error cleaning up the read log.
drivers/scsi/pm8001/pm8001_sas.c | 6 +++++- drivers/scsi/pm8001/pm80xx_hwi.c | 7 ++++--- drivers/scsi/pm8001/pm80xx_hwi.h | 2 ++ 3 files changed, 11 insertions(+), 4 deletions(-)
From: Viswas G Viswas.G@microsemi.com
commit 0b6df110b3d0c12562011fcd032cfb6ff16b6d56 upstream
when there's an error in 'ncq mode' the host has to read the ncq error log (10h) to clear the error state. however, the ccb that is setup for doing this doesn't setup the ccb so that the previous state is cleared. if the ccb was previously used for an IO n_elems is set and pm8001_ccb_task_free() treats this as the signal to go free a scatter-gather list (that's already been freed).
Signed-off-by: Deepak Ukey deepak.ukey@microsemi.com Signed-off-by: Viswas G Viswas.G@microsemi.com Acked-by: Jack Wang jinpu.wang@profitbricks.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com --- drivers/scsi/pm8001/pm80xx_hwi.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/pm8001/pm80xx_hwi.c b/drivers/scsi/pm8001/pm80xx_hwi.c index 8627feb80261..bd945d832eb8 100644 --- a/drivers/scsi/pm8001/pm80xx_hwi.c +++ b/drivers/scsi/pm8001/pm80xx_hwi.c @@ -1500,8 +1500,9 @@ static void pm80xx_send_read_log(struct pm8001_hba_info *pm8001_ha, ccb->ccb_tag = ccb_tag; ccb->task = task; ccb->n_elem = 0; - pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG; - pm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG; + pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG; // set this flag to indicate read log + pm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG; // set this flag to guard against 2nd RLE. Workaround + // till FW fix is available.
memset(&sata_cmd, 0, sizeof(sata_cmd)); circularQ = &pm8001_ha->inbnd_q_tbl[0];
On Mon, Mar 09, 2020 at 11:17:38AM +0100, Jack Wang wrote:
From: Viswas G Viswas.G@microsemi.com
commit 0b6df110b3d0c12562011fcd032cfb6ff16b6d56 upstream
when there's an error in 'ncq mode' the host has to read the ncq error log (10h) to clear the error state. however, the ccb that is setup for doing this doesn't setup the ccb so that the previous state is cleared. if the ccb was previously used for an IO n_elems is set and pm8001_ccb_task_free() treats this as the signal to go free a scatter-gather list (that's already been freed).
Signed-off-by: Deepak Ukey deepak.ukey@microsemi.com Signed-off-by: Viswas G Viswas.G@microsemi.com Acked-by: Jack Wang jinpu.wang@profitbricks.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com
drivers/scsi/pm8001/pm80xx_hwi.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
This commit showed up in 4.15, how can it also go into 4.19 again?
Totally confused,
greg k-h
On Mon, Mar 09, 2020 at 11:17:38AM +0100, Jack Wang wrote:
From: Viswas G Viswas.G@microsemi.com
commit 0b6df110b3d0c12562011fcd032cfb6ff16b6d56 upstream
when there's an error in 'ncq mode' the host has to read the ncq error log (10h) to clear the error state. however, the ccb that is setup for doing this doesn't setup the ccb so that the previous state is cleared. if the ccb was previously used for an IO n_elems is set and pm8001_ccb_task_free() treats this as the signal to go free a scatter-gather list (that's already been freed).
Signed-off-by: Deepak Ukey deepak.ukey@microsemi.com Signed-off-by: Viswas G Viswas.G@microsemi.com Acked-by: Jack Wang jinpu.wang@profitbricks.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com
drivers/scsi/pm8001/pm80xx_hwi.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/pm8001/pm80xx_hwi.c b/drivers/scsi/pm8001/pm80xx_hwi.c index 8627feb80261..bd945d832eb8 100644 --- a/drivers/scsi/pm8001/pm80xx_hwi.c +++ b/drivers/scsi/pm8001/pm80xx_hwi.c @@ -1500,8 +1500,9 @@ static void pm80xx_send_read_log(struct pm8001_hba_info *pm8001_ha, ccb->ccb_tag = ccb_tag; ccb->task = task; ccb->n_elem = 0;
- pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG;
- pm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG;
- pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG; // set this flag to indicate read log
- pm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG; // set this flag to guard against 2nd RLE. Workaround
// till FW fix is available.
Also, this isn't even the commit id referenced above :(
And there is trailing whitespace :(
On Tue, Mar 10, 2020 at 12:41 PM Greg KH gregkh@linuxfoundation.org wrote:
On Mon, Mar 09, 2020 at 11:17:38AM +0100, Jack Wang wrote:
From: Viswas G Viswas.G@microsemi.com
commit 0b6df110b3d0c12562011fcd032cfb6ff16b6d56 upstream
when there's an error in 'ncq mode' the host has to read the ncq error log (10h) to clear the error state. however, the ccb that is setup for doing this doesn't setup the ccb so that the previous state is cleared. if the ccb was previously used for an IO n_elems is set and pm8001_ccb_task_free() treats this as the signal to go free a scatter-gather list (that's already been freed).
Signed-off-by: Deepak Ukey deepak.ukey@microsemi.com Signed-off-by: Viswas G Viswas.G@microsemi.com Acked-by: Jack Wang jinpu.wang@profitbricks.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com
drivers/scsi/pm8001/pm80xx_hwi.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/pm8001/pm80xx_hwi.c b/drivers/scsi/pm8001/pm80xx_hwi.c index 8627feb80261..bd945d832eb8 100644 --- a/drivers/scsi/pm8001/pm80xx_hwi.c +++ b/drivers/scsi/pm8001/pm80xx_hwi.c @@ -1500,8 +1500,9 @@ static void pm80xx_send_read_log(struct pm8001_hba_info *pm8001_ha, ccb->ccb_tag = ccb_tag; ccb->task = task; ccb->n_elem = 0;
pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG;pm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG;
pm8001_ha_dev->id |= NCQ_READ_LOG_FLAG; // set this flag to indicate read logpm8001_ha_dev->id |= NCQ_2ND_RLE_FLAG; // set this flag to guard against 2nd RLE. Workaround// till FW fix is available.Also, this isn't even the commit id referenced above :(
And there is trailing whitespace :(
Sorry, I must mess it up, please ignore it.
From: Deepak Ukey deepak.ukey@microchip.com
commit 196ba6629cf95e51403337235d09742fcdc3febd upstream
Disabling the SATA drive interface cause kernel panic. When the drive Interface is disabled, device should be deregistered after aborting all pending I/Os. Also changed the port recovery timeout to 10000 ms for PM8006 controller.
Signed-off-by: Deepak Ukey deepak.ukey@microchip.com Signed-off-by: Viswas G Viswas.G@microchip.com Reviewed-by: Jack Wang jinpu.wang@cloud.ionos.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com --- drivers/scsi/pm8001/pm8001_sas.c | 6 +++++- drivers/scsi/pm8001/pm80xx_hwi.c | 2 +- drivers/scsi/pm8001/pm80xx_hwi.h | 2 ++ 3 files changed, 8 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/pm8001/pm8001_sas.c b/drivers/scsi/pm8001/pm8001_sas.c index 59feda261e08..5be4212312cb 100644 --- a/drivers/scsi/pm8001/pm8001_sas.c +++ b/drivers/scsi/pm8001/pm8001_sas.c @@ -866,6 +866,8 @@ static void pm8001_dev_gone_notify(struct domain_device *dev) spin_unlock_irqrestore(&pm8001_ha->lock, flags); pm8001_exec_internal_task_abort(pm8001_ha, pm8001_dev , dev, 1, 0); + while (pm8001_dev->running_req) + msleep(20); spin_lock_irqsave(&pm8001_ha->lock, flags); } PM8001_CHIP_DISP->dereg_dev_req(pm8001_ha, device_id); @@ -1238,8 +1240,10 @@ int pm8001_abort_task(struct sas_task *task) PM8001_MSG_DBG(pm8001_ha, pm8001_printk("Waiting for Port reset\n")); wait_for_completion(&completion_reset); - if (phy->port_reset_status) + if (phy->port_reset_status) { + pm8001_dev_gone_notify(dev); goto out; + }
/* * 4. SATA Abort ALL diff --git a/drivers/scsi/pm8001/pm80xx_hwi.c b/drivers/scsi/pm8001/pm80xx_hwi.c index bd945d832eb8..fd5f9892f3ac 100644 --- a/drivers/scsi/pm8001/pm80xx_hwi.c +++ b/drivers/scsi/pm8001/pm80xx_hwi.c @@ -604,7 +604,7 @@ static void update_main_config_table(struct pm8001_hba_info *pm8001_ha) pm8001_ha->main_cfg_tbl.pm80xx_tbl.port_recovery_timer &= 0x0000ffff; pm8001_ha->main_cfg_tbl.pm80xx_tbl.port_recovery_timer |= - 0x140000; + CHIP_8006_PORT_RECOVERY_TIMEOUT; } pm8001_mw32(address, MAIN_PORT_RECOVERY_TIMER, pm8001_ha->main_cfg_tbl.pm80xx_tbl.port_recovery_timer); diff --git a/drivers/scsi/pm8001/pm80xx_hwi.h b/drivers/scsi/pm8001/pm80xx_hwi.h index 7dd2699d0efb..bbe1747234ff 100644 --- a/drivers/scsi/pm8001/pm80xx_hwi.h +++ b/drivers/scsi/pm8001/pm80xx_hwi.h @@ -228,6 +228,8 @@ #define SAS_MAX_AIP 0x200000 #define IT_NEXUS_TIMEOUT 0x7D0 #define PORT_RECOVERY_TIMEOUT ((IT_NEXUS_TIMEOUT/100) + 30) +/* Port recovery timeout, 10000 ms for PM8006 controller */ +#define CHIP_8006_PORT_RECOVERY_TIMEOUT 0x640000
#ifdef __LITTLE_ENDIAN_BITFIELD struct sas_identify_frame_local {
On Mon, Mar 09, 2020 at 11:17:39AM +0100, Jack Wang wrote:
From: Deepak Ukey deepak.ukey@microchip.com
commit 196ba6629cf95e51403337235d09742fcdc3febd upstream
Disabling the SATA drive interface cause kernel panic. When the drive Interface is disabled, device should be deregistered after aborting all pending I/Os. Also changed the port recovery timeout to 10000 ms for PM8006 controller.
Signed-off-by: Deepak Ukey deepak.ukey@microchip.com Signed-off-by: Viswas G Viswas.G@microchip.com Reviewed-by: Jack Wang jinpu.wang@cloud.ionos.com Signed-off-by: Martin K. Petersen martin.petersen@oracle.com
drivers/scsi/pm8001/pm8001_sas.c | 6 +++++- drivers/scsi/pm8001/pm80xx_hwi.c | 2 +- drivers/scsi/pm8001/pm80xx_hwi.h | 2 ++ 3 files changed, 8 insertions(+), 2 deletions(-)
Now queued up, thanks.
greg k-h
linux-stable-mirror@lists.linaro.org
-
Greg KH -
Jack Wang -
Jinpu Wang