hisi_sas: Some misc patches for next
Subsystems (3)
cover hisi_sas: Some misc patches for next
This small series includes the following:
- Relocate the init device functionality to a more appropriate location
- Make the phy control function synchronous
- For faulty SATA disk, disable their phy to stop spinning in the error handler
- For previous patch we need to export a libsas phy control function
Thanks!
Luo Jiaxing (2):
scsi: libsas: Export sas_phy_enable()
scsi: hisi_sas: Disable SATA disk phy for severe I_T nexus reset
failure
Xiang Chen (2):
scsi: hisi_sas: Initialise devices in .slave_alloc callback
scsi: hisi_sas: Wait for phyup in hisi_sas_control_phy()
drivers/scsi/hisi_sas/hisi_sas.h | 1 + drivers/scsi/hisi_sas/hisi_sas_main.c | 89 ++++++++++++++++++++------ drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 13 +--- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 19 ++---- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 11 +--- drivers/scsi/libsas/sas_init.c | 3 +- include/scsi/libsas.h | 1 + 7 files changed, 84 insertions(+), 53 deletions(-)
Signature · 1 line
2.26.2
PATCH 1/4 scsi: hisi_sas: Initialise devices in .slave_alloc callback
From: Xiang Chen <[email protected]>
Put the driver custom SCSI device initialize procedure in the more appropriate callback from the SCSI midlayer, and not from the libsas dev found callback.
The .slave_alloc interface is called prior to sending any IO to the device, which is what we want.
--- drivers/scsi/hisi_sas/hisi_sas.h | 1 + drivers/scsi/hisi_sas/hisi_sas_main.c | 17 ++++++++++++++--- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 2 +- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 2 +- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 2 +- 5 files changed, 18 insertions(+), 6 deletions(-)
drivers/scsi/hisi_sas/hisi_sas.h +1 -0
@@ -649,6 +649,7 @@ extern int hisi_sas_probe(struct platform_device *pdev,extern int hisi_sas_remove (struct platform_device *pdev);extern int hisi_sas_slave_configure (struct scsi_device *sdev); +extern int hisi_sas_slave_alloc (struct scsi_device *sdev);extern int hisi_sas_scan_finished (struct Scsi_Host *shost,unsigned long time);extern void hisi_sas_scan_start (struct Scsi_Host *shost);extern int hisi_sas_host_reset (struct Scsi_Host *shost,int reset_type);
drivers/scsi/hisi_sas/hisi_sas_main.c +14 -3
@@ -756,6 +756,20 @@ static int hisi_sas_init_device(struct domain_device *device)return rc; } +int hisi_sas_slave_alloc (struct scsi_device *sdev) +{ +struct domain_device *ddev; +int rc; + + rc =sas_slave_alloc (sdev); +if (rc) +return rc; + ddev =sdev_to_domain_dev (sdev); + +return hisi_sas_init_device (ddev); +} +EXPORT_SYMBOL_GPL (hisi_sas_slave_alloc); +static int hisi_sas_dev_found (struct domain_device *device) {struct hisi_hba *hisi_hba =dev_to_hisi_hba (device);
@@ -802,9 +816,6 @@ static int hisi_sas_dev_found(struct domain_device *device)dev_info (dev,"dev[%d:%x] found\n" , sas_dev->device_id, sas_dev->dev_type); - rc =hisi_sas_init_device (device); -if (rc) - goto err_out; sas_dev->dev_status =HISI_SAS_DEV_NORMAL ;return 0 ;
drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +1 -1
@@ -1771,7 +1771,7 @@ static struct scsi_host_template sht_v1_hw = { .max_sectors = SCSI_DEFAULT_MAX_SECTORS, .eh_device_reset_handler = sas_eh_device_reset_handler, .eh_target_reset_handler = sas_eh_target_reset_handler, - .slave_alloc = sas_slave_alloc, + .slave_alloc = hisi_sas_slave_alloc, .target_destroy = sas_target_destroy, .ioctl = sas_ioctl, #ifdef CONFIG_COMPAT
drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +1 -1
@@ -3584,7 +3584,7 @@ static struct scsi_host_template sht_v2_hw = { .max_sectors = SCSI_DEFAULT_MAX_SECTORS, .eh_device_reset_handler = sas_eh_device_reset_handler, .eh_target_reset_handler = sas_eh_target_reset_handler, - .slave_alloc = sas_slave_alloc, + .slave_alloc = hisi_sas_slave_alloc, .target_destroy = sas_target_destroy, .ioctl = sas_ioctl, #ifdef CONFIG_COMPAT
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +1 -1
@@ -3157,7 +3157,7 @@ static struct scsi_host_template sht_v3_hw = { .max_sectors = SCSI_DEFAULT_MAX_SECTORS, .eh_device_reset_handler = sas_eh_device_reset_handler, .eh_target_reset_handler = sas_eh_target_reset_handler, - .slave_alloc = sas_slave_alloc, + .slave_alloc = hisi_sas_slave_alloc, .target_destroy = sas_target_destroy, .ioctl = sas_ioctl, #ifdef CONFIG_COMPAT
Signature · 1 line
2.26.2
PATCH 2/4 scsi: hisi_sas: Wait for phyup in hisi_sas_control_phy()
In reply to John GarryFrom: Xiang Chen <[email protected]>
When issuing a hardreset/linkreset/phy_set_linkrate from sysfs, the phy will be disabled and re-enabled for the directly attached scenario.
It takes some time for the phy to come back up after re-enabling the phy. If the controller becomes suspended while waiting for the phy to come back, the phy up may be lost (along with the disk).
To solve the problem, wait for the phy up to occur with a timeout. Indeed this is already done in hisi_sas_debug_I_T_nexus_reset() for local phys, so just relocate the functionality to hisi_sas_control_phy().
Since the HA workqueue is drained when suspending the controller, and the phy control function is called from the same workqueue, we can guarantee that the controller will not be suspended during this period.
--- drivers/scsi/hisi_sas/hisi_sas_main.c | 43 +++++++++++++++++++------- drivers/scsi/hisi_sas/hisi_sas_v1_hw.c | 11 ++----- drivers/scsi/hisi_sas/hisi_sas_v2_hw.c | 17 ++-------- drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 9 ++---- 4 files changed, 39 insertions(+), 41 deletions(-)
drivers/scsi/hisi_sas/hisi_sas_main.c +32 -11
@@ -1146,9 +1146,17 @@ static int hisi_sas_phy_set_linkrate(struct hisi_hba *hisi_hba, int phy_no,static int hisi_sas_control_phy (struct asd_sas_phy *sas_phy,enum phy_func func,void *funcdata) { +struct hisi_sas_phy *phy =container_of (sas_phy, + struct hisi_sas_phy, sas_phy);struct sas_ha_struct *sas_ha = sas_phy->ha;struct hisi_hba *hisi_hba = sas_ha->lldd_ha; +struct device *dev = hisi_hba->dev; +DECLARE_COMPLETION_ONSTACK (completion);int phy_no = sas_phy->id; +u8 sts = phy->phy_attached; +int ret =0 ; + + phy->reset_completion = &completion;switch (func) {case PHY_FUNC_HARD_RESET :
@@ -1163,21 +1171,35 @@ static int hisi_sas_control_phy(struct asd_sas_phy *sas_phy, enum phy_func func,case PHY_FUNC_DISABLE :hisi_sas_phy_enable (hisi_hba, phy_no,0 ); -break ; + goto out;case PHY_FUNC_SET_LINK_RATE : -return hisi_sas_phy_set_linkrate (hisi_hba, phy_no, funcdata); + ret =hisi_sas_phy_set_linkrate (hisi_hba, phy_no, funcdata); +break ; +case PHY_FUNC_GET_EVENTS :if (hisi_hba->hw->get_events) { hisi_hba->hw->get_events (hisi_hba, phy_no); -break ; + goto out; } fallthrough;case PHY_FUNC_RELEASE_SPINUP_HOLD :default : -return -EOPNOTSUPP ; + ret = -EOPNOTSUPP ; + goto out; } -return 0 ; + +if (sts && !wait_for_completion_timeout (&completion,2 *HZ )) { +dev_warn (dev,"phy%d wait phyup timed out for func %d\n" , + phy_no, func); +if (phy->in_reset) + ret = -ETIMEDOUT ; + } + +out: + phy->reset_completion =NULL ; + +return ret; }static void hisi_sas_task_done (struct sas_task *task)
@@ -1783,7 +1805,6 @@ static int hisi_sas_debug_I_T_nexus_reset(struct domain_device *device)struct hisi_sas_device *sas_dev = device->lldd_dev;struct hisi_hba *hisi_hba =dev_to_hisi_hba (device);struct sas_ha_struct *sas_ha = &hisi_hba->sha; -DECLARE_COMPLETION_ONSTACK (phyreset);int rc, reset_type;if (!local_phy->enabled) {
@@ -1796,8 +1817,11 @@ static int hisi_sas_debug_I_T_nexus_reset(struct domain_device *device) sas_ha->sas_phy[local_phy->number];struct hisi_sas_phy *phy = container_of(sas_phy, struct hisi_sas_phy, sas_phy); +unsigned long flags; + +spin_lock_irqsave (&phy->lock, flags); phy->in_reset =1 ; - phy->reset_completion = &phyreset; +spin_unlock_irqrestore (&phy->lock, flags); } reset_type = (sas_dev->dev_status == HISI_SAS_DEV_INIT ||
@@ -1811,17 +1835,14 @@ static int hisi_sas_debug_I_T_nexus_reset(struct domain_device *device) sas_ha->sas_phy[local_phy->number];struct hisi_sas_phy *phy = container_of(sas_phy, struct hisi_sas_phy, sas_phy); -int ret =wait_for_completion_timeout (&phyreset, -I_T_NEXUS_RESET_PHYUP_TIMEOUT );unsigned long flags;spin_lock_irqsave (&phy->lock, flags); - phy->reset_completion =NULL ; phy->in_reset =0 ;spin_unlock_irqrestore (&phy->lock, flags);/* report PHY down if timed out */ -if (!ret) +if (rc == -ETIMEDOUT )hisi_sas_phy_down (hisi_hba, sas_phy->id,0 ,GFP_KERNEL ); }else if (sas_dev->dev_status !=HISI_SAS_DEV_INIT ) {/*
drivers/scsi/hisi_sas/hisi_sas_v1_hw.c +2 -9
@@ -1327,7 +1327,6 @@ static irqreturn_t int_phyup_v1_hw(int irq_no, void *p)u32 *frame_rcvd = (u32 *)sas_phy->frame_rcvd;struct sas_identify_frame *id = (struct sas_identify_frame *)frame_rcvd;irqreturn_t res =IRQ_HANDLED ; -unsigned long flags; irq_value =hisi_sas_phy_read32 (hisi_hba, phy_no,CHL_INT2 );if (!(irq_value &CHL_INT2_SL_PHY_ENA_MSK )) {
@@ -1380,15 +1379,9 @@ static irqreturn_t int_phyup_v1_hw(int irq_no, void *p) phy->identify.target_port_protocols =SAS_PROTOCOL_SMP ;hisi_sas_notify_phy_event (phy,HISI_PHYE_PHY_UP ); - -spin_lock_irqsave (&phy->lock, flags); -if (phy->reset_completion) { - phy->in_reset =0 ; -complete (phy->reset_completion); - } -spin_unlock_irqrestore (&phy->lock, flags); - end: +if (phy->reset_completion) +complete (phy->reset_completion);hisi_sas_phy_write32 (hisi_hba, phy_no,CHL_INT2 ,CHL_INT2_SL_PHY_ENA_MSK );
drivers/scsi/hisi_sas/hisi_sas_v2_hw.c +3 -14
@@ -2641,7 +2641,6 @@ static int phy_up_v2_hw(int phy_no, struct hisi_hba *hisi_hba)struct device *dev = hisi_hba->dev;u32 *frame_rcvd = (u32 *)sas_phy->frame_rcvd;struct sas_identify_frame *id = (struct sas_identify_frame *)frame_rcvd; -unsigned long flags;hisi_sas_phy_write32 (hisi_hba, phy_no,PHYCTRL_PHY_ENA_MSK ,1 );
@@ -2696,14 +2695,9 @@ static int phy_up_v2_hw(int phy_no, struct hisi_hba *hisi_hba)set_link_timer_quirk (hisi_hba); }hisi_sas_notify_phy_event (phy,HISI_PHYE_PHY_UP ); -spin_lock_irqsave (&phy->lock, flags); -if (phy->reset_completion) { - phy->in_reset =0 ; -complete (phy->reset_completion); - } -spin_unlock_irqrestore (&phy->lock, flags); - end: +if (phy->reset_completion) +complete (phy->reset_completion);hisi_sas_phy_write32 (hisi_hba, phy_no,CHL_INT0 ,CHL_INT0_SL_PHY_ENABLE_MSK );hisi_sas_phy_write32 (hisi_hba, phy_no,PHYCTRL_PHY_ENA_MSK ,0 );
@@ -3204,7 +3198,6 @@ static irqreturn_t sata_int_v2_hw(int irq_no, void *p)u32 ent_tmp, ent_msk, ent_int, port_id, link_rate, hard_phy_linkrate;irqreturn_t res =IRQ_HANDLED ;u8 attached_sas_addr[SAS_ADDR_SIZE ] = {0 }; -unsigned long flags;int phy_no, offset;del_timer (&phy->timer);
@@ -3280,12 +3273,8 @@ static irqreturn_t sata_int_v2_hw(int irq_no, void *p) phy->identify.target_port_protocols =SAS_PROTOCOL_SATA ;hisi_sas_notify_phy_event (phy,HISI_PHYE_PHY_UP ); -spin_lock_irqsave (&phy->lock, flags); -if (phy->reset_completion) { - phy->in_reset =0 ; +if (phy->reset_completion)complete (phy->reset_completion); - } -spin_unlock_irqrestore (&phy->lock, flags); end:hisi_sas_write32 (hisi_hba,ENT_INT_SRC1 + offset, ent_tmp);hisi_sas_write32 (hisi_hba,ENT_INT_SRC_MSK1 + offset, ent_msk);
drivers/scsi/hisi_sas/hisi_sas_v3_hw.c +2 -7
@@ -1482,7 +1482,6 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba)struct hisi_sas_phy *phy = &hisi_hba->phy[phy_no];struct asd_sas_phy *sas_phy = &phy->sas_phy;struct device *dev = hisi_hba->dev; -unsigned long flags;del_timer (&phy->timer);hisi_sas_phy_write32 (hisi_hba, phy_no,PHYCTRL_PHY_ENA_MSK ,1 );
@@ -1564,13 +1563,9 @@ static irqreturn_t phy_up_v3_hw(int phy_no, struct hisi_hba *hisi_hba) phy->phy_attached =1 ;hisi_sas_notify_phy_event (phy,HISI_PHYE_PHY_UP ); res =IRQ_HANDLED ; -spin_lock_irqsave (&phy->lock, flags); -if (phy->reset_completion) { - phy->in_reset =0 ; -complete (phy->reset_completion); - } -spin_unlock_irqrestore (&phy->lock, flags); end: +if (phy->reset_completion) +complete (phy->reset_completion);hisi_sas_phy_write32 (hisi_hba, phy_no,CHL_INT0 ,CHL_INT0_SL_PHY_ENABLE_MSK );hisi_sas_phy_write32 (hisi_hba, phy_no,PHYCTRL_PHY_ENA_MSK ,0 );
Signature · 1 line
2.26.2
PATCH 3/4 scsi: libsas: Export sas_phy_enable()
In reply to John GarryFrom: Luo Jiaxing <[email protected]>
Export sas_phy_enable(), so LLDDs can directly use it to control remote phys.
We already do this for companion function sas_phy_reset().
--- drivers/scsi/libsas/sas_init.c | 3 ++- include/scsi/libsas.h | 1 + 2 files changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/libsas/sas_init.c +2 -1
@@ -254,7 +254,7 @@ static int transport_sas_phy_reset(struct sas_phy *phy, int hard_reset) } } -static int sas_phy_enable (struct sas_phy *phy,int enable) +int sas_phy_enable (struct sas_phy *phy,int enable) {int ret;enum phy_func cmd;
@@ -286,6 +286,7 @@ static int sas_phy_enable(struct sas_phy *phy, int enable) }return ret; } +EXPORT_SYMBOL_GPL (sas_phy_enable);int sas_phy_reset (struct sas_phy *phy,int hard_reset) {
include/scsi/libsas.h +1 -0
@@ -664,6 +664,7 @@ extern void sas_suspend_ha(struct sas_ha_struct *sas_ha);int sas_set_phy_speed (struct sas_phy *phy,struct sas_phy_linkrates *rates);int sas_phy_reset (struct sas_phy *phy,int hard_reset); +int sas_phy_enable (struct sas_phy *phy,int enable);extern int sas_queuecommand (struct Scsi_Host *,struct scsi_cmnd *);extern int sas_target_alloc (struct scsi_target *);extern int sas_slave_configure (struct scsi_device *);
Signature · 1 line
2.26.2
PATCH 4/4 scsi: hisi_sas: Disable SATA disk phy for severe I_T nexus reset failure
In reply to John GarryFrom: Luo Jiaxing <[email protected]>
If the softreset fails in the I_T reset, libsas will then continue to issue a controller reset to try to recover.
However a faulty disk may cause the softreset to fail, and resetting the controller will not help this scenario. Indeed, we will just continue the cycle of error handle handling to try to recover.
So if the softreset fails upon certain conditions, just disable the phy associated with the disk. The user needs to handle this problem.
--- drivers/scsi/hisi_sas/hisi_sas_main.c | 29 ++++++++++++++++++++++----- 1 file changed, 24 insertions(+), 5 deletions(-)
drivers/scsi/hisi_sas/hisi_sas_main.c +24 -5
@@ -1870,14 +1870,33 @@ static int hisi_sas_I_T_nexus_reset(struct domain_device *device) }hisi_sas_dereg_device (hisi_hba, device); -if (dev_is_sata (device)) { + rc =hisi_sas_debug_I_T_nexus_reset (device); +if (rc ==TMF_RESP_FUNC_COMPLETE &&dev_is_sata (device)) { +struct sas_phy *local_phy; + rc =hisi_sas_softreset_ata_disk (device); -if (rc ==TMF_RESP_FUNC_FAILED ) -return TMF_RESP_FUNC_FAILED ; +switch (rc) { +case -ECOMM : + rc = -ENODEV ; +break ; +case TMF_RESP_FUNC_FAILED : +case -EMSGSIZE : +case -EIO : + local_phy =sas_get_local_phy (device); + rc =sas_phy_enable (local_phy,0 ); +if (!rc) { + local_phy->enabled =0 ; +dev_err (dev,"Disabled local phy of ATA disk %016llx due to softreset fail (%d)\n" , +SAS_ADDR (device->sas_addr), rc); + rc = -ENODEV ; + } +sas_put_local_phy (local_phy); +break ; +default : +break ; + } } - rc =hisi_sas_debug_I_T_nexus_reset (device); -if ((rc ==TMF_RESP_FUNC_COMPLETE ) || (rc == -ENODEV ))hisi_sas_release_task (hisi_hba, device);
Signature · 1 line
2.26.2
Martin K. Petersen
AppliedJohn,
Quoted 6 lines
This small series includes the following:
- Relocate the init device functionality to a more appropriate location
- Make the phy control function synchronous
- For faulty SATA disk, disable their phy to stop spinning in the error handler
- For previous patch we need to export a libsas phy control function
Applied to 5.16/scsi-staging, thanks!
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Martin K. Petersen
AppliedQuoted 8 lines · John Garry
This small series includes the following:
- Relocate the init device functionality to a more appropriate location
- Make the phy control function synchronous
- For faulty SATA disk, disable their phy to stop spinning in the error handler
- For previous patch we need to export a libsas phy control function[...]
Applied to 5.16/scsi-queue, thanks!
[1/4] scsi: hisi_sas: Initialise devices in .slave_alloc callback
https://git.kernel.org/mkp/scsi/c/36c6b7613ef1 [2/4] scsi: hisi_sas: Wait for phyup in hisi_sas_control_phy() https://git.kernel.org/mkp/scsi/c/046ab7d0f594 [3/4] scsi: libsas: Export sas_phy_enable() https://git.kernel.org/mkp/scsi/c/00aeaf329a3a [4/4] scsi: hisi_sas: Disable SATA disk phy for severe I_T nexus reset failure https://git.kernel.org/mkp/scsi/c/21c7e972475e
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering