linux-kernel-mentees.lists.linuxfoundation.org archive mirror
 help / color / mirror / Atom feed
* [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management
@ 2020-11-02 16:47 Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume Vaibhav Gupta
                   ` (29 more replies)
  0 siblings, 30 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Linux Kernel Mentee: Remove Legacy Power Management.

The purpose of this patch series is to upgrade power management in xxxxxxxx
drivers. This has been done by upgrading .suspend() and .resume() callbacks.

The upgrade makes sure that the involvement of PCI Core does not change the
order of operations executed in a driver. Thus, does not change its behavior.

In general, drivers with legacy PM, .suspend() and .resume() make use of PCI
helper functions like pci_enable/disable_device_mem(), pci_set_power_state(),
pci_save/restore_state(), pci_enable/disable_device(), etc. to complete
their job.

The conversion requires the removal of those function calls, change the
callbacks' definition accordingly and make use of dev_pm_ops structure.

All patches are compile-tested only.

Test tools:
    - Compiler: gcc (GCC) 10.2.0
    - allmodconfig build: make -j$(nproc) W=1 all

Vaibhav Gupta (29):
  scsi: megaraid_sas: Drop PCI wakeup calls from .resume
  scsi: megaraid_sas: use generic power management
  scsi: megaraid_sas: update function description
  scsi: aacraid: Drop pci_enable_wake() from .resume
  scsi: aacraid: use generic power management
  scsi: aic7xxx: use generic power management
  scsi: aic79xx: use generic power management
  scsi: arcmsr: Drop PCI wakeup calls from .resume
  scsi: arcmsr: use generic power management
  scsi: esas2r: Drop PCI Wakeup calls from .resume
  scsi: esas2r: use generic power management
  scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume
  scsi: hisi_sas_v3_hw: Don't use PCI helper functions
  scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm
  scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume
  scsi: mpt3sas_scsih: use generic power management
  scsi: lpfc: use generic power management
  scsi: pm_8001: Drop PCI Wakeup calls from .resume
  scsi: pm_8001: use generic power management
  scsi: hpsa: use generic power management
  scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume
  scsi: 3w-9xxx: use generic power management
  scsi: 3w-sas: Drop PCI Wakeup calls from .resume
  scsi: 3w-sas: use generic power management
  scsi: mvumi: Drop PCI Wakeup calls from .resume
  scsi: mvumi: use generic power management
  scsi: mvumi: update function description
  scsi: pmcraid: Drop PCI Wakeup calls from .resume
  scsi: pmcraid: use generic power management

 drivers/scsi/3w-9xxx.c                    |  30 ++-----
 drivers/scsi/3w-sas.c                     |  32 ++-----
 drivers/scsi/aacraid/linit.c              |  34 ++------
 drivers/scsi/aic7xxx/aic79xx.h            |  12 +--
 drivers/scsi/aic7xxx/aic79xx_core.c       |   8 +-
 drivers/scsi/aic7xxx/aic79xx_osm_pci.c    |  43 +++-------
 drivers/scsi/aic7xxx/aic79xx_pci.c        |   6 +-
 drivers/scsi/aic7xxx/aic7xxx.h            |  10 +--
 drivers/scsi/aic7xxx/aic7xxx_core.c       |   6 +-
 drivers/scsi/aic7xxx/aic7xxx_osm_pci.c    |  46 +++-------
 drivers/scsi/aic7xxx/aic7xxx_pci.c        |   4 +-
 drivers/scsi/arcmsr/arcmsr_hba.c          |  29 ++-----
 drivers/scsi/esas2r/esas2r.h              |   5 +-
 drivers/scsi/esas2r/esas2r_init.c         |  48 +++--------
 drivers/scsi/esas2r/esas2r_main.c         |   3 +-
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c    |  41 ++-------
 drivers/scsi/hpsa.c                       |  12 +--
 drivers/scsi/lpfc/lpfc_init.c             | 100 +++++++---------------
 drivers/scsi/megaraid/megaraid_sas_base.c |  54 +++---------
 drivers/scsi/mpt3sas/mpt3sas_scsih.c      |  35 +++-----
 drivers/scsi/mvumi.c                      |  49 +++--------
 drivers/scsi/pm8001/pm8001_init.c         |  46 ++++------
 drivers/scsi/pmcraid.c                    |  44 +++-------
 23 files changed, 193 insertions(+), 504 deletions(-)

-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 02/29] scsi: megaraid_sas: use generic power management Vaibhav Gupta
                   ` (28 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in megasas_resume(), and
there is no corresponding pci_enable_wake(...., true) in megasas_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
megasas_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/megaraid/megaraid_sas_base.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c
index 41cd66fc7d81..47ffdada541d 100644
--- a/drivers/scsi/megaraid/megaraid_sas_base.c
+++ b/drivers/scsi/megaraid/megaraid_sas_base.c
@@ -7668,7 +7668,6 @@ megasas_resume(struct pci_dev *pdev)
 
 	host = instance->host;
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 
 	dev_info(&pdev->dev, "%s is called\n", __func__);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 02/29] scsi: megaraid_sas: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 03/29] scsi: megaraid_sas: update function description Vaibhav Gupta
                   ` (27 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/megaraid/megaraid_sas_base.c | 51 ++++++-----------------
 1 file changed, 13 insertions(+), 38 deletions(-)

diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c
index 47ffdada541d..bd330ea4063a 100644
--- a/drivers/scsi/megaraid/megaraid_sas_base.c
+++ b/drivers/scsi/megaraid/megaraid_sas_base.c
@@ -7593,25 +7593,23 @@ static void megasas_shutdown_controller(struct megasas_instance *instance,
 	megasas_return_cmd(instance, cmd);
 }
 
-#ifdef CONFIG_PM
 /**
  * megasas_suspend -	driver suspend entry point
- * @pdev:		PCI device structure
- * @state:		PCI power state to suspend routine
+ * @dev:		Device structure
  */
-static int
-megasas_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused
+megasas_suspend(struct device *dev)
 {
 	struct megasas_instance *instance;
 
-	instance = pci_get_drvdata(pdev);
+	instance = dev_get_drvdata(dev);
 
 	if (!instance)
 		return 0;
 
 	instance->unload = 1;
 
-	dev_info(&pdev->dev, "%s is called\n", __func__);
+	dev_info(dev, "%s is called\n", __func__);
 
 	/* Shutdown SR-IOV heartbeat timer */
 	if (instance->requestorId && !instance->skip_heartbeat_timer_del)
@@ -7641,47 +7639,29 @@ megasas_suspend(struct pci_dev *pdev, pm_message_t state)
 	if (instance->msix_vectors)
 		pci_free_irq_vectors(instance->pdev);
 
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
-
 	return 0;
 }
 
 /**
  * megasas_resume-      driver resume entry point
- * @pdev:               PCI device structure
+ * @dev:		Device structure
  */
-static int
-megasas_resume(struct pci_dev *pdev)
+static int __maybe_unused
+megasas_resume(struct device *dev)
 {
 	int rval;
 	struct Scsi_Host *host;
 	struct megasas_instance *instance;
 	u32 status_reg;
 
-	instance = pci_get_drvdata(pdev);
+	instance = dev_get_drvdata(dev);
 
 	if (!instance)
 		return 0;
 
 	host = instance->host;
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	dev_info(&pdev->dev, "%s is called\n", __func__);
-	/*
-	 * PCI prepping: enable device set bus mastering and dma mask
-	 */
-	rval = pci_enable_device_mem(pdev);
 
-	if (rval) {
-		dev_err(&pdev->dev, "Enable device failed\n");
-		return rval;
-	}
-
-	pci_set_master(pdev);
+	dev_info(dev, "%s is called\n", __func__);
 
 	/*
 	 * We expect the FW state to be READY
@@ -7807,14 +7787,8 @@ megasas_resume(struct pci_dev *pdev)
 fail_set_dma_mask:
 fail_ready_state:
 
-	pci_disable_device(pdev);
-
 	return -ENODEV;
 }
-#else
-#define megasas_suspend	NULL
-#define megasas_resume	NULL
-#endif
 
 static inline int
 megasas_wait_for_adapter_operational(struct megasas_instance *instance)
@@ -8572,6 +8546,8 @@ static const struct file_operations megasas_mgmt_fops = {
 	.llseek = noop_llseek,
 };
 
+static SIMPLE_DEV_PM_OPS(megasas_pm_ops, megasas_suspend, megasas_resume);
+
 /*
  * PCI hotplug support registration structure
  */
@@ -8581,8 +8557,7 @@ static struct pci_driver megasas_pci_driver = {
 	.id_table = megasas_pci_table,
 	.probe = megasas_probe_one,
 	.remove = megasas_detach_one,
-	.suspend = megasas_suspend,
-	.resume = megasas_resume,
+	.driver.pm = &megasas_pm_ops,
 	.shutdown = megasas_shutdown,
 };
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 03/29] scsi: megaraid_sas: update function description
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 02/29] scsi: megaraid_sas: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 04/29] scsi: aacraid: Drop pci_enable_wake() from .resume Vaibhav Gupta
                   ` (26 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Function parameter 'pdev 'is described as Generic Device Structure. It is
a PCI device structure.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/megaraid/megaraid_sas_base.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c
index bd330ea4063a..7384c8e9149d 100644
--- a/drivers/scsi/megaraid/megaraid_sas_base.c
+++ b/drivers/scsi/megaraid/megaraid_sas_base.c
@@ -7958,7 +7958,7 @@ static void megasas_detach_one(struct pci_dev *pdev)
 
 /**
  * megasas_shutdown -	Shutdown entry point
- * @pdev:		Generic device structure
+ * @pdev:		PCI device structure
  */
 static void megasas_shutdown(struct pci_dev *pdev)
 {
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 04/29] scsi: aacraid: Drop pci_enable_wake() from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (2 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 03/29] scsi: megaraid_sas: update function description Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 05/29] scsi: aacraid: use generic power management Vaibhav Gupta
                   ` (25 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in aac_resume(), and there is
no corresponding pci_enable_wake(...., true) in aac_suspend(). Either it
should do enable-wake the device in .suspend() or should not invoke
pci_enable_wake() at all.

Concluding that this is a bug and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
aac_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/aacraid/linit.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/aacraid/linit.c b/drivers/scsi/aacraid/linit.c
index 8f3772480582..8c4dcb5ab329 100644
--- a/drivers/scsi/aacraid/linit.c
+++ b/drivers/scsi/aacraid/linit.c
@@ -1938,7 +1938,6 @@ static int aac_resume(struct pci_dev *pdev)
 	int r;
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 	r = pci_enable_device(pdev);
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 05/29] scsi: aacraid: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (3 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 04/29] scsi: aacraid: Drop pci_enable_wake() from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 06/29] scsi: aic7xxx: " Vaibhav Gupta
                   ` (24 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Acked-by: Balsundar P <balsundar.p@microchip.com>
---
 drivers/scsi/aacraid/linit.c | 33 +++++++--------------------------
 1 file changed, 7 insertions(+), 26 deletions(-)

diff --git a/drivers/scsi/aacraid/linit.c b/drivers/scsi/aacraid/linit.c
index 8c4dcb5ab329..5177d387854a 100644
--- a/drivers/scsi/aacraid/linit.c
+++ b/drivers/scsi/aacraid/linit.c
@@ -1910,11 +1910,9 @@ static int aac_acquire_resources(struct aac_dev *dev)
 
 }
 
-#if (defined(CONFIG_PM))
-static int aac_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused aac_suspend(struct device *dev)
 {
-
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev);
 	struct aac_dev *aac = (struct aac_dev *)shost->hostdata;
 
 	scsi_host_block(shost);
@@ -1923,28 +1921,14 @@ static int aac_suspend(struct pci_dev *pdev, pm_message_t state)
 
 	aac_release_resources(aac);
 
-	pci_set_drvdata(pdev, shost);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
-
 	return 0;
 }
 
-static int aac_resume(struct pci_dev *pdev)
+static int __maybe_unused aac_resume(struct device *dev)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev);
 	struct aac_dev *aac = (struct aac_dev *)shost->hostdata;
-	int r;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	r = pci_enable_device(pdev);
-
-	if (r)
-		goto fail_device;
 
-	pci_set_master(pdev);
 	if (aac_acquire_resources(aac))
 		goto fail_device;
 	/*
@@ -1959,10 +1943,8 @@ static int aac_resume(struct pci_dev *pdev)
 fail_device:
 	printk(KERN_INFO "%s%d: resume failed.\n", aac->name, aac->id);
 	scsi_host_put(shost);
-	pci_disable_device(pdev);
 	return -ENODEV;
 }
-#endif
 
 static void aac_shutdown(struct pci_dev *dev)
 {
@@ -2107,15 +2089,14 @@ static struct pci_error_handlers aac_pci_err_handler = {
 	.resume			= aac_pci_resume,
 };
 
+static SIMPLE_DEV_PM_OPS(aac_pm_ops, aac_suspend, aac_resume);
+
 static struct pci_driver aac_pci_driver = {
 	.name		= AAC_DRIVERNAME,
 	.id_table	= aac_pci_tbl,
 	.probe		= aac_probe_one,
 	.remove		= aac_remove_one,
-#if (defined(CONFIG_PM))
-	.suspend	= aac_suspend,
-	.resume		= aac_resume,
-#endif
+	.driver.pm      = &aac_pm_ops,
 	.shutdown	= aac_shutdown,
 	.err_handler    = &aac_pci_err_handler,
 };
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 06/29] scsi: aic7xxx: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (4 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 05/29] scsi: aacraid: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 07/29] scsi: aic79xx: " Vaibhav Gupta
                   ` (23 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

use generic power management

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/aic7xxx/aic7xxx.h         | 10 ++----
 drivers/scsi/aic7xxx/aic7xxx_core.c    |  6 ++--
 drivers/scsi/aic7xxx/aic7xxx_osm_pci.c | 46 ++++++--------------------
 drivers/scsi/aic7xxx/aic7xxx_pci.c     |  4 +--
 4 files changed, 17 insertions(+), 49 deletions(-)

diff --git a/drivers/scsi/aic7xxx/aic7xxx.h b/drivers/scsi/aic7xxx/aic7xxx.h
index 88b90f9806c9..11a09798e6b5 100644
--- a/drivers/scsi/aic7xxx/aic7xxx.h
+++ b/drivers/scsi/aic7xxx/aic7xxx.h
@@ -1134,9 +1134,7 @@ const struct ahc_pci_identity	*ahc_find_pci_device(ahc_dev_softc_t);
 int			 ahc_pci_config(struct ahc_softc *,
 					const struct ahc_pci_identity *);
 int			 ahc_pci_test_register_access(struct ahc_softc *);
-#ifdef CONFIG_PM
-void			 ahc_pci_resume(struct ahc_softc *ahc);
-#endif
+void __maybe_unused	 ahc_pci_resume(struct ahc_softc *ahc);
 
 /*************************** EISA/VL Front End ********************************/
 struct aic7770_identity *aic7770_find_device(uint32_t);
@@ -1160,10 +1158,8 @@ int			 ahc_chip_init(struct ahc_softc *ahc);
 int			 ahc_init(struct ahc_softc *ahc);
 void			 ahc_intr_enable(struct ahc_softc *ahc, int enable);
 void			 ahc_pause_and_flushwork(struct ahc_softc *ahc);
-#ifdef CONFIG_PM
-int			 ahc_suspend(struct ahc_softc *ahc); 
-int			 ahc_resume(struct ahc_softc *ahc);
-#endif
+int __maybe_unused	 ahc_suspend(struct ahc_softc *ahc);
+int __maybe_unused	 ahc_resume(struct ahc_softc *ahc);
 void			 ahc_set_unit(struct ahc_softc *, int);
 void			 ahc_set_name(struct ahc_softc *, char *);
 void			 ahc_free(struct ahc_softc *ahc);
diff --git a/drivers/scsi/aic7xxx/aic7xxx_core.c b/drivers/scsi/aic7xxx/aic7xxx_core.c
index 725bb7f58054..4ef7af8c0f55 100644
--- a/drivers/scsi/aic7xxx/aic7xxx_core.c
+++ b/drivers/scsi/aic7xxx/aic7xxx_core.c
@@ -5590,8 +5590,7 @@ ahc_pause_and_flushwork(struct ahc_softc *ahc)
 	ahc->flags &= ~AHC_ALL_INTERRUPTS;
 }
 
-#ifdef CONFIG_PM
-int
+int __maybe_unused
 ahc_suspend(struct ahc_softc *ahc)
 {
 
@@ -5617,7 +5616,7 @@ ahc_suspend(struct ahc_softc *ahc)
 	return (0);
 }
 
-int
+int __maybe_unused
 ahc_resume(struct ahc_softc *ahc)
 {
 
@@ -5626,7 +5625,6 @@ ahc_resume(struct ahc_softc *ahc)
 	ahc_restart(ahc);
 	return (0);
 }
-#endif
 /************************** Busy Target Table *********************************/
 /*
  * Return the untagged transaction id for a given target/channel lun.
diff --git a/drivers/scsi/aic7xxx/aic7xxx_osm_pci.c b/drivers/scsi/aic7xxx/aic7xxx_osm_pci.c
index 9b293b1f0b71..a07e94fac673 100644
--- a/drivers/scsi/aic7xxx/aic7xxx_osm_pci.c
+++ b/drivers/scsi/aic7xxx/aic7xxx_osm_pci.c
@@ -121,47 +121,23 @@ static const struct pci_device_id ahc_linux_pci_id_table[] = {
 
 MODULE_DEVICE_TABLE(pci, ahc_linux_pci_id_table);
 
-#ifdef CONFIG_PM
-static int
-ahc_linux_pci_dev_suspend(struct pci_dev *pdev, pm_message_t mesg)
+static int __maybe_unused
+ahc_linux_pci_dev_suspend(struct device *dev)
 {
-	struct ahc_softc *ahc = pci_get_drvdata(pdev);
-	int rc;
-
-	if ((rc = ahc_suspend(ahc)))
-		return rc;
+	struct ahc_softc *ahc = dev_get_drvdata(dev);
 
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-
-	if (mesg.event & PM_EVENT_SLEEP)
-		pci_set_power_state(pdev, PCI_D3hot);
-
-	return rc;
+	return ahc_suspend(ahc);
 }
 
-static int
-ahc_linux_pci_dev_resume(struct pci_dev *pdev)
+static int __maybe_unused
+ahc_linux_pci_dev_resume(struct device *dev)
 {
-	struct ahc_softc *ahc = pci_get_drvdata(pdev);
-	int rc;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	if ((rc = pci_enable_device(pdev))) {
-		dev_printk(KERN_ERR, &pdev->dev,
-			   "failed to enable device after resume (%d)\n", rc);
-		return rc;
-	}
-
-	pci_set_master(pdev);
+	struct ahc_softc *ahc = dev_get_drvdata(dev);
 
 	ahc_pci_resume(ahc);
 
 	return (ahc_resume(ahc));
 }
-#endif
 
 static void
 ahc_linux_pci_dev_remove(struct pci_dev *pdev)
@@ -319,14 +295,14 @@ ahc_pci_write_config(ahc_dev_softc_t pci, int reg, uint32_t value, int width)
 	}
 }
 
+static SIMPLE_DEV_PM_OPS(ahc_linux_pci_dev_pm_ops,
+			 ahc_linux_pci_dev_suspend,
+			 ahc_linux_pci_dev_resume);
 
 static struct pci_driver aic7xxx_pci_driver = {
 	.name		= "aic7xxx",
 	.probe		= ahc_linux_pci_dev_probe,
-#ifdef CONFIG_PM
-	.suspend	= ahc_linux_pci_dev_suspend,
-	.resume		= ahc_linux_pci_dev_resume,
-#endif
+	.driver.pm	= &ahc_linux_pci_dev_pm_ops,
 	.remove		= ahc_linux_pci_dev_remove,
 	.id_table	= ahc_linux_pci_id_table
 };
diff --git a/drivers/scsi/aic7xxx/aic7xxx_pci.c b/drivers/scsi/aic7xxx/aic7xxx_pci.c
index 656f680c7802..dab3a6d12c4d 100644
--- a/drivers/scsi/aic7xxx/aic7xxx_pci.c
+++ b/drivers/scsi/aic7xxx/aic7xxx_pci.c
@@ -2008,8 +2008,7 @@ ahc_pci_chip_init(struct ahc_softc *ahc)
 	return (ahc_chip_init(ahc));
 }
 
-#ifdef CONFIG_PM
-void
+void __maybe_unused
 ahc_pci_resume(struct ahc_softc *ahc)
 {
 	/*
@@ -2040,7 +2039,6 @@ ahc_pci_resume(struct ahc_softc *ahc)
 		ahc_release_seeprom(&sd);
 	}
 }
-#endif
 
 static int
 ahc_aic785X_setup(struct ahc_softc *ahc)
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 07/29] scsi: aic79xx: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (5 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 06/29] scsi: aic7xxx: " Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 08/29] scsi: arcmsr: Drop PCI wakeup calls from .resume Vaibhav Gupta
                   ` (22 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/aic7xxx/aic79xx.h         | 12 +++----
 drivers/scsi/aic7xxx/aic79xx_core.c    |  8 ++---
 drivers/scsi/aic7xxx/aic79xx_osm_pci.c | 43 +++++++-------------------
 drivers/scsi/aic7xxx/aic79xx_pci.c     |  6 ++--
 4 files changed, 20 insertions(+), 49 deletions(-)

diff --git a/drivers/scsi/aic7xxx/aic79xx.h b/drivers/scsi/aic7xxx/aic79xx.h
index 9a515551641c..dd5dfd4f30a5 100644
--- a/drivers/scsi/aic7xxx/aic79xx.h
+++ b/drivers/scsi/aic7xxx/aic79xx.h
@@ -1330,10 +1330,8 @@ const struct	ahd_pci_identity *ahd_find_pci_device(ahd_dev_softc_t);
 int			  ahd_pci_config(struct ahd_softc *,
 					 const struct ahd_pci_identity *);
 int	ahd_pci_test_register_access(struct ahd_softc *);
-#ifdef CONFIG_PM
-void	ahd_pci_suspend(struct ahd_softc *);
-void	ahd_pci_resume(struct ahd_softc *);
-#endif
+void __maybe_unused	ahd_pci_suspend(struct ahd_softc *);
+void __maybe_unused	ahd_pci_resume(struct ahd_softc *);
 
 /************************** SCB and SCB queue management **********************/
 void		ahd_qinfifo_requeue_tail(struct ahd_softc *ahd,
@@ -1344,10 +1342,8 @@ struct ahd_softc	*ahd_alloc(void *platform_arg, char *name);
 int			 ahd_softc_init(struct ahd_softc *);
 void			 ahd_controller_info(struct ahd_softc *ahd, char *buf);
 int			 ahd_init(struct ahd_softc *ahd);
-#ifdef CONFIG_PM
-int			 ahd_suspend(struct ahd_softc *ahd);
-void			 ahd_resume(struct ahd_softc *ahd);
-#endif
+int __maybe_unused	 ahd_suspend(struct ahd_softc *ahd);
+void __maybe_unused	 ahd_resume(struct ahd_softc *ahd);
 int			 ahd_default_config(struct ahd_softc *ahd);
 int			 ahd_parse_vpddata(struct ahd_softc *ahd,
 					   struct vpd_config *vpd);
diff --git a/drivers/scsi/aic7xxx/aic79xx_core.c b/drivers/scsi/aic7xxx/aic79xx_core.c
index 98b02e7d38bb..78560a85b1e3 100644
--- a/drivers/scsi/aic7xxx/aic79xx_core.c
+++ b/drivers/scsi/aic7xxx/aic79xx_core.c
@@ -7866,11 +7866,9 @@ ahd_pause_and_flushwork(struct ahd_softc *ahd)
 	ahd->flags &= ~AHD_ALL_INTERRUPTS;
 }
 
-#ifdef CONFIG_PM
-int
+int __maybe_unused
 ahd_suspend(struct ahd_softc *ahd)
 {
-
 	ahd_pause_and_flushwork(ahd);
 
 	if (LIST_FIRST(&ahd->pending_scbs) != NULL) {
@@ -7881,15 +7879,13 @@ ahd_suspend(struct ahd_softc *ahd)
 	return (0);
 }
 
-void
+void __maybe_unused
 ahd_resume(struct ahd_softc *ahd)
 {
-
 	ahd_reset(ahd, /*reinit*/TRUE);
 	ahd_intr_enable(ahd, TRUE); 
 	ahd_restart(ahd);
 }
-#endif
 
 /************************** Busy Target Table *********************************/
 /*
diff --git a/drivers/scsi/aic7xxx/aic79xx_osm_pci.c b/drivers/scsi/aic7xxx/aic79xx_osm_pci.c
index 8b891a05d9e7..07b670b80f1b 100644
--- a/drivers/scsi/aic7xxx/aic79xx_osm_pci.c
+++ b/drivers/scsi/aic7xxx/aic79xx_osm_pci.c
@@ -74,11 +74,10 @@ static const struct pci_device_id ahd_linux_pci_id_table[] = {
 
 MODULE_DEVICE_TABLE(pci, ahd_linux_pci_id_table);
 
-#ifdef CONFIG_PM
-static int
-ahd_linux_pci_dev_suspend(struct pci_dev *pdev, pm_message_t mesg)
+static int __maybe_unused
+ahd_linux_pci_dev_suspend(struct device *dev)
 {
-	struct ahd_softc *ahd = pci_get_drvdata(pdev);
+	struct ahd_softc *ahd = dev_get_drvdata(dev);
 	int rc;
 
 	if ((rc = ahd_suspend(ahd)))
@@ -86,39 +85,20 @@ ahd_linux_pci_dev_suspend(struct pci_dev *pdev, pm_message_t mesg)
 
 	ahd_pci_suspend(ahd);
 
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-
-	if (mesg.event & PM_EVENT_SLEEP)
-		pci_set_power_state(pdev, PCI_D3hot);
-
 	return rc;
 }
 
-static int
-ahd_linux_pci_dev_resume(struct pci_dev *pdev)
+static int __maybe_unused
+ahd_linux_pci_dev_resume(struct device *dev)
 {
-	struct ahd_softc *ahd = pci_get_drvdata(pdev);
-	int rc;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	if ((rc = pci_enable_device(pdev))) {
-		dev_printk(KERN_ERR, &pdev->dev,
-			   "failed to enable device after resume (%d)\n", rc);
-		return rc;
-	}
-
-	pci_set_master(pdev);
+	struct ahd_softc *ahd = dev_get_drvdata(dev);
 
 	ahd_pci_resume(ahd);
 
 	ahd_resume(ahd);
 
-	return rc;
+	return 0;
 }
-#endif
 
 static void
 ahd_linux_pci_dev_remove(struct pci_dev *pdev)
@@ -224,13 +204,14 @@ ahd_linux_pci_dev_probe(struct pci_dev *pdev, const struct pci_device_id *ent)
 	return (0);
 }
 
+static SIMPLE_DEV_PM_OPS(ahd_linux_pci_dev_pm_ops,
+			 ahd_linux_pci_dev_suspend,
+			 ahd_linux_pci_dev_resume);
+
 static struct pci_driver aic79xx_pci_driver = {
 	.name		= "aic79xx",
 	.probe		= ahd_linux_pci_dev_probe,
-#ifdef CONFIG_PM
-	.suspend	= ahd_linux_pci_dev_suspend,
-	.resume		= ahd_linux_pci_dev_resume,
-#endif
+	.driver.pm	= &ahd_linux_pci_dev_pm_ops,
 	.remove		= ahd_linux_pci_dev_remove,
 	.id_table	= ahd_linux_pci_id_table
 };
diff --git a/drivers/scsi/aic7xxx/aic79xx_pci.c b/drivers/scsi/aic7xxx/aic79xx_pci.c
index 8397ae93f7dd..2f0bdb9225a4 100644
--- a/drivers/scsi/aic7xxx/aic79xx_pci.c
+++ b/drivers/scsi/aic7xxx/aic79xx_pci.c
@@ -377,8 +377,7 @@ ahd_pci_config(struct ahd_softc *ahd, const struct ahd_pci_identity *entry)
 	return ahd_pci_map_int(ahd);
 }
 
-#ifdef CONFIG_PM
-void
+void __maybe_unused
 ahd_pci_suspend(struct ahd_softc *ahd)
 {
 	/*
@@ -394,7 +393,7 @@ ahd_pci_suspend(struct ahd_softc *ahd)
 
 }
 
-void
+void __maybe_unused
 ahd_pci_resume(struct ahd_softc *ahd)
 {
 	ahd_pci_write_config(ahd->dev_softc, DEVCONFIG,
@@ -404,7 +403,6 @@ ahd_pci_resume(struct ahd_softc *ahd)
 	ahd_pci_write_config(ahd->dev_softc, CSIZE_LATTIME,
 			     ahd->suspend_state.pci_state.csize_lattime, /*bytes*/1);
 }
-#endif
 
 /*
  * Perform some simple tests that should catch situations where
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 08/29] scsi: arcmsr: Drop PCI wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (6 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 07/29] scsi: aic79xx: " Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 09/29] scsi: arcmsr: use generic power management Vaibhav Gupta
                   ` (21 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in arcmsr_resume(), and
there is no corresponding pci_enable_wake(...., true) in arcmsr_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
arcmsr_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/arcmsr/arcmsr_hba.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/arcmsr/arcmsr_hba.c b/drivers/scsi/arcmsr/arcmsr_hba.c
index e4fdb473b990..c7ba4cbd197b 100644
--- a/drivers/scsi/arcmsr/arcmsr_hba.c
+++ b/drivers/scsi/arcmsr/arcmsr_hba.c
@@ -1154,7 +1154,6 @@ static int arcmsr_resume(struct pci_dev *pdev)
 		(struct AdapterControlBlock *)host->hostdata;
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 	if (pci_enable_device(pdev)) {
 		pr_warn("%s: pci_enable_device error\n", __func__);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 09/29] scsi: arcmsr: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (7 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 08/29] scsi: arcmsr: Drop PCI wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 10/29] scsi: esas2r: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (20 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/arcmsr/arcmsr_hba.c | 28 +++++++++-------------------
 1 file changed, 9 insertions(+), 19 deletions(-)

diff --git a/drivers/scsi/arcmsr/arcmsr_hba.c b/drivers/scsi/arcmsr/arcmsr_hba.c
index c7ba4cbd197b..907f5af3bbd4 100644
--- a/drivers/scsi/arcmsr/arcmsr_hba.c
+++ b/drivers/scsi/arcmsr/arcmsr_hba.c
@@ -113,8 +113,8 @@ static int arcmsr_bios_param(struct scsi_device *sdev,
 static int arcmsr_queue_command(struct Scsi_Host *h, struct scsi_cmnd *cmd);
 static int arcmsr_probe(struct pci_dev *pdev,
 				const struct pci_device_id *id);
-static int arcmsr_suspend(struct pci_dev *pdev, pm_message_t state);
-static int arcmsr_resume(struct pci_dev *pdev);
+static int __maybe_unused arcmsr_suspend(struct device *dev);
+static int __maybe_unused arcmsr_resume(struct device *dev);
 static void arcmsr_remove(struct pci_dev *pdev);
 static void arcmsr_shutdown(struct pci_dev *pdev);
 static void arcmsr_iop_init(struct AdapterControlBlock *acb);
@@ -216,13 +216,14 @@ static struct pci_device_id arcmsr_device_id_table[] = {
 };
 MODULE_DEVICE_TABLE(pci, arcmsr_device_id_table);
 
+static SIMPLE_DEV_PM_OPS(arcmsr_pm_ops, arcmsr_suspend, arcmsr_resume);
+
 static struct pci_driver arcmsr_pci_driver = {
 	.name			= "arcmsr",
 	.id_table		= arcmsr_device_id_table,
 	.probe			= arcmsr_probe,
 	.remove			= arcmsr_remove,
-	.suspend		= arcmsr_suspend,
-	.resume			= arcmsr_resume,
+	.driver.pm		= &arcmsr_pm_ops,
 	.shutdown		= arcmsr_shutdown,
 };
 /*
@@ -1126,8 +1127,9 @@ static void arcmsr_free_irq(struct pci_dev *pdev,
 	pci_free_irq_vectors(pdev);
 }
 
-static int arcmsr_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused arcmsr_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *host = pci_get_drvdata(pdev);
 	struct AdapterControlBlock *acb =
 		(struct AdapterControlBlock *)host->hostdata;
@@ -1140,28 +1142,18 @@ static int arcmsr_suspend(struct pci_dev *pdev, pm_message_t state)
 	flush_work(&acb->arcmsr_do_message_isr_bh);
 	arcmsr_stop_adapter_bgrb(acb);
 	arcmsr_flush_adapter_cache(acb);
-	pci_set_drvdata(pdev, host);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
 	return 0;
 }
 
-static int arcmsr_resume(struct pci_dev *pdev)
+static int __maybe_unused arcmsr_resume(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *host = pci_get_drvdata(pdev);
 	struct AdapterControlBlock *acb =
 		(struct AdapterControlBlock *)host->hostdata;
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	if (pci_enable_device(pdev)) {
-		pr_warn("%s: pci_enable_device error\n", __func__);
-		return -ENODEV;
-	}
 	if (arcmsr_set_dma_mask(acb))
 		goto controller_unregister;
-	pci_set_master(pdev);
 	if (arcmsr_request_irq(pdev, acb) == FAILED)
 		goto controller_stop;
 	switch (acb->adapter_type) {
@@ -1206,9 +1198,7 @@ static int arcmsr_resume(struct pci_dev *pdev)
 	if (acb->adapter_type == ACB_ADAPTER_TYPE_F)
 		arcmsr_free_io_queue(acb);
 	arcmsr_unmap_pciregion(acb);
-	pci_release_regions(pdev);
 	scsi_host_put(host);
-	pci_disable_device(pdev);
 	return -ENODEV;
 }
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 10/29] scsi: esas2r: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (8 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 09/29] scsi: arcmsr: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 11/29] scsi: esas2r: use generic power management Vaibhav Gupta
                   ` (19 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in esas2r_resume(), and
there is no corresponding pci_enable_wake(...., true) in esas2r_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
esas2r_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/esas2r/esas2r_init.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/drivers/scsi/esas2r/esas2r_init.c b/drivers/scsi/esas2r/esas2r_init.c
index 09c5c24bf391..905a6c874789 100644
--- a/drivers/scsi/esas2r/esas2r_init.c
+++ b/drivers/scsi/esas2r/esas2r_init.c
@@ -676,10 +676,6 @@ int esas2r_resume(struct pci_dev *pdev)
 		       "pci_set_power_state(PCI_D0) "
 		       "called");
 	pci_set_power_state(pdev, PCI_D0);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_enable_wake(PCI_D0, 0) "
-		       "called");
-	pci_enable_wake(pdev, PCI_D0, 0);
 	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
 		       "pci_restore_state() called");
 	pci_restore_state(pdev);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 11/29] scsi: esas2r: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (9 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 10/29] scsi: esas2r: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 12/29] scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (18 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/esas2r/esas2r.h      |  5 ++--
 drivers/scsi/esas2r/esas2r_init.c | 44 +++++++++----------------------
 drivers/scsi/esas2r/esas2r_main.c |  3 +--
 3 files changed, 16 insertions(+), 36 deletions(-)

diff --git a/drivers/scsi/esas2r/esas2r.h b/drivers/scsi/esas2r/esas2r.h
index e30d2f1f5368..ed63f7a9ea54 100644
--- a/drivers/scsi/esas2r/esas2r.h
+++ b/drivers/scsi/esas2r/esas2r.h
@@ -996,8 +996,9 @@ void esas2r_adapter_tasklet(unsigned long context);
 irqreturn_t esas2r_interrupt(int irq, void *dev_id);
 irqreturn_t esas2r_msi_interrupt(int irq, void *dev_id);
 void esas2r_kickoff_timer(struct esas2r_adapter *a);
-int esas2r_suspend(struct pci_dev *pcid, pm_message_t state);
-int esas2r_resume(struct pci_dev *pcid);
+
+extern const struct dev_pm_ops esas2r_pm_ops;
+
 void esas2r_fw_event_off(struct esas2r_adapter *a);
 void esas2r_fw_event_on(struct esas2r_adapter *a);
 bool esas2r_nvram_write(struct esas2r_adapter *a, struct esas2r_request *rq,
diff --git a/drivers/scsi/esas2r/esas2r_init.c b/drivers/scsi/esas2r/esas2r_init.c
index 905a6c874789..eb7763af7089 100644
--- a/drivers/scsi/esas2r/esas2r_init.c
+++ b/drivers/scsi/esas2r/esas2r_init.c
@@ -640,49 +640,27 @@ void esas2r_kill_adapter(int i)
 	}
 }
 
-int esas2r_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused esas2r_suspend(struct device *dev)
 {
-	struct Scsi_Host *host = pci_get_drvdata(pdev);
-	u32 device_state;
+	struct Scsi_Host *host = dev_get_drvdata(dev);
 	struct esas2r_adapter *a = (struct esas2r_adapter *)host->hostdata;
 
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev), "suspending adapter()");
+	esas2r_log_dev(ESAS2R_LOG_INFO, dev, "suspending adapter()");
 	if (!a)
 		return -ENODEV;
 
 	esas2r_adapter_power_down(a, 1);
-	device_state = pci_choose_state(pdev, state);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_save_state() called");
-	pci_save_state(pdev);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_disable_device() called");
-	pci_disable_device(pdev);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_set_power_state() called");
-	pci_set_power_state(pdev, device_state);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev), "esas2r_suspend(): 0");
+	esas2r_log_dev(ESAS2R_LOG_INFO, dev, "esas2r_suspend(): 0");
 	return 0;
 }
 
-int esas2r_resume(struct pci_dev *pdev)
+static int __maybe_unused esas2r_resume(struct device *dev)
 {
-	struct Scsi_Host *host = pci_get_drvdata(pdev);
+	struct Scsi_Host *host = dev_get_drvdata(dev);
 	struct esas2r_adapter *a = (struct esas2r_adapter *)host->hostdata;
-	int rez;
-
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev), "resuming adapter()");
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_set_power_state(PCI_D0) "
-		       "called");
-	pci_set_power_state(pdev, PCI_D0);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_restore_state() called");
-	pci_restore_state(pdev);
-	esas2r_log_dev(ESAS2R_LOG_INFO, &(pdev->dev),
-		       "pci_enable_device() called");
-	rez = pci_enable_device(pdev);
-	pci_set_master(pdev);
+	int rez = 0;
+
+	esas2r_log_dev(ESAS2R_LOG_INFO, dev, "resuming adapter()");
 
 	if (!a) {
 		rez = -ENODEV;
@@ -726,11 +704,13 @@ int esas2r_resume(struct pci_dev *pdev)
 	}
 
 error_exit:
-	esas2r_log_dev(ESAS2R_LOG_CRIT, &(pdev->dev), "esas2r_resume(): %d",
+	esas2r_log_dev(ESAS2R_LOG_CRIT, dev, "esas2r_resume(): %d",
 		       rez);
 	return rez;
 }
 
+SIMPLE_DEV_PM_OPS(esas2r_pm_ops, esas2r_suspend, esas2r_resume);
+
 bool esas2r_set_degraded_mode(struct esas2r_adapter *a, char *error_str)
 {
 	set_bit(AF_DEGRADED_MODE, &a->flags);
diff --git a/drivers/scsi/esas2r/esas2r_main.c b/drivers/scsi/esas2r/esas2r_main.c
index 7b49e2e9fcde..aab3ea580e6b 100644
--- a/drivers/scsi/esas2r/esas2r_main.c
+++ b/drivers/scsi/esas2r/esas2r_main.c
@@ -346,8 +346,7 @@ static struct pci_driver
 	.id_table	= esas2r_pci_table,
 	.probe		= esas2r_probe,
 	.remove		= esas2r_remove,
-	.suspend	= esas2r_suspend,
-	.resume		= esas2r_resume,
+	.driver.pm	= &esas2r_pm_ops,
 };
 
 static int esas2r_probe(struct pci_dev *pcid,
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 12/29] scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (10 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 11/29] scsi: esas2r: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 13/29] scsi: hisi_sas_v3_hw: Don't use PCI helper functions Vaibhav Gupta
                   ` (17 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in hisi_sas_v3_resume(), and
there is no corresponding pci_enable_wake(...., true) in
hisi_sas_v3_suspend(). Either it should do enable-wake the device in
.suspend() or should not invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
hisi_sas_v3_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index 7133ca859b5e..f19f3db1ac6d 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -3492,7 +3492,6 @@ static int _resume_v3_hw(struct device *device)
 	dev_warn(dev, "resuming from operating state [D%d]\n",
 		 device_state);
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 	rc = pci_enable_device(pdev);
 	if (rc) {
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 13/29] scsi: hisi_sas_v3_hw: Don't use PCI helper functions
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (11 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 12/29] scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 14/29] scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm Vaibhav Gupta
                   ` (16 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers using new-framework/generic-framework should not handle standard
power management operations. These operations were performed by legacy
framework through PCI helper functions like pci_save/restore_state(),
pci_set_power_state(), etc.

Drivers should not use them now.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 17 +----------------
 1 file changed, 1 insertion(+), 16 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index f19f3db1ac6d..dfeb86c865d3 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -3440,7 +3440,6 @@ static int _suspend_v3_hw(struct device *device)
 	struct hisi_hba *hisi_hba = sha->lldd_ha;
 	struct device *dev = hisi_hba->dev;
 	struct Scsi_Host *shost = hisi_hba->shost;
-	pci_power_t device_state;
 	int rc;
 
 	if (!pdev->pm_cap) {
@@ -3466,12 +3465,7 @@ static int _suspend_v3_hw(struct device *device)
 
 	hisi_sas_init_mem(hisi_hba);
 
-	device_state = pci_choose_state(pdev, PMSG_SUSPEND);
-	dev_warn(dev, "entering operating state [D%d]\n",
-			device_state);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, device_state);
+	dev_warn(dev, "entering suspend state\n");
 
 	hisi_sas_release_tasks(hisi_hba);
 
@@ -3491,15 +3485,7 @@ static int _resume_v3_hw(struct device *device)
 
 	dev_warn(dev, "resuming from operating state [D%d]\n",
 		 device_state);
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	rc = pci_enable_device(pdev);
-	if (rc) {
-		dev_err(dev, "enable device failed during resume (%d)\n", rc);
-		return rc;
-	}
 
-	pci_set_master(pdev);
 	scsi_unblock_requests(shost);
 	clear_bit(HISI_SAS_REJECT_CMD_BIT, &hisi_hba->flags);
 
@@ -3507,7 +3493,6 @@ static int _resume_v3_hw(struct device *device)
 	rc = hw_init_v3_hw(hisi_hba);
 	if (rc) {
 		scsi_remove_host(shost);
-		pci_disable_device(pdev);
 		return rc;
 	}
 	hisi_hba->hw->phys_init(hisi_hba);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 14/29] scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (12 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 13/29] scsi: hisi_sas_v3_hw: Don't use PCI helper functions Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 15/29] scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (15 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Both runtime_suspend_v3_hw() and runtime_resume_v3_hw() do nothing else but
invoke suspend_v3_hw() and resume_v3_hw() respectively. This is the case of
unnecessary function calls. To use those functions for runtime pm as well,
simply use UNIVERSAL_DEV_PM_OPS.

make -j$(nproc) W=1, with CONFIG_PM disabled, throws '-Wunused-function'
warning for runtime_suspend_v3_hw() and runtime_resume_v3_hw(). After
dropping those function definitions, the warning was thrown for
suspend_v3_hw() and resume_v3_hw(). Hence, mark them as '__maybe_unused'.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 23 ++++++-----------------
 1 file changed, 6 insertions(+), 17 deletions(-)

diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
index dfeb86c865d3..9f0b4fe564cc 100644
--- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
+++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
@@ -3502,7 +3502,7 @@ static int _resume_v3_hw(struct device *device)
 	return 0;
 }
 
-static int suspend_v3_hw(struct device *device)
+static int __maybe_unused suspend_v3_hw(struct device *device)
 {
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct sas_ha_struct *sha = pci_get_drvdata(pdev);
@@ -3518,7 +3518,7 @@ static int suspend_v3_hw(struct device *device)
 	return rc;
 }
 
-static int resume_v3_hw(struct device *device)
+static int __maybe_unused resume_v3_hw(struct device *device)
 {
 	struct pci_dev *pdev = to_pci_dev(device);
 	struct sas_ha_struct *sha = pci_get_drvdata(pdev);
@@ -3541,21 +3541,10 @@ static const struct pci_error_handlers hisi_sas_err_handler = {
 	.reset_done	= hisi_sas_reset_done_v3_hw,
 };
 
-static int runtime_suspend_v3_hw(struct device *dev)
-{
-	return suspend_v3_hw(dev);
-}
-
-static int runtime_resume_v3_hw(struct device *dev)
-{
-	return resume_v3_hw(dev);
-}
-
-static const struct dev_pm_ops hisi_sas_v3_pm_ops = {
-	SET_SYSTEM_SLEEP_PM_OPS(suspend_v3_hw, resume_v3_hw)
-	SET_RUNTIME_PM_OPS(runtime_suspend_v3_hw,
-			   runtime_resume_v3_hw, NULL)
-};
+static UNIVERSAL_DEV_PM_OPS(hisi_sas_v3_pm_ops,
+			    suspend_v3_hw,
+			    resume_v3_hw,
+			    NULL);
 
 static struct pci_driver sas_v3_pci_driver = {
 	.name		= DRV_NAME,
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 15/29] scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (13 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 14/29] scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 16/29] scsi: mpt3sas_scsih: use generic power management Vaibhav Gupta
                   ` (14 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in scsih_resume(), and
there is no corresponding pci_enable_wake(...., true) in scsih_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
scsih_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/mpt3sas/mpt3sas_scsih.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
index 5f845d7094fc..2d201558b8fb 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
@@ -11140,7 +11140,6 @@ scsih_resume(struct pci_dev *pdev)
 		 pdev, pci_name(pdev), device_state);
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 	ioc->pdev = pdev;
 	r = mpt3sas_base_map_resources(ioc);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 16/29] scsi: mpt3sas_scsih: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (14 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 15/29] scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 17/29] scsi: lpfc: " Vaibhav Gupta
                   ` (13 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/mpt3sas/mpt3sas_scsih.c | 34 +++++++++++-----------------
 1 file changed, 13 insertions(+), 21 deletions(-)

diff --git a/drivers/scsi/mpt3sas/mpt3sas_scsih.c b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
index 2d201558b8fb..bc68544856b9 100644
--- a/drivers/scsi/mpt3sas/mpt3sas_scsih.c
+++ b/drivers/scsi/mpt3sas/mpt3sas_scsih.c
@@ -11084,20 +11084,18 @@ _scsih_probe(struct pci_dev *pdev, const struct pci_device_id *id)
 	return rv;
 }
 
-#ifdef CONFIG_PM
 /**
  * scsih_suspend - power management suspend main entry point
- * @pdev: PCI device struct
- * @state: PM state change to (usually PCI_D3)
+ * @dev: Device struct
  *
  * Return: 0 success, anything else error.
  */
-static int
-scsih_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused
+scsih_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *shost;
 	struct MPT3SAS_ADAPTER *ioc;
-	pci_power_t device_state;
 	int rc;
 
 	rc = _scsih_get_shost_and_ioc(pdev, &shost, &ioc);
@@ -11108,25 +11106,23 @@ scsih_suspend(struct pci_dev *pdev, pm_message_t state)
 	flush_scheduled_work();
 	scsi_block_requests(shost);
 	_scsih_nvme_shutdown(ioc);
-	device_state = pci_choose_state(pdev, state);
-	ioc_info(ioc, "pdev=0x%p, slot=%s, entering operating state [D%d]\n",
-		 pdev, pci_name(pdev), device_state);
+	ioc_info(ioc, "pdev=0x%p, slot=%s, entering operating state\n",
+		 pdev, pci_name(pdev));
 
-	pci_save_state(pdev);
 	mpt3sas_base_free_resources(ioc);
-	pci_set_power_state(pdev, device_state);
 	return 0;
 }
 
 /**
  * scsih_resume - power management resume main entry point
- * @pdev: PCI device struct
+ * @dev: Device struct
  *
  * Return: 0 success, anything else error.
  */
-static int
-scsih_resume(struct pci_dev *pdev)
+static int __maybe_unused
+scsih_resume(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *shost;
 	struct MPT3SAS_ADAPTER *ioc;
 	pci_power_t device_state = pdev->current_state;
@@ -11139,8 +11135,6 @@ scsih_resume(struct pci_dev *pdev)
 	ioc_info(ioc, "pdev=0x%p, slot=%s, previous operating state [D%d]\n",
 		 pdev, pci_name(pdev), device_state);
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
 	ioc->pdev = pdev;
 	r = mpt3sas_base_map_resources(ioc);
 	if (r)
@@ -11151,7 +11145,6 @@ scsih_resume(struct pci_dev *pdev)
 	mpt3sas_base_start_watchdog(ioc);
 	return 0;
 }
-#endif /* CONFIG_PM */
 
 /**
  * scsih_pci_error_detected - Called when a PCI error is detected.
@@ -11453,6 +11446,8 @@ static struct pci_error_handlers _mpt3sas_err_handler = {
 	.resume		= scsih_pci_resume,
 };
 
+static SIMPLE_DEV_PM_OPS(scsih_pm_ops, scsih_suspend, scsih_resume);
+
 static struct pci_driver mpt3sas_driver = {
 	.name		= MPT3SAS_DRIVER_NAME,
 	.id_table	= mpt3sas_pci_table,
@@ -11460,10 +11455,7 @@ static struct pci_driver mpt3sas_driver = {
 	.remove		= scsih_remove,
 	.shutdown	= scsih_shutdown,
 	.err_handler	= &_mpt3sas_err_handler,
-#ifdef CONFIG_PM
-	.suspend	= scsih_suspend,
-	.resume		= scsih_resume,
-#endif
+	.driver.pm	= &scsih_pm_ops,
 };
 
 /**
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 17/29] scsi: lpfc: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (15 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 16/29] scsi: mpt3sas_scsih: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 18/29] scsi: pm_8001: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (12 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/lpfc/lpfc_init.c | 100 +++++++++++-----------------------
 1 file changed, 33 insertions(+), 67 deletions(-)

diff --git a/drivers/scsi/lpfc/lpfc_init.c b/drivers/scsi/lpfc/lpfc_init.c
index ca25e54bb782..bf12328a5e45 100644
--- a/drivers/scsi/lpfc/lpfc_init.c
+++ b/drivers/scsi/lpfc/lpfc_init.c
@@ -12581,8 +12581,7 @@ lpfc_pci_remove_one_s3(struct pci_dev *pdev)
 
 /**
  * lpfc_pci_suspend_one_s3 - PCI func to suspend SLI-3 device for power mgmnt
- * @pdev: pointer to PCI device
- * @msg: power management message
+ * @dev_d: pointer to device
  *
  * This routine is to be called from the kernel's PCI subsystem to support
  * system Power Management (PM) to device with SLI-3 interface spec. When
@@ -12600,10 +12599,10 @@ lpfc_pci_remove_one_s3(struct pci_dev *pdev)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_suspend_one_s3(struct pci_dev *pdev, pm_message_t msg)
+static int __maybe_unused
+lpfc_pci_suspend_one_s3(struct device *dev_d)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev_d);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 
 	lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
@@ -12617,16 +12616,12 @@ lpfc_pci_suspend_one_s3(struct pci_dev *pdev, pm_message_t msg)
 	/* Disable interrupt from device */
 	lpfc_sli_disable_intr(phba);
 
-	/* Save device state to PCI config space */
-	pci_save_state(pdev);
-	pci_set_power_state(pdev, PCI_D3hot);
-
 	return 0;
 }
 
 /**
  * lpfc_pci_resume_one_s3 - PCI func to resume SLI-3 device for power mgmnt
- * @pdev: pointer to PCI device
+ * @dev_d: pointer to device
  *
  * This routine is to be called from the kernel's PCI subsystem to support
  * system Power Management (PM) to device with SLI-3 interface spec. When PM
@@ -12643,10 +12638,10 @@ lpfc_pci_suspend_one_s3(struct pci_dev *pdev, pm_message_t msg)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_resume_one_s3(struct pci_dev *pdev)
+static int __maybe_unused
+lpfc_pci_resume_one_s3(struct device *dev_d)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev_d);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 	uint32_t intr_mode;
 	int error;
@@ -12654,19 +12649,6 @@ lpfc_pci_resume_one_s3(struct pci_dev *pdev)
 	lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
 			"0452 PCI device Power Management resume.\n");
 
-	/* Restore device state from PCI config space */
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	/*
-	 * As the new kernel behavior of pci_restore_state() API call clears
-	 * device saved_state flag, need to save the restored state again.
-	 */
-	pci_save_state(pdev);
-
-	if (pdev->is_busmaster)
-		pci_set_master(pdev);
-
 	/* Startup the kernel thread for this host adapter. */
 	phba->worker_thread = kthread_run(lpfc_do_work, phba,
 					"lpfc_worker_%d", phba->brd_no);
@@ -13423,8 +13405,7 @@ lpfc_pci_remove_one_s4(struct pci_dev *pdev)
 
 /**
  * lpfc_pci_suspend_one_s4 - PCI func to suspend SLI-4 device for power mgmnt
- * @pdev: pointer to PCI device
- * @msg: power management message
+ * @dev_d: pointer to device
  *
  * This routine is called from the kernel's PCI subsystem to support system
  * Power Management (PM) to device with SLI-4 interface spec. When PM invokes
@@ -13442,10 +13423,10 @@ lpfc_pci_remove_one_s4(struct pci_dev *pdev)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_suspend_one_s4(struct pci_dev *pdev, pm_message_t msg)
+static int __maybe_unused
+lpfc_pci_suspend_one_s4(struct device *dev_d)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev_d);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 
 	lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
@@ -13460,16 +13441,12 @@ lpfc_pci_suspend_one_s4(struct pci_dev *pdev, pm_message_t msg)
 	lpfc_sli4_disable_intr(phba);
 	lpfc_sli4_queue_destroy(phba);
 
-	/* Save device state to PCI config space */
-	pci_save_state(pdev);
-	pci_set_power_state(pdev, PCI_D3hot);
-
 	return 0;
 }
 
 /**
  * lpfc_pci_resume_one_s4 - PCI func to resume SLI-4 device for power mgmnt
- * @pdev: pointer to PCI device
+ * @dev_d: pointer to device
  *
  * This routine is called from the kernel's PCI subsystem to support system
  * Power Management (PM) to device with SLI-4 interface spac. When PM invokes
@@ -13486,10 +13463,10 @@ lpfc_pci_suspend_one_s4(struct pci_dev *pdev, pm_message_t msg)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_resume_one_s4(struct pci_dev *pdev)
+static int __maybe_unused
+lpfc_pci_resume_one_s4(struct device *dev_d)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev_d);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 	uint32_t intr_mode;
 	int error;
@@ -13497,19 +13474,6 @@ lpfc_pci_resume_one_s4(struct pci_dev *pdev)
 	lpfc_printf_log(phba, KERN_INFO, LOG_INIT,
 			"0292 PCI device Power Management resume.\n");
 
-	/* Restore device state from PCI config space */
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	/*
-	 * As the new kernel behavior of pci_restore_state() API call clears
-	 * device saved_state flag, need to save the restored state again.
-	 */
-	pci_save_state(pdev);
-
-	if (pdev->is_busmaster)
-		pci_set_master(pdev);
-
 	 /* Startup the kernel thread for this host adapter. */
 	phba->worker_thread = kthread_run(lpfc_do_work, phba,
 					"lpfc_worker_%d", phba->brd_no);
@@ -13825,8 +13789,7 @@ lpfc_pci_remove_one(struct pci_dev *pdev)
 
 /**
  * lpfc_pci_suspend_one - lpfc PCI func to suspend dev for power management
- * @pdev: pointer to PCI device
- * @msg: power management message
+ * @dev: pointer to device
  *
  * This routine is to be registered to the kernel's PCI subsystem to support
  * system Power Management (PM). When PM invokes this method, it dispatches
@@ -13837,19 +13800,19 @@ lpfc_pci_remove_one(struct pci_dev *pdev)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_suspend_one(struct pci_dev *pdev, pm_message_t msg)
+static int __maybe_unused
+lpfc_pci_suspend_one(struct device *dev)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 	int rc = -ENODEV;
 
 	switch (phba->pci_dev_grp) {
 	case LPFC_PCI_DEV_LP:
-		rc = lpfc_pci_suspend_one_s3(pdev, msg);
+		rc = lpfc_pci_suspend_one_s3(dev);
 		break;
 	case LPFC_PCI_DEV_OC:
-		rc = lpfc_pci_suspend_one_s4(pdev, msg);
+		rc = lpfc_pci_suspend_one_s4(dev);
 		break;
 	default:
 		lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
@@ -13862,7 +13825,7 @@ lpfc_pci_suspend_one(struct pci_dev *pdev, pm_message_t msg)
 
 /**
  * lpfc_pci_resume_one - lpfc PCI func to resume dev for power management
- * @pdev: pointer to PCI device
+ * @dev: pointer to device
  *
  * This routine is to be registered to the kernel's PCI subsystem to support
  * system Power Management (PM). When PM invokes this method, it dispatches
@@ -13873,19 +13836,19 @@ lpfc_pci_suspend_one(struct pci_dev *pdev, pm_message_t msg)
  * 	0 - driver suspended the device
  * 	Error otherwise
  **/
-static int
-lpfc_pci_resume_one(struct pci_dev *pdev)
+static int __maybe_unused
+lpfc_pci_resume_one(struct device *dev)
 {
-	struct Scsi_Host *shost = pci_get_drvdata(pdev);
+	struct Scsi_Host *shost = dev_get_drvdata(dev);
 	struct lpfc_hba *phba = ((struct lpfc_vport *)shost->hostdata)->phba;
 	int rc = -ENODEV;
 
 	switch (phba->pci_dev_grp) {
 	case LPFC_PCI_DEV_LP:
-		rc = lpfc_pci_resume_one_s3(pdev);
+		rc = lpfc_pci_resume_one_s3(dev);
 		break;
 	case LPFC_PCI_DEV_OC:
-		rc = lpfc_pci_resume_one_s4(pdev);
+		rc = lpfc_pci_resume_one_s4(dev);
 		break;
 	default:
 		lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
@@ -14065,14 +14028,17 @@ static const struct pci_error_handlers lpfc_err_handler = {
 	.resume = lpfc_io_resume,
 };
 
+static SIMPLE_DEV_PM_OPS(lpfc_pci_pm_ops_one,
+			 lpfc_pci_suspend_one,
+			 lpfc_pci_resume_one);
+
 static struct pci_driver lpfc_driver = {
 	.name		= LPFC_DRIVER_NAME,
 	.id_table	= lpfc_id_table,
 	.probe		= lpfc_pci_probe_one,
 	.remove		= lpfc_pci_remove_one,
 	.shutdown	= lpfc_pci_remove_one,
-	.suspend        = lpfc_pci_suspend_one,
-	.resume		= lpfc_pci_resume_one,
+	.driver.pm	= &lpfc_pci_pm_ops_one,
 	.err_handler    = &lpfc_err_handler,
 };
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 18/29] scsi: pm_8001: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (16 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 17/29] scsi: lpfc: " Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 19/29] scsi: pm_8001: use generic power management Vaibhav Gupta
                   ` (11 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in pm8001_pci_resume(), and
there is no corresponding pci_enable_wake(...., true) in
pm8001_pci_suspend(). Either it should do enable-wake the device in
.suspend() or should not invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
pm8001_pci__resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
---
 drivers/scsi/pm8001/pm8001_init.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/pm8001/pm8001_init.c b/drivers/scsi/pm8001/pm8001_init.c
index 3cf3e58b6979..d47ed9a33173 100644
--- a/drivers/scsi/pm8001/pm8001_init.c
+++ b/drivers/scsi/pm8001/pm8001_init.c
@@ -1327,7 +1327,6 @@ static int pm8001_pci_resume(struct pci_dev *pdev)
 		"operating state [D%d]\n", pdev, pm8001_ha->name, device_state);
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 	rc = pci_enable_device(pdev);
 	if (rc) {
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 19/29] scsi: pm_8001: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (17 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 18/29] scsi: pm_8001: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: " Vaibhav Gupta
                   ` (10 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/pm8001/pm8001_init.c | 45 +++++++++++--------------------
 1 file changed, 16 insertions(+), 29 deletions(-)

diff --git a/drivers/scsi/pm8001/pm8001_init.c b/drivers/scsi/pm8001/pm8001_init.c
index d47ed9a33173..7a732fd09b38 100644
--- a/drivers/scsi/pm8001/pm8001_init.c
+++ b/drivers/scsi/pm8001/pm8001_init.c
@@ -1257,23 +1257,21 @@ static void pm8001_pci_remove(struct pci_dev *pdev)
 
 /**
  * pm8001_pci_suspend - power management suspend main entry point
- * @pdev: PCI device struct
- * @state: PM state change to (usually PCI_D3)
+ * @dev: Device struct
  *
  * Returns 0 success, anything else error.
  */
-static int pm8001_pci_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused pm8001_pci_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct sas_ha_struct *sha = pci_get_drvdata(pdev);
-	struct pm8001_hba_info *pm8001_ha;
+	struct pm8001_hba_info *pm8001_ha = sha->lldd_ha;
 	int  i, j;
-	u32 device_state;
-	pm8001_ha = sha->lldd_ha;
 	sas_suspend_ha(sha);
 	flush_workqueue(pm8001_wq);
 	scsi_block_requests(pm8001_ha->shost);
 	if (!pdev->pm_cap) {
-		dev_err(&pdev->dev, " PCI PM not supported\n");
+		dev_err(dev, " PCI PM not supported\n");
 		return -ENODEV;
 	}
 	PM8001_CHIP_DISP->interrupt_disable(pm8001_ha, 0xFF);
@@ -1296,24 +1294,21 @@ static int pm8001_pci_suspend(struct pci_dev *pdev, pm_message_t state)
 		for (j = 0; j < PM8001_MAX_MSIX_VEC; j++)
 			tasklet_kill(&pm8001_ha->tasklet[j]);
 #endif
-	device_state = pci_choose_state(pdev, state);
 	pm8001_printk("pdev=0x%p, slot=%s, entering "
-		      "operating state [D%d]\n", pdev,
-		      pm8001_ha->name, device_state);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, device_state);
+		"suspended state\n", pdev,
+		pm8001_ha->name);
 	return 0;
 }
 
 /**
  * pm8001_pci_resume - power management resume main entry point
- * @pdev: PCI device struct
+ * @dev: Device struct
  *
  * Returns 0 success, anything else error.
  */
-static int pm8001_pci_resume(struct pci_dev *pdev)
+static int __maybe_unused pm8001_pci_resume(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct sas_ha_struct *sha = pci_get_drvdata(pdev);
 	struct pm8001_hba_info *pm8001_ha;
 	int rc;
@@ -1326,16 +1321,6 @@ static int pm8001_pci_resume(struct pci_dev *pdev)
 	pm8001_printk("pdev=0x%p, slot=%s, resuming from previous "
 		"operating state [D%d]\n", pdev, pm8001_ha->name, device_state);
 
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-	rc = pci_enable_device(pdev);
-	if (rc) {
-		pm8001_printk("slot=%s Enable device failed during resume\n",
-			      pm8001_ha->name);
-		goto err_out_enable;
-	}
-
-	pci_set_master(pdev);
 	rc = pci_go_44(pdev);
 	if (rc)
 		goto err_out_disable;
@@ -1396,8 +1381,7 @@ static int pm8001_pci_resume(struct pci_dev *pdev)
 
 err_out_disable:
 	scsi_remove_host(pm8001_ha->shost);
-	pci_disable_device(pdev);
-err_out_enable:
+
 	return rc;
 }
 
@@ -1480,13 +1464,16 @@ static struct pci_device_id pm8001_pci_table[] = {
 	{} /* terminate list */
 };
 
+static SIMPLE_DEV_PM_OPS(pm8001_pci_pm_ops,
+			 pm8001_pci_suspend,
+			 pm8001_pci_resume);
+
 static struct pci_driver pm8001_pci_driver = {
 	.name		= DRV_NAME,
 	.id_table	= pm8001_pci_table,
 	.probe		= pm8001_pci_probe,
 	.remove		= pm8001_pci_remove,
-	.suspend	= pm8001_pci_suspend,
-	.resume		= pm8001_pci_resume,
+	.driver.pm	= &pm8001_pci_pm_ops,
 };
 
 /**
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (18 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 19/29] scsi: pm_8001: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-04 18:08   ` Don.Brace--- via Linux-kernel-mentees
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 21/29] scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (9 subsequent siblings)
  29 siblings, 1 reply; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/hpsa.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c
index 83ce4f11a589..e53364141fa3 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -9090,25 +9090,27 @@ static void hpsa_remove_one(struct pci_dev *pdev)
 	hpda_free_ctlr_info(h);				/* init_one 1 */
 }
 
-static int hpsa_suspend(__attribute__((unused)) struct pci_dev *pdev,
-	__attribute__((unused)) pm_message_t state)
+static int __maybe_unused hpsa_suspend(
+	__attribute__((unused)) struct device *dev)
 {
 	return -ENOSYS;
 }
 
-static int hpsa_resume(__attribute__((unused)) struct pci_dev *pdev)
+static int __maybe_unused hpsa_resume
+	(__attribute__((unused)) struct device *dev)
 {
 	return -ENOSYS;
 }
 
+static SIMPLE_DEV_PM_OPS(hpsa_pm_ops, hpsa_suspend, hpsa_resume);
+
 static struct pci_driver hpsa_pci_driver = {
 	.name = HPSA,
 	.probe = hpsa_init_one,
 	.remove = hpsa_remove_one,
 	.id_table = hpsa_pci_device_id,	/* id_table */
 	.shutdown = hpsa_shutdown,
-	.suspend = hpsa_suspend,
-	.resume = hpsa_resume,
+	.driver.pm = &hpsa_pm_ops,
 };
 
 /* Fill in bucket_map[], given nsgs (the max number of
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 21/29] scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (19 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: " Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 22/29] scsi: 3w-9xxx: use generic power management Vaibhav Gupta
                   ` (8 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	Don Brace, linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in twa_resume(), and
there is no corresponding pci_enable_wake(...., true) in twa_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
twa_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Acked-by: Don Brace <don.brace@microchip.com>
---
 drivers/scsi/3w-9xxx.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/3w-9xxx.c b/drivers/scsi/3w-9xxx.c
index 3337b1e80412..e458e99ff161 100644
--- a/drivers/scsi/3w-9xxx.c
+++ b/drivers/scsi/3w-9xxx.c
@@ -2230,7 +2230,6 @@ static int twa_resume(struct pci_dev *pdev)
 
 	printk(KERN_WARNING "3w-9xxx: Resuming host %d.\n", tw_dev->host->host_no);
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 
 	retval = pci_enable_device(pdev);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 22/29] scsi: 3w-9xxx: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (20 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 21/29] scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 23/29] scsi: 3w-sas: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (7 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/3w-9xxx.c | 29 +++++++----------------------
 1 file changed, 7 insertions(+), 22 deletions(-)

diff --git a/drivers/scsi/3w-9xxx.c b/drivers/scsi/3w-9xxx.c
index e458e99ff161..b4718a1b2bd6 100644
--- a/drivers/scsi/3w-9xxx.c
+++ b/drivers/scsi/3w-9xxx.c
@@ -2191,10 +2191,10 @@ static void twa_remove(struct pci_dev *pdev)
 	twa_device_extension_count--;
 } /* End twa_remove() */
 
-#ifdef CONFIG_PM
 /* This function is called on PCI suspend */
-static int twa_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused twa_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *host = pci_get_drvdata(pdev);
 	TW_Device_Extension *tw_dev = (TW_Device_Extension *)host->hostdata;
 
@@ -2214,31 +2214,19 @@ static int twa_suspend(struct pci_dev *pdev, pm_message_t state)
 	}
 	TW_CLEAR_ALL_INTERRUPTS(tw_dev);
 
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
-
 	return 0;
 } /* End twa_suspend() */
 
 /* This function is called on PCI resume */
-static int twa_resume(struct pci_dev *pdev)
+static int __maybe_unused twa_resume(struct device *dev)
 {
 	int retval = 0;
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *host = pci_get_drvdata(pdev);
 	TW_Device_Extension *tw_dev = (TW_Device_Extension *)host->hostdata;
 
 	printk(KERN_WARNING "3w-9xxx: Resuming host %d.\n", tw_dev->host->host_no);
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
 
-	retval = pci_enable_device(pdev);
-	if (retval) {
-		TW_PRINTK(tw_dev->host, TW_DRIVER, 0x39, "Enable device failed during resume");
-		return retval;
-	}
-
-	pci_set_master(pdev);
 	pci_try_set_mwi(pdev);
 
 	retval = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
@@ -2276,11 +2264,9 @@ static int twa_resume(struct pci_dev *pdev)
 
 out_disable_device:
 	scsi_remove_host(host);
-	pci_disable_device(pdev);
 
 	return retval;
 } /* End twa_resume() */
-#endif
 
 /* PCI Devices supported by this driver */
 static struct pci_device_id twa_pci_tbl[] = {
@@ -2296,16 +2282,15 @@ static struct pci_device_id twa_pci_tbl[] = {
 };
 MODULE_DEVICE_TABLE(pci, twa_pci_tbl);
 
+static SIMPLE_DEV_PM_OPS(twa_pm_ops, twa_suspend, twa_resume);
+
 /* pci_driver initializer */
 static struct pci_driver twa_driver = {
 	.name		= "3w-9xxx",
 	.id_table	= twa_pci_tbl,
 	.probe		= twa_probe,
 	.remove		= twa_remove,
-#ifdef CONFIG_PM
-	.suspend	= twa_suspend,
-	.resume		= twa_resume,
-#endif
+	.driver.pm	= &twa_pm_ops,
 	.shutdown	= twa_shutdown
 };
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 23/29] scsi: 3w-sas: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (21 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 22/29] scsi: 3w-9xxx: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 24/29] scsi: 3w-sas: use generic power management Vaibhav Gupta
                   ` (6 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in twl_resume(), and
there is no corresponding pci_enable_wake(...., true) in twl_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
twl_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/3w-sas.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/3w-sas.c b/drivers/scsi/3w-sas.c
index dda6fa857709..0b4888199699 100644
--- a/drivers/scsi/3w-sas.c
+++ b/drivers/scsi/3w-sas.c
@@ -1795,7 +1795,6 @@ static int twl_resume(struct pci_dev *pdev)
 
 	printk(KERN_WARNING "3w-sas: Resuming host %d.\n", tw_dev->host->host_no);
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 
 	retval = pci_enable_device(pdev);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 24/29] scsi: 3w-sas: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (22 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 23/29] scsi: 3w-sas: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 25/29] scsi: mvumi: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (5 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/3w-sas.c | 31 +++++++------------------------
 1 file changed, 7 insertions(+), 24 deletions(-)

diff --git a/drivers/scsi/3w-sas.c b/drivers/scsi/3w-sas.c
index 0b4888199699..b8f1848ecef2 100644
--- a/drivers/scsi/3w-sas.c
+++ b/drivers/scsi/3w-sas.c
@@ -1756,11 +1756,10 @@ static void twl_remove(struct pci_dev *pdev)
 	twl_device_extension_count--;
 } /* End twl_remove() */
 
-#ifdef CONFIG_PM
 /* This function is called on PCI suspend */
-static int twl_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused twl_suspend(struct device *dev)
 {
-	struct Scsi_Host *host = pci_get_drvdata(pdev);
+	struct Scsi_Host *host = dev_get_drvdata(dev);
 	TW_Device_Extension *tw_dev = (TW_Device_Extension *)host->hostdata;
 
 	printk(KERN_WARNING "3w-sas: Suspending host %d.\n", tw_dev->host->host_no);
@@ -1779,31 +1778,18 @@ static int twl_suspend(struct pci_dev *pdev, pm_message_t state)
 	/* Clear doorbell interrupt */
 	TWL_CLEAR_DB_INTERRUPT(tw_dev);
 
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
-
 	return 0;
 } /* End twl_suspend() */
 
 /* This function is called on PCI resume */
-static int twl_resume(struct pci_dev *pdev)
+static int __maybe_unused twl_resume(struct device *dev)
 {
 	int retval = 0;
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct Scsi_Host *host = pci_get_drvdata(pdev);
 	TW_Device_Extension *tw_dev = (TW_Device_Extension *)host->hostdata;
 
 	printk(KERN_WARNING "3w-sas: Resuming host %d.\n", tw_dev->host->host_no);
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	retval = pci_enable_device(pdev);
-	if (retval) {
-		TW_PRINTK(tw_dev->host, TW_DRIVER, 0x24, "Enable device failed during resume");
-		return retval;
-	}
-
-	pci_set_master(pdev);
 	pci_try_set_mwi(pdev);
 
 	retval = dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(64));
@@ -1841,11 +1827,9 @@ static int twl_resume(struct pci_dev *pdev)
 
 out_disable_device:
 	scsi_remove_host(host);
-	pci_disable_device(pdev);
 
 	return retval;
 } /* End twl_resume() */
-#endif
 
 /* PCI Devices supported by this driver */
 static struct pci_device_id twl_pci_tbl[] = {
@@ -1854,16 +1838,15 @@ static struct pci_device_id twl_pci_tbl[] = {
 };
 MODULE_DEVICE_TABLE(pci, twl_pci_tbl);
 
+static SIMPLE_DEV_PM_OPS(twl_pm_ops, twl_suspend, twl_resume);
+
 /* pci_driver initializer */
 static struct pci_driver twl_driver = {
 	.name		= "3w-sas",
 	.id_table	= twl_pci_tbl,
 	.probe		= twl_probe,
 	.remove		= twl_remove,
-#ifdef CONFIG_PM
-	.suspend	= twl_suspend,
-	.resume		= twl_resume,
-#endif
+	.driver.pm	= &twl_pm_ops,
 	.shutdown	= twl_shutdown
 };
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 25/29] scsi: mvumi: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (23 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 24/29] scsi: 3w-sas: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 26/29] scsi: mvumi: use generic power management Vaibhav Gupta
                   ` (4 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in mvumi_resume(), and
there is no corresponding pci_enable_wake(...., true) in mvumi_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
mvumi_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/mvumi.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/mvumi.c b/drivers/scsi/mvumi.c
index 0354898d7cac..7cd9c70e32dd 100644
--- a/drivers/scsi/mvumi.c
+++ b/drivers/scsi/mvumi.c
@@ -2595,7 +2595,6 @@ static int __maybe_unused mvumi_resume(struct pci_dev *pdev)
 	mhba = pci_get_drvdata(pdev);
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 
 	ret = pci_enable_device(pdev);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 26/29] scsi: mvumi: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (24 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 25/29] scsi: mvumi: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 27/29] scsi: mvumi: update function description Vaibhav Gupta
                   ` (3 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/mvumi.c | 46 +++++++++-----------------------------------
 1 file changed, 9 insertions(+), 37 deletions(-)

diff --git a/drivers/scsi/mvumi.c b/drivers/scsi/mvumi.c
index 7cd9c70e32dd..bbf0faac1f05 100644
--- a/drivers/scsi/mvumi.c
+++ b/drivers/scsi/mvumi.c
@@ -2568,46 +2568,26 @@ static void mvumi_shutdown(struct pci_dev *pdev)
 	mvumi_flush_cache(mhba);
 }
 
-static int __maybe_unused mvumi_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused mvumi_suspend(struct device *dev)
 {
-	struct mvumi_hba *mhba = NULL;
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct mvumi_hba *mhba = pci_get_drvdata(pdev);
 
-	mhba = pci_get_drvdata(pdev);
 	mvumi_flush_cache(mhba);
 
-	pci_set_drvdata(pdev, mhba);
 	mhba->instancet->disable_intr(mhba);
-	free_irq(mhba->pdev->irq, mhba);
 	mvumi_unmap_pci_addr(pdev, mhba->base_addr);
-	pci_release_regions(pdev);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
 
 	return 0;
 }
 
-static int __maybe_unused mvumi_resume(struct pci_dev *pdev)
+static int __maybe_unused mvumi_resume(struct device *dev)
 {
 	int ret;
-	struct mvumi_hba *mhba = NULL;
-
-	mhba = pci_get_drvdata(pdev);
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	ret = pci_enable_device(pdev);
-	if (ret) {
-		dev_err(&pdev->dev, "enable device failed\n");
-		return ret;
-	}
+	struct pci_dev *pdev = to_pci_dev(dev);
+	struct mvumi_hba *mhba = pci_get_drvdata(pdev);
 
-	ret = mvumi_pci_set_master(pdev);
 	ret = dma_set_mask(&pdev->dev, DMA_BIT_MASK(32));
-	if (ret)
-		goto fail;
-	ret = pci_request_regions(mhba->pdev, MV_DRIVER_NAME);
 	if (ret)
 		goto fail;
 	ret = mvumi_map_pci_addr(mhba->pdev, mhba->base_addr);
@@ -2627,12 +2607,6 @@ static int __maybe_unused mvumi_resume(struct pci_dev *pdev)
 		goto unmap_pci_addr;
 	}
 
-	ret = request_irq(mhba->pdev->irq, mvumi_isr_handler, IRQF_SHARED,
-				"mvumi", mhba);
-	if (ret) {
-		dev_err(&pdev->dev, "failed to register IRQ\n");
-		goto unmap_pci_addr;
-	}
 	mhba->instancet->enable_intr(mhba);
 
 	return 0;
@@ -2642,11 +2616,12 @@ static int __maybe_unused mvumi_resume(struct pci_dev *pdev)
 release_regions:
 	pci_release_regions(pdev);
 fail:
-	pci_disable_device(pdev);
 
 	return ret;
 }
 
+static SIMPLE_DEV_PM_OPS(mvumi_pm_ops, mvumi_suspend, mvumi_resume);
+
 static struct pci_driver mvumi_pci_driver = {
 
 	.name = MV_DRIVER_NAME,
@@ -2654,10 +2629,7 @@ static struct pci_driver mvumi_pci_driver = {
 	.probe = mvumi_probe_one,
 	.remove = mvumi_detach_one,
 	.shutdown = mvumi_shutdown,
-#ifdef CONFIG_PM
-	.suspend = mvumi_suspend,
-	.resume = mvumi_resume,
-#endif
+	.driver.pm = &mvumi_pm_ops,
 };
 
 module_pci_driver(mvumi_pci_driver);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 27/29] scsi: mvumi: update function description
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (25 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 26/29] scsi: mvumi: use generic power management Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 28/29] scsi: pmcraid: Drop PCI Wakeup calls from .resume Vaibhav Gupta
                   ` (2 subsequent siblings)
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

There is no "device" parameter in mvumi_shutdown(). Instead there is
"pdev" which is not described.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/mvumi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/scsi/mvumi.c b/drivers/scsi/mvumi.c
index bbf0faac1f05..6c82b0bc391a 100644
--- a/drivers/scsi/mvumi.c
+++ b/drivers/scsi/mvumi.c
@@ -2559,7 +2559,7 @@ static void mvumi_detach_one(struct pci_dev *pdev)
 
 /**
  * mvumi_shutdown -	Shutdown entry point
- * @device:		Generic device structure
+ * @pdev:		PCI device structure
  */
 static void mvumi_shutdown(struct pci_dev *pdev)
 {
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 28/29] scsi: pmcraid: Drop PCI Wakeup calls from .resume
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (26 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 27/29] scsi: mvumi: update function description Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 29/29] scsi: pmcraid: use generic power management Vaibhav Gupta
  2020-12-01  5:05 ` [Linux-kernel-mentees] [PATCH v4 00/29] scsi: " Martin K. Petersen
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

The driver calls pci_enable_wake(...., false) in pmcraid_resume(), and
there is no corresponding pci_enable_wake(...., true) in pmcraid_suspend().
Either it should do enable-wake the device in .suspend() or should not
invoke pci_enable_wake() at all.

Concluding that this driver doesn't support enable-wake and PCI core calls
pci_enable_wake(pci_dev, PCI_D0, false) during resume, drop it from
pmcraid_resume().

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/pmcraid.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index cbe5fab793eb..5c767cd8ffc3 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -5271,7 +5271,6 @@ static int pmcraid_resume(struct pci_dev *pdev)
 	int rc;
 
 	pci_set_power_state(pdev, PCI_D0);
-	pci_enable_wake(pdev, PCI_D0, 0);
 	pci_restore_state(pdev);
 
 	rc = pci_enable_device(pdev);
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* [Linux-kernel-mentees] [PATCH v4 29/29] scsi: pmcraid: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (27 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 28/29] scsi: pmcraid: Drop PCI Wakeup calls from .resume Vaibhav Gupta
@ 2020-11-02 16:47 ` Vaibhav Gupta
  2020-12-01  5:05 ` [Linux-kernel-mentees] [PATCH v4 00/29] scsi: " Martin K. Petersen
  29 siblings, 0 replies; 32+ messages in thread
From: Vaibhav Gupta @ 2020-11-02 16:47 UTC (permalink / raw)
  To: Bjorn Helgaas, Bjorn Helgaas, Bjorn Helgaas, Vaibhav Gupta,
	Adam Radford, James E.J. Bottomley, Martin K. Petersen,
	Adaptec OEM Raid Solutions, Hannes Reinecke, Bradley Grove,
	John Garry, Don Brace, Xiang Chen, James Smart, Dick Kennedy,
	Kashyap Desai, Sumit Saxena, Shivasharan S, Sathya Prakash,
	Sreekanth Reddy, Suganath Prabu Subramani, Jack Wang,
	Balsundar P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

Drivers should do only device-specific jobs. But in general, drivers using
legacy PCI PM framework for .suspend()/.resume() have to manage many PCI
PM-related tasks themselves which can be done by PCI Core itself. This
brings extra load on the driver and it directly calls PCI helper functions
to handle them.

Switch to the new generic framework by updating function signatures and
define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove
unnecessary calls to the PCI Helper functions along with the legacy
.suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
---
 drivers/scsi/pmcraid.c | 43 ++++++++++--------------------------------
 1 file changed, 10 insertions(+), 33 deletions(-)

diff --git a/drivers/scsi/pmcraid.c b/drivers/scsi/pmcraid.c
index 5c767cd8ffc3..834556ea21d2 100644
--- a/drivers/scsi/pmcraid.c
+++ b/drivers/scsi/pmcraid.c
@@ -5234,53 +5234,37 @@ static void pmcraid_remove(struct pci_dev *pdev)
 	return;
 }
 
-#ifdef CONFIG_PM
 /**
  * pmcraid_suspend - driver suspend entry point for power management
- * @pdev:   PCI device structure
- * @state:  PCI power state to suspend routine
+ * @dev:   Device structure
  *
  * Return Value - 0 always
  */
-static int pmcraid_suspend(struct pci_dev *pdev, pm_message_t state)
+static int __maybe_unused pmcraid_suspend(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct pmcraid_instance *pinstance = pci_get_drvdata(pdev);
 
 	pmcraid_shutdown(pdev);
 	pmcraid_disable_interrupts(pinstance, ~0);
 	pmcraid_kill_tasklets(pinstance);
-	pci_set_drvdata(pinstance->pdev, pinstance);
 	pmcraid_unregister_interrupt_handler(pinstance);
-	pci_save_state(pdev);
-	pci_disable_device(pdev);
-	pci_set_power_state(pdev, pci_choose_state(pdev, state));
 
 	return 0;
 }
 
 /**
  * pmcraid_resume - driver resume entry point PCI power management
- * @pdev: PCI device structure
+ * @dev: Device structure
  *
  * Return Value - 0 in case of success. Error code in case of any failure
  */
-static int pmcraid_resume(struct pci_dev *pdev)
+static int __maybe_unused pmcraid_resume(struct device *dev)
 {
+	struct pci_dev *pdev = to_pci_dev(dev);
 	struct pmcraid_instance *pinstance = pci_get_drvdata(pdev);
 	struct Scsi_Host *host = pinstance->host;
-	int rc;
-
-	pci_set_power_state(pdev, PCI_D0);
-	pci_restore_state(pdev);
-
-	rc = pci_enable_device(pdev);
-
-	if (rc) {
-		dev_err(&pdev->dev, "resume: Enable device failed\n");
-		return rc;
-	}
-
-	pci_set_master(pdev);
+	int rc = 0;
 
 	if (sizeof(dma_addr_t) == 4 ||
 	    dma_set_mask(&pdev->dev, DMA_BIT_MASK(64)))
@@ -5333,18 +5317,10 @@ static int pmcraid_resume(struct pci_dev *pdev)
 	scsi_host_put(host);
 
 disable_device:
-	pci_disable_device(pdev);
 
 	return rc;
 }
 
-#else
-
-#define pmcraid_suspend NULL
-#define pmcraid_resume  NULL
-
-#endif /* CONFIG_PM */
-
 /**
  * pmcraid_complete_ioa_reset - Called by either timer or tasklet during
  *				completion of the ioa reset
@@ -5832,6 +5808,8 @@ static int pmcraid_probe(struct pci_dev *pdev,
 	return -ENODEV;
 }
 
+static SIMPLE_DEV_PM_OPS(pmcraid_pm_ops, pmcraid_suspend, pmcraid_resume);
+
 /*
  * PCI driver structure of pmcraid driver
  */
@@ -5840,8 +5818,7 @@ static struct pci_driver pmcraid_driver = {
 	.id_table = pmcraid_pci_table,
 	.probe = pmcraid_probe,
 	.remove = pmcraid_remove,
-	.suspend = pmcraid_suspend,
-	.resume = pmcraid_resume,
+	.driver.pm = &pmcraid_pm_ops,
 	.shutdown = pmcraid_shutdown
 };
 
-- 
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply related	[flat|nested] 32+ messages in thread

* Re: [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: use generic power management
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: " Vaibhav Gupta
@ 2020-11-04 18:08   ` Don.Brace--- via Linux-kernel-mentees
  0 siblings, 0 replies; 32+ messages in thread
From: Don.Brace--- via Linux-kernel-mentees @ 2020-11-04 18:08 UTC (permalink / raw)
  To: vaibhavgupta40, helgaas, bhelgaas, bjorn, vaibhav.varodek,
	aradford, jejb, martin.petersen, aacraid, hare, linuxdrivers,
	john.garry, don.brace, chenxiang66, james.smart, dick.kennedy,
	kashyap.desai, sumit.saxena, shivasharan.srikanteshwara,
	sathya.prakash, sreekanth.reddy, suganath-prabu.subramani,
	jinpu.wang, Balsundar.P
  Cc: linux-scsi, MPT-FusionLinux.pdl, esc.storagedev, linux-kernel,
	linux-kernel-mentees, megaraidlinux.pdl

-----Original Message-----

Switch to the new generic framework by updating function signatures and define a "struct dev_pm_ops" variable to bind PM callbacks. Also, remove unnecessary calls to the PCI Helper functions along with the legacy .suspend & .resume bindings.

Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>

Acked-by: Don Brace <don.brace@microchip.com>

I thought that I had given my Acked-by on 10/27, but it must have been lost.

Thanks for your attention to hpsa,
Don
---
 drivers/scsi/hpsa.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

diff --git a/drivers/scsi/hpsa.c b/drivers/scsi/hpsa.c index 83ce4f11a589..e53364141fa3 100644
--- a/drivers/scsi/hpsa.c
+++ b/drivers/scsi/hpsa.c
@@ -9090,25 +9090,27 @@ static void hpsa_remove_one(struct pci_dev *pdev)
        hpda_free_ctlr_info(h);                         /* init_one 1 */
 }

-static int hpsa_suspend(__attribute__((unused)) struct pci_dev *pdev,
-       __attribute__((unused)) pm_message_t state)
+static int __maybe_unused hpsa_suspend(
+       __attribute__((unused)) struct device *dev)
 {
        return -ENOSYS;
 }

-static int hpsa_resume(__attribute__((unused)) struct pci_dev *pdev)
+static int __maybe_unused hpsa_resume
+       (__attribute__((unused)) struct device *dev)
 {
        return -ENOSYS;
 }

+static SIMPLE_DEV_PM_OPS(hpsa_pm_ops, hpsa_suspend, hpsa_resume);
+
 static struct pci_driver hpsa_pci_driver = {
        .name = HPSA,
        .probe = hpsa_init_one,
        .remove = hpsa_remove_one,
        .id_table = hpsa_pci_device_id, /* id_table */
        .shutdown = hpsa_shutdown,
-       .suspend = hpsa_suspend,
-       .resume = hpsa_resume,
+       .driver.pm = &hpsa_pm_ops,
 };

 /* Fill in bucket_map[], given nsgs (the max number of
--
2.28.0

_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply	[flat|nested] 32+ messages in thread

* Re: [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management
  2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
                   ` (28 preceding siblings ...)
  2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 29/29] scsi: pmcraid: use generic power management Vaibhav Gupta
@ 2020-12-01  5:05 ` Martin K. Petersen
  29 siblings, 0 replies; 32+ messages in thread
From: Martin K. Petersen @ 2020-12-01  5:05 UTC (permalink / raw)
  To: Bjorn Helgaas, Sreekanth Reddy, Adam Radford, Jack Wang,
	Bjorn Helgaas, Sumit Saxena, James E.J. Bottomley, James Smart,
	John Garry, Suganath Prabu Subramani, Xiang Chen,
	Hannes Reinecke, Kashyap Desai, Bjorn Helgaas,
	Adaptec OEM Raid Solutions, Balsundar P, Don Brace,
	Bradley Grove, Dick Kennedy, Vaibhav Gupta, Shivasharan S,
	Vaibhav Gupta, Sathya Prakash
  Cc: Martin K . Petersen, linux-scsi, MPT-FusionLinux.pdl,
	esc.storagedev, linux-kernel, linux-kernel-mentees,
	megaraidlinux.pdl

On Mon, 2 Nov 2020 22:17:01 +0530, Vaibhav Gupta wrote:

> Linux Kernel Mentee: Remove Legacy Power Management.
> 
> The purpose of this patch series is to upgrade power management in xxxxxxxx
> drivers. This has been done by upgrading .suspend() and .resume() callbacks.
> 
> The upgrade makes sure that the involvement of PCI Core does not change the
> order of operations executed in a driver. Thus, does not change its behavior.
> 
> [...]

Applied to 5.11/scsi-queue, thanks!

[01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/8ed9d987c6d9
[02/29] scsi: megaraid_sas: Use generic power management
        https://git.kernel.org/mkp/scsi/c/977001df0368
[03/29] scsi: megaraid_sas: Update function description
        https://git.kernel.org/mkp/scsi/c/498854102c1c
[04/29] scsi: aacraid: Drop pci_enable_wake() from .resume
        https://git.kernel.org/mkp/scsi/c/7dd222218076
[05/29] scsi: aacraid: Use generic power management
        https://git.kernel.org/mkp/scsi/c/7e380b5c27ea
[06/29] scsi: aic7xxx: Use generic power management
        https://git.kernel.org/mkp/scsi/c/6897b9a177df
[07/29] scsi: aic79xx: Use generic power management
        https://git.kernel.org/mkp/scsi/c/ec199a8df698
[08/29] scsi: arcmsr: Drop PCI wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/b7b862d75b49
[09/29] scsi: arcmsr: Use generic power management
        https://git.kernel.org/mkp/scsi/c/756ebbe73fc4
[10/29] scsi: esas2r: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/996360c141a6
[11/29] scsi: esas2r: Use generic power management
        https://git.kernel.org/mkp/scsi/c/5f2d8c365050
[12/29] scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/17b5e4d14837
[13/29] scsi: hisi_sas_v3_hw: Don't use PCI helper functions
        https://git.kernel.org/mkp/scsi/c/027e508aea45
[14/29] scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm
        https://git.kernel.org/mkp/scsi/c/71c8f15e1dbc
[15/29] scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/eaf148359d00
[16/29] scsi: mpt3sas_scsih: Use generic power management
        https://git.kernel.org/mkp/scsi/c/17287305a526
[17/29] scsi: lpfc: Use generic power management
        https://git.kernel.org/mkp/scsi/c/ef6fa16b5d4a
[18/29] scsi: pm_8001: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/66aea31a2d26
[19/29] scsi: pm_8001: Use generic power management
        https://git.kernel.org/mkp/scsi/c/47c37c4dbf93
[20/29] scsi: hpsa: Use generic power management
        https://git.kernel.org/mkp/scsi/c/e5b79ebfb854
[21/29] scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/3a09951ae366
[22/29] scsi: 3w-9xxx: Use generic power management
        https://git.kernel.org/mkp/scsi/c/d53ae6bbeb71
[23/29] scsi: 3w-sas: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/7ea03ab79e00
[24/29] scsi: 3w-sas: Use generic power management
        https://git.kernel.org/mkp/scsi/c/99769d8d9109
[25/29] scsi: mvumi: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/bd7463cdbe1a
[26/29] scsi: mvumi: Use generic power management
        https://git.kernel.org/mkp/scsi/c/0572edbc32c5
[27/29] scsi: mvumi: Update function description
        https://git.kernel.org/mkp/scsi/c/53fdec73c14f
[28/29] scsi: pmcraid: Drop PCI Wakeup calls from .resume
        https://git.kernel.org/mkp/scsi/c/0aea8a8f3a77
[29/29] scsi: pmcraid: Use generic power management
        https://git.kernel.org/mkp/scsi/c/ac85cca31637

-- 
Martin K. Petersen	Oracle Linux Engineering
_______________________________________________
Linux-kernel-mentees mailing list
Linux-kernel-mentees@lists.linuxfoundation.org
https://lists.linuxfoundation.org/mailman/listinfo/linux-kernel-mentees

^ permalink raw reply	[flat|nested] 32+ messages in thread

end of thread, other threads:[~2020-12-01  5:19 UTC | newest]

Thread overview: 32+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-02 16:47 [Linux-kernel-mentees] [PATCH v4 00/29] scsi: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 01/29] scsi: megaraid_sas: Drop PCI wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 02/29] scsi: megaraid_sas: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 03/29] scsi: megaraid_sas: update function description Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 04/29] scsi: aacraid: Drop pci_enable_wake() from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 05/29] scsi: aacraid: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 06/29] scsi: aic7xxx: " Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 07/29] scsi: aic79xx: " Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 08/29] scsi: arcmsr: Drop PCI wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 09/29] scsi: arcmsr: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 10/29] scsi: esas2r: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 11/29] scsi: esas2r: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 12/29] scsi: hisi_sas_v3_hw: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 13/29] scsi: hisi_sas_v3_hw: Don't use PCI helper functions Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 14/29] scsi: hisi_sas_v3_hw: Remove extra function calls for runtime pm Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 15/29] scsi: mpt3sas_scsih: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 16/29] scsi: mpt3sas_scsih: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 17/29] scsi: lpfc: " Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 18/29] scsi: pm_8001: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 19/29] scsi: pm_8001: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 20/29] scsi: hpsa: " Vaibhav Gupta
2020-11-04 18:08   ` Don.Brace--- via Linux-kernel-mentees
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 21/29] scsi: 3w-9xxx: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 22/29] scsi: 3w-9xxx: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 23/29] scsi: 3w-sas: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 24/29] scsi: 3w-sas: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 25/29] scsi: mvumi: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 26/29] scsi: mvumi: use generic power management Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 27/29] scsi: mvumi: update function description Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 28/29] scsi: pmcraid: Drop PCI Wakeup calls from .resume Vaibhav Gupta
2020-11-02 16:47 ` [Linux-kernel-mentees] [PATCH v4 29/29] scsi: pmcraid: use generic power management Vaibhav Gupta
2020-12-01  5:05 ` [Linux-kernel-mentees] [PATCH v4 00/29] scsi: " Martin K. Petersen

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).