diff mbox series

[v2,1/4] PCI: Check PCI_PM_CTRL in pci_dev_wait()

Message ID 20240710205838.2413465-2-superm1@kernel.org (mailing list archive)
State Superseded
Headers show
Series Verify devices transition from D3cold to D0 | expand

Commit Message

Mario Limonciello July 10, 2024, 8:58 p.m. UTC
From: Mario Limonciello <mario.limonciello@amd.com>

A device that has gone through a reset may return a value in PCI_COMMAND
but that doesn't mean it's finished transitioning to D0.  On devices that
support power management explicitly check PCI_PM_CTRL on everything but
system resume to ensure the transition happened.

Devices that don't support power management and system resume will
continue to use PCI_COMMAND.

Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
---
 drivers/pci/pci.c | 27 ++++++++++++++++++++-------
 1 file changed, 20 insertions(+), 7 deletions(-)

Comments

Ilpo Järvinen July 11, 2024, 3:07 p.m. UTC | #1
On Wed, 10 Jul 2024, superm1@kernel.org wrote:

> From: Mario Limonciello <mario.limonciello@amd.com>
> 
> A device that has gone through a reset may return a value in PCI_COMMAND
> but that doesn't mean it's finished transitioning to D0.  On devices that
> support power management explicitly check PCI_PM_CTRL on everything but
> system resume to ensure the transition happened.
> 
> Devices that don't support power management and system resume will
> continue to use PCI_COMMAND.
> 
> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
> ---
>  drivers/pci/pci.c | 27 ++++++++++++++++++++-------
>  1 file changed, 20 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
> index 35fb1f17a589c..4ad02ad640518 100644
> --- a/drivers/pci/pci.c
> +++ b/drivers/pci/pci.c
> @@ -1270,21 +1270,34 @@ static int pci_dev_wait(struct pci_dev *dev, char *reset_type, int timeout)
>  	 * the read (except when CRS SV is enabled and the read was for the
>  	 * Vendor ID; in that case it synthesizes 0x0001 data).
>  	 *
> -	 * Wait for the device to return a non-CRS completion.  Read the
> -	 * Command register instead of Vendor ID so we don't have to
> -	 * contend with the CRS SV value.
> +	 * Wait for the device to return a non-CRS completion.  On devices
> +	 * that support PM control and on waits that aren't part of system
> +	 * resume read the PM control register to ensure the device has
> +	 * transitioned to D0.  On devices that don't support PM control,
> +	 * or during system resume read the command register to instead of
> +	 * Vendor ID so we don't have to contend with the CRS SV value.
>  	 */
>  	for (;;) {
> -		u32 id;
>  
>  		if (pci_dev_is_disconnected(dev)) {
>  			pci_dbg(dev, "disconnected; not waiting\n");
>  			return -ENOTTY;
>  		}
>  
> -		pci_read_config_dword(dev, PCI_COMMAND, &id);
> -		if (!PCI_POSSIBLE_ERROR(id))
> -			break;
> +		if (dev->pm_cap && strcmp(reset_type, "resume") != 0) {

Comparing to a string makes me feel reset_type should be changed to
something that allows direct compare and those values only mapped into 
string while printing it.
Mario Limonciello July 11, 2024, 3:10 p.m. UTC | #2
On 7/11/2024 10:07, Ilpo Järvinen wrote:
> On Wed, 10 Jul 2024, superm1@kernel.org wrote:
> 
>> From: Mario Limonciello <mario.limonciello@amd.com>
>>
>> A device that has gone through a reset may return a value in PCI_COMMAND
>> but that doesn't mean it's finished transitioning to D0.  On devices that
>> support power management explicitly check PCI_PM_CTRL on everything but
>> system resume to ensure the transition happened.
>>
>> Devices that don't support power management and system resume will
>> continue to use PCI_COMMAND.
>>
>> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com>
>> ---
>>   drivers/pci/pci.c | 27 ++++++++++++++++++++-------
>>   1 file changed, 20 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
>> index 35fb1f17a589c..4ad02ad640518 100644
>> --- a/drivers/pci/pci.c
>> +++ b/drivers/pci/pci.c
>> @@ -1270,21 +1270,34 @@ static int pci_dev_wait(struct pci_dev *dev, char *reset_type, int timeout)
>>   	 * the read (except when CRS SV is enabled and the read was for the
>>   	 * Vendor ID; in that case it synthesizes 0x0001 data).
>>   	 *
>> -	 * Wait for the device to return a non-CRS completion.  Read the
>> -	 * Command register instead of Vendor ID so we don't have to
>> -	 * contend with the CRS SV value.
>> +	 * Wait for the device to return a non-CRS completion.  On devices
>> +	 * that support PM control and on waits that aren't part of system
>> +	 * resume read the PM control register to ensure the device has
>> +	 * transitioned to D0.  On devices that don't support PM control,
>> +	 * or during system resume read the command register to instead of
>> +	 * Vendor ID so we don't have to contend with the CRS SV value.
>>   	 */
>>   	for (;;) {
>> -		u32 id;
>>   
>>   		if (pci_dev_is_disconnected(dev)) {
>>   			pci_dbg(dev, "disconnected; not waiting\n");
>>   			return -ENOTTY;
>>   		}
>>   
>> -		pci_read_config_dword(dev, PCI_COMMAND, &id);
>> -		if (!PCI_POSSIBLE_ERROR(id))
>> -			break;
>> +		if (dev->pm_cap && strcmp(reset_type, "resume") != 0) {
> 
> Comparing to a string makes me feel reset_type should be changed to
> something that allows direct compare and those values only mapped into
> string while printing it.
> 

Thanks, that's a great suggestion.  I'll add a patch earlier in the 
series to make an enum of the types instead and a mapping function for 
them to get the string as needed.
diff mbox series

Patch

diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index 35fb1f17a589c..4ad02ad640518 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -1270,21 +1270,34 @@  static int pci_dev_wait(struct pci_dev *dev, char *reset_type, int timeout)
 	 * the read (except when CRS SV is enabled and the read was for the
 	 * Vendor ID; in that case it synthesizes 0x0001 data).
 	 *
-	 * Wait for the device to return a non-CRS completion.  Read the
-	 * Command register instead of Vendor ID so we don't have to
-	 * contend with the CRS SV value.
+	 * Wait for the device to return a non-CRS completion.  On devices
+	 * that support PM control and on waits that aren't part of system
+	 * resume read the PM control register to ensure the device has
+	 * transitioned to D0.  On devices that don't support PM control,
+	 * or during system resume read the command register to instead of
+	 * Vendor ID so we don't have to contend with the CRS SV value.
 	 */
 	for (;;) {
-		u32 id;
 
 		if (pci_dev_is_disconnected(dev)) {
 			pci_dbg(dev, "disconnected; not waiting\n");
 			return -ENOTTY;
 		}
 
-		pci_read_config_dword(dev, PCI_COMMAND, &id);
-		if (!PCI_POSSIBLE_ERROR(id))
-			break;
+		if (dev->pm_cap && strcmp(reset_type, "resume") != 0) {
+			u16 pmcsr;
+
+			pci_read_config_word(dev, dev->pm_cap + PCI_PM_CTRL, &pmcsr);
+			if (!PCI_POSSIBLE_ERROR(pmcsr) &&
+			    (pmcsr & PCI_PM_CTRL_STATE_MASK) == PCI_D0)
+				break;
+		} else {
+			u32 id;
+
+			pci_read_config_dword(dev, PCI_COMMAND, &id);
+			if (!PCI_POSSIBLE_ERROR(id))
+				break;
+		}
 
 		if (delay > timeout) {
 			pci_warn(dev, "not ready %dms after %s; giving up\n",