[PATCH] PCI: PM: Also move to D0 before calling pci_legacy_resume_early()

[PATCH] PCI: PM: Also move to D0 before calling pci_legacy_resume_early()

[Dexuan Cui]

In pci_legacy_suspend_late(), the device state is moved to PCI_UNKNOWN.
In pci_pm_thaw_noirq(), the state is supposed to be moved back to PCI_D0,
but the current code misses the pci_legacy_resume_early() path, so the
state remains in PCI_UNKNOWN in that path, and during hiberantion this
causes an error for the Mellanox VF driver, which fails to enable
MSI-X: pci_msi_supported() is false due to dev->current_state != PCI_D0:

mlx4_core a6d1:00:02.0: Detected virtual function - running in slave mode
mlx4_core a6d1:00:02.0: Sending reset
mlx4_core a6d1:00:02.0: Sending vhcr0
mlx4_core a6d1:00:02.0: HCA minimum page size:512
mlx4_core a6d1:00:02.0: Timestamping is not supported in slave mode
mlx4_core a6d1:00:02.0: INTx is not supported in multi-function mode, aborting
PM: dpm_run_callback(): pci_pm_thaw+0x0/0xd7 returns -95
PM: Device a6d1:00:02.0 failed to thaw: error -95

Signed-off-by: Dexuan Cui <decui@microsoft.com>
---
 drivers/pci/pci-driver.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
index 36dbe960306b..27dfc68db9e7 100644
--- a/drivers/pci/pci-driver.c
+++ b/drivers/pci/pci-driver.c
@@ -1074,15 +1074,16 @@ static int pci_pm_thaw_noirq(struct device *dev)
 			return error;
 	}
 
-	if (pci_has_legacy_pm_support(pci_dev))
-		return pci_legacy_resume_early(dev);
-
 	/*
 	 * pci_restore_state() requires the device to be in D0 (because of MSI
 	 * restoration among other things), so force it into D0 in case the
 	 * driver's "freeze" callbacks put it into a low-power state directly.
 	 */
 	pci_set_power_state(pci_dev, PCI_D0);
+
+	if (pci_has_legacy_pm_support(pci_dev))
+		return pci_legacy_resume_early(dev);
+
 	pci_restore_state(pci_dev);
 
 	if (drv && drv->pm && drv->pm->thaw_noirq)
-- 
2.19.1

Re: [PATCH] PCI: PM: Also move to D0 before calling pci_legacy_resume_early()

[Bjorn Helgaas]

On Thu, Aug 08, 2019 at 06:46:51PM +0000, Dexuan Cui wrote:
> 
> In pci_legacy_suspend_late(), the device state is moved to PCI_UNKNOWN.
> In pci_pm_thaw_noirq(), the state is supposed to be moved back to PCI_D0,
> but the current code misses the pci_legacy_resume_early() path, so the
> state remains in PCI_UNKNOWN in that path, and during hiberantion this
> causes an error for the Mellanox VF driver, which fails to enable
> MSI-X: pci_msi_supported() is false due to dev->current_state != PCI_D0:

s/hiberantion/hibernation/

Actually, it sounds more like "during *resume*, this causes an error",
so maybe you want s/hiberantion/resume/ instead?

> mlx4_core a6d1:00:02.0: Detected virtual function - running in slave mode
> mlx4_core a6d1:00:02.0: Sending reset
> mlx4_core a6d1:00:02.0: Sending vhcr0
> mlx4_core a6d1:00:02.0: HCA minimum page size:512
> mlx4_core a6d1:00:02.0: Timestamping is not supported in slave mode
> mlx4_core a6d1:00:02.0: INTx is not supported in multi-function mode, aborting
> PM: dpm_run_callback(): pci_pm_thaw+0x0/0xd7 returns -95
> PM: Device a6d1:00:02.0 failed to thaw: error -95
> 
> Signed-off-by: Dexuan Cui <decui@microsoft.com>
> ---
>  drivers/pci/pci-driver.c | 7 ++++---
>  1 file changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/pci/pci-driver.c b/drivers/pci/pci-driver.c
> index 36dbe960306b..27dfc68db9e7 100644
> --- a/drivers/pci/pci-driver.c
> +++ b/drivers/pci/pci-driver.c
> @@ -1074,15 +1074,16 @@ static int pci_pm_thaw_noirq(struct device *dev)
>  			return error;
>  	}
>  
> -	if (pci_has_legacy_pm_support(pci_dev))
> -		return pci_legacy_resume_early(dev);
> -
>  	/*
>  	 * pci_restore_state() requires the device to be in D0 (because of MSI
>  	 * restoration among other things), so force it into D0 in case the
>  	 * driver's "freeze" callbacks put it into a low-power state directly.
>  	 */
>  	pci_set_power_state(pci_dev, PCI_D0);
> +
> +	if (pci_has_legacy_pm_support(pci_dev))
> +		return pci_legacy_resume_early(dev);
> +
>  	pci_restore_state(pci_dev);
>  
>  	if (drv && drv->pm && drv->pm->thaw_noirq)
> -- 
> 2.19.1
> 

RE: [PATCH] PCI: PM: Also move to D0 before calling pci_legacy_resume_early()

[Dexuan Cui]

> From: Bjorn Helgaas <helgaas@kernel.org>
> Sent: Thursday, August 8, 2019 12:19 PM
> To: Dexuan Cui <decui@microsoft.com>
> 
> On Thu, Aug 08, 2019 at 06:46:51PM +0000, Dexuan Cui wrote:
> >
> > In pci_legacy_suspend_late(), the device state is moved to PCI_UNKNOWN.
> > In pci_pm_thaw_noirq(), the state is supposed to be moved back to PCI_D0,
> > but the current code misses the pci_legacy_resume_early() path, so the
> > state remains in PCI_UNKNOWN in that path, and during hiberantion this
> > causes an error for the Mellanox VF driver, which fails to enable
> > MSI-X: pci_msi_supported() is false due to dev->current_state != PCI_D0:
> 
> s/hiberantion/hibernation/

Thanks for spoting this typo! :-)
 
> Actually, it sounds more like "during *resume*, this causes an error",
> so maybe you want s/hiberantion/resume/ instead?

Yes, it's during "resume", and to be more accurate, it happens during the
"resume" of the "thaw" phase: when we run "echo disk > /sys/power/state",
first the kernel "freezes" all the devices and create a hibernation image, then
the kernel "thaws" the devices including the disk/NIC, and writes the memory
to the disk and powers down. This patch fixes the error message for the
Mellanox VF driver in this phase. 

When the system starts again, a fresh kernel starts to run, and when the kernel
detects that a hibernation image was saved, the kernel "quiesces" the devices,
and "restores" the devices from the saved image. Here device_resume_noirq()
-> ... -> pci_pm_restore_noirq() -> pci_pm_default_resume_early() ->
pci_power_up() moves the device states back to PCI_D0. This path is not broken
and doesn't need my patch. 

Thanks,
-- Dexuan