The imx6q suspend/resume is broken on 3.16-rc due to PCIe

[Shawn Guo]

Hi Lucas,

I found that imx6q suspend/resume is broken on 3.16-rc, and it works
just fine if PCIe support is built out.  However the issue was not
exposed on 3.15, so I suspect the PCIe driver change merged in 3.16
is part of the cause.

$ echo mem > /sys/power/state
PM: Syncing filesystems ... done.
Freezing user space processes ... (elapsed 0.002 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.002 seconds) done.
sd 0:0:0:0: [sda] Synchronizing SCSI cache
sd 0:0:0:0: [sda] Stopping disk
PM: suspend of devices complete after 85.695 msecs
PM: suspend devices took 0.100 seconds
PM: late suspend of devices complete after 6.786 msecs
PM: noirq suspend of devices complete after 6.228 msecs
Disabling non-boot CPUs ...
CPU1: shutdown
CPU2: shutdown
CPU3: shutdown
Enabling non-boot CPUs ...
CPU1: Booted secondary processor
CPU1 is up
CPU2: Booted secondary processor
CPU2 is up
CPU3: Booted secondary processor
CPU3 is up

The issue is actually with resume procedure.  I noticed that Richard
added the following change in FSL kernel tree, which seems to fix the
issue.

@@ -229,6 +232,22 @@ static int imx6_suspend_finish(unsigned long val)

 static int imx6_pm_enter(suspend_state_t state)
 {
+       struct regmap *gpr;
+
+       /*
+        * L2 can exit by 'reset' or Inband beacon (from remote EP)
+        * toggling phy_powerdown has same effect as 'inband beacon'
+        * So, toggle bit18 of GPR1, to fix errata
+        * "PCIe PCIe does not support L2 Power Down"
+        */
+       gpr = syscon_regmap_lookup_by_compatible("fsl,imx6q-iomuxc-gpr");
+       if (IS_ERR(gpr)) {
+               pr_err("failed to find fsl,imx6q-iomux-gpr regmap\n");
+               return PTR_ERR(gpr);
+       }
+       regmap_update_bits(gpr, IOMUXC_GPR1, IMX6Q_GPR1_PCIE_TEST_PD,
+                       IMX6Q_GPR1_PCIE_TEST_PD);
+
        switch (state) {
        case PM_SUSPEND_STANDBY:
                imx6_set_lpm(STOP_POWER_ON);
@@ -258,6 +277,15 @@ static int imx6_pm_enter(suspend_state_t state)
                return -EINVAL;
        }

+       /*
+        * L2 can exit by 'reset' or Inband beacon (from remote EP)
+        * toggling phy_powerdown has same effect as 'inband beacon'
+        * So, toggle bit18 of GPR1, to fix errata
+        * "PCIe PCIe does not support L2 Power Down"
+        */
+       regmap_update_bits(gpr, IOMUXC_GPR1, IMX6Q_GPR1_PCIE_TEST_PD,
+                       !IMX6Q_GPR1_PCIE_TEST_PD);
+
        return 0;
 }
 
And I copied the details of the errata as below.

ERR005723 PCIe: PCIe does not support L2 power down

Description:
When PCIe works as Root Complex, it can exit L2 mode only through reset.
Since PCIe does not have a dedicated reset control bit, it cannot exit
L2 mode.

Projected Impact:
PCIe does not support L2 power down.

Workarounds:
The PCIe can be put into PDDQ mode to save PCIe PHY power and wake up
only by the OOB (Out of Band) wakeup signal (since wakeup by a beacon
from link partner is not supported) driven from the link partner (End
Point). This signal could be used as a GPIO interrupt to exit this mode.
The limitation of this workaround is that the link partner cannot be put
into L2 mode.

As I have no knowledge of how PCIe works, I provided all the info I
found here to kick the discussion.  Hopefully, we can find a proper fix
to get suspend/resume back to work.

Shawn