2022-09-01 02:33:06

by Yinbo Zhu

[permalink] [raw]
Subject: [PATCH v3] usb: ohci-platform: fix usb disconnect issue after s4

Avoid retaining bogus hardware states during resume-from-hibernation.
Previously we had reset the hardware as part of preparing to reinstate
the snapshot image. But we can do better now with the new PM framework,
since we know exactly which resume operations are from hibernation.

According to the commit 'cd1965db054e ("USB: ohci: move ohci_pci_{
suspend,resume} to ohci-hcd.c")' and commit '6ec4beb5c701 ("USB: new
flag for resume-from-hibernation")', the flag "hibernated" is for
resume-from-hibernation and it should be true when usb resume from disk.

When this flag "hibernated" is set, the drivers will reset the hardware
to get rid of any existing state and make sure resume from hibernation
re-enumerates everything for ohci.

Signed-off-by: Yinbo Zhu <[email protected]>
---
Change in v3:
Cancel helper routines.

drivers/usb/host/ohci-platform.c | 28 ++++++++++++++++++++++++----
1 file changed, 24 insertions(+), 4 deletions(-)

diff --git a/drivers/usb/host/ohci-platform.c b/drivers/usb/host/ohci-platform.c
index 0adae6265127..822a0f927e62 100644
--- a/drivers/usb/host/ohci-platform.c
+++ b/drivers/usb/host/ohci-platform.c
@@ -289,7 +289,7 @@ static int ohci_platform_suspend(struct device *dev)
return ret;
}

-static int ohci_platform_resume(struct device *dev)
+static int ohci_platform_resume_common(struct device *dev, bool hibernated)
{
struct usb_hcd *hcd = dev_get_drvdata(dev);
struct usb_ohci_pdata *pdata = dev_get_platdata(dev);
@@ -301,7 +301,7 @@ static int ohci_platform_resume(struct device *dev)
return err;
}

- ohci_resume(hcd, false);
+ ohci_resume(hcd, hibernated);

pm_runtime_disable(dev);
pm_runtime_set_active(dev);
@@ -309,6 +309,16 @@ static int ohci_platform_resume(struct device *dev)

return 0;
}
+
+static int ohci_platform_resume(struct device *dev)
+{
+ return ohci_platform_resume_common(dev, false);
+}
+
+static int ohci_platform_restore(struct device *dev)
+{
+ return ohci_platform_resume_common(dev, true);
+}
#endif /* CONFIG_PM_SLEEP */

static const struct of_device_id ohci_platform_ids[] = {
@@ -325,8 +335,16 @@ static const struct platform_device_id ohci_platform_table[] = {
};
MODULE_DEVICE_TABLE(platform, ohci_platform_table);

-static SIMPLE_DEV_PM_OPS(ohci_platform_pm_ops, ohci_platform_suspend,
- ohci_platform_resume);
+#ifdef CONFIG_PM_SLEEP
+static const struct dev_pm_ops ohci_platform_pm_ops = {
+ .suspend = ohci_platform_suspend,
+ .resume = ohci_platform_resume,
+ .freeze = ohci_platform_suspend,
+ .thaw = ohci_platform_resume,
+ .poweroff = ohci_platform_suspend,
+ .restore = ohci_platform_restore,
+};
+#endif

static struct platform_driver ohci_platform_driver = {
.id_table = ohci_platform_table,
@@ -335,7 +353,9 @@ static struct platform_driver ohci_platform_driver = {
.shutdown = usb_hcd_platform_shutdown,
.driver = {
.name = "ohci-platform",
+#ifdef CONFIG_PM_SLEEP
.pm = &ohci_platform_pm_ops,
+#endif
.of_match_table = ohci_platform_ids,
.probe_type = PROBE_PREFER_ASYNCHRONOUS,
}
--
2.31.1


2022-09-01 14:51:06

by Alan Stern

[permalink] [raw]
Subject: Re: [PATCH v3] usb: ohci-platform: fix usb disconnect issue after s4

On Thu, Sep 01, 2022 at 09:54:46AM +0800, Yinbo Zhu wrote:
> Avoid retaining bogus hardware states during resume-from-hibernation.
> Previously we had reset the hardware as part of preparing to reinstate
> the snapshot image. But we can do better now with the new PM framework,
> since we know exactly which resume operations are from hibernation.
>
> According to the commit 'cd1965db054e ("USB: ohci: move ohci_pci_{
> suspend,resume} to ohci-hcd.c")' and commit '6ec4beb5c701 ("USB: new
> flag for resume-from-hibernation")', the flag "hibernated" is for
> resume-from-hibernation and it should be true when usb resume from disk.
>
> When this flag "hibernated" is set, the drivers will reset the hardware
> to get rid of any existing state and make sure resume from hibernation
> re-enumerates everything for ohci.

The patch title mentions "usb disconnect issue", but the description
says nothing about it. Please explain what this disconnect issue is.

Alan Stern