Hi Felipe,
On 27/06/14 16:54, Felipe Balbi wrote:
Hi,
On Wed, Jun 18, 2014 at 06:01:08PM +0100, Srinivas Kandagatla wrote:
Use case is when the phy is configured in host mode and a usb device is
attached to board before bootup. On bootup, with the existing code and
runtime pm enabled, the driver would decrement the pm usage count
without checking the current state of the phy. This pm usage count
decrement would trigger the runtime pm which than would abort the
usb enumeration which was in progress. In my case a usb stick gets
detected and then immediatly the driver goes to low power mode which is
not correct.
log:
[ 1.631412] msm_hsusb_host 12520000.usb: EHCI Host Controller
[ 1.636556] msm_hsusb_host 12520000.usb: new USB bus registered, assigned bus number 1
[ 1.642563] msm_hsusb_host 12520000.usb: irq 220, io mem 0x12520000
[ 1.658197] msm_hsusb_host 12520000.usb: USB 2.0 started, EHCI 1.00
[ 1.659473] hub 1-0:1.0: USB hub found
[ 1.663415] hub 1-0:1.0: 1 port detected
...
[ 1.973352] usb 1-1: new high-speed USB device number 2 using msm_hsusb_host
[ 2.107707] usb-storage 1-1:1.0: USB Mass Storage device detected
[ 2.108993] scsi0 : usb-storage 1-1:1.0
[ 2.678341] msm_otg 12520000.phy: USB in low power mode
[ 3.168977] usb 1-1: USB disconnect, device number 2
This issue was detected on IFC6410 board.
This patch fixes the intial runtime pm trigger by checking the phy
state and decrementing the pm use count only when the phy state is IDLE.
Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx>
---
drivers/usb/phy/phy-msm-usb.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/usb/phy/phy-msm-usb.c b/drivers/usb/phy/phy-msm-usb.c
index 3bb559d..78cc870 100644
--- a/drivers/usb/phy/phy-msm-usb.c
+++ b/drivers/usb/phy/phy-msm-usb.c
@@ -1229,7 +1229,9 @@ static void msm_otg_sm_work(struct work_struct *w)
motg->chg_state = USB_CHG_STATE_UNDEFINED;
motg->chg_type = USB_INVALID_CHARGER;
}
- pm_runtime_put_sync(otg->phy->dev);
+
+ if (otg->phy->state == OTG_STATE_B_IDLE)
+ pm_runtime_put_sync(otg->phy->dev);
instead, you might as well just use autosuspend.
autosuspend is a good idea and will provide a delay before rumtime
suspend, however the bug which is fixed here still needs to be fixed anyway.
runtime PM in this driver is not that great, the driver just increments
the pm use count if the phy is configured in host or deivce mode. This
patch fixes a bug in its state-machine which decrements pm use-count
without checking the state.
As this is a PHY driver, am not sure moving to autosuspend means we
would end up just adding some static inactivity delay? Is it really
going to add any value to this PHY driver?
--srini
--
To unsubscribe from this list: send the line "unsubscribe linux-arm-msm" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
