Hi! On Thu, Feb 11, 2021 at 01:03:54PM +0100, Ahmad Fatoum wrote: >Hello, > >On 18.01.21 20:09, Michael Grzeschik wrote: >> Since we don't power cycle the ports on start since patch "19bb0b2a usb: >> hub: Do not power-cycle usb devices on init" it is possible that the >> device on this port is already active from a previous enumeration. This >> way barebox will never get any change in USB_PORT_STAT_C_CONNECTION bit >> change. >> >> Although the device will probably still work fine after the following >> port reset, the current code will always miss reenumerating these still >> plugged devices. This patch fixes this by ignoring the check for >> STAT_C_CONNECTION bit and only go for USB_PORT_STAT_CONNECTION which >> should be enough. > >This breaks USB enumeration for me on an i.MX6Q. The USB flash driver behind >the hub is detected twice, which barebox doesn't like at all: > >barebox@Embest MarS Board i.MX6Dual:/ usb -s >usb: USB: scanning bus for devices... >usb1: Bus 001 Device 001: ID 0000:0000 EHCI Host Controller >usb1-0: Bus 001 Device 002: ID 1a40:0101 USB 2.0 Hub [MTT] >usb1-0-1: Bus 001 Device 003: ID 058f:6387 Mass Storage >Using index 0 for the new disk >usb1-0-1: Bus 001 Device 004: ID 058f:6387 Mass Storage >ERROR: register_device: already registered usb1-0-1 >ERROR: usb1-0-1: Failed to register device: Invalid argument >usb: 3 USB Device(s) found >Bus 001 Device 001: ID 0000:0000 EHCI Host Controller >Bus 001 Device 002: ID 1a40:0101 USB 2.0 Hub [MTT] >Bus 001 Device 003: ID 058f:6387 Mass Storage > >This seems to corrupt some internal barebox state as reading >/dev/disk0.0 will hang. Reading /dev/disk0 is fine however. > >With this patch reverted it however works: > >usb: USB: scanning bus for devices... >usb1: Bus 001 Device 001: ID 0000:0000 EHCI Host Controller >usb1-0: Bus 001 Device 002: ID 1a40:0101 USB 2.0 Hub [MTT] >usb1-0-1: Bus 001 Device 003: ID 058f:6387 Mass Storage >Using index 0 for the new disk >usb: 3 USB Device(s) found >Bus 001 Device 001: ID 0000:0000 EHCI Host Controller >Bus 001 Device 002: ID 1a40:0101 USB 2.0 Hub [MTT] >Bus 001 Device 003: ID 058f:6387 Mass Storage > >The USB drive is not self-powered. It doesn't make a difference >whether it's done directly after cold reset or: > usb -s > boot bnet > usb -s > >Cheers, >Ahmad I have reproduced that. My latest findings are that, when I debugged this, usb core on my board I was debugging was behaving uncommon. It was not setting the USB_PORT_STAT_C_CONNECTION in the first place. But the further patches I send did solve this. So this whole issue was due to an uncorrect configured over-current polarity setup. With my other patches applied I see this bit USB_PORT_STAT_C_CONNECTION set even without the power-cycle. My suggestion is to _revert_ this patch, as it is solving a bogus problem not related to the whole usb-core state machine. Thanks for testing! Michael >> >> Signed-off-by: Michael Grzeschik >> --- >> drivers/usb/core/hub.c | 3 +-- >> 1 file changed, 1 insertion(+), 2 deletions(-) >> >> diff --git a/drivers/usb/core/hub.c b/drivers/usb/core/hub.c >> index 01653d8c20..d1112248ee 100644 >> --- a/drivers/usb/core/hub.c >> +++ b/drivers/usb/core/hub.c >> @@ -336,8 +336,7 @@ static void usb_scan_port(struct usb_device_scan *usb_scan) >> dev_dbg(&dev->dev, "port%d: Status 0x%04x Change 0x%04x\n", >> port + 1, portstatus, portchange); >> >> - if (!(portchange & USB_PORT_STAT_C_CONNECTION) || >> - !(portstatus & USB_PORT_STAT_CONNECTION)) { >> + if (!(portstatus & USB_PORT_STAT_CONNECTION)) { >> if (get_time_ns() >= hub->connect_timeout) { >> dev_dbg(&dev->dev, "port%d: timeout\n", port + 1); >> /* Remove this device from scanning list */ >> -- Pengutronix e.K. | | Steuerwalder Str. 21 | http://www.pengutronix.de/ | 31137 Hildesheim, Germany | Phone: +49-5121-206917-0 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |