On Mon, Mar 14, 2016 at 6:59 PM, Bin Liu <b-liu@xxxxxx> wrote: > Hi, > > On Mon, Mar 14, 2016 at 06:17:26PM +0100, Yegor Yefremov wrote: >> On Mon, Mar 14, 2016 at 6:08 PM, Bin Liu <b-liu@xxxxxx> wrote: >> > Hi, >> > >> > On Mon, Mar 14, 2016 at 03:54:36PM +0100, Yegor Yefremov wrote: >> >> On Mon, Mar 14, 2016 at 3:42 PM, Bin Liu <b-liu@xxxxxx> wrote: >> >> > Hi, >> >> > >> >> > On Mon, Mar 14, 2016 at 03:33:51PM +0100, Yegor Yefremov wrote: >> >> >> On Mon, Mar 14, 2016 at 3:01 PM, Bin Liu <b-liu@xxxxxx> wrote: >> >> >> > Hi Yegor, >> >> >> > >> >> >> > On Mon, Mar 14, 2016 at 02:40:34PM +0100, Yegor Yefremov wrote: >> >> >> >> On Mon, Mar 14, 2016 at 1:11 PM, Ladislav Michl <ladis@xxxxxxxxxxxxxx> wrote: >> >> >> >> > On Mon, Mar 14, 2016 at 11:25:59AM +0100, Yegor Yefremov wrote: >> >> >> >> > >> >> >> >> > Hi Yegor, >> >> >> >> > >> >> >> >> >> Hi Bin, >> >> >> >> >> >> >> >> >> >> On Wed, Jan 27, 2016 at 11:01 AM, Ladislav Michl <ladis@xxxxxxxxxxxxxx> wrote: >> >> >> >> >> > Any chance it will actually happen? MUSB is still unusable for me and >> >> >> >> >> > doesn't work properly since TI's 2.6.37 kernel (oldest one I have >> >> >> >> >> > tried)... >> >> >> >> >> >> >> >> >> >> The issue with musb_h_tx_flush_fifo still persists. Kernel 4.5.0-rc6. >> >> >> >> >> Just activating SImCOM SIM7100E produces following error: >> >> >> >> >> >> >> >> >> >> # usb 1-1.3: new high-speed USB device number 6 using musb-hdrc >> >> >> >> >> usbcore: registered new interface driver option >> >> >> >> >> usbserial: USB Serial support registered for GSM modem (1-port) >> >> >> >> >> option 1-1.3:1.0: GSM modem (1-port) converter detected >> >> >> >> >> usb 1-1.3: GSM modem (1-port) converter now attached to ttyUSB4 >> >> >> >> >> option 1-1.3:1.1: GSM modem (1-port) converter detected >> >> >> >> >> usb 1-1.3: GSM modem (1-port) converter now attached to ttyUSB5 >> >> >> >> >> option 1-1.3:1.2: GSM modem (1-port) converter detected >> >> >> >> >> usb 1-1.3: GSM modem (1-port) converter now attached to ttyUSB6 >> >> >> >> >> option 1-1.3:1.3: GSM modem (1-port) converter detected >> >> >> >> >> usb 1-1.3: GSM modem (1-port) converter now attached to ttyUSB7 >> >> >> >> >> option 1-1.3:1.4: GSM modem (1-port) converter detected >> >> >> >> >> usb 1-1.3: GSM modem (1-port) converter now attached to ttyUSB8 >> >> >> >> >> qmi_wwan 1-1.3:1.5: cdc-wdm0: USB WDM device >> >> >> >> >> qmi_wwan 1-1.3:1.5 wwan0: register 'qmi_wwan' at >> >> >> >> >> usb-musb-hdrc.0.auto-1.3, WWAN/QMI device, b2:67:d8:31:8f:b6 >> >> >> >> >> usbcore: registered new interface driver qmi_wwan >> >> >> >> >> ------------[ cut here ]------------ >> >> >> >> >> WARNING: CPU: 0 PID: 868 at drivers/usb/musb/musb_host.c:139 >> >> >> >> >> musb_h_tx_flush_fifo+0xc4/0xe4() >> >> >> >> >> musb-hdrc musb-hdrc.0.auto: Could not flush host TX2 fifo: csr: 2403 >> >> >> >> >> Modules linked in: qmi_wwan option usb_wwan ftdi_sio usbserial >> >> >> >> >> musb_dsps musb_am335x >> >> >> >> >> CPU: 0 PID: 868 Comm: ModemManager Not tainted 4.5.0-rc6 #47 >> >> >> >> >> Hardware name: Generic AM33XX (Flattened Device Tree) >> >> >> >> >> [<c0017e34>] (unwind_backtrace) from [<c0014088>] (show_stack+0x10/0x14) >> >> >> >> >> [<c0014088>] (show_stack) from [<c02c5b24>] (dump_stack+0xb0/0xe4) >> >> >> >> >> [<c02c5b24>] (dump_stack) from [<c003bca4>] (warn_slowpath_common+0x7c/0xb8) >> >> >> >> >> [<c003bca4>] (warn_slowpath_common) from [<c003bd10>] >> >> >> >> >> (warn_slowpath_fmt+0x30/0x40) >> >> >> >> >> [<c003bd10>] (warn_slowpath_fmt) from [<c041ec10>] >> >> >> >> >> (musb_h_tx_flush_fifo+0xc4/0xe4) >> >> >> >> >> [<c041ec10>] (musb_h_tx_flush_fifo) from [<c041fd54>] >> >> >> >> >> (musb_cleanup_urb+0xbc/0x130) >> >> >> >> >> [<c041fd54>] (musb_cleanup_urb) from [<c0420404>] (musb_urb_dequeue+0xf0/0x128) >> >> >> >> >> [<c0420404>] (musb_urb_dequeue) from [<c03fa82c>] (usb_hcd_unlink_urb+0x50/0x84) >> >> >> >> >> [<c03fa82c>] (usb_hcd_unlink_urb) from [<c03fba70>] (usb_kill_urb+0x4c/0xc4) >> >> >> >> >> [<c03fba70>] (usb_kill_urb) from [<bf02f85c>] >> >> >> >> >> (usb_wwan_close+0xbc/0xf0 [usb_wwan]) >> >> >> >> >> [<bf02f85c>] (usb_wwan_close [usb_wwan]) from [<c0322ce8>] >> >> >> >> >> (tty_port_shutdown+0x90/0x9c) >> >> >> >> >> [<c0322ce8>] (tty_port_shutdown) from [<c03233cc>] (tty_port_close+0x24/0x4c) >> >> >> >> >> [<c03233cc>] (tty_port_close) from [<c031a258>] (tty_release+0xec/0x478) >> >> >> >> >> [<c031a258>] (tty_release) from [<c014ef38>] (__fput+0x80/0x1d0) >> >> >> >> >> [<c014ef38>] (__fput) from [<c005a8d8>] (task_work_run+0x90/0xc4) >> >> >> >> >> [<c005a8d8>] (task_work_run) from [<c0013a4c>] (do_work_pending+0x8c/0xb4) >> >> >> >> >> [<c0013a4c>] (do_work_pending) from [<c000f8e8>] (slow_work_pending+0xc/0x20) >> >> >> >> >> ---[ end trace fb2fc0420e2c4594 ]--- >> >> >> >> >> >> >> >> >> >> Could you please take a look at this? >> >> >> >> >> >> >> >> >> >> Yegor >> >> >> >> > >> >> >> >> > I cannot reproduce it anymore, as commits >> >> >> >> > 4c2ba0c67394514f3f75c660c9f5d02e66a7efd4 >> >> >> >> > bba40e6948b94cba71965285fbac31bd078c024a >> >> >> >> > fixed issue for me (using Inventra DMA on DM3730) - but those two are part >> >> >> >> > of 4.5.0-rc6 ... >> >> >> >> >> >> >> >> You are lucky! I've tried 4.5.0 and I still have this issue, though >> >> >> >> I'm on am335x and not on DM3730: >> >> >> > >> >> >> > How do you trigger the issue? Plug the modem then unplug it? >> >> >> > Can you please try the way in [1] to increase the loops to see if the >> >> >> > fifo could be flushed in a longer time? >> >> >> > >> >> >> > Regards, >> >> >> > -Bin. >> >> >> > >> >> >> > [1] http://www.spinics.net/lists/linux-usb/msg137242.html >> >> >> >> >> >> Do you mean something like this? >> >> >> >> >> >> diff --git a/drivers/usb/musb/musb_host.c b/drivers/usb/musb/musb_host.c >> >> >> index 58487a4..f6d6822 100644 >> >> >> --- a/drivers/usb/musb/musb_host.c >> >> >> +++ b/drivers/usb/musb/musb_host.c >> >> >> @@ -112,7 +112,7 @@ static void musb_h_tx_flush_fifo(struct musb_hw_ep *ep) >> >> >> struct musb *musb = ep->musb; >> >> >> void __iomem *epio = ep->regs; >> >> >> u16 csr; >> >> >> - int retries = 1000; >> >> >> + int retries = 10000; >> >> > >> >> > Yes. >> >> > >> >> >> >> >> >> csr = musb_readw(epio, MUSB_TXCSR); >> >> >> while (csr & MUSB_TXCSR_FIFONOTEMPTY) { >> >> >> >> >> >> No, it doesn't help. I'm triggering this issue via enabling it, i.e. >> >> >> our hardware can power on/off mPCIe slot via GPIO pin. When the system >> >> >> starts mPCI slot is disabled. The I enable it after booting into user >> >> >> space and the produces the error. It also happens if I insert USB >> >> > >> >> > Ok. I can imagine the device is no longer functional due to the fifo >> >> > flush issue. >> >> >> >> I've inserted the mPCIe modem into the system, where mPCIe is alsways >> >> on and the issue is still there. Some modems like Huawei MU609 are >> >> working, but when using with the latest usb_modeswitch and compiling >> >> all those drivers as modules (option, qmi_wwan etc.) I can also >> >> provoke the FIFO flush issue. >> > >> > What are the version number of old and new musb_modeswitch? Mine is >> > 2.0.1 but I didn't see such issue on my modems. >> >> This is not the usb_modeswitch, that makes trouble, but the fact, that >> the latest usb_modeswitch_data (it requires usb_modeswitch 2.3.0) has >> an entry for Huawei MU609/909 and tries to switch the modem in > > Oh, I see. > >> Configuration 1 instead of the default one. For usb_modeswitch to >> function properly all those drivers must be compiled as modules (at >> least this is the experience I've made with it). I'll try to find a > > I use modules in most of the times. > >> kernel configuration, that would trigger this behavior. >> >> What modems do you have? > > Alcatel HSPA Data Card, X230L > > I cannot find others at the moment. I just moved recently. OK :-) >> >> >> modem as a dongle, where usb_modeswitch first has to switch it from >> >> >> CD-ROM mode into modem mode. >> >> > >> >> > Any detail of the usb modem dongle? I will try to find one to take a >> >> > look. I have a few usb modems but none of them triggers this problem. >> >> >> >> TP-LINK MA260 (http://www.tp-link.com/en/products/details/cat-5090_MA260.html) >> > >> > It seems no where in US selling it. >> >> Can you get SIMcom 7100E and mPCIe adapter? > > I will do a search for it later. Btw if you do get hold of a mPCIe modem, you can have our Baltos iR5221 for testing. Ask Felipe Balbi. He knows, where you can find one sample. Yegor -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
