On Thu, 13 Feb 2014, Stanislaw Gruszka wrote: > Hello, > > We have lot of "clear_halt for a busy endpoint" warnings reported. > > [ 546.758270] WARNING: at drivers/usb/host/ehci-hcd.c:1051 ehci_endpoint_reset+0x113/0x120() > [ 546.758272] clear_halt for a busy endpoint > ... > [ 546.758352] [<ffffffff8105cdec>] warn_slowpath_fmt+0x4c/0x50 > [ 546.758355] [<ffffffff81478cd3>] ehci_endpoint_reset+0x113/0x120 > [ 546.758359] [<ffffffff8145f655>] usb_hcd_reset_endpoint+0x25/0x70 > [ 546.758362] [<ffffffff81461568>] usb_reset_endpoint+0x28/0x40 > [ 546.758365] [<ffffffff814615ee>] usb_clear_halt+0x6e/0x80 > [ 546.758369] [<ffffffff8146c7ea>] usbdev_do_ioctl+0xb8a/0x1130 > [ 546.758372] [<ffffffff8146cdbe>] usbdev_ioctl+0xe/0x20 > [ 546.758378] [<ffffffff811aaef5>] do_vfs_ioctl+0x305/0x520 > [ 546.758380] [<ffffffff811ab191>] sys_ioctl+0x81/0xa0 > [ 546.758384] [<ffffffff8164dd19>] system_call_fastpath+0x16/0x1b > > https://bugzilla.redhat.com/show_bug.cgi?id=928143 > > It is triggered when user space driver call CLEAR_HALT ioctl when > endpoint is not actually halted. That's not quite right. The warning is triggered when the userspace driver (or any other driver) does an endpoint reset (which is part of a Clear-Halt) while there are URBs queued for the endpoint. Whether the endpoint is halted or not doesn't matter. > Schould we request to fix that in > user space, or can we remove this warning like on below patch (call > CLEAR_HALT ioctl, does not seems to do any harm except printing a > warning) ? Doing an endpoint reset while URBs are queued is a bug, because the results are not defined. It should be fixed in the userspace program. Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
