eyal reizer wrote: > Hello David, > > I found what was the cause for the previous crash. > It was the init sequence. > > 1. in 'devconnect.c' in the function 'wusbhc_devconnect_ack()' the > driver calls: > > wusbhc->set_ptk(wusbhc, idx, 0, NULL, 0); > > This is done before the device info structure was filled (by calling > wusbhc_cack_add which is done a couple of steps later. > This caused an assert in the firmware of the HWA as the device address > was not known to it yet. > When i reversed the order of the calls the crash did not happen anymore. > What do you think, is it OK to do it? Yes, can you produce a patch, please? (You should probably get the firmware bug fixed too.) > 2. After solving this, the connection process continued and then was > halted when the driver tried to set the RPIPE descriptor with a value of > 0 written to the MaxBurst field of the RPIPE descriptor. This caused > another assert in the HWA firmware. I guess this is for remote device's ep0. bMaxBurst should 1 and bMaxSequence should be 2. Can you produce a patch for this? > When changing this value from 0 to something else the process continued. > The change i did was in rpipe_aim() function in wa-rpipe.c > What is your view on this change? > > > 3. Now the system doesn't crash anymore but i see the following error: > > [ 169.976034] hwa-hc 5-3.3:1.0: DN CONNECT: device 3f 04 6d 4f ce c2 64 > fa df 36 c8 c2 f8 d4 e2 9a @ 0 (self-beacon) wants to reconnect > [ 169.976037] Eyal -> wusbhc_handle_dn_connect -> dnc = 0 > [ 169.976041] hwa-hc 5-3.3:1.0: DN CONNECT: device 3f 04 6d 4f ce c2 64 > fa df 36 c8 c2 f8 d4 e2 9a @ 0 (self-beacon) wants to reconnect > [ 169.976044] Eyal -> wusbhc_handle_dn_connect -> dnc = 0 > [ 169.976047] hwa-hc 5-3.3:1.0: DN CONNECT: device 3f 04 6d 4f ce c2 64 > fa df 36 c8 c2 f8 d4 e2 9a @ 0 (self-beacon) wants to reconnect > [ 170.272026] usb 6-1: new variable speed Wireless USB device using > hwa-hc and address 2 > [ 170.282785] xfer f0f57c40#0: REQ submit failed: -2 This seems to be because the urb for the segment is queued to a endpoint that doesn't exist. Doesn't make any sense to me, I'm completely unfamiliar with the WA code. David -- David Vrabel, Senior Software Engineer, Drivers CSR, Churchill House, Cambridge Business Park, Tel: +44 (0)1223 692562 Cowley Road, Cambridge, CB4 0WZ http://www.csr.com/ -- 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
