On 21.06.2017 10:06, rocko r wrote:
Does the Samsung 840 EVO SSD support UAS (USB attached SCSI?) i.e. does
it use the uas driver instead of mass-storage?
I don't think UAS is an issue here: I tried disabling it for the SSD
using a quirk (ie with the line "options usb_storage
quirks=174c:1053:u" in the file /etc/modprobe.d/usb_storage.conf) and
the kernel still locked up.
Could you take full logs from boot with xhci debugging and tracing both enabled?
See attached. The SSD is /dev/sda and /media/cat is where it tries to
mount /dev/sda1, and there's a NULL dereference at 32.736157 that
might be the cause of the lockup.
Thanks, looks like UAS and xhci streams are still in use.
This is most likely a simple off by one error. stream ID 0 is not
valid, and turns out we actually have no stream ring allocated for
ep->stream_info->stream_rings[0]
In all other places we we start stream_id indexing from 1, except in this case
when bailing out for a suddenly removed (hotplug) xHC host.
does this help:
diff --git a/drivers/usb/host/xhci-ring.c b/drivers/usb/host/xhci-ring.c
index 03f63f5..e555a9c 100644
--- a/drivers/usb/host/xhci-ring.c
+++ b/drivers/usb/host/xhci-ring.c
@@ -845,11 +845,11 @@ static void xhci_kill_endpoint_urbs(struct xhci_hcd *xhci,
(ep->ep_state & EP_GETTING_NO_STREAMS)) {
int stream_id;
- for (stream_id = 0; stream_id < ep->stream_info->num_streams;
+ for (stream_id = 1; stream_id < ep->stream_info->num_streams;
stream_id++) {
xhci_dbg_trace(xhci, trace_xhci_dbg_cancel_urb,
"Killing URBs for slot ID %u, ep index %u, stream %u",
- slot_id, ep_index, stream_id + 1);
+ slot_id, ep_index, stream_id);
xhci_kill_ring_urbs(xhci,
ep->stream_info->stream_rings[stream_id]);
}
Thanks
Mathias
--
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
