https://bugzilla.kernel.org/show_bug.cgi?id=215740 --- Comment #19 from Alan Stern (stern@xxxxxxxxxxxxxxxxxxx) --- Created attachment 305456 --> https://bugzilla.kernel.org/attachment.cgi?id=305456&action=edit Diagnostic to find DMA overlap The patches are prepared against a 6.6-rc3 tree (it hasn't been updated since then), so they might not match 6.6 exactly. At least we're getting some good data from the patch: > DMA-API: dwc3 dwc3.0.auto: cacheline tracking EEXIST, overlapping mappings > aren't supported, len 4 ... > Map error: 128857248 128857248 len 4 The fact that the addresses and lengths agree indicate that the patch works right, and len = 4 means that the mapping error is indeed for the buffer containing the hub status data. By the way, do you have any USB hubs besides this one on your system? The next question is what other mapping already exists for that buffer? This next patch may be able to tell us. It's supposed to produce a second stack dump when the overlap occurs, for the call that created the first mapping. -- You may reply to this email to add a comment. You are receiving this mail because: You are watching the assignee of the bug.
