On 10/17/23 12:13, Bart Van Assche wrote: > > On 10/17/23 10:09, Bob Pearson wrote: >> I don't yet understand the logic of the srp driver to fix this but >> the problem is not in the rxe driver as far as I can tell. > Is there any information available that supports this conclusion? I > think the KASAN output that I shared shows that there is an issue in > the RXE driver. > > Thanks, > > Bart. > Should have mentioned that the last set of tests in srp/002 have much longer writes than the earlier ones which require a lot more processing and thus time. My belief is that the completion logic in srp is faulty but works if the underlying transport is fast but not if it is slow. Bob
