>>>>> "Izumi" == Izumi Tsutsui <tsutsui@ceres.dti.ne.jp> writes:

    Izumi> The previous patch contains some redundant changes which I
    Izumi> have not confirmed. Could you try this one?

    Izumi> I tried full build by "build.sh -E build" and a test script
    Izumi> which reads/writes ~80MB files on wd0 more than 1 hour, and
    Izumi> both of them worked fine without problem on my RaQ2 with
    Izumi> 256MB RAM. (I have not applied chuq's patch yet, BTW)

I've tested your new patch w/ and w/o chuq's separate "pmap_unwire:
wiring ..." patch on my qube2 (256MB RAM, 160 GB IDE harddisk):

- Chuq's patch primarily seems to 'control' the "pmap_unwire: ..."
  message display. Data corruptions and frozen systems (see below) do
  not appear (cor)related to this patch.

- Your new patch seems to perform a bit better than your old one:
  System perforamnce is now the same as with a stock kernel. The old
  patch decreases system performance about 10%.

- I'm still able to provoke data corruptions and frozen systems, but
  it's much much better than w/o your patches.

  Concering data corruption: Your old patch seems to be more reliable
  than your new one. The more processes run in parallel the more
  likely the data corruptions appear. My test script reads (sha1),
  compares and copy/verifies several 100 MB files. In parallel I've
  untarred a huge *.tgz file.

  Concerning the frozen system: I'm not able to enter ddb, no reaction
  to serial break signal (while 'ddb.fromconsole = 1').

So, it's much much better, but (as expected?) not perfect.

First thanks for your effort, and anything else I can test?

Markus.