Ahh, thanks. Yes, the write-back of shared memory file mapping was broken
in 2.1.89 - the "offset" was wrong and the write-back data would as a
result essentially go to pretty much random places. That was the bug fixed
by -5, but I can imagine that the corruption done by -4 would have
confused innd further.
Thanks for the report - you had me nervous for a while with the first "-5
doesn't work either" one ;)
I hope this didn't get to bite anybody else, and that nobody lost any data
over it. Sorry,
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu