>Try to log into another VC and:
>-----------------------------------------------
>Welcome to Linux 2.1.124.
>
>wolverine login: ds
>/bin/login: can't map '/lib/libc.so.5'
>/bin/login: can't map '/lib/libc.so.5'
>/bin/login: can't load library 'libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't load library 'libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't load library 'libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't map '/lib/libc.so.5'
>/bin/sh: can't load library 'libc.so.5'
>Unable to load interpreter
>INIT: Id "c3" respawning too fast: disabled for 5 minutes
>--------------------------------------
If after that message everything worked fine, yes this should be the
normal behavior (I think ;-).
>I will do a few more tests and if I find anything usefull I will report
>it.
Right now I am changing some things in the patch (not that the old one was
completly wrong but there was very good reasons to run shrink_mmap() for
more than num_phys_pages because the page aging (I didn' t thought at
that some days ago)).
Note that the really important part of the patch are the
__get_free_pages() changes and the kswapd changes. Other things are minor
issues. Without that changes the system die, with that changes the system
live.
And swapoff -a seems to have problems. It' s not safe (even if the system
doesn' t deadlock anymore, but there _seems_ to be a race between swapoff
and the swap cache if during swapoff the system is swapping out a lot (I
have to think about it reading the code though so I am not sure right
now...)).
Andrea[s] Arcangeli
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/