diff options
author | kraxel@bytesex.org[kaf24] <kraxel@bytesex.org[kaf24]> | 2005-05-20 23:16:46 +0000 |
---|---|---|
committer | kraxel@bytesex.org[kaf24] <kraxel@bytesex.org[kaf24]> | 2005-05-20 23:16:46 +0000 |
commit | 43e04538d1498ea63ab2231af1707188ef3b53e4 (patch) | |
tree | aa78f2d6ffd353aa0db5acdadfbb6bb9de130bfd /buildconfigs | |
parent | 366aeecb5bc89e4061a8ecd164796a083c9f54c2 (diff) | |
download | xen-43e04538d1498ea63ab2231af1707188ef3b53e4.tar.gz xen-43e04538d1498ea63ab2231af1707188ef3b53e4.tar.bz2 xen-43e04538d1498ea63ab2231af1707188ef3b53e4.zip |
bitkeeper revision 1.1159.258.150 (428e6fdeb6W1XbDj2YkcF53xkdUozA)
[PATCH] Re: PAE
> We're taking quite a chunk out of lowmem, though....
Uhm, well, no, we don't. We should though, fix is below.
Current code in unstable takes the address space away from the
vmalloc area, not the lowmem area. vmalloc space is 128 MB by
default, so the non-PAE hypervisor hole fits in and it works
nevertheless. The larger PAE mode hypervisor hole doesn't fit
in, so it breaks when you use enougth memory (somewhere around
800-900 MB). I think that is the bug Scott Parish trapped into.
Gerd
Diffstat (limited to 'buildconfigs')
0 files changed, 0 insertions, 0 deletions