|
Description
|
Given a superblock with blatantly-bad values (ncg == 0, sbsize == 0(!),
ncyl == 0, cpg == 0, etc), the kernel should not attempt to mount it. At
the least, it should perform the same basic sanity checks that fsck does
before the latter says to try again with a different superblock.
An example bad superblock, with the specific problems mentioned above, is
attached. The observed symptom is that a read-only mount of the filesystem
will hang in a kernel memory allocation from ufs_getsummaryinfo().
|