In patch-2.4.16:kernel/sysctl.c
is this turdlet lurking in wait for a 3am kernel hacker.
{VM_MAX_READAHEAD, "max-readahead",
&vm_max_readahead,sizeof(int), 0644, NULL, &proc_dointvec},
While "VM_MAX_READAHEAD" will always be a separate object to
"vm_max_readahead", it will not pass what Kernighan and Plaugher
describe as the 'telephone test'.
"grep -i VM_MIN_READAHEAD" for the same threat to sanity.
cu johnf