sys_personality: change sys_personality() to accept "unsigned int" instead of u_long
authorOleg Nesterov <oleg@redhat.com>
Fri, 4 Jun 2010 21:14:58 +0000 (14:14 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 4 Jun 2010 22:21:45 +0000 (15:21 -0700)
commit485d527686850d68a0e9006dd9904f19f122485e
tree8400c646135bb4ce68f137004298e1be7fdbd913
parentd6d03f9158516b50d0d343158e3f33bcff1e4ca5
sys_personality: change sys_personality() to accept "unsigned int" instead of u_long

task_struct->pesonality is "unsigned int", but sys_personality() paths use
"unsigned long pesonality".  This means that every assignment or
comparison is not right.  In particular, if this argument does not fit
into "unsigned int" __set_personality() changes the caller's personality
and then sys_personality() returns -EINVAL.

Turn this argument into "unsigned int" and avoid overflows.  Obviously,
this is the user-visible change, we just ignore the upper bits.  But this
can't break the sane application.

There is another thing which can confuse the poorly written applications.
User-space thinks that this syscall returns int, not long.  This means
that the returned value can be negative and look like the error code.  But
note that libc won't be confused and thus errno won't be set, and with
this patch the user-space can never get -1 unless sys_personality() really
fails.  And, most importantly, the negative RET != -1 is only possible if
that app previously called personality(RET).

Pointed-out-by: Wenming Zhang <wezhang@redhat.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
include/linux/personality.h
include/linux/syscalls.h
kernel/exec_domain.c