uprobes: ensure that uprobe->offset and ->ref_ctr_offset are properly aligned
authorOleg Nesterov <oleg@redhat.com>
Mon, 4 May 2020 16:47:25 +0000 (18:47 +0200)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 9 Jun 2020 16:49:24 +0000 (09:49 -0700)
commit013b2deba9a6b80ca02f4fafd7dedf875e9b4450
tree5bd2542332362d7ea7e39c5243a0845515bfc789
parent8b4d37db9a566deaf22065ba1ba9b19c9fb964b4
uprobes: ensure that uprobe->offset and ->ref_ctr_offset are properly aligned

uprobe_write_opcode() must not cross page boundary; prepare_uprobe()
relies on arch_uprobe_analyze_insn() which should validate "vaddr" but
some architectures (csky, s390, and sparc) don't do this.

We can remove the BUG_ON() check in prepare_uprobe() and validate the
offset early in __uprobe_register(). The new IS_ALIGNED() check matches
the alignment check in arch_prepare_kprobe() on supported architectures,
so I think that all insns must be aligned to UPROBE_SWBP_INSN_SIZE.

Another problem is __update_ref_ctr() which was wrong from the very
beginning, it can read/write outside of kmap'ed page unless "vaddr" is
aligned to sizeof(short), __uprobe_register() should check this too.

Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Reviewed-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Acked-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Sven Schnelle <svens@linux.ibm.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/events/uprobes.c