binder: Handle start==NULL in binder_update_page_range()

commit 2a9edd056ed4fbf9d2e797c3fc06335af35bccc4 upstream.

The old loop wouldn't stop when reaching `start` if `start==NULL`, instead
continuing backwards to index -1 and crashing.

Luckily you need to be highly privileged to map things at NULL, so it's not
a big problem.

Fix it by adjusting the loop so that the loop variable is always in bounds.

This patch is deliberately minimal to simplify backporting, but IMO this
function could use a refactor. The jump labels in the second loop body are
horrible (the error gotos should be jumping to free_range instead), and
both loops would look nicer if they just iterated upwards through indices.
And the up_read()+mmput() shouldn't be duplicated like that.

Fixes: 457b9a6f09 ("Staging: android: add binder driver")
Change-Id: I4ad9023a3a40e1a6afdeb01a0bcee6a12e667a47
Signed-off-by: Jann Horn <jannh@google.com>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/r/20191018205631.248274-3-jannh@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[bwh: Backported to 3.16: There is no continue statement in the loop,
 so we only need to check the exit condition at the bottom]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
This commit is contained in:
Jann Horn 2019-10-18 22:56:31 +02:00 committed by syphyr
parent 407425bb9b
commit e30c8f0424
1 changed files with 3 additions and 3 deletions

View File

@ -738,8 +738,7 @@ static int __binder_update_page_range(struct binder_proc *proc, int allocate,
return 0;
free_range:
for (page_addr = end - PAGE_SIZE; page_addr >= start;
page_addr -= PAGE_SIZE) {
for (page_addr = end - PAGE_SIZE; 1; page_addr -= PAGE_SIZE) {
page = &proc->pages[(page_addr - proc->buffer) / PAGE_SIZE];
if (vma)
zap_page_range(vma, (uintptr_t)page_addr +
@ -750,7 +749,8 @@ err_map_kernel_failed:
__free_page(*page);
*page = NULL;
err_alloc_page_failed:
;
if (page_addr == start)
break;
}
err_no_vma:
if (mm) {