Skip to content

Commit 1d3d0f8

Browse files
keesgregkh
authored andcommitted
fs/exec.c: account for argv/envp pointers
commit 98da7d08850fb8bdeb395d6368ed15753304aa0c upstream. When limiting the argv/envp strings during exec to 1/4 of the stack limit, the storage of the pointers to the strings was not included. This means that an exec with huge numbers of tiny strings could eat 1/4 of the stack limit in strings and then additional space would be later used by the pointers to the strings. For example, on 32-bit with a 8MB stack rlimit, an exec with 1677721 single-byte strings would consume less than 2MB of stack, the max (8MB / 4) amount allowed, but the pointers to the strings would consume the remaining additional stack space (1677721 * 4 == 6710884). The result (1677721 + 6710884 == 8388605) would exhaust stack space entirely. Controlling this stack exhaustion could result in pathological behavior in setuid binaries (CVE-2017-1000365). [akpm@linux-foundation.org: additional commenting from Kees] Fixes: b6a2fea ("mm: variable length argument support") Link: http://lkml.kernel.org/r/20170622001720.GA32173@beast Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Rik van Riel <riel@redhat.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Qualys Security Advisory <qsa@qualys.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
1 parent 22da7ca commit 1d3d0f8

1 file changed

Lines changed: 24 additions & 4 deletions

File tree

fs/exec.c

Lines changed: 24 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -206,8 +206,26 @@ static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
206206

207207
if (write) {
208208
unsigned long size = bprm->vma->vm_end - bprm->vma->vm_start;
209+
unsigned long ptr_size;
209210
struct rlimit *rlim;
210211

212+
/*
213+
* Since the stack will hold pointers to the strings, we
214+
* must account for them as well.
215+
*
216+
* The size calculation is the entire vma while each arg page is
217+
* built, so each time we get here it's calculating how far it
218+
* is currently (rather than each call being just the newly
219+
* added size from the arg page). As a result, we need to
220+
* always add the entire size of the pointers, so that on the
221+
* last call to get_arg_page() we'll actually have the entire
222+
* correct size.
223+
*/
224+
ptr_size = (bprm->argc + bprm->envc) * sizeof(void *);
225+
if (ptr_size > ULONG_MAX - size)
226+
goto fail;
227+
size += ptr_size;
228+
211229
acct_arg_size(bprm, size / PAGE_SIZE);
212230

213231
/*
@@ -225,13 +243,15 @@ static struct page *get_arg_page(struct linux_binprm *bprm, unsigned long pos,
225243
* to work from.
226244
*/
227245
rlim = current->signal->rlim;
228-
if (size > ACCESS_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4) {
229-
put_page(page);
230-
return NULL;
231-
}
246+
if (size > READ_ONCE(rlim[RLIMIT_STACK].rlim_cur) / 4)
247+
goto fail;
232248
}
233249

234250
return page;
251+
252+
fail:
253+
put_page(page);
254+
return NULL;
235255
}
236256

237257
static void put_arg_page(struct page *page)

0 commit comments

Comments
 (0)