From: Dan Carpenter <dan.carpenter@oracle.com>
To: Hillf Danton <hdanton@sina.com>
Cc: syzbot <syzbot+005037419ebdf14e1d87@syzkaller.appspotmail.com>,
igormtorrente@gmail.com, linux-kernel@vger.kernel.org,
linux-media@vger.kernel.org, syzkaller-bugs@googlegroups.com
Subject: Re: [syzbot] KASAN: use-after-free Read in em28xx_close_extension
Date: Tue, 27 Jul 2021 17:14:55 +0300 [thread overview]
Message-ID: <20210727141455.GM1931@kadam> (raw)
In-Reply-To: <20210727100151.2051-1-hdanton@sina.com>
On Tue, Jul 27, 2021 at 06:01:51PM +0800, Hillf Danton wrote:
> Along the probe path,
>
> em28xx_usb_probe
> dev = kzalloc(sizeof(*dev), GFP_KERNEL);
> retval = em28xx_init_dev(dev, udev, intf, nr);
> em28xx_init_extension(dev);
> em28xx_ir_init(struct em28xx *dev)
> kref_get(&dev->ref);
>
> kref_init(&dev->ref);
Good detective work.
I've created a Smatch check to try find these. It uses the fact that
Smatch creates a bunch of fake assignments to set all the struct members
of "dev" to zero. Then it uses the modification hook to find the
kref_init(). Those are sort of new uses for those hooks so that's quite
fun.
I'll test it out overnight and see how it works.
drivers/media/usb/em28xx/em28xx-cards.c:4086 em28xx_usb_probe() warn: kref has already been modifed (see line 3979)
regards,
dan carpenter
#include "smatch.h"
#include "smatch_slist.h"
#include "smatch_extra.h"
static int my_id;
STATE(fresh);
static int get_line(struct sm_state *sm)
{
struct sm_state *tmp;
int line = 0;
FOR_EACH_PTR(sm->possible, tmp) {
if (tmp->state == &undefined &&
tmp->line > line)
line = tmp->line;
} END_FOR_EACH_PTR(tmp);
if (!line)
return sm->line;
return line;
}
static void match_modify(struct sm_state *sm, struct expression *mod_expr)
{
if (sm->state != &fresh &&
mod_expr &&
mod_expr->type == EXPR_CALL &&
sym_name_is("kref_init", mod_expr->fn))
sm_warning("kref has already been modifed (see line %d)", get_line(sm));
set_state(my_id, sm->name, sm->sym, &undefined);
}
static bool is_alloc(struct expression *expr)
{
static struct expression *ignore, *alloc_expr;
struct expression *right;
if (!expr || expr->type != EXPR_ASSIGNMENT || expr->op != '=')
return false;
if (expr == ignore)
return false;
if (expr == alloc_expr)
return true;
right = strip_expr(expr->right);
if (right->type == EXPR_CALL &&
(sym_name_is("kzalloc", right->fn) ||
sym_name_is("kmalloc", right->fn))) {
alloc_expr = expr;
return true;
}
ignore = expr;
return false;
}
static void match_assign(struct expression *expr)
{
char *name;
if (!is_alloc(get_faked_expression()))
return;
name = expr_to_str(expr->left);
if (name && strstr(name, "refcount.refs.counter"))
set_state_expr(my_id, expr->left, &fresh);
free_string(name);
}
void check_kref_init_too_late(int id)
{
my_id = id;
add_hook(&match_assign, ASSIGNMENT_HOOK_AFTER);
add_modification_hook(my_id, &match_modify);
}
next prev parent reply other threads:[~2021-07-27 14:15 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-14 14:27 [syzbot] KASAN: use-after-free Read in em28xx_close_extension syzbot
[not found] ` <20210727100151.2051-1-hdanton@sina.com>
2021-07-27 14:14 ` Dan Carpenter [this message]
2021-07-27 15:13 ` Dongliang Mu
2021-07-27 15:33 ` Dan Carpenter
2021-07-28 13:39 ` Dan Carpenter
2021-10-01 8:55 ` syzbot
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20210727141455.GM1931@kadam \
--to=dan.carpenter@oracle.com \
--cc=hdanton@sina.com \
--cc=igormtorrente@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-media@vger.kernel.org \
--cc=syzbot+005037419ebdf14e1d87@syzkaller.appspotmail.com \
--cc=syzkaller-bugs@googlegroups.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).