Skip to contentSkip to conversation

tty: fix data race in flush_to_ldisc

PATCH Superseded Dmitry Vyukov 3 participants 6 messages Sep 17, 2015 TTY layer and serial drivers

v2 of 5 Superseded Latest is v5 →

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 tty: fix data race in flush_to_ldisc

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

The data race was found with KernelThreadSanitizer (KTSAN).

---

Changed since first version:
- remove WRITE_ONCE when updating port->itty

---
 drivers/tty/tty_buffer.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/tty/tty_buffer.c +1 -1
index 5a3fa89..23de97d 100644
@@ -467,7 +467,7 @@ static void flush_to_ldisc(struct work_struct *work)
 	struct tty_struct *tty;
 	struct tty_ldisc *disc;
 
-	tty = port->itty;
+	tty = READ_ONCE(port->itty);
 	if (tty == NULL)
 		return;
 
Signature · 1 line
2.6.0.rc0.131.gf624c3d
maintainer
#1 lore raw
Quoted 11 lines · Dmitry Vyukov

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

The data race was found with KernelThreadSanitizer (KTSAN).

You sent 3 patches here, but no hint as to what order they need to be applied in. Please resend them as a patch series (i.e. 1/3, 2/3, 3/3) so they can be applied correctly.

thanks,

greg k-h

#2 lore raw
Quoted 9 lines · Greg KH
Quoted 8 lines · Dmitry Vyukov

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

Quoted 7 lines
Quoted 3 lines

The data race was found with KernelThreadSanitizer (KTSAN).

You sent 3 patches here, but no hint as to what order they need to be applied in. Please resend them as a patch series (i.e. 1/3, 2/3, 3/3) so they can be applied correctly.

Greg,

I don't think these 3 patches are dependent on each other; I think they can be applied in any order.

Regards,
Peter Hurley

#3 lore raw
Quoted 14 lines · Peter Hurley
Quoted 13 lines · Greg KH
Quoted 11 lines · Dmitry Vyukov

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

The data race was found with KernelThreadSanitizer (KTSAN).

Quoted 8 lines
Quoted 3 lines

You sent 3 patches here, but no hint as to what order they need to be applied in. Please resend them as a patch series (i.e. 1/3, 2/3, 3/3) so they can be applied correctly.

Greg,

I don't think these 3 patches are dependent on each other; I think they can be applied in any order.

Yes, these patches are independent and can be applied in any order, and any subset of them can be applied.
I can send them as patch series if necessary, though.

Signature · 11 lines
Dmitry Vyukov, Software Engineer, [email protected]
Google Germany GmbH, Dienerstraße 12, 80331, München
Geschäftsführer: Graham Law, Christine Elizabeth Flores
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Diese E-Mail ist vertraulich. Wenn Sie nicht der richtige Adressat
sind, leiten Sie diese bitte nicht weiter, informieren Sie den
Absender und löschen Sie die E-Mail und alle Anhänge. Vielen Dank.
This e-mail is confidential. If you are not the right addressee please
do not forward it, please inform the sender, and please erase this
e-mail including any attachments. Thanks.
maintainer
#4 lore raw
Quoted 21 lines · Dmitry Vyukov
Quoted 20 lines · Peter Hurley
Quoted 9 lines · Greg KH
Quoted 8 lines · Dmitry Vyukov

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

Quoted 7 lines
Quoted 3 lines

The data race was found with KernelThreadSanitizer (KTSAN).

You sent 3 patches here, but no hint as to what order they need to be applied in. Please resend them as a patch series (i.e. 1/3, 2/3, 3/3) so they can be applied correctly.

Greg,

Quoted 7 lines
Quoted 2 lines

I don't think these 3 patches are dependent on each other; I think they can be applied in any order.

Yes, these patches are independent and can be applied in any order, and any subset of them can be applied.
I can send them as patch series if necessary, though.

Please do, that makes it easier for me.

thanks,

greg k-h

maintainer
#5 lore raw
In reply to Peter Hurley
Quoted 14 lines · Peter Hurley
Quoted 13 lines · Greg KH
Quoted 11 lines · Dmitry Vyukov

flush_to_ldisc reads port->itty and checks that it is not NULL, concurrently release_tty sets port->itty to NULL. It is possible that flush_to_ldisc loads port->itty once, ensures that it is not NULL, but then reloads it again and uses. The second load can already return NULL, which will cause a crash.

Use READ_ONCE to read port->itty.

The data race was found with KernelThreadSanitizer (KTSAN).

Quoted 8 lines
Quoted 3 lines

You sent 3 patches here, but no hint as to what order they need to be applied in. Please resend them as a patch series (i.e. 1/3, 2/3, 3/3) so they can be applied correctly.

Greg,

I don't think these 3 patches are dependent on each other; I think they can be applied in any order.

How do I know that? :)