[PATCH] isdn: hisax: isac: fix a possible concurrency use-after-free bug in ISAC_l1hw()

[PATCH] isdn: hisax: isac: fix a possible concurrency use-after-free bug in ISAC_l1hw()

[Young Xiao]

In drivers/isdn/hisax/isac.c, the function isac_interrupt() and
ISAC_l1hw() may be concurrently executed.

ISAC_l1hw()
    line 499: if (!cs->tx_skb)

isac_interrupt()
    line 250: dev_kfree_skb_irq(cs->tx_skb);

Thus, a possible concurrency use-after-free bug may occur in ISAC_l1hw().

To fix these bugs, the calls to spin_lock_irqsave() and
spin_unlock_irqrestore() are added in HFCPCI_l1hw(), to protect the
access to cs->tx_skb.

See commit 7418e6520f22 ("isdn: hisax: hfc_pci: Fix a possible concurrency
use-after-free bug in HFCPCI_l1hw()") for details.

Signed-off-by: Young Xiao <92siuyang@gmail.com>
---
 drivers/isdn/hisax/isac.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/isdn/hisax/isac.c b/drivers/isdn/hisax/isac.c
index bd40e06..60dd805 100644
--- a/drivers/isdn/hisax/isac.c
+++ b/drivers/isdn/hisax/isac.c
@@ -496,11 +496,13 @@ ISAC_l1hw(struct PStack *st, int pr, void *arg)
 		if (cs->debug & L1_DEB_LAPD)
 			debugl1(cs, "-> PH_REQUEST_PULL");
 #endif
+		spin_lock_irqsave(&cs->lock, flags);
 		if (!cs->tx_skb) {
 			test_and_clear_bit(FLG_L1_PULL_REQ, &st->l1.Flags);
 			st->l1.l1l2(st, PH_PULL | CONFIRM, NULL);
 		} else
 			test_and_set_bit(FLG_L1_PULL_REQ, &st->l1.Flags);
+		spin_unlock_irqrestore(&cs->lock, flags);
 		break;
 	case (HW_RESET | REQUEST):
 		spin_lock_irqsave(&cs->lock, flags);
-- 
2.7.4

Re: [PATCH] isdn: hisax: isac: fix a possible concurrency use-after-free bug in ISAC_l1hw()

[David Miller]

From: Young Xiao <92siuyang@gmail.com>
Date: Wed, 29 May 2019 16:48:59 +0800

> In drivers/isdn/hisax/isac.c, the function isac_interrupt() and
> ISAC_l1hw() may be concurrently executed.
> 
> ISAC_l1hw()
>     line 499: if (!cs->tx_skb)
> 
> isac_interrupt()
>     line 250: dev_kfree_skb_irq(cs->tx_skb);
> 
> Thus, a possible concurrency use-after-free bug may occur in ISAC_l1hw().
 ...
> +		spin_lock_irqsave(&cs->lock, flags);
>  		if (!cs->tx_skb) {
>  			test_and_clear_bit(FLG_L1_PULL_REQ, &st->l1.Flags);
>  			st->l1.l1l2(st, PH_PULL | CONFIRM, NULL);
>  		} else
>  			test_and_set_bit(FLG_L1_PULL_REQ, &st->l1.Flags);
> +		spin_unlock_irqrestore(&cs->lock, flags);

Nothing in this code accesses the cs->tx_skb object.  It is just a logic test
upon whether it is NULL or not.

I'm not applying stuff like this, sorry.

You have to show how something can actually go wrong when fixing a bug.