Skip to contentSkip to conversation

aoe: fix the potential use-after-free problem in more places

PATCH In review Lee, Chun-Yi 3 participants 4 messages Apr 10, 2024 – May 13, 2024

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 aoe: fix the potential use-after-free problem in more places

From: Chun-Yi Lee <[email protected]>

For fixing CVE-2023-6270, f98364e92662 patch moved dev_put() from aoecmd_cfg_pkts() to tx() to avoid that the tx() runs into use-after-free.

But Nicolai Stange found more places in aoe have potential use-after-free problem with tx(). e.g. revalidate(), aoecmd_ata_rw(), resend(), probe() and aoecmd_cfg_rsp(). Those functions also use aoenet_xmit() to push packet to tx queue. So they should also use dev_hold() to increase the refcnt of skb->dev.

This patch adds dev_hold() to those functions and also uses dev_put() when the skb_clone() returns NULL.

Fixes f98364e92662 ("aoe: fix the potential use-after-free problem in

aoecmd_cfg_pkts")

---
 drivers/block/aoe/aoecmd.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/block/aoe/aoecmd.c +12 -3
index cc9077b588d7..fcedbad8e3be 100644
@@ -361,6 +361,7 @@ ata_rw_frameinit(struct frame *f)
 	}
 
 	ah->cmdstat = ATA_CMD_PIO_READ | writebit | extbit;
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 }
 
@@ -401,7 +402,8 @@ aoecmd_ata_rw(struct aoedev *d)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 	return 1;
 }
 
@@ -483,10 +485,13 @@ resend(struct aoedev *d, struct frame *f)
 	memcpy(h->dst, t->addr, sizeof h->dst);
 	memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src);
 
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 	skb = skb_clone(skb, GFP_ATOMIC);
-	if (skb == NULL)
+	if (skb == NULL) {
+		dev_put(t->ifp->nd);
 		return;
+	}
 	f->sent = ktime_get();
 	__skb_queue_head_init(&queue);
 	__skb_queue_tail(&queue, skb);
@@ -617,7 +622,8 @@ probe(struct aoetgt *t)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 }
 
 static long
@@ -1395,6 +1401,7 @@ aoecmd_ata_id(struct aoedev *d)
 	ah->cmdstat = ATA_CMD_ID_ATA;
 	ah->lba3 = 0xa0;
 
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 
 	d->rttavg = RTTAVG_INIT;
@@ -1404,6 +1411,8 @@ aoecmd_ata_id(struct aoedev *d)
 	skb = skb_clone(skb, GFP_ATOMIC);
 	if (skb)
 		f->sent = ktime_get();
+	else
+		dev_put(t->ifp->nd);
 
 	return skb;
 }
Signature · 1 line
2.35.3
#1 lore raw

Hi Justin,

Quoted 11 lines · Chun-Yi

From: Chun-Yi Lee <[email protected]>

For fixing CVE-2023-6270, f98364e92662 patch moved dev_put() from aoecmd_cfg_pkts() to tx() to avoid that the tx() runs into use-after-free.

But Nicolai Stange found more places in aoe have potential use-after-free problem with tx(). e.g. revalidate(), aoecmd_ata_rw(), resend(), probe() and aoecmd_cfg_rsp(). Those functions also use aoenet_xmit() to push packet to tx queue. So they should also use dev_hold() to increase the refcnt of skb->dev.

Quoted 8 lines

This patch adds dev_hold() to those functions and also uses dev_put() when the skb_clone() returns NULL.

Fixes f98364e92662 ("aoe: fix the potential use-after-free problem in

aoecmd_cfg_pkts")

Do you have any suggestion for this patch? or I missed anything?

Thanks a lot!
Joey Lee

Quoted 70 lines
---
 drivers/block/aoe/aoecmd.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)
drivers/block/aoe/aoecmd.c +12 -3
index cc9077b588d7..fcedbad8e3be 100644
@@ -361,6 +361,7 @@ ata_rw_frameinit(struct frame *f)
 	}
 
 	ah->cmdstat = ATA_CMD_PIO_READ | writebit | extbit;
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 }
 
@@ -401,7 +402,8 @@ aoecmd_ata_rw(struct aoedev *d)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 	return 1;
 }
 
@@ -483,10 +485,13 @@ resend(struct aoedev *d, struct frame *f)
 	memcpy(h->dst, t->addr, sizeof h->dst);
 	memcpy(h->src, t->ifp->nd->dev_addr, sizeof h->src);
 
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 	skb = skb_clone(skb, GFP_ATOMIC);
-	if (skb == NULL)
+	if (skb == NULL) {
+		dev_put(t->ifp->nd);
 		return;
+	}
 	f->sent = ktime_get();
 	__skb_queue_head_init(&queue);
 	__skb_queue_tail(&queue, skb);
@@ -617,7 +622,8 @@ probe(struct aoetgt *t)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 }
 
 static long
@@ -1395,6 +1401,7 @@ aoecmd_ata_id(struct aoedev *d)
 	ah->cmdstat = ATA_CMD_ID_ATA;
 	ah->lba3 = 0xa0;
 
+	dev_hold(t->ifp->nd);
 	skb->dev = t->ifp->nd;
 
 	d->rttavg = RTTAVG_INIT;
@@ -1404,6 +1411,8 @@ aoecmd_ata_id(struct aoedev *d)
 	skb = skb_clone(skb, GFP_ATOMIC);
 	if (skb)
 		f->sent = ktime_get();
+	else
+		dev_put(t->ifp->nd);
 
 	return skb;
 }
Signature · 1 line
2.35.3
#2 lore raw
In reply to Lee, Chun-Yi
Quoted 1 line

For fixing CVE-2023-6270, f98364e92662 patch moved dev_put() from

…

Please add a subject for the mentioned commit hash.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n99

Quoted 2 lines

This patch adds dev_hold() to those functions and also uses dev_put() when the skb_clone() returns NULL.

Please improve this change description with a corresponding imperative wording.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n94

…

Quoted 2 lines
Fixes f98364e92662 ("aoe: fix the potential use-after-free problem in

aoecmd_cfg_pkts")

I suggest to omit a line break for this tag.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n145

…

Quoted 1 line

+++ b/drivers/block/aoe/aoecmd.c

…

Quoted 10 lines
drivers/block/aoe/aoecmd.c +2 -1
@@ -401,7 +402,8 @@ aoecmd_ata_rw(struct aoedev *d)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 	return 1;
 }
 

…

drivers/block/aoe/aoecmd.c probe(struct aoetgt *t) · line 629

Quoted 10 lines
drivers/block/aoe/aoecmd.c +2 -1
@@ -617,7 +622,8 @@ probe(struct aoetgt *t)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 }
 
 static long

…

Should curly brackets be used for both if branches in these function implementations?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/coding-style.rst?h=v6.9-rc6#n213

Regards,
Markus

#3 lore raw

Hi Markus,

Thanks for your review! I will send v2 patch.

Joey Lee

Quoted 53 lines · Markus Elfring
Quoted 1 line

For fixing CVE-2023-6270, f98364e92662 patch moved dev_put() from

…

Please add a subject for the mentioned commit hash.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n99

Quoted 2 lines

This patch adds dev_hold() to those functions and also uses dev_put() when the skb_clone() returns NULL.

Please improve this change description with a corresponding imperative wording.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n94

…

Quoted 2 lines
Fixes f98364e92662 ("aoe: fix the potential use-after-free problem in

aoecmd_cfg_pkts")

I suggest to omit a line break for this tag.
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/submitting-patches.rst?h=v6.9-rc6#n145

…

Quoted 1 line

+++ b/drivers/block/aoe/aoecmd.c

…

Quoted 10 lines
drivers/block/aoe/aoecmd.c +2 -1
@@ -401,7 +402,8 @@ aoecmd_ata_rw(struct aoedev *d)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 	return 1;
 }
 

…

drivers/block/aoe/aoecmd.c probe(struct aoetgt *t) · line 629

Quoted 10 lines
drivers/block/aoe/aoecmd.c +2 -1
@@ -617,7 +622,8 @@ probe(struct aoetgt *t)
 		__skb_queue_head_init(&queue);
 		__skb_queue_tail(&queue, skb);
 		aoenet_xmit(&queue);
-	}
+	} else
+		dev_put(f->t->ifp->nd);
 }
 
 static long

…

Should curly brackets be used for both if branches in these function implementations?
https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/process/coding-style.rst?h=v6.9-rc6#n213

Regards,
Markus