Skip to contentSkip to conversation

FAILED: patch "[PATCH] dm verity: fix crash on bufio buffer that was allocated with" failed to apply to 4.14-stable tree

PATCH [email protected] 3 participants 4 messages Sep 24, 2018 – Jan 10, 2019

Subsystems (2)
lore
Share & tools
lore

[email protected]

author maintainer
#0 lore raw

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

Quoted 1 line

From e4b069e0945fa14c71cf8b5b89f8b1b2aa68dbc2 Mon Sep 17 00:00:00 2001

From: Mikulas Patocka <[email protected]>
Date: Wed, 22 Aug 2018 12:45:51 -0400
Subject: [PATCH] dm verity: fix crash on bufio buffer that was allocated with
vmalloc

Since commit d1ac3ff008fb ("dm verity: switch to using asynchronous hash crypto API") dm-verity uses asynchronous crypto calls for verification, so that it can use hardware with asynchronous processing of crypto operations.

These asynchronous calls don't support vmalloc memory, but the buffer data can be allocated with vmalloc if dm-bufio is short of memory and uses a reserved buffer that was preallocated in dm_bufio_client_create().

Fix verity_hash_update() so that it deals with vmalloc'd memory correctly.

Fixes d1ac3ff008fb ("dm verity: switch to using asynchronous hash crypto API")
drivers/md/dm-verity-target.c +20 -4
index 12decdbd722d..fc65f0dedf7f 100644
@@ -99,10 +99,26 @@ static int verity_hash_update(struct dm_verity *v, struct ahash_request *req,
 {
 	struct scatterlist sg;
 
-	sg_init_one(&sg, data, len);
-	ahash_request_set_crypt(req, &sg, NULL, len);
-
-	return crypto_wait_req(crypto_ahash_update(req), wait);
+	if (likely(!is_vmalloc_addr(data))) {
+		sg_init_one(&sg, data, len);
+		ahash_request_set_crypt(req, &sg, NULL, len);
+		return crypto_wait_req(crypto_ahash_update(req), wait);
+	} else {
+		do {
+			int r;
+			size_t this_step = min_t(size_t, len, PAGE_SIZE - offset_in_page(data));
+			flush_kernel_vmap_range((void *)data, this_step);
+			sg_init_table(&sg, 1);
+			sg_set_page(&sg, vmalloc_to_page(data), this_step, offset_in_page(data));
+			ahash_request_set_crypt(req, &sg, NULL, this_step);
+			r = crypto_wait_req(crypto_ahash_update(req), wait);
+			if (unlikely(r))
+				return r;
+			data += this_step;
+			len -= this_step;
+		} while (len);
+		return 0;
+	}
 }
 
 /*
#1 lore raw

Hi Greg,

Quoted 5 lines · On Sun, Sep 23, 2018 at 09:17:16PM +0200, [email protected] wrote:

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

The attached backported patch will apply to 4.14-stable tree. But I will want to for an ack.

--
Regards
Sudip

From 3027f3f3f9c61c4b870c23096050dd948ac2ceb1 Mon Sep 17 00:00:00 2001
From: Mikulas Patocka <[email protected]>
Date: Wed, 22 Aug 2018 12:45:51 -0400
Subject: [PATCH] dm verity: fix crash on bufio buffer that was allocated with
vmalloc

commit e4b069e0945fa14c71cf8b5b89f8b1b2aa68dbc2 upstream

Since commit d1ac3ff008fb ("dm verity: switch to using asynchronous hash crypto API") dm-verity uses asynchronous crypto calls for verification, so that it can use hardware with asynchronous processing of crypto operations.

These asynchronous calls don't support vmalloc memory, but the buffer data can be allocated with vmalloc if dm-bufio is short of memory and uses a reserved buffer that was preallocated in dm_bufio_client_create().

Fix verity_hash_update() so that it deals with vmalloc'd memory correctly.

Fixes d1ac3ff008fb ("dm verity: switch to using asynchronous hash crypto API")
---
 drivers/md/dm-verity-target.c | 24 ++++++++++++++++++++----
 1 file changed, 20 insertions(+), 4 deletions(-)
drivers/md/dm-verity-target.c +20 -4
index bda3caca23ca..8573c70a1880 100644
@@ -139,10 +139,26 @@ static int verity_hash_update(struct dm_verity *v, struct ahash_request *req,
 {
 	struct scatterlist sg;
 
-	sg_init_one(&sg, data, len);
-	ahash_request_set_crypt(req, &sg, NULL, len);
-
-	return verity_complete_op(res, crypto_ahash_update(req));
+	if (likely(!is_vmalloc_addr(data))) {
+		sg_init_one(&sg, data, len);
+		ahash_request_set_crypt(req, &sg, NULL, len);
+		return verity_complete_op(res, crypto_ahash_update(req));
+	} else {
+		do {
+			int r;
+			size_t this_step = min_t(size_t, len, PAGE_SIZE - offset_in_page(data));
+			flush_kernel_vmap_range((void *)data, this_step);
+			sg_init_table(&sg, 1);
+			sg_set_page(&sg, vmalloc_to_page(data), this_step, offset_in_page(data));
+			ahash_request_set_crypt(req, &sg, NULL, this_step);
+			r = verity_complete_op(res, crypto_ahash_update(req));
+			if (unlikely(r))
+				return r;
+			data += this_step;
+			len -= this_step;
+		} while (len);
+		return 0;
+	}
 }
 
 /*
Signature · 1 line
2.11.0
maintainer Acked-by
#2 lore raw
Quoted 9 lines · Sudip Mukherjee

Hi Greg,

Quoted 5 lines · On Sun, Sep 23, 2018 at 09:17:16PM +0200, [email protected] wrote:

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Quoted 6 lines

The attached backported patch will apply to 4.14-stable tree. But I will want to for an ack.

--
Regards
Sudip

The patch looks correct.

Mikulas

author maintainer
#3 lore raw
Quoted 16 lines · Mikulas Patocka
Quoted 12 lines · Sudip Mukherjee

Hi Greg,

Quoted 5 lines · On Sun, Sep 23, 2018 at 09:17:16PM +0200, [email protected] wrote:

The patch below does not apply to the 4.14-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

The attached backported patch will apply to 4.14-stable tree. But I will want to for an ack.

Quoted 7 lines
Quoted 3 lines

--
Regards
Sudip

The patch looks correct.

Thanks, now queued up.

greg k-h