linux-kernel - [GIT PULL] s390 fixes for 6.5-rc3

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <20230722160215.7417-A-hca@linux.ibm.com>
Date:   Sat, 22 Jul 2023 18:02:15 +0200
From:   Heiko Carstens <hca@...ux.ibm.com>
To:     Linus Torvalds <torvalds@...ux-foundation.org>
Cc:     Alexander Gordeev <agordeev@...ux.ibm.com>,
        Vasily Gorbik <gor@...ux.ibm.com>, linux-s390@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: [GIT PULL] s390 fixes for 6.5-rc3

Hello Linus,

please pull a couple of s390 fixes for 6.5-rc3.

Thanks,
Heiko

The following changes since commit fdf0eaf11452d72945af31804e2a1048ee1b574c:

  Linux 6.5-rc2 (2023-07-16 15:10:37 -0700)

are available in the Git repository at:

  git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux.git tags/s390-6.5-3

for you to fetch changes up to 4cfca532ddc3474b3fc42592d0e4237544344b1a:

  s390/zcrypt: fix reply buffer calculations for CCA replies (2023-07-20 16:48:56 +0200)

----------------------------------------------------------------
s390 fixes for 6.5-rc3

- Fix per vma lock fault handling: add missing !(fault & VM_FAULT_ERROR)
  check to fault handler to prevent error handling for return values that
  don't indicate an error

- Use kfree_sensitive() instead of kfree() in paes crypto code to clear
  memory that may contain keys before freeing it

- Fix reply buffer size calculation for CCA replies in zcrypt device driver

----------------------------------------------------------------
Harald Freudenberger (1):
      s390/zcrypt: fix reply buffer calculations for CCA replies

Sven Schnelle (1):
      s390/mm: fix per vma lock fault handling

Wang Ming (1):
      s390/crypto: use kfree_sensitive() instead of kfree()

 arch/s390/crypto/paes_s390.c          |  2 +-
 arch/s390/mm/fault.c                  |  2 ++
 drivers/s390/crypto/zcrypt_msgtype6.c | 33 +++++++++++++++++++++++----------
 3 files changed, 26 insertions(+), 11 deletions(-)

diff --git a/arch/s390/crypto/paes_s390.c b/arch/s390/crypto/paes_s390.c
index d29a9d908797..38349150c96e 100644
--- a/arch/s390/crypto/paes_s390.c
+++ b/arch/s390/crypto/paes_s390.c
@@ -103,7 +103,7 @@ static inline void _free_kb_keybuf(struct key_blob *kb)
 {
 	if (kb->key && kb->key != kb->keybuf
 	    && kb->keylen > sizeof(kb->keybuf)) {
-		kfree(kb->key);
+		kfree_sensitive(kb->key);
 		kb->key = NULL;
 	}
 }
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index dbe8394234e2..2f123429a291 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -421,6 +421,8 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access)
 	vma_end_read(vma);
 	if (!(fault & VM_FAULT_RETRY)) {
 		count_vm_vma_lock_event(VMA_LOCK_SUCCESS);
+		if (likely(!(fault & VM_FAULT_ERROR)))
+			fault = 0;
 		goto out;
 	}
 	count_vm_vma_lock_event(VMA_LOCK_RETRY);
diff --git a/drivers/s390/crypto/zcrypt_msgtype6.c b/drivers/s390/crypto/zcrypt_msgtype6.c
index 67fd2ec9c5a1..e668ff5eb384 100644
--- a/drivers/s390/crypto/zcrypt_msgtype6.c
+++ b/drivers/s390/crypto/zcrypt_msgtype6.c
@@ -1101,23 +1101,36 @@ static long zcrypt_msgtype6_send_cprb(bool userspace, struct zcrypt_queue *zq,
 				      struct ica_xcRB *xcrb,
 				      struct ap_message *ap_msg)
 {
-	int rc;
 	struct response_type *rtype = ap_msg->private;
 	struct {
 		struct type6_hdr hdr;
 		struct CPRBX cprbx;
 		/* ... more data blocks ... */
 	} __packed * msg = ap_msg->msg;
+	unsigned int max_payload_size;
+	int rc, delta;
 
-	/*
-	 * Set the queue's reply buffer length minus 128 byte padding
-	 * as reply limit for the card firmware.
-	 */
-	msg->hdr.fromcardlen1 = min_t(unsigned int, msg->hdr.fromcardlen1,
-				      zq->reply.bufsize - 128);
-	if (msg->hdr.fromcardlen2)
-		msg->hdr.fromcardlen2 =
-			zq->reply.bufsize - msg->hdr.fromcardlen1 - 128;
+	/* calculate maximum payload for this card and msg type */
+	max_payload_size = zq->reply.bufsize - sizeof(struct type86_fmt2_msg);
+
+	/* limit each of the two from fields to the maximum payload size */
+	msg->hdr.fromcardlen1 = min(msg->hdr.fromcardlen1, max_payload_size);
+	msg->hdr.fromcardlen2 = min(msg->hdr.fromcardlen2, max_payload_size);
+
+	/* calculate delta if the sum of both exceeds max payload size */
+	delta = msg->hdr.fromcardlen1 + msg->hdr.fromcardlen2
+		- max_payload_size;
+	if (delta > 0) {
+		/*
+		 * Sum exceeds maximum payload size, prune fromcardlen1
+		 * (always trust fromcardlen2)
+		 */
+		if (delta > msg->hdr.fromcardlen1) {
+			rc = -EINVAL;
+			goto out;
+		}
+		msg->hdr.fromcardlen1 -= delta;
+	}
 
 	init_completion(&rtype->work);
 	rc = ap_queue_message(zq->queue, ap_msg);