From: Harald Freudenberger <freude@linux.ibm.com>
Date: Mon, 23 Mar 2020 14:32:04 +0100
Subject: s390/zcrypt: use kvmalloc instead of kmalloc for 256k alloc
Git-commit: 34515df25d7e5ae19f66eadfb1351a3178344f36
Patch-mainline: v5.7-rc1
References: bsc#1169003 LTC#185029

Tests showed that it may happen that a 256k kmalloc may fail
due to a temporary shortage on 256k slab entries.

The find functions for cca and ep11 use a 256k array to fetch the
states of all possible crypto cards and their domains in one
piece. With the patch now kvmalloc is used to allocate this
temporary memory as there is no need to have this memory area
physical continuously.

Signed-off-by: Harald Freudenberger <freude@linux.ibm.com>
Reviewed-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
[ ptesarik: Support for EP11 crypto cards was not added until
  v5.6-rc1, so the corresponding hunks are removed. ]
Acked-by: Petr Tesarik <ptesarik@suse.com>
---
 drivers/s390/crypto/zcrypt_ccamisc.c |    8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

--- a/drivers/s390/crypto/zcrypt_ccamisc.c
+++ b/drivers/s390/crypto/zcrypt_ccamisc.c
@@ -1568,9 +1568,9 @@ static int findcard(u64 mkvp, u16 *pcard
 		return -EINVAL;
 
 	/* fetch status of all crypto cards */
-	device_status = kmalloc_array(MAX_ZDEV_ENTRIES_EXT,
-				      sizeof(struct zcrypt_device_status_ext),
-				      GFP_KERNEL);
+	device_status = kvmalloc_array(MAX_ZDEV_ENTRIES_EXT,
+				       sizeof(struct zcrypt_device_status_ext),
+				       GFP_KERNEL);
 	if (!device_status)
 		return -ENOMEM;
 	zcrypt_device_status_mask_ext(device_status);
@@ -1640,7 +1640,7 @@ static int findcard(u64 mkvp, u16 *pcard
 	} else
 		rc = -ENODEV;
 
-	kfree(device_status);
+	kvfree(device_status);
 	return rc;
 }