Hi Vitaly,
On 8/12/21 9:59 PM, Vitaly Chikunov wrote:
Tianjia,
On Thu, Aug 12, 2021 at 09:17:46PM +0800, Tianjia Zhang wrote:
The position of the return value test statement of crypto_aead_setkey()
is wrong, adjust to make it work properly.
This commit message does not explain anything. It's nearly equivalent to
"fix".
will update in next version.
Signed-off-by: Tianjia Zhang <tianjia.zhang@xxxxxxxxxxxxxxxxx>
---
crypto/tcrypt.c | 13 ++++++-------
1 file changed, 6 insertions(+), 7 deletions(-)
diff --git a/crypto/tcrypt.c b/crypto/tcrypt.c
index d73a42fdaa9b..73c97e085baf 100644
--- a/crypto/tcrypt.c
+++ b/crypto/tcrypt.c
@@ -612,6 +612,12 @@ static void test_aead_speed(const char *algo, int enc, unsigned int secs,
}
}
ret = crypto_aead_setkey(tfm, key, *keysize);
Perhaps, you would say that return value of crypto_aead_setkey was lost.
Yes, good idea.
+ if (ret) {
+ pr_err("setkey() failed flags=%x\n",
+ crypto_aead_get_flags(tfm));
+ goto out;
+ }
+
ret = crypto_aead_setauthsize(tfm, authsize);
But, isn't now return value of crypto_aead_setauthsize is lost?
Thanks,
Yes, it is needed.
I will move crypto_aead_setauthsize() call out of the loop, only need to
call it once after load transform.
Best regards,
Tianjia
iv_len = crypto_aead_ivsize(tfm);
@@ -622,15 +628,8 @@ static void test_aead_speed(const char *algo, int enc, unsigned int secs,
printk(KERN_INFO "test %u (%d bit key, %d byte blocks): ",
i, *keysize * 8, bs);
-
memset(tvmem[0], 0xff, PAGE_SIZE);
- if (ret) {
- pr_err("setkey() failed flags=%x\n",
- crypto_aead_get_flags(tfm));
- goto out;
- }
-
sg_init_aead(sg, xbuf, bs + (enc ? 0 : authsize),
assoc, aad_size);
--
2.19.1.3.ge56e4f7
