Hi,
On Sat, Mar 26, 2011 at 2:26 AM, Nikos Mavrogiannopoulos
<n...@gnutls.org> wrote:
> On 03/25/2011 12:07 AM, Vladimir Zapolskiy wrote:
>> This change adds setting up of alignment mask for buffers according to
>> the value, which is obtained from the kernel algorithm definition.
>>
>> Signed-off-by: Vladimir Zapolskiy <vzapols...@gmail.com>
>> ---
>> examples/hashcrypt_speed.c | 10 ++++++----
>> examples/sha_speed.c | 10 ++++++----
>> examples/speed.c | 12 ++++++++++--
>> 3 files changed, 22 insertions(+), 10 deletions(-)
>>
>> diff --git a/examples/hashcrypt_speed.c b/examples/hashcrypt_speed.c
>> index f5e0e40..2344577 100644
>> --- a/examples/hashcrypt_speed.c
>> +++ b/examples/hashcrypt_speed.c
>> @@ -80,7 +80,7 @@ int hash_data(struct session_op *sess, int fdc, int
>> chunksize, int alignmask)
>> uint8_t mac[AALG_MAX_RESULT_LEN];
>>
>> if (alignmask) {
>> - if (posix_memalign((void **)&buffer, alignmask, chunksize)) {
>> + if (posix_memalign((void **)&buffer, alignmask + 1,
>> chunksize)) {
>
>
> Hello Vladimir,
> This part of the patch I don't understand. Why do you add 1 to alignmask?
>
man 3 posix_memalign, the alignment argument must be a power of two.
Best wishes,
Vladimir
_______________________________________________
Cryptodev-linux-devel mailing list
Cryptodev-linux-devel@gna.org
https://mail.gna.org/listinfo/cryptodev-linux-devel
