In function sg_split, the second sg_calculate_split will return -EINVAL
when in_mapped_nents is 0.
Indeed there is no need to do second sg_calculate_split and sg_split_mapped
when in_mapped_nents is 0, as in_mapped_nents indicates no mapped entry in
original sgl.
Signed-off-by: Zhou Wang <[email protected]>
---
lib/sg_split.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/lib/sg_split.c b/lib/sg_split.c
index b063410..d4181c8 100644
--- a/lib/sg_split.c
+++ b/lib/sg_split.c
@@ -178,11 +178,13 @@ int sg_split(struct scatterlist *in, const int in_mapped_nents,
* The order of these 3 calls is important and should be kept.
*/
sg_split_phys(splitters, nb_splits);
- ret = sg_calculate_split(in, in_mapped_nents, nb_splits, skip,
- split_sizes, splitters, true);
- if (ret < 0)
- goto err;
- sg_split_mapped(splitters, nb_splits);
+ if (in_mapped_nents) {
+ ret = sg_calculate_split(in, in_mapped_nents, nb_splits, skip,
+ split_sizes, splitters, true);
+ if (ret < 0)
+ goto err;
+ sg_split_mapped(splitters, nb_splits);
+ }
for (i = 0; i < nb_splits; i++) {
out[i] = splitters[i].out_sg;
--
2.8.1
Zhou Wang <[email protected]> writes:
> In function sg_split, the second sg_calculate_split will return -EINVAL
> when in_mapped_nents is 0.
>
> Indeed there is no need to do second sg_calculate_split and sg_split_mapped
> when in_mapped_nents is 0, as in_mapped_nents indicates no mapped entry in
> original sgl.
>
> Signed-off-by: Zhou Wang <[email protected]>
That's right. May I know what is the usecase which led you to this fix ?
Acked-by: Robert Jarzmik <[email protected]>
--
Robert
On 2019/4/19 18:43, Robert Jarzmik wrote:
> Zhou Wang <[email protected]> writes:
>
>> In function sg_split, the second sg_calculate_split will return -EINVAL
>> when in_mapped_nents is 0.
>>
>> Indeed there is no need to do second sg_calculate_split and sg_split_mapped
>> when in_mapped_nents is 0, as in_mapped_nents indicates no mapped entry in
>> original sgl.
>>
>> Signed-off-by: Zhou Wang <[email protected]>
> That's right. May I know what is the usecase which led you to this fix ?
I posted a patchset to support HiSilicon compression engine:
https://lkml.org/lkml/2019/1/23/358. As Herber Xu suggested, it should use
new crypto acomp API which uses sgl as data input and output. However, our
hardware can not handle compress head for input/output data, so I need split
input/output sgl firstly which is not mapped sgl.
New version of HiSilicon compression engine driver has not been posted to
community yet. I will do this later.
>
> Acked-by: Robert Jarzmik <[email protected]>
Many thanks for your review,
Zhou
>
> --
> Robert
>
> .
>