diff mbox

[rdma-core] libhns: Update the size and member order of hns_roce_create_qp structure

Message ID 1506761477-47350-1-git-send-email-oulijun@huawei.com (mailing list archive)
State Changes Requested
Headers show

Commit Message

Lijun Ou Sept. 30, 2017, 8:51 a.m. UTC 
In order to keep the size and member order with the
hns_roce_ib_create_qp sturcture of the kernel driver,
it need to update hns_roce_create_qp's structure
members. Otherwise, it will run fail.

Signed-off-by: Lijun Ou <oulijun@huawei.com>
Signed-off-by: Wei Hu (Xavier) <xavier.huwei@huawei.com>
---
 providers/hns/hns_roce_u_abi.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Jason Gunthorpe Sept. 30, 2017, 3:45 p.m. UTC | #1 
On Sat, Sep 30, 2017 at 04:51:17PM +0800, Lijun Ou wrote:
> In order to keep the size and member order with the
> hns_roce_ib_create_qp sturcture of the kernel driver,
> it need to update hns_roce_create_qp's structure
> members. Otherwise, it will run fail.

What?

Did you break the uapi? What kernel patch does this match with?

Generally structs in the u_abi.h headers cannot be changed like this.

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Leon Romanovsky Sept. 30, 2017, 4:05 p.m. UTC | #2 
On Sat, Sep 30, 2017 at 09:45:18AM -0600, Jason Gunthorpe wrote:
> On Sat, Sep 30, 2017 at 04:51:17PM +0800, Lijun Ou wrote:
> > In order to keep the size and member order with the
> > hns_roce_ib_create_qp sturcture of the kernel driver,
> > it need to update hns_roce_create_qp's structure
> > members. Otherwise, it will run fail.
>
> What?
>
> Did you break the uapi? What kernel patch does this match with?
>
> Generally structs in the u_abi.h headers cannot be changed like this.

Jason,

It was my suggestion to change rdma-core, because before they tried to
change kernel.

https://patchwork.kernel.org/patch/9975329/


>
> Jason
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
Leon Romanovsky Sept. 30, 2017, 4:07 p.m. UTC | #3 
On Sat, Sep 30, 2017 at 07:05:15PM +0300, Leon Romanovsky wrote:
> On Sat, Sep 30, 2017 at 09:45:18AM -0600, Jason Gunthorpe wrote:
> > On Sat, Sep 30, 2017 at 04:51:17PM +0800, Lijun Ou wrote:
> > > In order to keep the size and member order with the
> > > hns_roce_ib_create_qp sturcture of the kernel driver,
> > > it need to update hns_roce_create_qp's structure
> > > members. Otherwise, it will run fail.
> >
> > What?
> >
> > Did you break the uapi? What kernel patch does this match with?
> >
> > Generally structs in the u_abi.h headers cannot be changed like this.
>
> Jason,
>
> It was my suggestion to change rdma-core, because before they tried to
> change kernel.
>
> https://patchwork.kernel.org/patch/9975329/
>

And according to the git log on providers/hns/hns_roce_u_abi.h and
include/uapi/rdma/hns-abi.h, it never worked.

>
> >
> > Jason
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
Jason Gunthorpe Sept. 30, 2017, 4:29 p.m. UTC | #4 
On Sat, Sep 30, 2017 at 07:05:15PM +0300, Leon Romanovsky wrote:
> On Sat, Sep 30, 2017 at 09:45:18AM -0600, Jason Gunthorpe wrote:
> > On Sat, Sep 30, 2017 at 04:51:17PM +0800, Lijun Ou wrote:
> > > In order to keep the size and member order with the
> > > hns_roce_ib_create_qp sturcture of the kernel driver,
> > > it need to update hns_roce_create_qp's structure
> > > members. Otherwise, it will run fail.
> >
> > What?
> >
> > Did you break the uapi? What kernel patch does this match with?
> >
> > Generally structs in the u_abi.h headers cannot be changed like this.
> 
> Jason,
> 
> It was my suggestion to change rdma-core, because before they tried to
> change kernel.
> 
> https://patchwork.kernel.org/patch/9975329/

Ugh, better commit messages please.

Also, the patch is wrong, it needs to exactly duplicate the current
kernel struct word for word. You can send some thing else once the
kernel side is accepted.

How did we get to a point where hns in rdma-core doesn't work with the
kernel code? Is huawei not testing this stuff?

Jason
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Leon Romanovsky Oct. 1, 2017, 5:28 a.m. UTC | #5 
On Sat, Sep 30, 2017 at 10:29:53AM -0600, Jason Gunthorpe wrote:
> How did we get to a point where hns in rdma-core doesn't work with the
> kernel code? Is huawei not testing this stuff?

You probably know the answer by yourself now.

Thanks

>
> Jason
Wei Hu (Xavier) Oct. 1, 2017, 2:39 p.m. UTC | #6 
On 2017/10/1 0:29, Jason Gunthorpe wrote:
> On Sat, Sep 30, 2017 at 07:05:15PM +0300, Leon Romanovsky wrote:
>> On Sat, Sep 30, 2017 at 09:45:18AM -0600, Jason Gunthorpe wrote:
>>> On Sat, Sep 30, 2017 at 04:51:17PM +0800, Lijun Ou wrote:
>>>> In order to keep the size and member order with the
>>>> hns_roce_ib_create_qp sturcture of the kernel driver,
>>>> it need to update hns_roce_create_qp's structure
>>>> members. Otherwise, it will run fail.
>>> What?
>>>
>>> Did you break the uapi? What kernel patch does this match with?
>>>
>>> Generally structs in the u_abi.h headers cannot be changed like this.
>> Jason,
>>
>> It was my suggestion to change rdma-core, because before they tried to
>> change kernel.
>>
>> https://patchwork.kernel.org/patch/9975329/
> Ugh, better commit messages please.
>
> Also, the patch is wrong, it needs to exactly duplicate the current
> kernel struct word for word. You can send some thing else once the
> kernel side is accepted.
>
> How did we get to a point where hns in rdma-core doesn't work with the
> kernel codse? Is huawei not testing this stuff?
Hi, Jason
     In fact testing is mandatory in our development process. sorry for 
this fault.
     We will do our best to avoid the occurrence of this kind of problem.
     And we will send patch v2 at your suggestion. Thanks.

     Regards
Wei Hu
> Jason
> --
> To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/providers/hns/hns_roce_u_abi.h b/providers/hns/hns_roce_u_abi.h
index e78f967..276d29d 100644
--- a/providers/hns/hns_roce_u_abi.h
+++ b/providers/hns/hns_roce_u_abi.h
@@ -61,9 +61,10 @@  struct hns_roce_create_cq_resp {
 struct hns_roce_create_qp {
 	struct ibv_create_qp		ibv_cmd;
 	__u64				buf_addr;
+	__u64				rsv;
 	__u8				log_sq_bb_count;
 	__u8				log_sq_stride;
-	__u8				reserved[5];
+	__u8				reserved[6];
 };
 
 #endif /* _HNS_ROCE_U_ABI_H */