mbox series

[V2,00/20] Cleanup for siw

Message ID 20231013020053.2120-1-guoqing.jiang@linux.dev (mailing list archive)
Headers show
Series Cleanup for siw | expand

Message

Guoqing Jiang Oct. 13, 2023, 2 a.m. UTC
V2 changes:
1. address W=1 warning in patch 12 and 19 per the report from lkp.
2. add one more patch (20th).

Hi,

This series aim to cleanup siw code, please review and comment!

Thanks,
Guoqing

Guoqing Jiang (20):
  RDMA/siw: Introduce siw_get_page
  RDMA/siw: Introduce siw_srx_update_skb
  RDMA/siw: Use iov.iov_len in kernel_sendmsg
  RDMA/siw: Remove goto lable in siw_mmap
  RDMA/siw: Remove rcu from siw_qp
  RDMA/siw: No need to check term_info.valid before call
    siw_send_terminate
  RDMA/siw: Also goto out_sem_up if pin_user_pages returns 0
  RDMA/siw: Factor out siw_generic_rx helper
  RDMA/siw: Introduce SIW_STAG_MAX_INDEX
  RDMA/siw: Add one parameter to siw_destroy_cpulist
  RDMA/siw: Introduce siw_cep_set_free_and_put
  RDMA/siw: Introduce siw_free_cm_id
  RDMA/siw: Simplify siw_qp_id2obj
  RDMA/siw: Simplify siw_mem_id2obj
  RDMA/siw: Cleanup siw_accept
  RDMA/siw: Remove siw_sk_assign_cm_upcalls
  RDMA/siw: Fix typo
  RDMA/siw: Only check attrs->cap.max_send_wr in siw_create_qp
  RDMA/siw: Introduce siw_destroy_cep_sock
  RDMA/siw: Update comments for siw_qp_sq_process

 drivers/infiniband/sw/siw/siw.h       |   9 +-
 drivers/infiniband/sw/siw/siw_cm.c    | 154 +++++++++++---------------
 drivers/infiniband/sw/siw/siw_main.c  |  30 +++--
 drivers/infiniband/sw/siw/siw_mem.c   |  22 ++--
 drivers/infiniband/sw/siw/siw_qp.c    |   2 +-
 drivers/infiniband/sw/siw/siw_qp_rx.c |  84 ++++++--------
 drivers/infiniband/sw/siw/siw_qp_tx.c |  39 +++----
 drivers/infiniband/sw/siw/siw_verbs.c |  23 +---
 8 files changed, 144 insertions(+), 219 deletions(-)


base-commit: 964168970cef5f5b738fae047e6de2107842feb7

Comments

Leon Romanovsky Oct. 24, 2023, 2:09 p.m. UTC | #1
On Fri, Oct 13, 2023 at 10:00:33AM +0800, Guoqing Jiang wrote:
> V2 changes:
> 1. address W=1 warning in patch 12 and 19 per the report from lkp.
> 2. add one more patch (20th).
> 
> Hi,
> 
> This series aim to cleanup siw code, please review and comment!
> 
> Thanks,
> Guoqing
> 
> Guoqing Jiang (20):

Bernard, did you return from the vacation?

Thanks
Bernard Metzler Oct. 24, 2023, 5:22 p.m. UTC | #2
> -----Original Message-----
> From: Leon Romanovsky <leon@kernel.org>
> Sent: Tuesday, October 24, 2023 4:10 PM
> To: Guoqing Jiang <guoqing.jiang@linux.dev>; Bernard Metzler
> <BMT@zurich.ibm.com>
> Cc: jgg@ziepe.ca; linux-rdma@vger.kernel.org
> Subject: [EXTERNAL] Re: [PATCH V2 00/20] Cleanup for siw
> 
> On Fri, Oct 13, 2023 at 10:00:33AM +0800, Guoqing Jiang wrote:
> > V2 changes:
> > 1. address W=1 warning in patch 12 and 19 per the report from lkp.
> > 2. add one more patch (20th).
> >
> > Hi,
> >
> > This series aim to cleanup siw code, please review and comment!
> >
> > Thanks,
> > Guoqing
> >
> > Guoqing Jiang (20):
> 
> Bernard, did you return from the vacation?
> 

Hi Leon, yes. I have it on my list. I hope to get to it tomorrow. 
Sorry for the delay!

Bernard.
Leon Romanovsky Oct. 24, 2023, 6:19 p.m. UTC | #3
On Tue, Oct 24, 2023 at 05:22:38PM +0000, Bernard Metzler wrote:
> 
> 
> > -----Original Message-----
> > From: Leon Romanovsky <leon@kernel.org>
> > Sent: Tuesday, October 24, 2023 4:10 PM
> > To: Guoqing Jiang <guoqing.jiang@linux.dev>; Bernard Metzler
> > <BMT@zurich.ibm.com>
> > Cc: jgg@ziepe.ca; linux-rdma@vger.kernel.org
> > Subject: [EXTERNAL] Re: [PATCH V2 00/20] Cleanup for siw
> > 
> > On Fri, Oct 13, 2023 at 10:00:33AM +0800, Guoqing Jiang wrote:
> > > V2 changes:
> > > 1. address W=1 warning in patch 12 and 19 per the report from lkp.
> > > 2. add one more patch (20th).
> > >
> > > Hi,
> > >
> > > This series aim to cleanup siw code, please review and comment!
> > >
> > > Thanks,
> > > Guoqing
> > >
> > > Guoqing Jiang (20):
> > 
> > Bernard, did you return from the vacation?
> > 
> 
> Hi Leon, yes. I have it on my list. I hope to get to it tomorrow.
> Sorry for the delay!

Thanks

> 
> Bernard.
Bernard Metzler Oct. 25, 2023, 1:37 p.m. UTC | #4
> -----Original Message-----
> From: Guoqing Jiang <guoqing.jiang@linux.dev>
> Sent: Friday, October 13, 2023 4:01 AM
> To: Bernard Metzler <BMT@zurich.ibm.com>; jgg@ziepe.ca; leon@kernel.org
> Cc: linux-rdma@vger.kernel.org
> Subject: [EXTERNAL] [PATCH V2 00/20] Cleanup for siw
> 
> V2 changes:
> 1. address W=1 warning in patch 12 and 19 per the report from lkp.
> 2. add one more patch (20th).
> 
> Hi,
> 
> This series aim to cleanup siw code, please review and comment!
> 
> Thanks,
> Guoqing
> 
> Guoqing Jiang (20):
>   RDMA/siw: Introduce siw_get_page
>   RDMA/siw: Introduce siw_srx_update_skb
>   RDMA/siw: Use iov.iov_len in kernel_sendmsg
>   RDMA/siw: Remove goto lable in siw_mmap
>   RDMA/siw: Remove rcu from siw_qp
>   RDMA/siw: No need to check term_info.valid before call
>     siw_send_terminate
>   RDMA/siw: Also goto out_sem_up if pin_user_pages returns 0
>   RDMA/siw: Factor out siw_generic_rx helper
>   RDMA/siw: Introduce SIW_STAG_MAX_INDEX
>   RDMA/siw: Add one parameter to siw_destroy_cpulist
>   RDMA/siw: Introduce siw_cep_set_free_and_put
>   RDMA/siw: Introduce siw_free_cm_id
>   RDMA/siw: Simplify siw_qp_id2obj
>   RDMA/siw: Simplify siw_mem_id2obj
>   RDMA/siw: Cleanup siw_accept
>   RDMA/siw: Remove siw_sk_assign_cm_upcalls
>   RDMA/siw: Fix typo
>   RDMA/siw: Only check attrs->cap.max_send_wr in siw_create_qp
>   RDMA/siw: Introduce siw_destroy_cep_sock
>   RDMA/siw: Update comments for siw_qp_sq_process
> 
>  drivers/infiniband/sw/siw/siw.h       |   9 +-
>  drivers/infiniband/sw/siw/siw_cm.c    | 154 +++++++++++---------------
>  drivers/infiniband/sw/siw/siw_main.c  |  30 +++--
>  drivers/infiniband/sw/siw/siw_mem.c   |  22 ++--
>  drivers/infiniband/sw/siw/siw_qp.c    |   2 +-
>  drivers/infiniband/sw/siw/siw_qp_rx.c |  84 ++++++--------
>  drivers/infiniband/sw/siw/siw_qp_tx.c |  39 +++----
>  drivers/infiniband/sw/siw/siw_verbs.c |  23 +---
>  8 files changed, 144 insertions(+), 219 deletions(-)
> 
> 
> base-commit: 964168970cef5f5b738fae047e6de2107842feb7
> --
> 2.35.3

Hi Guoqing,
Thanks for the effort! I like most of it.

And, sorry, I saw I started my review with version 1 of your
patches. Luckily it does not have functional differences to
v2.

But I expect a version 3 anyway.

I currently do not have access to physical machines to check and
run the more complex patches. I hope to do that (patch 08,10,12,15)
tomorrow when back in office.

Thanks,
Bernard.
Guoqing Jiang Oct. 26, 2023, 7:02 a.m. UTC | #5
Hi Bernard,

On 10/25/23 21:37, Bernard Metzler wrote:
>> -----Original Message-----
>> From: Guoqing Jiang<guoqing.jiang@linux.dev>
>> Sent: Friday, October 13, 2023 4:01 AM
>> To: Bernard Metzler<BMT@zurich.ibm.com>;jgg@ziepe.ca;leon@kernel.org
>> Cc:linux-rdma@vger.kernel.org
>> Subject: [EXTERNAL] [PATCH V2 00/20] Cleanup for siw
>>
>> V2 changes:
>> 1. address W=1 warning in patch 12 and 19 per the report from lkp.
>> 2. add one more patch (20th).
>>
>> Hi,
>>
>> This series aim to cleanup siw code, please review and comment!
>>
>> Thanks,
>> Guoqing
>>
>> Guoqing Jiang (20):
>>    RDMA/siw: Introduce siw_get_page
>>    RDMA/siw: Introduce siw_srx_update_skb
>>    RDMA/siw: Use iov.iov_len in kernel_sendmsg
>>    RDMA/siw: Remove goto lable in siw_mmap
>>    RDMA/siw: Remove rcu from siw_qp
>>    RDMA/siw: No need to check term_info.valid before call
>>      siw_send_terminate
>>    RDMA/siw: Also goto out_sem_up if pin_user_pages returns 0
>>    RDMA/siw: Factor out siw_generic_rx helper
>>    RDMA/siw: Introduce SIW_STAG_MAX_INDEX
>>    RDMA/siw: Add one parameter to siw_destroy_cpulist
>>    RDMA/siw: Introduce siw_cep_set_free_and_put
>>    RDMA/siw: Introduce siw_free_cm_id
>>    RDMA/siw: Simplify siw_qp_id2obj
>>    RDMA/siw: Simplify siw_mem_id2obj
>>    RDMA/siw: Cleanup siw_accept
>>    RDMA/siw: Remove siw_sk_assign_cm_upcalls
>>    RDMA/siw: Fix typo
>>    RDMA/siw: Only check attrs->cap.max_send_wr in siw_create_qp
>>    RDMA/siw: Introduce siw_destroy_cep_sock
>>    RDMA/siw: Update comments for siw_qp_sq_process
>>
>>   drivers/infiniband/sw/siw/siw.h       |   9 +-
>>   drivers/infiniband/sw/siw/siw_cm.c    | 154 +++++++++++---------------
>>   drivers/infiniband/sw/siw/siw_main.c  |  30 +++--
>>   drivers/infiniband/sw/siw/siw_mem.c   |  22 ++--
>>   drivers/infiniband/sw/siw/siw_qp.c    |   2 +-
>>   drivers/infiniband/sw/siw/siw_qp_rx.c |  84 ++++++--------
>>   drivers/infiniband/sw/siw/siw_qp_tx.c |  39 +++----
>>   drivers/infiniband/sw/siw/siw_verbs.c |  23 +---
>>   8 files changed, 144 insertions(+), 219 deletions(-)
>>
>>
>> base-commit: 964168970cef5f5b738fae047e6de2107842feb7
>> --
>> 2.35.3
> Hi Guoqing,
> Thanks for the effort! I like most of it.
>
> And, sorry, I saw I started my review with version 1 of your
> patches. Luckily it does not have functional differences to
> v2.

Thanks a lot for your review! I have repliedĀ  them separately.

> But I expect a version 3 anyway.

No problem, will do it.

> I currently do not have access to physical machines to check and
> run the more complex patches. I hope to do that (patch 08,10,12,15)
> tomorrow when back in office.

More check and running test with physical machines would be great :).
I have run some tests (blktests and xfstests) in VM against nvme host
which connected to nvme target through siw.

Thanks,
Guoqing