From: Leon Romanovsky <leon@kernel.org>
To: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: mustafa.ismail@intel.com, shiraz.saleem@intel.com,
dledford@redhat.com, jgg@ziepe.ca, linux-rdma@vger.kernel.org,
linux-kernel@vger.kernel.org, kernel-janitors@vger.kernel.org
Subject: Re: [PATCH] RDMA/irdma: Improve the way 'cqp_request' structures are cleaned when they are recycled
Date: Tue, 20 Jul 2021 15:23:08 +0300 [thread overview]
Message-ID: <YPbALA/P5+NsC7MO@unreal> (raw)
In-Reply-To: <7f93f2a2c2fd18ddfeb99339d175b85ffd1c6398.1626713915.git.christophe.jaillet@wanadoo.fr>
On Mon, Jul 19, 2021 at 07:02:15PM +0200, Christophe JAILLET wrote:
> A set of IRDMA_CQP_SW_SQSIZE_2048 (i.e. 2048) 'cqp_request' are
> pre-allocated and zeroed in 'irdma_create_cqp()' (hw.c). These
> structures are managed with the 'cqp->cqp_avail_reqs' list which keeps
> track of available entries.
>
> In 'irdma_free_cqp_request()' (utils.c), when an entry is recycled and goes
> back to the 'cqp_avail_reqs' list, some fields are reseted.
>
> However, one of these fields, 'compl_info', is initialized within
> 'irdma_alloc_and_get_cqp_request()'.
>
> Move the corresponding memset to 'irdma_free_cqp_request()' so that the
> clean-up is done in only one place. This makes the logic more easy to
> understand.
I'm not so sure. The function irdma_alloc_and_get_cqp_request() returns
prepared cqp_request and all users expect that it will returned cleaned
one. The reliance on some other place to clear part of the structure is
prone to errors.
Thanks
>
> This also saves this memset in the case that the 'cqp_avail_reqs' list is
> empty and a new 'cqp_request' structure must be allocated. This memset is
> useless, because the structure is already kzalloc'ed.
>
> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
> ---
> drivers/infiniband/hw/irdma/utils.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/infiniband/hw/irdma/utils.c b/drivers/infiniband/hw/irdma/utils.c
> index 5bbe44e54f9a..66711024d38b 100644
> --- a/drivers/infiniband/hw/irdma/utils.c
> +++ b/drivers/infiniband/hw/irdma/utils.c
> @@ -445,7 +445,6 @@ struct irdma_cqp_request *irdma_alloc_and_get_cqp_request(struct irdma_cqp *cqp,
>
> cqp_request->waiting = wait;
> refcount_set(&cqp_request->refcnt, 1);
> - memset(&cqp_request->compl_info, 0, sizeof(cqp_request->compl_info));
>
> return cqp_request;
> }
> @@ -475,6 +474,7 @@ void irdma_free_cqp_request(struct irdma_cqp *cqp,
> cqp_request->request_done = false;
> cqp_request->callback_fcn = NULL;
> cqp_request->waiting = false;
> + memset(&cqp_request->compl_info, 0, sizeof(cqp_request->compl_info));
>
> spin_lock_irqsave(&cqp->req_lock, flags);
> list_add_tail(&cqp_request->list, &cqp->cqp_avail_reqs);
> --
> 2.30.2
>
next prev parent reply other threads:[~2021-07-20 12:23 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-19 17:02 [PATCH] RDMA/irdma: Improve the way 'cqp_request' structures are cleaned when they are recycled Christophe JAILLET
2021-07-20 12:23 ` Leon Romanovsky [this message]
2021-07-20 13:05 ` Christophe JAILLET
2021-07-21 5:11 ` Leon Romanovsky
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=YPbALA/P5+NsC7MO@unreal \
--to=leon@kernel.org \
--cc=christophe.jaillet@wanadoo.fr \
--cc=dledford@redhat.com \
--cc=jgg@ziepe.ca \
--cc=kernel-janitors@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-rdma@vger.kernel.org \
--cc=mustafa.ismail@intel.com \
--cc=shiraz.saleem@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).