From: Varun Prakash <varun@chelsio.com>
To: Sagi Grimberg <sagi@grimberg.me>
Cc: hch@lst.de, kbusch@kernel.org, linux-nvme@lists.infradead.org
Subject: Re: [PATCH v2 2/2] nvme-tcp: send H2CData PDUs based on MAXH2CDATA
Date: Tue, 23 Nov 2021 14:03:49 +0530 [thread overview]
Message-ID: <20211123083346.GA1704@chelsio.com> (raw)
In-Reply-To: <f6e46674-5856-1f0d-045c-6ced393bda70@grimberg.me>
On Mon, Nov 22, 2021 at 12:35:11PM +0200, Sagi Grimberg wrote:
> >@@ -933,6 +934,7 @@ static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
> > {
> > struct nvme_tcp_queue *queue = req->queue;
> > int req_data_len = req->data_len;
> >+ u32 h2cdata_left = req->h2cdata_left;
>
> No need for the local variable, the reference only happens when the
> pdu data transfer is completed.
As this function also executes for inline data it is possible that completion
gets processed and request gets allocated for new cmd before
if (req->h2cdata_left) check, nvme_tcp_setup_cmd_pdu() will set
req->h2cdata_left to 0 so it will work but it does not look correct
to me. IMO we should use local variable here.
>
> > while (true) {
> > struct page *page = nvme_tcp_req_cur_page(req);
> >@@ -977,7 +979,13 @@ static int nvme_tcp_try_send_data(struct nvme_tcp_request *req)
> > req->state = NVME_TCP_SEND_DDGST;
> > req->offset = 0;
> > } else {
> >- nvme_tcp_done_send_req(queue);
> >+ if (h2cdata_left) {
> >+ nvme_tcp_setup_h2c_data_pdu(req);
> >+ req->state = NVME_TCP_SEND_H2C_PDU;
> >+ req->offset = 0;
>
> >@@ -1028,16 +1036,21 @@ static int nvme_tcp_try_send_data_pdu(struct nvme_tcp_request *req)
> > {
> > struct nvme_tcp_queue *queue = req->queue;
> > struct nvme_tcp_data_pdu *pdu = req->pdu;
> >+ struct page *page = virt_to_page(pdu);
> >+ size_t offset = offset_in_page(pdu) + req->offset;
> > u8 hdgst = nvme_tcp_hdgst_len(queue);
> > int len = sizeof(*pdu) - req->offset + hdgst;
> >+ int flags = MSG_DONTWAIT | MSG_MORE | MSG_SENDPAGE_NOTLAST;
> > int ret;
>
> Not sure the local variables are needed, but OK, I'm fine with that.
I will remove these local variables in next patch.
> >@@ -1057,6 +1070,7 @@ static int nvme_tcp_try_send_ddgst(struct nvme_tcp_request *req)
> > {
> > struct nvme_tcp_queue *queue = req->queue;
> > size_t offset = req->offset;
> >+ u32 h2cdata_left = req->h2cdata_left;
>
> No need for local variable.
Same explanation as above for local variable.
next prev parent reply other threads:[~2021-11-23 8:34 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-11-22 10:25 [PATCH v2 2/2] nvme-tcp: send H2CData PDUs based on MAXH2CDATA Varun Prakash
2021-11-22 10:35 ` Sagi Grimberg
2021-11-23 8:33 ` Varun Prakash [this message]
2021-11-23 9:09 ` Sagi Grimberg
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=20211123083346.GA1704@chelsio.com \
--to=varun@chelsio.com \
--cc=hch@lst.de \
--cc=kbusch@kernel.org \
--cc=linux-nvme@lists.infradead.org \
--cc=sagi@grimberg.me \
/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