From: sagi@grimberg.me (Sagi Grimberg)
Subject: [PATCH v3 3/3] nvmet-rdma: support 16K inline data
Date: Thu, 31 May 2018 00:45:42 +0300 [thread overview]
Message-ID: <2d10d67d-9504-b66b-b795-b69d536e39ad@grimberg.me> (raw)
In-Reply-To: <01000163b1bd094e-8a191404-8725-40c0-9af5-c4b69f324a1d-000000@email.amazonses.com>
>> @@ -200,17 +204,17 @@ static int nvmet_rdma_alloc_cmd(struct nvmet_rdma_device *ndev,
>> c->sge[0].length = sizeof(*c->nvme_cmd);
>> c->sge[0].lkey = ndev->pd->local_dma_lkey;
>>
>> - if (!admin) {
>> + if (!admin && inline_data_size) {
>> c->inline_page = alloc_pages(GFP_KERNEL,
>> - get_order(NVMET_RDMA_INLINE_DATA_SIZE));
>> + get_order(inline_data_size));
>
> Now we do higher order allocations here. This means that the allocation
> can fail if system memory is highly fragmented. And the allocations can no
> longer be satisfied from the per cpu caches. So allocation performance
> will drop.
That was my first thought as well. I'm not too keen on having
higher-order allocations on this, not at all. nvmet-rdma will
allocate a whole bunch of those. I think we should try to be
good citizens. I don't think its too complicated to do is it?
next prev parent reply other threads:[~2018-05-30 21:45 UTC|newest]
Thread overview: 39+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-29 18:26 [PATCH v3 0/3] NVMF/RDMA 16K Inline Support Steve Wise
2018-05-29 18:25 ` [PATCH v3 1/3] nvme-rdma: correctly check for target keyed sgl support Steve Wise
2018-05-29 20:23 ` Ruhl, Michael J
2018-05-30 14:39 ` Steve Wise
2018-05-30 15:11 ` Steve Wise
2018-05-30 21:37 ` Sagi Grimberg
2018-05-31 17:02 ` hch
2018-05-31 17:17 ` Steve Wise
2018-05-31 17:25 ` hch
2018-06-01 13:08 ` Steve Wise
2018-06-03 11:57 ` Sagi Grimberg
2018-06-03 18:27 ` Steve Wise
2018-06-04 12:01 ` Sagi Grimberg
2018-06-04 12:11 ` Christoph Hellwig
2018-06-04 12:17 ` Steve Wise
2018-06-04 13:52 ` Max Gurtovoy
2018-06-04 14:21 ` Steve Wise
2018-06-04 14:29 ` Max Gurtovoy
2018-06-04 14:31 ` Steve Wise
2018-06-04 14:37 ` Max Gurtovoy
2018-06-04 14:45 ` Steve Wise
2018-05-31 17:00 ` hch
2018-05-29 18:25 ` [PATCH v3 2/3] nvme-rdma: support up to 4 segments of inline data Steve Wise
2018-05-30 21:42 ` Sagi Grimberg
2018-05-30 21:46 ` Steve Wise
2018-05-29 18:25 ` [PATCH v3 3/3] nvmet-rdma: support 16K " Steve Wise
2018-05-30 15:49 ` Christopher Lameter
2018-05-30 16:46 ` Steve Wise
2018-05-30 17:02 ` Christopher Lameter
2018-05-30 21:45 ` Sagi Grimberg [this message]
2018-05-30 21:52 ` Steve Wise
2018-05-30 22:13 ` Sagi Grimberg
2018-05-30 22:26 ` Steve Wise
2018-06-03 8:39 ` Max Gurtovoy
2018-06-03 18:25 ` Steve Wise
2018-06-04 13:58 ` Max Gurtovoy
2018-06-04 14:18 ` Steve Wise
2018-06-05 8:52 ` Max Gurtovoy
2018-06-05 14:28 ` Steve Wise
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=2d10d67d-9504-b66b-b795-b69d536e39ad@grimberg.me \
--to=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;
as well as URLs for NNTP newsgroup(s).