From: Chuck Lever <chuck.lever@oracle.com>
To: Bruce Fields <bfields@fieldses.org>
Cc: Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH v1 19/42] SUNRPC: Fix xdr_get_next_encode_buffer() page boundary handling
Date: Wed, 3 Mar 2021 18:19:33 +0000 [thread overview]
Message-ID: <D8625219-C758-44C0-A74E-272019E12C2E@oracle.com> (raw)
In-Reply-To: <20210303165251.GB1282@fieldses.org>
> On Mar 3, 2021, at 11:52 AM, Bruce Fields <bfields@fieldses.org> wrote:
>
> On Wed, Mar 03, 2021 at 03:43:28PM +0000, Chuck Lever wrote:
>> Hi Bruce-
>>
>> Thanks for your careful review of this series!
>>
>>
>>> On Mar 2, 2021, at 5:11 PM, J. Bruce Fields <bfields@fieldses.org> wrote:
>>>
>>> On Mon, Mar 01, 2021 at 10:17:13AM -0500, Chuck Lever wrote:
>>>> The description of commit 2825a7f90753 ("nfsd4: allow encoding
>>>> across page boundaries") states:
>>>>
>>>>> Also we can't handle a new operation starting close to the end of
>>>>> a page.
>>>>
>>>> But does not detail why this is the case.
>>>
>>> That wasn't every helpful of me, sorry.
>>>
>>>> Subtracting the scratch buffer's "shift" value from the remaining
>>>> stream space seems to make reserving space close to the end of the
>>>> buf->pages array reliable.
>>>
>>> So, why is that?
>>>
>>> Thinking this through:
>>>
>>> When somebody asks for a buffer that would straddle a page boundary,
>>> with frag1bytes at the end of this page and frag2bytes at the start of
>>> the next page, we instead give them a buffer starting at start of the
>>> next page. That gives them a nice contiguous buffer to write into.
>>> When they're done using it, we fix things up by copying what they wrote
>>> back to where it should be.
>>>
>>> That means we're temporarily wasting frag1bytes of space. So, I don't
>>> know, maybe that's the logic behind subtracing frag1bytes from
>>> space_left.
>>>
>>> It means you may end up with xdr->end frag1bytes short of the next page.
>
> Wait, let me try that again:
>
> p = page_address(*xdr->page_ptr);
> xdr->p = (void *)p + frag2bytes;
> space_left = xdr->buf->buflen - xdr->buf->len - frag1bytes;
> xdr->end = (void *)p + min_t(int, space_left, PAGE_SIZE);
>
> If you've still got a lot of buffer space left, then that'll put
> xdr->end frag2bytes past the end of a page, won't it?
So far I haven't heard any complaints from the kernel. I'll
instrument the code and see what comes up.
>>> I'm not sure that's right.
>>
>> Why would that not be OK? the next call to xdr_get_next_encode_buffer()
>> should do the right thing and bounce the new encoded data from the
>> next page into this one again.
>>
>> So far I have not encountered any problems. Would such a problem show
>> up with some frequency under normal use, or would it be especially
>> subtle?
>
> I mainly just want to make sure we've got a coherent idea what this code
> is doing....
Agreed, that's a good thing.
> For testing: large replies that aren't just read data are readdir and
> getacl. So reading large directories with lots of variably-named files
> might be good. Also pynfs could easily send a single compound with lots
> of variable-sized reads, that might be interesting.
I've run the full git regression suite over NFSv3 many many times with
this patch applied. That includes unpacking via tar, a full build from
scratch, and then running thousands of individual tests.
So that doesn't exercise a particular corner case, but it does reflect
a broad variety of directory operations.
> Constructing a compound that will result in xdr_reserve_space calls that
> exactly hit the various corner cases may be hard. But maybe if we just
> send a bunch of compounds that vary over some range we can still
> guarantee hitting those cases.
>
> --b.
--
Chuck Lever
next prev parent reply other threads:[~2021-03-04 0:29 UTC|newest]
Thread overview: 58+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-03-01 15:15 [PATCH v1 00/42] NFSv2/3 XDR encoder overhaul Chuck Lever
2021-03-01 15:15 ` [PATCH v1 01/42] NFSD: Extract the svcxdr_init_encode() helper Chuck Lever
2021-03-02 16:47 ` J. Bruce Fields
2021-03-02 16:57 ` Chuck Lever
2021-03-01 15:15 ` [PATCH v1 02/42] NFSD: Update the GETATTR3res encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:15 ` [PATCH v1 03/42] NFSD: Update the NFSv3 ACCESS3res " Chuck Lever
2021-03-01 15:15 ` [PATCH v1 04/42] NFSD: Update the NFSv3 LOOKUP3res " Chuck Lever
2021-03-01 15:15 ` [PATCH v1 05/42] NFSD: Update the NFSv3 wccstat result " Chuck Lever
2021-03-01 15:15 ` [PATCH v1 06/42] NFSD: Update the NFSv3 READLINK3res " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 07/42] NFSD: Update the NFSv3 READ3res encode " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 08/42] NFSD: Update the NFSv3 WRITE3res encoder " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 09/42] NFSD: Update the NFSv3 CREATE family of encoders " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 10/42] NFSD: Update the NFSv3 RENAMEv3res encoder " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 11/42] NFSD: Update the NFSv3 LINK3res " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 12/42] NFSD: Update the NFSv3 FSSTAT3res " Chuck Lever
2021-03-02 20:45 ` J. Bruce Fields
2021-03-02 20:57 ` Chuck Lever
2021-03-01 15:16 ` [PATCH v1 13/42] NFSD: Update the NFSv3 FSINFO3res " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 14/42] NFSD: Update the NFSv3 PATHCONF3res " Chuck Lever
2021-03-03 7:13 ` Christoph Hellwig
2021-03-03 15:34 ` Chuck Lever
2021-03-01 15:16 ` [PATCH v1 15/42] NFSD: Update the NFSv3 COMMIT3res " Chuck Lever
2021-03-01 15:16 ` [PATCH v1 16/42] NFSD: Add a helper that encodes NFSv3 directory offset cookies Chuck Lever
2021-03-01 15:17 ` [PATCH v1 17/42] NFSD: Count bytes instead of pages in the NFSv3 READDIR encoder Chuck Lever
2021-03-01 15:17 ` [PATCH v1 18/42] NFSD: Update the NFSv3 READDIR3res encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:17 ` [PATCH v1 19/42] SUNRPC: Fix xdr_get_next_encode_buffer() page boundary handling Chuck Lever
2021-03-02 22:11 ` J. Bruce Fields
2021-03-03 15:43 ` Chuck Lever
2021-03-03 16:52 ` Bruce Fields
2021-03-03 18:19 ` Chuck Lever [this message]
2021-03-03 18:25 ` Bruce Fields
2021-03-03 18:27 ` Chuck Lever
2021-03-03 18:30 ` Chuck Lever
2021-03-03 18:38 ` Bruce Fields
2021-03-03 18:42 ` Chuck Lever
2021-03-01 15:17 ` [PATCH v1 20/42] NFSD: Update NFSv3 READDIR entry encoders to use struct xdr_stream Chuck Lever
2021-03-01 15:17 ` [PATCH v1 21/42] NFSD: Remove unused NFSv3 directory entry encoders Chuck Lever
2021-03-01 15:17 ` [PATCH v1 22/42] NFSD: Reduce svc_rqst::rq_pages churn during READDIR operations Chuck Lever
2021-03-01 15:17 ` [PATCH v1 23/42] NFSD: Update the NFSv2 stat encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:17 ` [PATCH v1 24/42] NFSD: Update the NFSv2 attrstat " Chuck Lever
2021-03-01 15:17 ` [PATCH v1 25/42] NFSD: Update the NFSv2 diropres " Chuck Lever
2021-03-01 15:17 ` [PATCH v1 26/42] NFSD: Update the NFSv2 READLINK result " Chuck Lever
2021-03-01 15:18 ` [PATCH v1 27/42] NFSD: Update the NFSv2 READ " Chuck Lever
2021-03-01 15:18 ` [PATCH v1 28/42] NFSD: Update the NFSv2 STATFS " Chuck Lever
2021-03-01 15:18 ` [PATCH v1 29/42] NFSD: Add a helper that encodes NFSv3 directory offset cookies Chuck Lever
2021-03-01 15:18 ` [PATCH v1 30/42] NFSD: Count bytes instead of pages in the NFSv2 READDIR encoder Chuck Lever
2021-03-01 15:18 ` [PATCH v1 31/42] NFSD: Update the NFSv2 READDIR result encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:18 ` [PATCH v1 32/42] NFSD: Update the NFSv2 READDIR entry " Chuck Lever
2021-03-01 15:18 ` [PATCH v1 33/42] NFSD: Remove unused NFSv2 directory entry encoders Chuck Lever
2021-03-01 15:18 ` [PATCH v1 34/42] NFSD: Add an xdr_stream-based encoder for NFSv2/3 ACLs Chuck Lever
2021-03-01 15:18 ` [PATCH v1 35/42] NFSD: Update the NFSv2 GETACL result encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:18 ` [PATCH v1 36/42] NFSD: Update the NFSv2 SETACL " Chuck Lever
2021-03-01 15:19 ` [PATCH v1 37/42] NFSD: Update the NFSv2 ACL GETATTR " Chuck Lever
2021-03-01 15:19 ` [PATCH v1 38/42] NFSD: Update the NFSv2 ACL ACCESS " Chuck Lever
2021-03-01 15:19 ` [PATCH v1 39/42] NFSD: Clean up after updating NFSv2 ACL encoders Chuck Lever
2021-03-01 15:19 ` [PATCH v1 40/42] NFSD: Update the NFSv3 GETACL result encoder to use struct xdr_stream Chuck Lever
2021-03-01 15:19 ` [PATCH v1 41/42] NFSD: Update the NFSv3 SETACL " Chuck Lever
2021-03-01 15:19 ` [PATCH v1 42/42] NFSD: Clean up after updating NFSv3 ACL encoders Chuck Lever
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=D8625219-C758-44C0-A74E-272019E12C2E@oracle.com \
--to=chuck.lever@oracle.com \
--cc=bfields@fieldses.org \
--cc=linux-nfs@vger.kernel.org \
/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).