From: "Morten Brørup" <mb@smartsharesystems.com>
To: "Emil Berg" <emil.berg@ericsson.com>,
"Mattias Rönnblom" <hofors@lysator.liu.se>,
bruce.richardson@intel.com, dev@dpdk.org
Cc: <stephen@networkplumber.org>, <stable@dpdk.org>,
<bugzilla@dpdk.org>, <olivier.matz@6wind.com>
Subject: RE: [PATCH v4] net: fix checksum with unaligned buffer
Date: Mon, 27 Jun 2022 15:22:30 +0200 [thread overview]
Message-ID: <98CBD80474FA8B44BF855DF32C47DC35D87187@smartserver.smartshare.dk> (raw)
In-Reply-To: <AM8PR07MB76660E55460287468718E80598B99@AM8PR07MB7666.eurprd07.prod.outlook.com>
> From: Emil Berg [mailto:emil.berg@ericsson.com]
> Sent: Monday, 27 June 2022 14.51
>
> > From: Emil Berg
> > Sent: den 27 juni 2022 14:46
> >
> > > From: Mattias Rönnblom <hofors@lysator.liu.se>
> > > Sent: den 27 juni 2022 14:28
> > >
> > > On 2022-06-23 14:51, Morten Brørup wrote:
> > > >> From: Morten Brørup [mailto:mb@smartsharesystems.com]
> > > >> Sent: Thursday, 23 June 2022 14.39
> > > >>
> > > >> With this patch, the checksum can be calculated on an unaligned
> buffer.
> > > >> I.e. the buf parameter is no longer required to be 16 bit
> aligned.
> > > >>
> > > >> The checksum is still calculated using a 16 bit aligned pointer,
> so
> > > >> the compiler can auto-vectorize the function's inner loop.
> > > >>
> > > >> When the buffer is unaligned, the first byte of the buffer is
> > > >> handled separately. Furthermore, the calculated checksum of the
> > > >> buffer is byte shifted before being added to the initial
> checksum,
> > > >> to compensate for the checksum having been calculated on the
> buffer
> > > >> shifted by one byte.
> > > >>
> > > >> v4:
> > > >> * Add copyright notice.
> > > >> * Include stdbool.h (Emil Berg).
> > > >> * Use RTE_PTR_ADD (Emil Berg).
> > > >> * Fix one more typo in commit message. Is 'unligned' even a
> word?
> > > >> v3:
> > > >> * Remove braces from single statement block.
> > > >> * Fix typo in commit message.
> > > >> v2:
> > > >> * Do not assume that the buffer is part of an aligned packet
> buffer.
> > > >>
> > > >> Bugzilla ID: 1035
> > > >> Cc: stable@dpdk.org
> > > >>
> > > >> Signed-off-by: Morten Brørup <mb@smartsharesystems.com>
> > > >> ---
> > > >> lib/net/rte_ip.h | 32 +++++++++++++++++++++++++++-----
> > > >> 1 file changed, 27 insertions(+), 5 deletions(-)
> > > >>
> > > >> diff --git a/lib/net/rte_ip.h b/lib/net/rte_ip.h index
> > > >> b502481670..738d643da0 100644
> > > >> --- a/lib/net/rte_ip.h
> > > >> +++ b/lib/net/rte_ip.h
> > > >> @@ -3,6 +3,7 @@
> > > >> * The Regents of the University of California.
> > > >> * Copyright(c) 2010-2014 Intel Corporation.
> > > >> * Copyright(c) 2014 6WIND S.A.
> > > >> + * Copyright(c) 2022 SmartShare Systems.
> > > >> * All rights reserved.
> > > >> */
> > > >>
> > > >> @@ -15,6 +16,7 @@
> > > >> * IP-related defines
> > > >> */
> > > >>
> > > >> +#include <stdbool.h>
> > > >> #include <stdint.h>
> > > >>
> > > >> #ifdef RTE_EXEC_ENV_WINDOWS
> > > >> @@ -162,20 +164,40 @@ __rte_raw_cksum(const void *buf, size_t
> len,
> > > >> uint32_t sum)
> > > >> {
> > > >> /* extend strict-aliasing rules */
> > > >> typedef uint16_t __attribute__((__may_alias__)) u16_p;
> > > >> - const u16_p *u16_buf = (const u16_p *)buf;
> > > >> - const u16_p *end = u16_buf + len / sizeof(*u16_buf);
> > > >> + const u16_p *u16_buf;
> > > >> + const u16_p *end;
> > > >> + uint32_t bsum = 0;
> > > >> + const bool unaligned = (uintptr_t)buf & 1;
> > > >> +
> > > >> + /* if buffer is unaligned, keeping it byte order
> independent */
> > > >> + if (unlikely(unaligned)) {
> > > >> + uint16_t first = 0;
> > > >> + if (unlikely(len == 0))
> > > >> + return 0;
> > > >> + ((unsigned char *)&first)[1] = *(const unsigned
> > > char *)buf;
> > > >> + bsum += first;
> > > >> + buf = RTE_PTR_ADD(buf, 1);
> > > >> + len--;
> > > >> + }
> > > >>
> > > >> + /* aligned access for compiler auto-vectorization */
> > >
> > > The compiler will be able to auto vectorize even unaligned
> accesses,
> > > just with different instructions. From what I can tell, there's no
> > > performance impact, at least not on the x86_64 systems I tried on.
> > >
> > > I think you should remove the first special case conditional and
> use
> > > memcpy() instead of the cumbersome __may_alias__ construct to
> retrieve
> > > the data.
> > >
> >
> > Here:
> > https://www.agner.org/optimize/instruction_tables.pdf
> > it lists the latency of vmovdqa (aligned) as 6 cycles and the latency
> for
> > vmovdqu (unaligned) as 7 cycles. So I guess there can be some
> difference.
> > Although in practice I'm not sure what difference it makes. I've not
> seen any
> > difference in runtime between the two versions.
> >
>
> Correction to my comment:
> Those stats are for some older CPU. For some newer CPUs such as Tiger
> Lake the stats seem to be the same regardless of aligned or unaligned.
>
I agree that the memcpy method is more elegant and easy to read.
However, we would need to performance test the modified checksum function with a large number of CPUs to prove that we don't introduce a performance regression on any CPU architecture still supported by DPDK. And Emil already found a CPU where it costs 1 extra cycle per 16 bytes, which adds up to a total of ca. 91 extra cycles on a 1460 byte TCP packet.
So I opted for a solution with zero changes to the inner loop, so no performance retesting is required (for the previously supported use cases, where the buffer is aligned).
I have previously submitted a couple of patches to fix some minor bugs in the mempool cache functions [1] and [2], while also refactoring the functions for readability. But after having incorporated various feedback, nobody wants to proceed with the patches, probably due to fear of performance regressions. I didn't want to risk the same with this patch.
[1] http://inbox.dpdk.org/dev/98CBD80474FA8B44BF855DF32C47DC35D8712B@smartserver.smartshare.dk/
[2] http://inbox.dpdk.org/dev/98CBD80474FA8B44BF855DF32C47DC35D86FBB@smartserver.smartshare.dk/
next prev parent reply other threads:[~2022-06-27 13:22 UTC|newest]
Thread overview: 74+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-15 7:16 [Bug 1035] __rte_raw_cksum() crash with misaligned pointer bugzilla
2022-06-15 14:40 ` Morten Brørup
2022-06-16 5:44 ` Emil Berg
2022-06-16 6:27 ` Morten Brørup
2022-06-16 6:32 ` Emil Berg
2022-06-16 6:44 ` Morten Brørup
2022-06-16 13:58 ` Mattias Rönnblom
2022-06-16 14:36 ` Morten Brørup
2022-06-17 7:32 ` Morten Brørup
2022-06-17 8:45 ` [PATCH] net: fix checksum with unaligned buffer Morten Brørup
2022-06-17 9:06 ` Morten Brørup
2022-06-17 12:17 ` Emil Berg
2022-06-20 10:37 ` Emil Berg
2022-06-20 10:57 ` Morten Brørup
2022-06-21 7:16 ` Emil Berg
2022-06-21 8:05 ` Morten Brørup
2022-06-21 8:23 ` Bruce Richardson
2022-06-21 9:35 ` Morten Brørup
2022-06-22 6:26 ` Emil Berg
2022-06-22 9:18 ` Bruce Richardson
2022-06-22 11:26 ` Morten Brørup
2022-06-22 12:25 ` Emil Berg
2022-06-22 14:01 ` Morten Brørup
2022-06-22 14:03 ` Emil Berg
2022-06-23 5:21 ` Emil Berg
2022-06-23 7:01 ` Morten Brørup
2022-06-23 11:39 ` Emil Berg
2022-06-23 12:18 ` Morten Brørup
2022-06-22 13:44 ` [PATCH v2] " Morten Brørup
2022-06-22 13:54 ` [PATCH v3] " Morten Brørup
2022-06-23 12:39 ` [PATCH v4] " Morten Brørup
2022-06-23 12:51 ` Morten Brørup
2022-06-27 7:56 ` Emil Berg
2022-06-27 10:54 ` Morten Brørup
2022-06-27 12:28 ` Mattias Rönnblom
2022-06-27 12:46 ` Emil Berg
2022-06-27 12:50 ` Emil Berg
2022-06-27 13:22 ` Morten Brørup [this message]
2022-06-27 17:22 ` Mattias Rönnblom
2022-06-27 20:21 ` Morten Brørup
2022-06-28 6:28 ` Mattias Rönnblom
2022-06-30 16:28 ` Morten Brørup
2022-07-07 15:21 ` Stanisław Kardach
2022-07-07 18:34 ` [PATCH 1/2] app/test: add cksum performance test Mattias Rönnblom
2022-07-07 18:34 ` [PATCH 2/2] net: have checksum routines accept unaligned data Mattias Rönnblom
2022-07-07 21:44 ` Morten Brørup
2022-07-08 12:43 ` Mattias Rönnblom
2022-07-08 12:56 ` [PATCH v2 1/2] app/test: add cksum performance test Mattias Rönnblom
2022-07-08 12:56 ` [PATCH v2 2/2] net: have checksum routines accept unaligned data Mattias Rönnblom
2022-07-08 14:44 ` Ferruh Yigit
2022-07-11 9:53 ` Olivier Matz
2022-07-11 10:53 ` Mattias Rönnblom
2022-07-11 9:47 ` [PATCH v2 1/2] app/test: add cksum performance test Olivier Matz
2022-07-11 10:42 ` Mattias Rönnblom
2022-07-11 11:33 ` Olivier Matz
2022-07-11 12:11 ` [PATCH v3 " Mattias Rönnblom
2022-07-11 12:11 ` [PATCH v3 2/2] net: have checksum routines accept unaligned data Mattias Rönnblom
2022-07-11 13:25 ` Olivier Matz
2022-08-08 9:25 ` Mattias Rönnblom
2022-09-20 12:09 ` Mattias Rönnblom
2022-09-20 16:10 ` Thomas Monjalon
2022-07-11 13:20 ` [PATCH v3 1/2] app/test: add cksum performance test Olivier Matz
2022-07-08 13:02 ` [PATCH 2/2] net: have checksum routines accept unaligned data Morten Brørup
2022-07-08 13:52 ` Mattias Rönnblom
2022-07-08 14:10 ` Bruce Richardson
2022-07-08 14:30 ` Morten Brørup
2022-06-30 17:41 ` [PATCH v4] net: fix checksum with unaligned buffer Stephen Hemminger
2022-06-30 17:45 ` Stephen Hemminger
2022-07-01 4:11 ` Emil Berg
2022-07-01 16:50 ` Morten Brørup
2022-07-01 17:04 ` Stephen Hemminger
2022-07-01 20:46 ` Morten Brørup
2022-06-16 14:09 ` [Bug 1035] __rte_raw_cksum() crash with misaligned pointer Mattias Rönnblom
2022-10-10 10:40 ` bugzilla
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=98CBD80474FA8B44BF855DF32C47DC35D87187@smartserver.smartshare.dk \
--to=mb@smartsharesystems.com \
--cc=bruce.richardson@intel.com \
--cc=bugzilla@dpdk.org \
--cc=dev@dpdk.org \
--cc=emil.berg@ericsson.com \
--cc=hofors@lysator.liu.se \
--cc=olivier.matz@6wind.com \
--cc=stable@dpdk.org \
--cc=stephen@networkplumber.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).