patches for DPDK stable branches
 help / color / mirror / Atom feed
From: christian.ehrhardt@canonical.com
To: "Mattias Rönnblom" <mattias.ronnblom@ericsson.com>
Cc: Olivier Matz <olivier.matz@6wind.com>, dpdk stable <stable@dpdk.org>
Subject: patch 'net: accept unaligned data in checksum routines' has been queued to stable release 19.11.14
Date: Fri, 11 Nov 2022 12:38:19 +0100	[thread overview]
Message-ID: <20221111113904.1549618-2-christian.ehrhardt@canonical.com> (raw)
In-Reply-To: <20221111113904.1549618-1-christian.ehrhardt@canonical.com>

Hi,

FYI, your patch has been queued to stable release 19.11.14

Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
It will be pushed if I get no objections before 11/18/22. So please
shout if anyone has objections.

Also note that after the patch there's a diff of the upstream commit vs the
patch applied to the branch. This will indicate if there was any rebasing
needed to apply to the stable branch. If there were code changes for rebasing
(ie: not only metadata diffs), please double check that the rebase was
correctly done.

Queued patches are on a temporary branch at:
https://github.com/cpaelzer/dpdk-stable-queue

This queued commit can be viewed at:
https://github.com/cpaelzer/dpdk-stable-queue/commit/8803ccb7fdf7f7c76463b871d36eeb685fae38b8

Thanks.

Christian Ehrhardt <christian.ehrhardt@canonical.com>

---
From 8803ccb7fdf7f7c76463b871d36eeb685fae38b8 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Mattias=20R=C3=B6nnblom?= <mattias.ronnblom@ericsson.com>
Date: Mon, 11 Jul 2022 14:11:32 +0200
Subject: [PATCH] net: accept unaligned data in checksum routines
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

[ upstream commit 1c9a7fba5c90e0422b517404499ed106f647bcff ]

__rte_raw_cksum() (used by rte_raw_cksum() among others) accessed its
data through an uint16_t pointer, which allowed the compiler to assume
the data was 16-bit aligned. This in turn would, with certain
architectures and compiler flag combinations, result in code with SIMD
load or store instructions with restrictions on data alignment.

This patch keeps the old algorithm, but data is read using memcpy()
instead of direct pointer access, forcing the compiler to always
generate code that handles unaligned input. The __may_alias__ GCC
attribute is no longer needed.

The data on which the Internet checksum functions operates are almost
always 16-bit aligned, but there are exceptions. In particular, the
PDCP protocol header may (literally) have an odd size.

Performance impact seems to range from none to a very slight
regression.

Bugzilla ID: 1035
Fixes: 6006818cfb26 ("net: new checksum functions")

Signed-off-by: Mattias Rönnblom <mattias.ronnblom@ericsson.com>
Acked-by: Olivier Matz <olivier.matz@6wind.com>
---
 lib/librte_net/rte_ip.h | 17 ++++++++++-------
 1 file changed, 10 insertions(+), 7 deletions(-)

diff --git a/lib/librte_net/rte_ip.h b/lib/librte_net/rte_ip.h
index 279e88ee39..9ad8f0f3f4 100644
--- a/lib/librte_net/rte_ip.h
+++ b/lib/librte_net/rte_ip.h
@@ -119,18 +119,21 @@ struct rte_ipv4_hdr {
 static inline uint32_t
 __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 void *end;
 
-	for (; u16_buf != end; ++u16_buf)
-		sum += *u16_buf;
+	for (end = RTE_PTR_ADD(buf, RTE_ALIGN_FLOOR(len, sizeof(uint16_t)));
+	     buf != end; buf = RTE_PTR_ADD(buf, sizeof(uint16_t))) {
+		uint16_t v;
+
+		memcpy(&v, buf, sizeof(uint16_t));
+		sum += v;
+	}
 
 	/* if length is odd, keeping it byte order independent */
 	if (unlikely(len % 2)) {
 		uint16_t left = 0;
-		*(unsigned char *)&left = *(const unsigned char *)end;
+
+		memcpy(&left, end, 1);
 		sum += left;
 	}
 
-- 
2.38.1

---
  Diff of the applied patch vs upstream commit (please double-check if non-empty:
---
--- -	2022-11-11 12:35:04.873728886 +0100
+++ 0002-net-accept-unaligned-data-in-checksum-routines.patch	2022-11-11 12:35:04.697191643 +0100
@@ -1 +1 @@
-From 1c9a7fba5c90e0422b517404499ed106f647bcff Mon Sep 17 00:00:00 2001
+From 8803ccb7fdf7f7c76463b871d36eeb685fae38b8 Mon Sep 17 00:00:00 2001
@@ -8,0 +9,2 @@
+[ upstream commit 1c9a7fba5c90e0422b517404499ed106f647bcff ]
+
@@ -29 +30,0 @@
-Cc: stable@dpdk.org
@@ -34 +35 @@
- lib/net/rte_ip.h | 17 ++++++++++-------
+ lib/librte_net/rte_ip.h | 17 ++++++++++-------
@@ -37,5 +38,5 @@
-diff --git a/lib/net/rte_ip.h b/lib/net/rte_ip.h
-index b502481670..ecd250e9be 100644
---- a/lib/net/rte_ip.h
-+++ b/lib/net/rte_ip.h
-@@ -160,18 +160,21 @@ rte_ipv4_hdr_len(const struct rte_ipv4_hdr *ipv4_hdr)
+diff --git a/lib/librte_net/rte_ip.h b/lib/librte_net/rte_ip.h
+index 279e88ee39..9ad8f0f3f4 100644
+--- a/lib/librte_net/rte_ip.h
++++ b/lib/librte_net/rte_ip.h
+@@ -119,18 +119,21 @@ struct rte_ipv4_hdr {

  reply	other threads:[~2022-11-11 11:39 UTC|newest]

Thread overview: 47+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-11 11:38 patch 'vhost: fix virtqueue use after free on NUMA reallocation' " christian.ehrhardt
2022-11-11 11:38 ` christian.ehrhardt [this message]
2022-11-11 11:38 ` patch 'app/testpmd: restore ixgbe bypass commands' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/bonding: fix array overflow in Rx burst' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/bonding: fix double slave link status query' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/nfp: improve HW info header log readability' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/mvneta: fix build with GCC 12' " christian.ehrhardt
2022-11-11 11:38 ` patch 'malloc: fix storage size for some allocations' " christian.ehrhardt
2022-11-11 11:38 ` patch 'event/sw: fix device name in dump' " christian.ehrhardt
2022-11-11 11:38 ` patch 'eventdev/eth_tx: add spinlock for adapter start/stop' " christian.ehrhardt
2022-11-11 11:38 ` patch 'eventdev/eth_tx: fix adapter stop' " christian.ehrhardt
2022-11-11 11:38 ` patch 'test/ipsec: skip if no compatible device' " christian.ehrhardt
2022-11-11 11:38 ` patch 'test/crypto: fix debug messages' " christian.ehrhardt
2022-11-11 11:38 ` patch 'test/ipsec: fix build with GCC 12' " christian.ehrhardt
2022-11-11 11:38 ` patch 'ipsec: " christian.ehrhardt
2022-11-11 11:38 ` patch 'crypto/qat: " christian.ehrhardt
2022-11-11 11:38 ` patch 'eventdev: fix name of Rx conf type in documentation' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/i40e: fix VF representor release' " christian.ehrhardt
2022-11-11 11:38 ` patch 'common/iavf: avoid copy in async mode' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/ice/base: fix media type of PHY 10G SFI C2C' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/ice/base: fix array overflow in add switch recipe' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/virtio: fix crash when configured twice' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/mlx4: fix Verbs FD leak in secondary process' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/mlx5: fix meter profile delete after disable' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/ice: check illegal packet sizes' " christian.ehrhardt
2022-11-11 11:38 ` patch 'common/sfc_efx/base: fix maximum Tx data count' " christian.ehrhardt
2022-11-11 11:38 ` patch 'examples/vm_power_manager: use safe list iterator' " christian.ehrhardt
2022-11-11 11:38 ` patch 'gro: fix chain index for more than 2 packets' " christian.ehrhardt
2022-11-11 11:38 ` patch 'timer: fix stopping all timers' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/qede/base: fix 32-bit build with GCC 12' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/tap: fix overflow of network interface index' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/memif: fix crash with different number of Rx/Tx queues' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/hns3: fix crash when secondary process access FW' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/bonding: fix Tx hash for TCP' " christian.ehrhardt
2022-11-11 11:38 ` patch 'eal: fix data race in multi-process support' " christian.ehrhardt
2022-11-11 11:38 ` patch 'pdump: do not allow enable/disable in primary process' " christian.ehrhardt
2022-11-11 11:38 ` patch 'event/sw: fix flow ID init in self test' " christian.ehrhardt
2022-11-11 11:38 ` patch 'event/sw: fix log " christian.ehrhardt
2022-11-11 11:38 ` patch 'eventdev/eth_tx: fix queue delete' " christian.ehrhardt
2022-11-11 11:38 ` patch 'license: fix paths' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/bonding: fix descriptor limit reporting' " christian.ehrhardt
2022-11-11 11:38 ` patch 'net/ixgbe: fix broadcast Rx on VF after promisc removal' " christian.ehrhardt
2022-11-11 11:39 ` patch 'net/ixgbe: fix unexpected VLAN Rx in promisc mode on VF' " christian.ehrhardt
2022-11-11 11:39 ` patch 'examples/l2fwd-crypto: fix typo in error message' " christian.ehrhardt
2022-11-11 11:39 ` patch 'test/crypto: fix wireless auth digest segment' " christian.ehrhardt
2022-11-11 11:39 ` patch 'test/crypto: fix PDCP vectors' " christian.ehrhardt
2022-11-11 11:39 ` patch 'crypto/qat: fix null hash algorithm digest size' " christian.ehrhardt

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=20221111113904.1549618-2-christian.ehrhardt@canonical.com \
    --to=christian.ehrhardt@canonical.com \
    --cc=mattias.ronnblom@ericsson.com \
    --cc=olivier.matz@6wind.com \
    --cc=stable@dpdk.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).