[PATCH] app/test: use memcpy in ipsec test

[PATCH] app/test: use memcpy in ipsec test

[Stephen Hemminger]

This test has tables of data that get copied with rte_memcpy.
But when compiled without always inline the compiler gets confused
by the inlining of rte_memcpy and thinks that it is possible for AVX
code to reference past the input data.

Workaround is to use memcpy() which is better for this test anyway
since regular memcpy has more static checking from compiler and
analyzers.

Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
---
 app/test/test_ipsec.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/app/test/test_ipsec.c b/app/test/test_ipsec.c
index 139c1e8dec..b5a430996d 100644
--- a/app/test/test_ipsec.c
+++ b/app/test/test_ipsec.c
@@ -10,7 +10,6 @@
 #include <rte_hexdump.h>
 #include <rte_mbuf.h>
 #include <rte_malloc.h>
-#include <rte_memcpy.h>
 #include <rte_cycles.h>
 #include <rte_bus_vdev.h>
 #include <rte_ip.h>
@@ -559,7 +558,7 @@ setup_test_string(struct rte_mempool *mpool, const char *string,
 			return NULL;
 		}
 		if (string != NULL)
-			rte_memcpy(dst, string, t_len);
+			memcpy(dst, string, t_len);
 		else
 			memset(dst, 0, t_len);
 	}
@@ -604,22 +603,22 @@ setup_test_string_tunneled(struct rte_mempool *mpool, const char *string,
 	/* copy outer IP and ESP header */
 	ipv4_outer.total_length = rte_cpu_to_be_16(t_len);
 	ipv4_outer.packet_id = rte_cpu_to_be_16(seq);
-	rte_memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
+	memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
 	dst += sizeof(ipv4_outer);
 	m->l3_len = sizeof(ipv4_outer);
-	rte_memcpy(dst, &esph, sizeof(esph));
+	memcpy(dst, &esph, sizeof(esph));
 	dst += sizeof(esph);
 
 	if (string != NULL) {
 		/* copy payload */
-		rte_memcpy(dst, string, len);
+		memcpy(dst, string, len);
 		dst += len;
 		/* copy pad bytes */
-		rte_memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
+		memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
 			sizeof(esp_pad_bytes)));
 		dst += padlen;
 		/* copy ESP tail header */
-		rte_memcpy(dst, &espt, sizeof(espt));
+		memcpy(dst, &espt, sizeof(espt));
 	} else
 		memset(dst, 0, t_len);
 
-- 
2.53.0

RE: [PATCH] app/test: use memcpy in ipsec test

[Konstantin Ananyev]

> 
> This test has tables of data that get copied with rte_memcpy.
> But when compiled without always inline the compiler gets confused
> by the inlining of rte_memcpy and thinks that it is possible for AVX
> code to reference past the input data.
> 
> Workaround is to use memcpy() which is better for this test anyway
> since regular memcpy has more static checking from compiler and
> analyzers.
> 
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> ---
>  app/test/test_ipsec.c | 13 ++++++-------
>  1 file changed, 6 insertions(+), 7 deletions(-)
> 
> diff --git a/app/test/test_ipsec.c b/app/test/test_ipsec.c
> index 139c1e8dec..b5a430996d 100644
> --- a/app/test/test_ipsec.c
> +++ b/app/test/test_ipsec.c
> @@ -10,7 +10,6 @@
>  #include <rte_hexdump.h>
>  #include <rte_mbuf.h>
>  #include <rte_malloc.h>
> -#include <rte_memcpy.h>
>  #include <rte_cycles.h>
>  #include <rte_bus_vdev.h>
>  #include <rte_ip.h>
> @@ -559,7 +558,7 @@ setup_test_string(struct rte_mempool *mpool, const
> char *string,
>  			return NULL;
>  		}
>  		if (string != NULL)
> -			rte_memcpy(dst, string, t_len);
> +			memcpy(dst, string, t_len);
>  		else
>  			memset(dst, 0, t_len);
>  	}
> @@ -604,22 +603,22 @@ setup_test_string_tunneled(struct rte_mempool
> *mpool, const char *string,
>  	/* copy outer IP and ESP header */
>  	ipv4_outer.total_length = rte_cpu_to_be_16(t_len);
>  	ipv4_outer.packet_id = rte_cpu_to_be_16(seq);
> -	rte_memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
> +	memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
>  	dst += sizeof(ipv4_outer);
>  	m->l3_len = sizeof(ipv4_outer);
> -	rte_memcpy(dst, &esph, sizeof(esph));
> +	memcpy(dst, &esph, sizeof(esph));
>  	dst += sizeof(esph);
> 
>  	if (string != NULL) {
>  		/* copy payload */
> -		rte_memcpy(dst, string, len);
> +		memcpy(dst, string, len);
>  		dst += len;
>  		/* copy pad bytes */
> -		rte_memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
> +		memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
>  			sizeof(esp_pad_bytes)));
>  		dst += padlen;
>  		/* copy ESP tail header */
> -		rte_memcpy(dst, &espt, sizeof(espt));
> +		memcpy(dst, &espt, sizeof(espt));
>  	} else
>  		memset(dst, 0, t_len);
> 
> --

Acked-by: Konstantin Ananyev <konstantin.ananyev@huawei.com>

> 2.53.0

RE: [PATCH] app/test: use memcpy in ipsec test

[Morten Brørup]

> > This test has tables of data that get copied with rte_memcpy.
> > But when compiled without always inline the compiler gets confused
> > by the inlining of rte_memcpy and thinks that it is possible for AVX
> > code to reference past the input data.
> >
> > Workaround is to use memcpy() which is better for this test anyway
> > since regular memcpy has more static checking from compiler and
> > analyzers.
> >
> > Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>
> > ---
> >  app/test/test_ipsec.c | 13 ++++++-------
> >  1 file changed, 6 insertions(+), 7 deletions(-)
> >
> > diff --git a/app/test/test_ipsec.c b/app/test/test_ipsec.c
> > index 139c1e8dec..b5a430996d 100644
> > --- a/app/test/test_ipsec.c
> > +++ b/app/test/test_ipsec.c
> > @@ -10,7 +10,6 @@
> >  #include <rte_hexdump.h>
> >  #include <rte_mbuf.h>
> >  #include <rte_malloc.h>
> > -#include <rte_memcpy.h>
> >  #include <rte_cycles.h>
> >  #include <rte_bus_vdev.h>
> >  #include <rte_ip.h>
> > @@ -559,7 +558,7 @@ setup_test_string(struct rte_mempool *mpool,
> const
> > char *string,
> >  			return NULL;
> >  		}
> >  		if (string != NULL)
> > -			rte_memcpy(dst, string, t_len);
> > +			memcpy(dst, string, t_len);
> >  		else
> >  			memset(dst, 0, t_len);
> >  	}
> > @@ -604,22 +603,22 @@ setup_test_string_tunneled(struct rte_mempool
> > *mpool, const char *string,
> >  	/* copy outer IP and ESP header */
> >  	ipv4_outer.total_length = rte_cpu_to_be_16(t_len);
> >  	ipv4_outer.packet_id = rte_cpu_to_be_16(seq);
> > -	rte_memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
> > +	memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));

How about:
*dst = ipv4_outer;

Don't know if it applies here.

> >  	dst += sizeof(ipv4_outer);
> >  	m->l3_len = sizeof(ipv4_outer);
> > -	rte_memcpy(dst, &esph, sizeof(esph));
> > +	memcpy(dst, &esph, sizeof(esph));
> >  	dst += sizeof(esph);
> >
> >  	if (string != NULL) {
> >  		/* copy payload */
> > -		rte_memcpy(dst, string, len);
> > +		memcpy(dst, string, len);
> >  		dst += len;
> >  		/* copy pad bytes */
> > -		rte_memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
> > +		memcpy(dst, esp_pad_bytes, RTE_MIN(padlen,
> >  			sizeof(esp_pad_bytes)));
> >  		dst += padlen;
> >  		/* copy ESP tail header */
> > -		rte_memcpy(dst, &espt, sizeof(espt));
> > +		memcpy(dst, &espt, sizeof(espt));

Also here:
*dst = espt;

Also don't know if it applies here.

> >  	} else
> >  		memset(dst, 0, t_len);
> >
> > --
> 
> Acked-by: Konstantin Ananyev <konstantin.ananyev@huawei.com>
> 
> > 2.53.0

With or without suggested changes:
Acked-by: Morten Brørup <mb@smartsharesystems.com>

If you are curious too...
Does the compiler still get confused about AVX rte_memcpy (without this patch), if applying the rte_memcpy patch?
https://patchwork.dpdk.org/project/dpdk/patch/20260521185631.116046-1-mb@smartsharesystems.com/

Re: [PATCH] app/test: use memcpy in ipsec test

[Stephen Hemminger]

On Fri, 29 May 2026 22:45:00 +0200
Morten Brørup <mb@smartsharesystems.com> wrote:

> > > @@ -604,22 +603,22 @@ setup_test_string_tunneled(struct rte_mempool
> > > *mpool, const char *string,
> > >  	/* copy outer IP and ESP header */
> > >  	ipv4_outer.total_length = rte_cpu_to_be_16(t_len);
> > >  	ipv4_outer.packet_id = rte_cpu_to_be_16(seq);
> > > -	rte_memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
> > > +	memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));  
> 
> How about:
> *dst = ipv4_outer;
> 
> Don't know if it applies here.

Good idea but dst is char *.
I suppose could use a cast but at that point the good
properties of assignment disappear.

Didn't want to go changing other code.

Re: [PATCH] app/test: use memcpy in ipsec test

[Stephen Hemminger]

On Fri, 29 May 2026 22:45:00 +0200
Morten Brørup <mb@smartsharesystems.com> wrote:

> If you are curious too...
> Does the compiler still get confused about AVX rte_memcpy (without this patch), if applying the rte_memcpy patch?
> https://patchwork.dpdk.org/project/dpdk/patch/20260521185631.116046-1-mb@smartsharesystems.com/

No still fails.

This is if __rte_always_inline is defined as just inline as an experiment.
Compiler gets confused in virtio_net because of matching conditions doing
initialization in virtio_net.

Also, has issue with rte_memcpy.

ninja: Entering directory `build'
[2352/3763] Compiling C object lib/librte_vhost.a.p/vhost_virtio_net.c.o
../lib/vhost/virtio_net.c: In function ‘desc_to_mbuf’:
../lib/vhost/virtio_net.c:3025:34: warning: ‘pkts_info’ may be used uninitialized [-Wmaybe-uninitialized]
 3025 |                         pkts_info[slot_idx].nethdr = *hdr;
      |                                  ^
../lib/vhost/virtio_net.c:2915:37: note: ‘pkts_info’ was declared here
 2915 |         struct async_inflight_info *pkts_info;
      |                                     ^~~~~~~~~
[3487/3763] Compiling C object app/dpdk-test.p/test_test_ipsec.c.o
In file included from /usr/lib/gcc/x86_64-linux-gnu/15/include/immintrin.h:43,
                 from ../lib/eal/x86/include/rte_rtm.h:8,
                 from ../lib/eal/x86/include/rte_spinlock.h:9,
                 from ../lib/mempool/rte_mempool.h:44,
                 from ../lib/mbuf/rte_mbuf.h:39,
                 from ../app/test/test_ipsec.c:11:
In function ‘_mm256_loadu_si256’,
    inlined from ‘rte_mov32’ at ../lib/eal/x86/include/rte_memcpy.h:119:9,
    inlined from ‘rte_mov64’ at ../lib/eal/x86/include/rte_memcpy.h:158:2,
    inlined from ‘rte_mov128’ at ../lib/eal/x86/include/rte_memcpy.h:170:2,
    inlined from ‘rte_memcpy_generic_more_than_64’ at ../lib/eal/x86/include/rte_memcpy.h:389:4,
    inlined from ‘rte_memcpy’ at ../lib/eal/x86/include/rte_memcpy.h:715:10,
    inlined from ‘setup_test_string_tunneled.constprop’ at ../app/test/test_ipsec.c:615:3:
/usr/lib/gcc/x86_64-linux-gnu/15/include/avxintrin.h:873:10: warning: array subscript ‘__m256i_u[3]’ is partly outside array bounds of ‘const char[108]’ [-Warray-bounds=]
  873 |   return *__P;
      |          ^~~~
../app/test/test_ipsec.c: In function ‘setup_test_string_tunneled.constprop’:
../app/test/test_ipsec.c:527:12: note: at offset 96 into object ‘null_plain_data’ of size 108
  527 | const char null_plain_data[] =
      |            ^~~~~~~~~~~~~~~
[3763/3763] Linking target app/dpdk-test

RE: [PATCH] app/test: use memcpy in ipsec test

[Morten Brørup]

> From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> Sent: Saturday, 30 May 2026 00.53
> 
> On Fri, 29 May 2026 22:45:00 +0200
> Morten Brørup <mb@smartsharesystems.com> wrote:
> 
> > > > @@ -604,22 +603,22 @@ setup_test_string_tunneled(struct
> rte_mempool
> > > > *mpool, const char *string,
> > > >  	/* copy outer IP and ESP header */
> > > >  	ipv4_outer.total_length = rte_cpu_to_be_16(t_len);
> > > >  	ipv4_outer.packet_id = rte_cpu_to_be_16(seq);
> > > > -	rte_memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
> > > > +	memcpy(dst, &ipv4_outer, sizeof(ipv4_outer));
> >
> > How about:
> > *dst = ipv4_outer;
> >
> > Don't know if it applies here.
> 
> Good idea but dst is char *.
> I suppose could use a cast but at that point the good
> properties of assignment disappear.
> 
> Didn't want to go changing other code.

Agree. Better stick with memcpy() than type cast.

RE: [PATCH] app/test: use memcpy in ipsec test

[Morten Brørup]

> From: Stephen Hemminger [mailto:stephen@networkplumber.org]
> Sent: Saturday, 30 May 2026 00.59
> 
> On Fri, 29 May 2026 22:45:00 +0200
> Morten Brørup <mb@smartsharesystems.com> wrote:
> 
> > If you are curious too...
> > Does the compiler still get confused about AVX rte_memcpy (without
> this patch), if applying the rte_memcpy patch?
> > https://patchwork.dpdk.org/project/dpdk/patch/20260521185631.116046-
> 1-mb@smartsharesystems.com/
> 
> No still fails.
> 
> This is if __rte_always_inline is defined as just inline as an
> experiment.
> Compiler gets confused in virtio_net because of matching conditions
> doing
> initialization in virtio_net.
> 
> Also, has issue with rte_memcpy.
> 
> ninja: Entering directory `build'
> [2352/3763] Compiling C object
> lib/librte_vhost.a.p/vhost_virtio_net.c.o
> ../lib/vhost/virtio_net.c: In function ‘desc_to_mbuf’:
> ../lib/vhost/virtio_net.c:3025:34: warning: ‘pkts_info’ may be used
> uninitialized [-Wmaybe-uninitialized]
>  3025 |                         pkts_info[slot_idx].nethdr = *hdr;
>       |                                  ^
> ../lib/vhost/virtio_net.c:2915:37: note: ‘pkts_info’ was declared here
>  2915 |         struct async_inflight_info *pkts_info;
>       |                                     ^~~~~~~~~
> [3487/3763] Compiling C object app/dpdk-test.p/test_test_ipsec.c.o
> In file included from /usr/lib/gcc/x86_64-linux-
> gnu/15/include/immintrin.h:43,
>                  from ../lib/eal/x86/include/rte_rtm.h:8,
>                  from ../lib/eal/x86/include/rte_spinlock.h:9,
>                  from ../lib/mempool/rte_mempool.h:44,
>                  from ../lib/mbuf/rte_mbuf.h:39,
>                  from ../app/test/test_ipsec.c:11:
> In function ‘_mm256_loadu_si256’,
>     inlined from ‘rte_mov32’ at
> ../lib/eal/x86/include/rte_memcpy.h:119:9,
>     inlined from ‘rte_mov64’ at
> ../lib/eal/x86/include/rte_memcpy.h:158:2,
>     inlined from ‘rte_mov128’ at
> ../lib/eal/x86/include/rte_memcpy.h:170:2,
>     inlined from ‘rte_memcpy_generic_more_than_64’ at
> ../lib/eal/x86/include/rte_memcpy.h:389:4,
>     inlined from ‘rte_memcpy’ at
> ../lib/eal/x86/include/rte_memcpy.h:715:10,
>     inlined from ‘setup_test_string_tunneled.constprop’ at
> ../app/test/test_ipsec.c:615:3:
> /usr/lib/gcc/x86_64-linux-gnu/15/include/avxintrin.h:873:10: warning:
> array subscript ‘__m256i_u[3]’ is partly outside array bounds of ‘const
> char[108]’ [-Warray-bounds=]
>   873 |   return *__P;
>       |          ^~~~
> ../app/test/test_ipsec.c: In function
> ‘setup_test_string_tunneled.constprop’:
> ../app/test/test_ipsec.c:527:12: note: at offset 96 into object
> ‘null_plain_data’ of size 108
>   527 | const char null_plain_data[] =
>       |            ^~~~~~~~~~~~~~~
> [3763/3763] Linking target app/dpdk-test

Interesting experiment. Thanks for sharing.

RE: [EXTERNAL] [PATCH] app/test: use memcpy in ipsec test

[Akhil Goyal]

> This test has tables of data that get copied with rte_memcpy.
> But when compiled without always inline the compiler gets confused
> by the inlining of rte_memcpy and thinks that it is possible for AVX
> code to reference past the input data.
> 
> Workaround is to use memcpy() which is better for this test anyway
> since regular memcpy has more static checking from compiler and
> analyzers.
> 
> Signed-off-by: Stephen Hemminger <stephen@networkplumber.org>

Applied to dpdk-next-crypto
Thanks.