[PATCH] ethdev: fix 32-bit build with GCC-13

Ori Kam orika at nvidia.com
Wed Nov 1 09:12:59 CET 2023


Hi

> -----Original Message-----
> From: Ruifeng Wang <ruifeng.wang at arm.com>
> Sent: Wednesday, November 1, 2023 9:16 AM
> 
> aarch32 build with gcc-13.0.1 generated following warning:
> 
> In function 'memcpy',
>     inlined from 'rte_memcpy' at
> ../lib/eal/arm/include/rte_memcpy_32.h:296:9,
>     inlined from 'rte_flow_conv_action_conf' at ../lib/ethdev/rte_flow.c:726:20,
>     inlined from 'rte_flow_conv_actions' at ../lib/ethdev/rte_flow.c:936:10:
> warning: '__builtin_memcpy' specified bound 4294967264 exceeds maximum
> object size 2147483647 [-Wstringop-overflow=]
> 
> The issue is due to possible wrapping in unsigned arithmetic.
> The 'size' can be 0. 'off' is 32. When 'tmp' is equal to (unsigned)-32,
> the copy length is more than half the address space. Hence the warning.
> 
> Casted variables to 64-bit to avoid wrapping.
> 
> Fixes: 063911ee1df4 ("ethdev: add flow API object converter")
> Cc: adrien.mazarguil at 6wind.com
> Cc: stable at dpdk.org
> 
> Reported-by: Luca Boccassi <bluca at debian.org>
> Signed-off-by: Ruifeng Wang <ruifeng.wang at arm.com>
> ---
>  lib/ethdev/rte_flow.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/lib/ethdev/rte_flow.c b/lib/ethdev/rte_flow.c
> index 3a67f1aaba..2a5a057195 100644
> --- a/lib/ethdev/rte_flow.c
> +++ b/lib/ethdev/rte_flow.c
> @@ -722,7 +722,7 @@ rte_flow_conv_action_conf(void *buf, const size_t
> size,
>  		if (src.rss->key_len && src.rss->key) {
>  			off = RTE_ALIGN_CEIL(off, sizeof(*dst.rss->key));
>  			tmp = sizeof(*src.rss->key) * src.rss->key_len;
> -			if (size >= off + tmp)
> +			if (size >= (uint64_t)off + (uint64_t)tmp)
>  				dst.rss->key = rte_memcpy
>  					((void *)((uintptr_t)dst.rss + off),
>  					 src.rss->key, tmp);
> @@ -731,7 +731,7 @@ rte_flow_conv_action_conf(void *buf, const size_t
> size,
>  		if (src.rss->queue_num) {
>  			off = RTE_ALIGN_CEIL(off, sizeof(*dst.rss->queue));
>  			tmp = sizeof(*src.rss->queue) * src.rss->queue_num;
> -			if (size >= off + tmp)
> +			if (size >= (uint64_t)off + (uint64_t)tmp)
>  				dst.rss->queue = rte_memcpy
>  					((void *)((uintptr_t)dst.rss + off),
>  					 src.rss->queue, tmp);
> --
> 2.25.1

Acked-by: Ori Kam <orika at nvidia.com>
Best,
Ori


More information about the stable mailing list