|SUCCESS| net/ark: fix null dereference on allocation failure

dpdklab at iol.unh.edu dpdklab at iol.unh.edu
Wed Jun 3 18:40:20 CEST 2026


Test-Label: ai-code-review
Test-Status: SUCCESS
https://dpdk.org/patch/164871

_AI Code Review_

=== Patch Review: series.patch (via Claude) ===
Review date: 2026-06-03

Reviewing patch: net/ark: fix null dereference on allocation failure

commit bf73ee28f4 ("net/ark: support single function with multiple port")
Author: Denis Sergeev <denserg.edu at gmail.com>

    net/ark: fix null dereference on allocation failure

This patch adds a NULL check after rte_zmalloc_socket() to prevent
undefined behavior when memory allocation fails.

> diff --git a/drivers/net/ark/ark_ethdev.c b/drivers/net/ark/ark_ethdev.c
> index 8b25ed948f..546a44704f 100644
> --- a/drivers/net/ark/ark_ethdev.c
> +++ b/drivers/net/ark/ark_ethdev.c
> @@ -445,6 +445,12 @@ ark_dev_init(struct rte_eth_dev *dev)
>  					   sizeof(struct ark_adapter),
>  					   RTE_CACHE_LINE_SIZE,
>  					   rte_socket_id());
> +		if (eth_dev->data->dev_private == NULL) {
> +			ARK_PMD_LOG(ERR,
> +				    "Memory allocation for dev_private failed!"
> +				    " Exiting.\n");

The coding style guide prohibits unnecessary exclamation marks and
discourages overly dramatic error messages. Can this be changed to:
"Failed to allocate memory for dev_private\n"

Also, does the newline character need to be on a separate string literal?
The ARK_PMD_LOG macro likely expands to include its own newline handling,
making the trailing "\n" within the format string potentially redundant
or creating double newlines.

> +			goto error;
> +		}
>  
>  		memcpy(eth_dev->data->dev_private, ark,
>  		       sizeof(struct ark_adapter));

Does this error path properly clean up all previously allocated resources
before reaching the 'error' label? Looking at the function context, if
this allocation fails after rte_eth_dev_probing_finish() was called or
other resources were allocated, are those properly released by the error
label handler?


More information about the test-report mailing list